FISCAL POLICY RULES AND PUBLIC CAPITAL FORMATION IN AUSTRALIA · PDF file ·...

FISCAL POLICY RULES AND PUBLIC CAPITAL FORMATION IN AUSTRALIA

Kerry Ann Carne B.Bus (Prof. Acc.), Grad. Dip. App. Ec., C.P.A., M. App. Ec.

Griffith Business School, Griffith University

Submitted in fulfillment of the requirements of the degree of Doctor of Philosophy

February 2007

ABSTRACT

Subsequent to the abandonment of the Bretton-Woods agreement, many governments

experienced worsening fiscal outcomes and subsequent heightening debt levels over

several decades. Many have recently adopted rules-based fiscal policy regimes in an

attempt to correct this. The experience of capital formation by Australian national

and sub-national governments is therefore examined before and after their adoption of

fiscal policy rules. Applying non-parametric and parametric methods to data drawn

from public policy documents, the degree to which the examined governments

complied with the constraints imposed on fiscal measures by adoption of fiscal policy

rules was ascertained. The Australian governments have generally, though not always,

met fiscal constraints imposed by their fiscal policy rules. The absence of penalties

for non-compliance may have contributed to the occasional exceptions to this high

level of compliance. However, intentionality of compliance where observed cannot

be ascertained due to the virtually simultaneous adoption of accrual-based financial

reporting frameworks, and resulting informational effect, and other possible causal

factors.

The degree of compliance varied with the type of fiscal policy rule. In order,

constraints imposed by net debt, followed by net worth and budgetary balance rules

were most frequently met. Possible causes include the significantly enhanced

information set available to governments after adoption of accrual-based financial

reporting networks and the significance attributed by governments to their credit

ratings.

Attention was then focused on the experience of public capital formation and whether

it changed at the date of adoption of net worth fiscal policy rules. The

Commonwealth, Victorian, Queensland and Western Australian Governments

increased the level, growth rate or output elasticity of their investment when they

adopted a fiscal policy rule requiring, at a minimum, that they maintain their net

worth. The Victorian Government’s experience showed the effects of unique

infrastructure financing arrangements. The investment experience of the New South

Wales Government cannot be modelled effectively in this way due to its practice of

transfer of assets to other levels of government during the period of the study.

One potential implication of the research findings reported in this thesis is that the

usual macroeconomic assumption of exogeneity of government expenditures may be

too strong in circumstances where governments have adopted such fiscal policy rules.

Specifically, it appears necessary to review the general assumption that only certain

elements of government expenditures, those that are related to automatic stabilisers,

are business cycle dependent. That is, other government expenditures, those usually

considered to be independent of levels of economic activity, may no longer be able to

be considered to be so when certain institutional arrangements, such as fiscal policy

rules, exist. Instead, constraints imposed by adoption of fiscal policy rules appear

likely to assume the position of determining upper or lower bounds on certain fiscal

measures.

Further, consistent with the literature on supply-side effects of public capital

formation, the jurisdictions experiencing increased growth of public capital formation

subsequent to adoption of fiscal policy rules are those which have experienced higher

growth rates than other jurisdictions. This indicates the existence of a number of

interesting directions for further research.

STATEMENT OF ORIGINALITY This work has not previously been submitted for a degree or diploma in any university. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.

/ /

ACKNOWLEDGEMENTS Grateful thanks to Christine Smith, Tom Nguyen and John Forster, each of whom at various times assumed the responsibilities of principal supervisor, for guidance, encouragement and direction. Thanks also to Peter Crossman for early advice. Thanks also to my family for their support and encouragement throughout the duration of this work. Completion of this dissertation was made possible through support provided by an Australian Research Council funded APAI scholarship. The industry partner associated with this project was the Office of Economic and Statistical Research, Queensland Treasury.

TABLE OF CONTENTS

CHAPTER 1 INTRODUCTION 1.1 TYPES OF FISCAL DECISION-MAKING FRAMEWORKS 1 1.2 RECENT DEVELOPMENTS IN FISCAL POLICY RULES

LITERATURE AND IDENTIFICATION OF RESEARCH GAP 5 1.3 THE SIGNIFICANCE OF FISCAL DECISION-MAKING

FRAMEWORKS 7 1.4 MOTIVATION AND CONTRIBUTION OF THIS STUDY 9 1.5 STRUCTURE OF THE DISSERTATION 11 CHAPTER 2 REVIEW OF RELEVANT THEORETICAL AND EMPIRICAL LITERATURE 2.1 INTRODUCTION 15 2.2 DEMAND-SIDE IMPACTS OF CHANGES IN GOVERNMENT

EXPENDITURE 2.2.1 Neoclassical theory 17 2.2.2 Keynesian theory 19 2.2.3 New classical models - impact of rational expectations 25

2.3 SUPPLY-SIDE EFFECTS OF GOVERNMENT INVESTMENT ON PRIVATE PRODUCTION 26

2.4 THE DEFICIT BIAS 30 2.5 INSTITUTIONAL ARRANGEMENTS AND PUBLIC CAPITAL FORMATION 33 2.5.1 Governmental reporting and budgeting practices 34 2.5.2 Credit agencies’ public debt rating methodologies 41 2.6 EMPIRICAL INVESTIGATIONS OF THE EFFECTIVENESS OF FISCAL

POLICY RULES 48 2.7 SUMMARY 56 CHAPTER 3 FINANCIAL MANAGEMENT FRAMEWORKS AND FISCAL POLICY RULES 3.1 INTRODUCTION 58 3.2 IMPACTS OF FISCAL POLICY RULES 59 3.2.1 The role of penalties 59 3.2.2 Intentional and unintentional compliance 69

3.2.3 What constitutes compliance – prospective or retrospective measurement 68

3.2.4 Bracketing 71 3.3 AUSTRALIAN GOVERNMENTAL FINANCIAL REPORTING

FRAMEWORKS 72 3.3.1 Budgetary Uniform Presentation Framework 72 3.3.2 Cash and accrual bases of accounting 73

3.3.3 Government Finance Statistics (GFS) reporting framework 74 3.3.4 General purpose financial reporting 79 3.3.5 Differences between GFS and GPFR 80

3.4 FISCAL MEASURES 84 3.4.1 The net debt measure 85 3.4.2 Budgetary balance measures 86

3.4.3 Net worth measures 90 3.5 INTERDEPENDENCE OF FISCAL POLICY RULES 91 3.6 CONCLUSION 93 CHAPTER 4 FISCAL POLICY RULE ADOPTION BY AUSTRALIAN NATIONAL AND STATE GOVERNMENTS 4.1 INTRODUCTION 97 4.2 AUSTRALIAN FISCAL POLICY RULES REGIMES 98 4.2.1 Commonwealth Government 100 4.2.2 New South Wales Government 108 4.2.3 Victorian Government 115 4.2.4 Queensland Government 121 4.2.5 South Australian Government 126 4.2.6 Western Australian Government 132 4.2.7 Tasmanian Government 138 4.3 SUMMARY 142 CHAPTER 5 METHODS AND DATA 5.1 INTRODUCTION 145 5.2 FISCAL POLICY RULES AND MODIFICATION OF THE ASSUMED

EXOGENEITY OF GOVERNMENT 146 5.3 ANALYTICAL TECHNIQUES 149

5.3.1 Overview 149 5.3.2 Model specification and rationale 152

5.4 THE DATA 161 5.4.1 Fiscal outcomes 161 5.4.2 Gross fixed capital formation 162 5.4.3 Gross Domestic Product/Gross State Product 166 5.4.4 Issues relating to the bracketed period 167

5.5 SUMMARY 173

CHAPTER 6 COMPLIANCE AND PUBLIC CAPITAL FORMATION 6.1 INTRODUCTION 174 6.2 COMPLIANCE WITH FISCAL POLICY RULES 177

6.2.1 Budgetary balance rules 177 6.2.2 Net worth rules 181 6.2.3 Net debt rules 184 6.2.4 Summary 187

6.3 TRENDS IN PUBLIC CAPITAL FORMATION 188 6.4 NON-DEPENDENCE OF POSITIVE GROWTH IN PUBLIC CAPITAL

FORMATION AND NET WORTH FISCAL POLICY RULES 193 6.5 PROBABILITY OF INCREASED PUBLIC CAPITAL FORMATION

AFTER ADOPTION OF FISCAL POLICY RULES 196 6.6 PARAMETRIC ESTIMATION OF FISCAL POLICY RULES 197

6.6.1 Fiscal policy rules and the relationship between public capital formation and economic activity 197

6.6.2 Fiscal policy rules and the relationship between public capital formation and the passage of time 209

6.7 VARIATION IN PUBLIC CAPITAL FORMATION AND FISCAL POLICY RULES ADOPTION 214 6.7.1 Defining the event and outcome of interest 215 6.7.2 Identifying the period of examination 215 6.7.3 Determining the selection criteria for inclusion in the study 216 6.7.4 Selecting a method for measuring normal and abnormal outcomes 216 6.7.5 Designing the testing framework for abnormal outcomes 219 6.7.6 Results 220

6.8 CONCLUSION 223 6.8.1 Did governments comply with rules? 223 6.8.2 Did public capital formation vary after adoption of rules? 224

CHAPTER 7 CONCLUSION 7.1 INTRODUCTION 227 7.2 RESEARCH DIRECTIONS 231

LIST OF TABLES

Table 1.1 Major types of fiscal policy rules 4 Table 2.1 Moody’s long-term credit ratings 43 Table 2.2 Moody’s financial terms, definitions and ratios 47 Table 2.3 Standard & Poor’s ratios 49 Table 3.1 Possible fiscal outcomes following adoption of fiscal policy rules 64 Table 3.2 Integrated nature of GFS stocks and flows 76 Table 3.3 Outline of GFS Operating Statement 78 Table 3.4 Outline of GFS Statement of Stocks and Flows 78 Table 3.5 Reconciliation of GFS Net Operating Balance with Operating

Surplus/Deficit 83 Table 3.6 Reconciliation of GFS Net Worth with AAS 31 Net Assets 84 Table 3.7 Gross Domestic/State Product, Chain volume measures - percentage

change from previous year 89 Table 4.1 Commonwealth Government fiscal policy rules 100 Table 4.2 NSW Government fiscal policy rules 108 Table 4.3 Victorian Government fiscal policy rules 115 Table 4.4 Queensland Government fiscal policy rules 121 Table 4.5 South Australian government fiscal policy rules 127 Table 4.6 West Australian Government fiscal policy rules 132 Table 4.7 Key adoption and bracketing dates 143 Table 5.1 Dates of adoption of fiscal policy rules and interregnum 168 Table 6.1 Measurement of budgetary balance fiscal policy rules 179 Table 6.2 Budgetary balance fiscal outcomes (general government sector) 180 Table 6.3 Measurement of net worth fiscal policy rules 183 Table 6.4 Net worth fiscal outcomes 183 Table 6.5 Measurement of net debt fiscal policy rules 186 Table 6.6 Net debt outcomes (general government) 186 Table 6.7 Commonwealth Government results 201 Table 6.8 New South Wales Government results 202 Table 6.9 Victorian Government results 203 Table 6.10 Queensland Government results 204 Table 6.11 Western Australian Government results 205 Table 6.12 Summary of results 206 Table 6.13 Results of trend analysis 213 Table 6.14 Calculated values of the test statistic used in event analyses 222 Table 6.15 Critical values of the test statistic used in event analyses 222

LIST OF FIGURES

Figure 3.1 Measurement of fiscal outcomes in the context of a net debt fiscal policy rule 94 Figure 3.2 Measurement of fiscal outcomes in the context of a budgetary balance fiscal policy rule 95 Figure 3.3 Measurement of fiscal outcomes in the context of a net worth fiscal policy rule 96 Figure 5.1 Ratio of GFCF to GDP, Commonwealth Government 169 Figure 5.2 Ratio of GFCF to GSP, New South Wales Government 170 Figure 5.3 Ratio of GFCF to GSP, Victorian Government 171 Figure 5.4 Ratio of GFCF to GSP, Queensland Government 172 Figure 5.5 Ratio of GFCF to GSP, Western Australian Government 173 Figure 6.1 Trends in the ratio of GFCF to GDP/GSP, by jurisdiction 190 Figure 6.2 GFCF over time, by jurisdiction 192 Figure 6.3 Levels of GDP/GSP, by jurisdiction 193 Figure 6.4 Event study timelines 217

Chapter 1: Introduction

CHAPTER 1

INTRODUCTION

1.1 TYPES OF FISCAL DECISION-MAKING FRAMEWORKS

Governments of developed countries face perennial demands for increased service

provision and reduced taxation. The recent phenomenon of population ageing and the

challenge of meeting increased demands for services such as health care, while

drawing funding from a shrinking taxpayer base, seem likely to exacerbate these

pressures. These factors, taken together, may place a premium on fiscal management

in developed economies in future decades. In addition, governments of developing

countries face continual challenges to improve the living standards of their

populations.

One obvious contributor to a solution to the challenge of maintaining living standards

in the face of emerging demographic change, or of improving living standards in

lesser developed countries, is increased productivity. One way in which this could be

achieved is by improving the quality of investment and financing decision-making in

the public sector. In this context, it is useful to consider the frameworks used by

governments in their fiscal decision-making processes.

Fiscal decision-making frameworks can be seen to fall into two categories -

discretionary approaches and rules-based approaches. That is, fiscal policy rules may

be generally seen as an alternative approach to discretionary policy. The two

approaches can be distinguished on two main bases - the basis of authority for, and

the intended period of observance of, the fiscal intentions that comprise them.

One emerging definition of discretionary approaches are those approaches to fiscal

1


decision-making that do not incorporate explicitly-stated intentions that are both

enshrined in some authoritative form and intended to apply beyond the life of the

current government.

In contrast, the International Monetary Fund (IMF) (2001) at paragraph 105 defines

fiscal policy rules as “forms of agreement (usually in law) that restrict the fiscal

policy action of government”. Similarly, Kopits and Symansky (1998) at page 2

define such rules as “permanent constraints on fiscal policy, typically defined in

terms of an indicator of overall fiscal performance”. These definitions are accepted

for the purposes of this research, and are further explained below.

Kopits and Symansky (1998) posit that the intended period of observance is crucial to

identification of a fiscal rule. An intention that the rule be applied on a permanent

basis by successive governments is said to be an essential criterion. Therefore it is

necessary to consider the intended period of application when determining whether a

fiscal intention, however expressed, may be identified as a fiscal rule.

However, discretionary approaches may also demonstrate time consistency. Where a

discretionary approach is applied with consistency over a considerable period of time,

differences between the rule based and discretionary approaches become less

identifiable and tend to reside in the method of expression and establishment of the

approach. A “norm” or “principle”, if not enshrined in some authoritative form such

as legislation or international treaty, may be identified as a discretionary approach

even if, in retrospect, it has been observed over a considerable period of time. In

other words, it can be argued that it is the intended period of observance of a fiscal

intention, at the time of its establishment, as well as the method of its establishment,

that provides guidance in identifying whether that intention is part of a discretionary

2


or rules-based approach to fiscal decision-making.

Fiscal policy rules, as stated above, may be seen as forms of agreement, either

between governments (as in international treaties such as the Growth and Stability

Pact between European Union member countries, also known as the Maastricht Treaty)

or between a government and its constituency. Where enshrined in some lasting and

authoritative form, an explicitly stated intention of government regarding its future

fiscal activities, in a democratic society, represents an agreement of the latter type1.

There is a high degree of heterogeneity of fiscal policy rules. Rules vary widely in

their target variable, institutional coverage and methods of implementation. In

addition, local terminology varies, with few governments using the term ‘rule’.

Table 1.1 presents major types of fiscal policy rules.

1 The democratic process provides a means whereby the electorate may exercise its voting power to terminate the agreement should it be found unacceptable. In addition, it is elementary that the existence of this voting power, and the threat of its use, tends to restrict a government to actions that it expects the electorate will find agreeable.

3


Table 1.1 Major types of fiscal policy rules

Source: Adapted from Kopits and Symansky (1998)

Rules may cover fiscal aggregates (such as budget deficit, borrowings or debt) or

subcomponents of these aggregates. One frequently-occurring rule is the so-called

‘Golden Rule’, which requires that borrowing be undertaken only to finance

investment.

Following Kopits and Symansky (1998), rules may be distinguished by their degree of

stringency, precision and enforcement of the statutory instrument. Relevant issues

include whether rules require ex ante approval only or, additionally, ex post

observance (i.e. whether both prospective and retrospective financial statements

reflect adherence to the stated constraint), what penalties exist (tangible or intangible,

reputation, financial or judicial) and the degree of specificity of the stated constraint.

Type of rule Form

Balance between overall revenue and expenditure (that is, prohibition on government borrowing); or limit on government deficit as a proportion of GDP.

Balance between structural (or cyclically adjusted) revenue and expenditure; or limit on structural (or cyclically adjusted) deficit as a proportion of GDP.

Balanced-budget or deficit rules

Balance between current revenue and current expenditure (that is, borrowing permitted only to finance capital expenditure).

Prohibition on government borrowing from domestic sources Borrowing rules

Prohibition on government borrowing from the central bank; or limit on such borrowing as a proportion of past government revenue or expenditure.

A limit on the stock of gross (or net) government liabilities as a proportion of GDP

Debt or reserve rules Target stock of reserves of extra budgetary contingency

funds (such as social security funds) as a proportion of annual benefit payments

4


1.2 RECENT DEVELOPMENTS IN FISCAL POLICY RULES

LITERATURE AND IDENTIFICATION OF RESEARCH GAP

Two interesting developments have recently occurred in the economics and public

finance literature. The first is the detailed assessment of the effectiveness of fiscal

policy rules in achieving fiscal sustainability by both the IMF and the Organisation

for Economic Co-operation and Development (OECD). Kopits and Symansky (1998)

for the IMF concluded that fiscal policy rules could provide a useful contribution to

the achievement of fiscal sustainability if the rules had certain characteristics and

were supported by a complementary program of underlying structural change. The

OECD (2003) adopted a similar view.

The second is the emergence of a body of literature that indicates that public capital

stocks have important impacts on private production. Gramlich (1994) notes that

Aschauer first linked public infrastructure investment and productivity in 1989, and

that supporting work by Aschauer and following work by others (e.g. Nourzad (2000),

Otto and Voss (1994 and 1996)) has further developed on this theme.

An important research issue arising from the above insights is how fiscal policy rules

have affected productivity and growth through their effect on public assets

accumulation. This thesis makes a contribution to exploring this research issue in that

it investigates the nature and extent of any relationship between fiscal policy rules and

the management of debt and accumulation of public assets in Australia.

This general research issue can be further disaggregated into specific elements. One

element is the question of how effective rules have been in terms of achievement of

their stated targets or constraints. As a logical progression, the next element, in an

investigation of the effect of fiscal policy rules on productivity and growth, is how

5


appropriate have targets or constraints, established under fiscal policy rules, been to

fostering economic growth.

One aspect of the latter is how rules have affected capital formation. While, in an

economic sense, capital formation refers to accumulation of real assets, in a financial

sense, capital formation refers to accumulation of both assets and liabilities. Past

research on the effectiveness of fiscal policy rules has mainly focused on examining

how rules have contributed to achievement of fiscal sustainability (which has been

interpreted as meaning reductions in debt levels) and on macroeconomic effects via

fiscal multipliers that have varied widely in their coverage, but none of which have

focussed on private capital production elasticities. In other words, the effect of fiscal

policy rules on the assets dimension of public capital formation, and thence on private

sector production, has not yet been considered, either explicitly or in isolation from

other effects. It is this research gap this dissertation seeks to make a contribution

towards filling.

The finding of a link between levels of public asset holdings and growth, together

with recently-developed awareness of the effectiveness of fiscal policy rules in

managing one aspect of the liabilities dimension of public capital formation, raises the

question of the effect of fiscal policy rules on the assets dimension of public capital

formation. Specifically, this refers to the amount and composition of public assets

accumulation. In the case where sub-national governments are examined, the

question also arises as to the relative effect of national and sub-national fiscal policy

rules.

Awareness of the answers to these questions would inform an assessment of how

different fiscal policy rules may be expected to impact, through their effect on capital

6


formation, on productivity and growth. Provision of an enriched information set with

respect to this issue may assist governments to choose improved decision-making

frameworks and, eventually, lead to improved fiscal policy, with subsequent positive

macroeconomic effects. Improved information should also be of interest to

international agencies tasked with assisting governments of developing and transition

countries to meet their full growth potential.

1.3 THE SIGNIFICANCE OF FISCAL DECISION-MAKING

FRAMEWORKS

In order to best formulate fiscal policy, it is necessary to understand how the decision-

making frameworks adopted by governments affect the fiscal policy stances

eventually adopted. In order to then extend the analysis to gain an understanding of

how fiscal policy decision-making frameworks have ultimately impacted the

macroeconomy, it is necessary to understand how those stances, when implemented,

have affected productivity and growth. In essence, it is necessary to understand the

linkage between the decision-making framework (in this case, fiscal policy rules) and

the effect on the economy.

Fiscal policy rules impact productivity and economic growth in four main ways.

Firstly, fiscal policy rules determine levels of government outlays in the short run.

The impact of government outlays on the economy is the subject of extensive

scholarly investigation. Mainstream economics of long standing teaches that

government outlays comprises a part of aggregate demand, and that demand, price

and output levels can be manipulated at the aggregate level through fiscal policy. The

potential for a ‘crowding out’ effect of government outlays and the importance of

sources of finance (such as use of public debt), in affecting demand for money and

7


interest rates, are also well recognised. New growth theory provides a basis for

recognition that the indirect effects of government service provision (e.g. education)

may also impact on economic growth.

Secondly, it is also well accepted that fiscal policy rules influence the capacity of

government to stabilise output over business cycles by adopting counter-cyclical

policies, effected through automatic stabilisers, customised measures or both. Thus

government may offset price pressures that threaten continued output growth by

adopting a contractionary fiscal policy stance (running a budget surplus, both ex ante

and ex post) or may stimulate a sluggish economy by adopting an expansionary fiscal

stance (running a budget deficit). The latter requires that the government’s fiscal

position is sustainable and that capacity exists to finance such a policy position, either

by drawing down reserves or by accessing debt, which in turn requires a sufficiently

conservative existing debt position to allow for increases without overly burdensome

costs. Fiscal sustainability is a major motivator to adoption of fiscal policy rules.

Thirdly, fiscal policy rules affect the composition of government outlays (between

consumption expenditure and investment) and hence affect the level of public capital

stock. In other words, fiscal policy rules affect the investment decisions made by

governments.

As previously stated, a literature has recently developed that indicates the importance

of public capital formation stocks on private production. For example, Nourzad (2000)

notes that, while the textbook version of Keynesian fiscal policy does not distinguish

between consumption spending and investment by governments, the two types of

spending are likely to have different effects on the economy. Specifically, while

government consumption spending affects the economy through its contribution

8


to aggregate demand in the short run, government investment also affects the

economy, not only through aggregate demand in the short run, but also, in the longer

run, through aggregate supply. This is because government-provided infrastructure is

used by firms in their production processes. Use of roads, highways and bridges by

private sector transport providers is a common example. This is what Aschauer (2000)

calls the direct effect of government investment on private sector output growth, the

indirect effect being that the complementary nature of private and public capital in

private-sector productive activity is such that an increase in the public capital stock

raises the return to private capital, which spurs the rate of growth of private capital

stock. The literature is further surveyed in Chapter 2.

Fourthly, fiscal policy rules affect the use of debt by government to finance their

activities. In other words, they affect the financing decisions made by governments.

Kopits and Symansky (1998) trace the effect of fiscal policy rules on debt levels as an

indicator of sustainability. However, some rules go beyond targeting sustainable debt

levels, and target very low levels of debt or impose restrictions on the uses to which

borrowings may be put. Such rules have been criticised (for example by Buiter (2001)

and Quiggin (1995)) as potentially welfare-reducing.

1.4 MOTIVATION AND CONTRIBUTION OF THIS STUDY

The research question to be addressed in this dissertation is whether, in the context of

Australian national and sub-national (i.e. State) government, adoption of fiscal policy

rules coincided with changes in the rate of accumulation of public assets. In

accordance with economics concepts, the latter is termed ‘public capital formation’

within this study.

Consideration of the international experience of fiscal policy rules, as stated above,

9


has focussed on the contribution of rules to achievement of fiscal sustainability, where

this was considered to depend primarily on debt reduction. However, rules may also

have an effect on the accumulation of public assets. It is therefore intended to

investigate the nature and extent of any relationship between fiscal policy rules and

accumulation of public assets.

While fiscal policy rules are unlikely to prove the only causal factor in public assets

accumulation, fiscal policy rules provide the framework within which governments

undertake their investment and financing decision-making. Thus, fiscal policy rules

have both direct and indirect impacts on the financial decisions of governments.

However, the proposed research does not exhaust all possible areas of investigation.

For example, little focus is placed on the degree of success enjoyed by governments

in meeting targets or constraints imposed by fiscal policy rules. While coverage is

provided of a variety of data generated by fiscal decisions taken under a fiscal policy

rules regime, this is focussed on understanding more completely the effect on capital

formation i.e. a ‘filling out the picture’ approach is taken. For example, in order to

understand the effect of a fiscal policy rule aimed at maintaining net worth, data may

be provided on complementary ratios such as public investment as a percentage of

Gross Domestic Product (GDP) or Gross State Product (GSP) as well as on aggregate

net worth levels.

Importantly, no effort is made to identify other determinants of public investment.

For example, the political economy aspects of investment decision-making are not

presented or explored.

Capital formation aspects of fiscal policy rules are also not the only growth-relevant

aspect of government outlays. Endogenous growth theory provides a basis on which

10


to relate government expenditure on such services as education and health care to

productivity and growth. This area also is not investigated in this research.

Finally, this research does not attempt to assess the on-going issue of the respective

merits of discretionary versus rules-based approaches to fiscal policy. Rather, an

assessment of the effectiveness of rules-based regimes is presented, in order to allow

commentary on governments’ success or otherwise in meeting the constraints that

comprise those rules-based regimes. In addition, the analysis in this thesis does not

extend to the macroeconomic impacts of fiscal policy but rather, focuses on any

impact fiscal policy rules may be observed to have had on public capital formation.

1.5 STRUCTURE OF THE DISSERTATION

In Chapter 2, the relevant theoretical and empirical literature is surveyed. In

particular, the various influences on public capital formation and its impact on the

macroeconomy are described. The focus is on what economic theory has to say about

the effect of government expenditure on the macroeconomy, including how

government expenditure affects the demand and supply sides of the macroeconomy,

the influence of institutional arrangements and the effectiveness of fiscal policy rules

in influencing the level and composition of government expenditure.

Neo-classical and Keynesian theories of the government budget, the deficit bias, the

Ricardian equivalence theorem and the impact of rational expectations on the

effectiveness of fiscal policy are explored for their demand-side impacts. Then

follows description of the supply-side effects of public capital formation on private

production. Institutional arrangements surrounding public investment are discussed,

including governmental budgeting processes and the methods by which international

credit rating agencies assess and assign ratings in a practical supplement to the

11


theoretical treatment contained in the previous section. The significance of

governmental budgeting practices is that these practices form the means of decision-

making regarding the level and composition of governmental expenditure and

investment. The significance of credit rating agencies’ methodologies is that, through

their focus on public debt, their ratings have become important elements of Australian

governments’ fiscal strategies in the context of common adoption of credit ratings and

net debt level targets. The empirical literature on fiscal policy rules is then surveyed.

This literature investigates the effectiveness of fiscal policy rules in influencing

government fiscal outcomes and attaining their ultimate goal of economic growth by

enhancing the fiscal sustainability of adoptive governments. This provides a basis for

the reporting of results of analysis undertaken in Chapter 6.

Chapter 3 provides a basis for identification and understanding of fiscal measures (or

outcomes) which may be constrained by adoption of fiscal policy rules. It outlines the

range of fiscal outcomes that may follow adoption of fiscal policy rules, including the

possible role of factors not under the control of government. This enables distinction

of intentional from unintentional fiscal outcomes and thus intentional from

unintentional compliance. In order to generate understanding of the arrangements

under which are reported the fiscal outcomes subsequent to adoption of fiscal policy

rules, governmental financial reporting frameworks employed in the Australian

context are then described. The intended effect of various fiscal policy rules on key

fiscal measures are then specified and described in order to identify a link between

fiscal policy announcements and measures of fiscal outcomes. An important aspect

of an accrual reporting framework, namely its integrated nature and the resulting

interdependencies between fiscal aggregates, is then described. The implication of

this is that interdependencies exist between fiscal policy rules, and hence that certain

12


fiscal policy rules may imply constraints on fiscal measures additional to the stated

measure. Understanding of this is necessary to enable interpretation of the effects of

fiscal policy rules adopted by Australian governments.

Chapter 4 presents a brief history of rules-based fiscal regimes internationally and a

detailed coverage of fiscal strategies adopted by Australian national and State

jurisdictions. The latter provides a basis for identifying the date of adoption of fiscal

policy rules (as defined) relating to budgetary balance, debt reduction and net worth.

The precise targets adopted and the sector (general government or total public sector)

to which they apply are identified in order to provide a foundation for empirical

investigation of compliance and public capital formation impacts in Chapters 5 and 6.

Chapter 5 details methods and data used in this thesis to explore empirically

Australian governments’ fiscal outcomes after adoption of fiscal policy rules. This

chapter describes the logical process undertaken to arrive at the research question of

“What happens to public investment when governments adopt fiscal policy rules?”

The expectation is that public investment will increase as a result of adoption of a net

worth fiscal policy rule, if, as assumed, governments comply with the constraint

imposed on their fiscal measures by adoption of that rule and if the composition of net

worth remains unchanged.

Investigation of the research question uses several techniques. These bring multiple

lenses to focus on the research question, strengthening the overall analysis by

overlapping examination which in part mitigates weaknesses of any individual

technique, and permits development of an overall picture from a number of different

perspectives.

The results of application of these different approaches are provided in Chapter 6.

13


The primary finding and major contribution of this thesis is that adoption of net worth

fiscal policy rules has, in the Australian context, coincided with changed rates of

public capital formation. This implies that adoption of fiscal policy rules may have

had an influence on governmental investment decision-making. This is significant

when considered in the context of private production elasticities related to public

capital stocks and the resulting changes in levels of economic activity.

14

Chapter 2: Review of Relevant Theoretical and Empirical Literature

CHAPTER 2

REVIEW OF RELEVANT THEORETICAL AND EMPIRICAL

LITERATURE

2.1 INTRODUCTION

This chapter describes the various influences on public capital formation and its

impact on the macroeconomy. The line of enquiry is this: Firstly, what does the

theoretical literature have to say about the effects of government expenditure on the

macroeconomy? Specifically, how does government expenditure affect, separately,

the demand and supply sides of the macroeconomy? Secondly, what does the

empirical literature have to say about levels of government expenditure? Thirdly,

how do institutional arrangements influence government expenditure? Lastly, how

effective are particular institutional arrangements i.e. adoption of fiscal policy rules,

in influencing the level and composition of government expenditure?

In pursuing this line of enquiry, Section 2.2 presents neo-classical and Keynesian

theories of the government budget including the Ricardian equivalence theorem and

the impact of rational expectations on the effectiveness of fiscal policy and the deficit

bias. This section provides theoretical background to this study and establishes the

mainstream, currently-accepted view of how government expenditure broadly impacts

the macroeconomy. The focus is on demand-side impacts.

Section 2.3 provides coverage of supply-side effects of public capital formation on

private production. Findings of this literature indicate that public capital formation

has important impacts on private production, providing a rationale for this study and

justification for the research question. This section further supplements the coverage

of demand-side effects of fiscal policy contained in the previous section, providing a

15


more complete view of the effects of government expenditure on the macroeconomy.

Section 2.4 discusses the deficit bias. This behavioural construct has arisen from

empirical observation of governments’ fiscal outcomes over time. It provides a

political economy supplement to the economics literature and contributes further to

our in understanding of the many determinants of government expenditure.

In Section 2.5, institutional arrangements surrounding public investment are

presented, including the process of governmental budgeting and, in particular, the

methods by which international credit rating agencies assess and assign ratings.

Governmental budgeting practices are significant because it is through these practices

that decisions are reached regarding the level and composition of governmental

expenditure and investment. Credit rating agencies’ methodologies are significant

because public debt ratings have become important elements of Australian

governments’ fiscal strategies. Most jurisdictions targeted particular credit ratings

and adopted net debt fiscal policy rules as a means of achieving the targeted rating.

To the extent that net debt iwas a focal point of fiscal strategy, public capital

formation may have reduced as available funds awere used to retire debt or

accumulate financial assets. This section provides a practical supplement to the

theoretical treatment contained in the previous section.

Section 2.6 reviews the empirical literature on fiscal policy rules. This literature

broadly investigates how effective fiscal policy rules have been in influencing

government fiscal outcomes. The current state of thinking is summarized regarding

the efficacy of fiscal policy rules in attaining their ultimate goal of economic growth

by enhancing the fiscal sustainability of adoptive governments. This provides a basis

for evaluation and comparison of results reported in Chapter 6 of this study. Section

2.7 summarises.

16


2.2 DEMAND-SIDE IMPACTS OF CHANGES IN GOVERNMENT

EXPENDITURE

2.2.1 Neoclassical theory

The neoclassical theory of the budget examines a closed economy in which the

government may be characterised as a ‘benevolent social planner’ which maximises

the utility of a representative agent who consumes, works and saves. The agent’s

utility function depends on private consumption and leisure but does not depend on

the government’s provision of public goods. Neoclassical theory of the government

budget stresses tax smoothing, wherein budget deficits may be incurred when needed

to cover temporary increases in government spending, with an emphasis on

maintaining constant tax rates to minimise distortions.

An intertemporal budget constraint is imposed and implies that the present value of

the government’s (exogenously determined and stochastic) spending must equal the

present value of taxes2. The level of taxes is therefore determined by the

intertemporal budget constraint. Deficits occur when spending is high and surpluses

occur when spending is low. Time horizons are identical (between agent and

government) and infinite (abstracting from intergenerational aspects and from finite

government terms). The government finances spending in each period by imposing

taxes on labour income, which are distortionary as they affect labour supply. The

2 Some more complex versions of this model allow debt financing of deficits. For example, Barro’s 1979 model allows issuing non-state contingent (i.e. default free) debt to assist smooth tax rates and Lucas and Stokey’s 1983 model allows state contingent debt (Loukoianova and Vahey (2002)).

17


model is based on the idea that the government minimises the distortion from taxation

by allocating taxes over time, rather than varying taxation with expenditure.

The individual utility function is concave and has decreasing marginal utilities.

Therefore a policy of constant taxes dominates a balanced budget policy because the

additional tax distortions of the present more than offset the welfare gains of the lower

tax rates of the future (Loukoianova and Vahey (2002)).

Another element of neoclassical theory is the Ricardian equivalence theorem.

Ricardian equivalence, as the name implies, was first proposed by Ricardo3 and

revived by Barro (Hakes and McCormick (1996)). In its perfect form, Ricardian

equivalence is based on strong assumptions of long time horizons, infinitely-lived

families, perfect foresight, absence of liquidity constraints, perfect capital markets and

altruism. Ricardian consumers, aware of the government’s intertemporal budget

constraint, will anticipate that a (lump sum) tax cut that is financed by issuing

government debt will result in imposition of higher future taxes, and will therefore

consider their permanent income to be unaffected and will not change consumption.

Any intended expansionary effect of the debt-financed tax cut will be nullified and

equivalence will exist between taxes and debt. The fiscal multiplier will be zero.

Private saving will fully offset public dissaving and, as national savings therefore

remain unaltered, the interest rate will not respond to variations in the stance of fiscal

policy (Barro (1984), Hakes and McCormick (1996), Kotlikoff, Razin and Rosenthal

(1990)).

3 However, Hakes and McCormick (1996) argue that Adam Smith presented many of the same arguments as Ricardo, framed the equivalence issue in the same manner as modern scholars (asking whether government bonds are net wealth, or whether tax rebates are saved) and furthermore reached qualitatively similar conclusions to Ricardo, though using a less mathematically rigorous approach.

18


Barro showed how intergenerational altruism can economically link current and future

generations by current generations choosing to reduce consumption in response to

increased debt, anticipating future increased taxation, and leave bequests to future

generations (interior transfers) with which to pay that taxation. The implication is that

current generations thereby neutralize intergenerational redistribution by the

government.

Many authors criticise the realism of the underlying assumptions noted above and

argue that the practical significance of Ricardian equivalence in its perfect form is

therefore limited. See for example, Kotlikoff, Razin and Rosenthal (1990), Khalid

(1996) and Barskey, Mankiw and Zeldes (1986).

2.2.2 Keynesian theory

Keynesian models stress macroeconomic stabilisation i.e. moderation of the business

cycle as the goal of fiscal policy, indicating a use of expansionary policy during

episodes of economic contraction and contractionary policy during episodes of

economic expansion. These models adopt a focus on demand side management

wherein government can influence aggregate demand directly by changing its

expenditure level. This is the textbook expenditure approach to measurement of

Gross Domestic Product (GDP), expressed as: GDP = C + I + G + X – M where C =

consumption expenditure, I = private investment expenditure, G = government

expenditure, X = expenditure on exports and M = expenditure on imports.

Government expenditure is thus considered to be exogeneous to the model.

Hemming, Kell and Mahfouz (2002) note that the Keynesian approach, in its simplest

form, assumes price rigidity and excess capacity, so that output is determined by

aggregate demand, and an increase/decrease in fiscal outlays has a multiplier effect on

19


aggregate demand and output. Change in government expenditure in turn changes

income levels, which in turn changes expenditure in a series of successive ‘rounds’ of

expenditure and income. The tendency by households to additional consumption

expenditure in response to additional disposable income earned is termed the marginal

propensity to consume (MPC). If all income is either spent or saved, the MPC is that

proportion of each dollar of additional disposable income that is spent4. The

multiplier in this model is positive, exceeds one and its size depends on the

responsiveness of consumption to disposable income (Barro (1984), Stiglitz (1997)

and Gans, King, Stonecash and Mankew (2003)).

Thus the basic formula for the simple multiplier is : Multiplier = 1/(1 - MPC).The

nature of the fiscal change affects the size of the multiplier. Since consumption is

affected by individuals’ disposable income, taxes reduce consumption. The multiplier

is larger for an increase in spending than for a reduction in taxes. This is because an

increase in government expenditure represents a direct injection into the expenditure

flow, a ‘first round’ of increased expenditure which is compounded by successive

rounds as the increased expenditure increases incomes, which in turn are spent. In

contrast, a reduction in taxes works indirectly to affect an increase in expenditure by

increasing disposable income.

In addition, the type of taxation regime in existence has an effect on the size of the

multiplier. Proportional or consumption taxes will result in a smaller multiplier than

will lump sum taxes (Stiglitz (1997)).

Then the basic formula for the simple multiplier with taxes (fixed for simplicity at a

4 The Marginal Propensity to Save (MPS) is that proportion of each dollar of additional disposable income that is saved. The MPC plus the MPS equals one.

20


given fraction, t, of income) is: Multiplier = 1/(1- MPC*(1 – t)). A spending increase,

that is fully funded by a tax increase, results in a balanced budget multiplier of exactly

one5. When an increase in government expenditures is exactly matched by an

increase in taxes, the increased taxes reduce disposable income and thus private

consumption, while the expenditure increases disposable income and thus private

consumption. However, the direct injection of expenditure into the circular flow,

compared with the indirect effect of taxation increase (wherein consumption is

reduced via the reduction of disposable income), means that the two effects are not

equal. The increase in consumption is almost completely offset by the increased tax

effect, with the eventual amount of the expansionary effect equalling the amount of

the increase in government expenditure (Stiglitz (1997)).

Extensions to the simple Keynesian view include crowding out (both through induced

changes in interest and exchange rates and through direct substitution of public for

private provision of goods and services) and increases in imports to meet any increase

in domestic demand. These influences affect the size of fiscal multipliers but not their

sign. The extent of crowding out depends on the sensitivity of investment and money

demand to interest rates, openness of the economy, whether the exchange rate is fixed

or flexible and price rigidity.

In a closed economy and in the absence of accommodative monetary policy, an

increase in government expenditures will cause interest rates to rise as the increase in

5 Barro (1984) separates consumption expenditure effects of taxation and government expenditure increases into two classes, one where consumers substitute consumption of publicly-produced goods and services for privately-produced goods and services, and one where consumers reduce consumption generally as a result of reduced disposable income arising from increased taxes. The substitution effect is said to be unaffected by, and occur additional to, the disposable income effect. Barro then posits that a balanced budget multiplier may be less than one. This more common analysis of the balanced budget multiplier does not distinguish between the two consumption effects and essentially ignores the substitution effect.

21


national income causes increased demand for money6 (Stiglitz (1997), Gans et al

(2003)). The standard IS-LM model, (see for example Barro (1984)), holds that

private investment responds negatively to changes in interest rates. Thus public

expenditure will ‘crowd out’ private investment. Of course, a reduction in money

demand may be offset by monetary policy easing (unless the demand for money is

infinitely sensitive to interest rates, when monetary policy would be rendered

ineffective).

In addition, the financing method adopted is of some importance. If a fiscal

expansion is financed by build up of public debt, risk premia on interest rates may

increase to reflect the mounting risk of default or inflation, and reduced private

investment and other expenditure may result.

The permanency of a fiscal policy shift is significant. For example, a temporary fiscal

expansion financed by debt will be more effective than a permanent expansion

because the increase in risk premia will be less, reflecting the lower risk of

undermining debt sustainability. Policy credibility will impact this process – if the

government lacks a track record of fiscal prudence, and there is little faith in the

temporary nature of a spending increase or tax cut, interest rates are more likely to

incorporate increased risk premia. This supports the proposition that fiscal policy

rules are a more effective means of attaining fiscal sustainability than are

discretionary policy positions, since by the definition adopted in this study, fiscal

policy rules are intended to apply not only for the term of power of the adoptive

government but for the future as well i.e. for the term of all successive governments.

6 This is in the long run. Gans et al (2003) reverses the order of analysis in the short run, with the price level ‘stuck at some level’, the interest rate adjusting to balance the supply and demand for money and the level of output responding to changes in aggregate demand, which is in part determined by the interest rate.

22


These effects may be exacerbated in an open economy with highly mobile capital. If

a fiscal expansion is viewed as a risk to debt sustainability, fears may also be raised

that there will be future balance of payments problems. This may lead to reduced

foreign investment and to capital outflows, adversely impacting economy activity

(Hemming, Kell and Mahfouz (2002)).

Under a flexible exchange rate regime, capital inflows (attracted by higher domestic

interest rates) cause an appreciation of the exchange rate, leading to reduced

competitiveness of domestic products on external markets and offsetting the increase

in aggregate demand by reducing exports. Where capital is perfectly mobile,

complete crowding out will occur, rendering fiscal policy ineffective (unless the fiscal

expansion is financed by drawing down reserves). The extent of crowding out will

vary with sensitivity of investment and money demand to interest rates, and with the

relationship between money demand and income. If investment is an increasing

function of current income, multiplier-accelerator models show that quite large fiscal

multipliers occur despite the crowding out effect.

Under a fixed exchange rate regime, a smaller increase in interest rates will occur than

in a closed economy, and where capital is perfectly mobile, the money supply will

increase sufficiently to offset all of the threatened increase in domestic interest rates,

reducing crowding out and increasing the effectiveness of fiscal policy (and the size

of the multiplier).

Increased government expenditure is deemed productive, that is, it results in increased

provision of public goods and services. These public goods and services may

substitute for some private consumption, for example when the government provides

free services such as libraries. Then people reduce expenditure where those services

are a close substitute for services they might have bought from private sector

23


producers. To the extent that this impacts the marginal propensity to consume and

increases savings, direct crowding out occurs and the fiscal multiplier is reduced.

This analysis abstracts from any effect on the marginal products of labour and capital,

assuming no direct effect on work effort or investment demand. Changes in the real

interest rate are assumed to be the sole mechanism by which changes in government

expenditure impact on investment (Barro (1984)). This assertion is contradicted by

the findings of a number of writers regarding the existence of private production

elasticities in relation to public investment (see Aschauer (1988) and Otto and Voss

(1994, 1996)). These studies are surveyed in Section 2.4.

Neo-Keynesian models allow for price change and show that price flexibility narrows

the range of values taken by fiscal multipliers and limits the effects of the exchange

rate regime. For example, if prices are flexible, in a closed economy, a fiscal

expansion will put upward pressure on prices, which will partly offset the short term

increase in aggregate demand, reinforcing the crowding out effect induced by interest

rate rises and reducing the multiplier.

However, in an open economy with a flexible exchange rate, the extent of crowding

out depends on the response of domestic prices to exchange rate changes. If domestic

prices and exchange rates move together, crowding out will be lower (and multipliers

higher) than with price rigidity, as prices reduce with appreciation of the exchange

rate. With a fixed exchange rate, more crowding out will occur than with price

rigidity, as the current account deteriorates in response to exchange rate appreciation-

induced price increases.

The marginal propensity to import (MPI) equals the proportion of each additional

dollar of income that is spent on imports. The size of the MPI affects the size of the

24


fiscal multiplier. Any increase in income (such as that resulting from an increase in

government expenditure) does not enter the expenditure flow creating additional

rounds of expenditure if it is spend on imported goods and thus becomes income of

citizens of foreign countries. Thus imports are a third leakage (additional to savings

and taxes) from the circular income flow (Stiglitz (1997)).

Wealth effects may contribute to crowding out by affecting aggregate demand. If

changes in interest rates, the exchange rate and prices affect forward-looking

consumers’ perceptions of their financial wealth, wealth-dependent consumption will

also respond. This effect is variously known as the Pigou effect (after its discoverer)

and the real balance effect.

Dynamic effects of fiscal policy relate to the relative offsetting effects of the above

and below factors over time on the value of the fiscal multiplier. For example,

Auerbach and Kotlikoff (1987) note that if indirect (negative) effects such as

crowding out take longer to impact the multiplier than do direct (positive) effects of

increased expenditure, fiscal multipliers are likely to be larger in the short term than

in the long term.

2.2.3 New classical models - impact of rational expectations

In rational expectations models the distinction between temporary and permanent

policy changes is particularly important. Specifically, a permanent fiscal expansion

may lead to an expectation by households and firms that initial increases in interest

rates and appreciation of the exchange rate will become permanent and increase over

time, adding to crowding out, and possibly turning the fiscal multiplier negative. The

significance of this is that, when households hold rational expectations, the longer-

term effects of fiscal policy will matter even in the short term (Hemming, Kell and

25


Mahfouz (2002)). This is likely to strengthen the impact of fiscal policy rules on the

credibility of the adoptive government’s fiscal management, and thus on debt premia,

in the period immediately following their adoption.

2.3 SUPPLY-SIDE EFFECTS OF GOVERNMENT INVESTMENT ON

PRIVATE PRODUCTION

This section reviews the literature that focuses on how public investment impacts on

the macroeconomy. Barro (1984) notes that, while we could model public services as

the output from the government’s production function, this is typically not done.

Macromodels generally simplify matters by neglecting production in the public sector

and implicitly assume that the government buys only final goods and services in the

commodities markets, effectively subcontracting all of its production to the private

sector. Hence public investment, publicly owned capital and government

employment are always zero. To the extent that this is recognized to be unrealistic,

public sector technology and management capability are implicitly assumed to be

identical to that of private producers. This has been combined with an historical

attachment by the public sector to cash accounting. Arguably, until the emergence of

the literature linking public capital with private sector production, this lead to a lack

of awareness of the significance of public capital in determining private production

levels, composition and location.

To the extent that a fiscal expansion benefits supply, fiscal multipliers will increase in

size. Supply side effects of fiscal policy thus are of particular relevance to a study of

the effect of fiscal policy on public capital formation. This is because the ultimate

rationale of such a study is to gain an understanding of the impact of public capital

26


formation on private sector production and, in turn, on productivity and growth.

Dadgostar and Mirabelli (1998), in a review of studies of the effect of public capital

expenditure on the economy, state that the empirical research on the effect of fiscal

policy on GDP comprises three major groups. The first group focus on the

relationship between real GNP or employment and aggregate demand management.

Well-known economists such as Grossman and Barro have contributed to this group

of studies.

The second group focus on the sensitivity of the composition of national income to

the choice between tax and debt financing of current government expenditure i.e.

whether government expenditure directly substitutes for private consumption and

saving. Dadgostar and& Mirabelli (1998) report mixed findings, specifically that

Kochin in 1974, Ernest Tanner in 1979 and Kormendi in 1983 carried out empirical

studies that support the Ricardian equivalence proposition and generally concluded

that the method of financing for a given stream of government expenditure (tax or

debt) does not have any significant relationship with aggregate demand. Further,

Dadgostar and& Mirabelli (1998) report that Erenburg in 1993 found for the US that

deficit spending has no significant effect on private investment. Dadgostar and

Marabelli (1998) also find for Canada that the level of government debt has no effect

on private sector investment spending in both one and two period lagged models.

They further state that Feldstein in 1982 and Aschauer (1988) have found some

degree of substitution between government spending and private consumption.

However, none of these studies tested the relationship between public capital

expenditure, investment and productivity. This is the focus of the third group of

studies, which consider the supply side effects of government capital expenditure on

output and productivity. Aschauer (1988) finds that private sector investment

27


spending is enhanced by expected increases in government infrastructure spending.

Aschauer (1989) finds a correlation between ‘core infrastructure’ (highways, water

and sewer systems, mass transit and airports) and various measures of private sector

productivity. Aschauer (1997) finds the relationship between US public capital and

economic growth is not linear - at a certain level, the tax burden associated with

financing and maintaining public capital reduces the returns for private industry and

reduces growth. An optimal public-private ratio was estimated at 61% and the

recommendation was to reduce government consumption spending and increase

government investment in order to reach this level.

Dadgostar and Marabelli (1998) find for Canada that the effect of government

investment expenditure on private sector investment spending was statistically

significant and positive. In particular, the composition of government expenditure

affects the marginal efficiency of private capital and leads to increases in the level of

private investment.

Nourzad (2000) finds a positive and statistically significant effect of public capital on

labour productivity in 24 developed and developing economies during the period

1976 to 1989. Nourzad (2000) cites conflicting findings by a number of authors such

as Eisner, Hulte and Schwab, Tatom, Holtz-Eakin, Barth and Bradley, Barro and

Levine and Renelt , and concludes that the use of differing methodologies has

contributed to this lack of unanimity.

Otto and Voss (1994) find that public capital had a significant and positive impact on

private production and private total factor productivity in Australia over the period

1966-67 to 1989-90. Using a Cobb-Douglas production function for the private

economy, assuming exogenous technological change and differentiating between

private and public capital, output elasticities for public capital of about 0.40 are found.

28


Otto and Voss (1996) use data of greater frequency and econometric techniques to

account for non-stationary series. They confirm the complementarity of public and

private capital and revise their estimate of the elasticity of Australian public capital to

approximately 0.17, roughly half that of private capital. This means that for every

dollar invested in public capital, private output increases by 17 cents.

Hemming, Kell and Mahfouz (2002) note that the analysis of fiscal multipliers

traditionally focuses on short-term demand-side effects, while supply-side impacts are

seen to be longer-term in nature. However, there are certain exceptions to this. The

extent of crowding out, and thus the size of the multiplier, may depend on both

demand-side and supply-side effects even in the short term. For example, consider

the situation where a fiscal expansion is injected into an economy already operating at

full capacity and there is no ability to increase productive capacity in the short term.

In this case, the fiscal expansion has to be totally crowded out and will result in

contraction in the private sector. In the longer term, this may be addressed by

policies that promote supply-side responses, such as accumulation of public assets

that increase the marginal efficiency of private capital. Additionally, supply-side

effects of fiscal policy may increase short-term demand if these effects promote

expectations that longer-term growth will increase.

Tax mix is relevant to supply-side effects. Changes to labour income taxes may affect

the supply of labour and changes to capital taxes may affect domestic (and

internationally-sourced) savings and investment. Hemming, Kell and Mahfouz

(2002) note that conclusions have yet to be reached on the empirics of the impact of

29


tax changes on the supply of labour and capital, and thus on growth.

This section has surveyed the literature on the effects of public expenditure on private

production. While differing findings have resulted from the use of different

methodologies, this literature generally indicates that public capital formation has

significant impacts on private production.

The repeated experience of large fiscal imbalances and rapid build-up of public debt

in a number of OECD countries during the 1970s and 1980s represents a systematic

deviation from both the neoclassical and Keynesian principles of fiscal policy. This

experience led to concerns that political forces, together with budget compilation and

implementation procedures, cause an inherent bias towards deficits. The term ‘deficit

bias’ was coined to describe this phenomenon, which is described below.

2.4 THE DEFICIT BIAS

A ‘deficit bias’ is said to occur when democratically-elected governments consistently

run deficits regardless of the state of the business cycle – in particular failing to

reverse deficits incurred in economic contractions during subsequent periods of

expansion. It is in the context of the ‘deficit bias’ that fiscal rules have emerged.

Alesina and Perotti (1995) classify political economy models of fiscal policy into

models based on ‘fiscal illusion’ (in the spirit of public choice literature), models of

debt as a strategic variable, models of distributional conflict, models of

intergenerational redistribution, models of geographically dispersed interests and

models emphasizing the effects of budgetary institutions. These models indicate a

number of possible reasons for a tendency to deficits. Alesina and Perotti (1995) use

this classification of models in an attempt to answer the questions of why large and

30


persistent deficits have occurred during the decades of the 1970s and 1980s, but not

previously, and only in certain countries, but not in others.

The public choice literatures typically rests on key assumptions of opportunistic

policy makers (who use fiscal deficits to increase their electoral chances) and naïve

voters (who do not punish the policy makers because they either overestimate the

benefit of current expenditures or underestimate the future cost of current expenditure

i.e. the tax burden). Thus voters suffer from fiscal illusion as to the consequences of

re-electing policy makers who repeatedly enact deficits.

Strategic models of debt focus on the use of debt accumulated to finance budget

deficits by current governments as a means of constraining the actions of future

governments. Thus the tendency to deficits arises due to the conflict between

spending priorities of different political parties. In such a case, the current

government does not fully internalise the cost of running a current budget deficit,

because the constraint imposed on future spending may be borne by a different

government. The extent of electoral uncertainty is also a factor in determining the

extent of the deficit bias present in such a situation, both in its impact on governments

uncertain of their continued incumbency in power and in its impact on voters

uncertain of who will govern in the future. Such voters will generally favour current

politicians that promise them large current transfers, even if those transfers are debt

financed (Alesina and Perotti (1995)).

Models of distributional conflict emphasize the effect of conflict between different

social groups such as political parties, interest groups or members of a governing

coalition. Such conflict can delay adoption of fiscal probity measures due to lack of

agreement on how to share the burden of the fiscal adjustment. These models predict

that greater difficulty in implementing fiscal adjustment measures will be experienced

31


by fragmented or divided governments and polarized societies than where more

consensus exists.

Intergenerational redistribution models take the perspective that public debt

redistributes the tax burden across time and hence debt-financed public deficits may

effectively act as a form of intergenerational transfer. Alesina and Perotti (1995) cite

Cukierman and Meltzer, and Tabellini as highlighting cases where public deficits

were used as a form of negative bequest to transfer resources from richer future

generations to the poorer current generation.

Models emphasizing geographically-based constituencies stress that representatives of

one constituency may overestimate the benefits of public expenditure in their district

compared with its financing costs. This is because those costs are typically borne by

all taxpayers rather than by the legislator’s constituents only, whereas the benefits of

that expenditure largely accrue to the constituents. These models focus more closely

on institutional detail, are closer to the political science literature and study the

interaction between the organisation of legislatures and fiscal decisions.

Models emphasizing the effects of budgetary institutions explain the persistence of

fiscal imbalances and/or excessive levels of government expenditure in the context of

whether a deficit bias can be corrected by institutional reform such as appropriately

designed fiscal policy rules. One difficulty with such rules is the necessary trade-off

between simplicity of rules and the ability to retain policy flexibility. For example,

Alesina and Perotti (1995) cite Corsetii and Roubini as highlighting the trade-off

between deficit bias and the margin for stabilisation in the context of a closed and

open economy. They conclude that allowing the government a margin for

stabilisation policy does not contribute to worsening the deficit bias. Further they

conclude that the deficit bias is enhanced in an open economy because the

32


government, in a model without default risk, faces an infinitely elastic supply of funds

at the given world interest rate and additional borrowing to finance more expenditure

does not attract the higher cost of higher interest rates. In addition, some US states

with balanced budget rules run smaller budget deficits, while state budgets having

fiscal restraints show significantly reduced cyclical responsiveness. Milesi-Ferretti

(1997) indicates that the extent of decentralisation of the budget process has an impact.

Specifically, decentralized budgetary procedures are more likely to lead to deficits

bias due to failure by spending ministers to appropriately internalise the cost of their

expenditure (in contrast to the extent to which a prime minister or finance minister

would likely internalise the cost of expenditures, due to their differing

responsibilities).

As previously stated, fiscal policy rules were adopted by a large number of

governments in an apparent attempt to reverse the tendency during the 19870s and

1980s to run budgetary deficits and thus, to accumulate high levels of public debt.

However, the contention that a government deficit has an expansionary impact on the

macroeconomy is based on an assumption that consumers and investors will behave in

a particular way which could be characterised as suffering from a form of fiscal

illusion. The Ricardian equivalence theorem, described in the previous section,

challenges this assumption.

2.5 INSTITUTIONAL ARRANGEMENTS AND PUBLIC CAPITAL

FORMATION

This section outlines the role of institutional arrangements in determining fiscal

33


outcomes. In particular, the significance of governmental budgetary practices in

Australia is discussed. That institutional factors such as budgetary practices are

significant is supported by Poterba (1996) who points to the development of a search

for politico-economic explanations for deficit policies, observed in the form of large

and persistent budget deficits across nations, which are not obviously related to short-

term spending needs and thus are not explained by the existing literature.

Another important factor, impacting on Australian governments’ public capital

formation, is the methodology by which credit ratings agencies assign ratings to

public debt. Most Australian jurisdictions surveyed not only adopted fiscal policy

rules targeting net debt reduction, they also targeted a particular credit rating. In such

cases, the net debt target was a means by which to achieve the targeted credit rating.

For other jurisdictions an explicit credit rating target was not adopted but the rule

relating to net debt was likely to have been set at least in part to ensure maintenance

of a good credit rating. This focus on cash flow, with its implications for capacity of

governments to meet debt obligations as they become due, is likely to have impacted

on governments’ decisions to acquire financial assets rather than real assets.

2.5.1 Governmental reporting and budgeting practices

Arrangements governing public resource allocation constitute the background to the

operation of fiscal policy, whether policy is determined under a discretionary regime

or under a regime based on fiscal policy rules. A significant literature exists on the

conduct of government in Australia, a review of which is outside the scope of this

paper. Rather, this section focuses on the mechanisms by which Australian state and

national governments allocate resources and report performance.

34


Public resource allocation is performed primarily during the budgeting process.

Various authors have observed that government budgets are much more than merely

prospective financial reports (e.g. Kelly and Wanna (2004), Guthrie (1998), Boxall

(1998) and Robinson (2002b)). Kelly and Wanna (2004) at page 95, state:

“Government budgets are much more than financial documents; they are

strategic and resource allocation plans, economic projections and policy

statements, statements of government intention expressing policy

priorities, and means of resolving societal conflict. Whether in

parliamentary or presidential systems, budgets articulate the political

relations between budget guardians and spenders, between policy-specific

and whole-of-government objectives, and between the executive and

legislature.”

Australian governments’ budgets are produced before the beginning of the July – June

financial year, usually in May7. The national government budget is typically

produced first, as it provides tangible evidence of intended national government

grants to State governments, which comprise an important part of State government

revenue. State government budgets are then produced. Some governments also

produce amended budgets containing new initiatives during a mid-year review

process, at which progress toward budgeted goals and six-monthly financial reports

are reviewed.

Governments monitor progress toward budgeted financial statements, typically on a

monthly basis throughout the budgeted year. They also report estimates of actual

financial results (termed ‘estimated actuals’) or audited financial results (termed

7 However, the month of budget presentation sometimes varies as a result of unusual circumstances, such as when an election is held close to the usual time of production of a budget, as occurred in Queensland in 1999.

35


‘actuals’) for the budgeted year in budgetary financial reports produced for the

following year. These retrospective financial statements are produced in the same

format as the budgeted financial statements, that is, they are produced in accordance

with the Government Financial Statistics (GFS) framework, which are also provided

by governments to the Australian Bureau of Statistics (ABS) and form inputs to ABS

national accounting.

Also, retrospective general purpose financial reports (GPFR) are produced and a

variety of non-financial reporting is undertaken. In particular, two frameworks are

used by Australian governments to present financial reports – GFS and GPFR. These

frameworks are described in detail in Chapter 3. Movements are underway to

harmonise the two frameworks to provide a common reporting framework (Kaufmann

2002).

In addition, in producing financial reports, two conceptual bases have historically

been used – the cash basis and the accruals basis. The accruals basis is in current use

by all Australian governments, having replaced the previously-used cash basis around

the end of the 1990s. The differences between these two bases and their implications

for an empirical investigation relating to fiscal policy rules are analysed in detail in

Chapter 3.

Under the Westminister system of government, the budget is in fact an Act of

Parliament (usually termed an Appropriation Act) which authorizes payments

proposed by the government according to its policies and priorities (Government of

New South Wales 1998-99 Budget Paper 3). Following Churchill (1992), the

Westminister system of government effectively forced the public sector to report on a

cash accounting basis in order to parallel the budgeting process with its focus on a

cash or ‘funds’ basis. However, with increasing pressure on government to adopt

36


operational principles more in keeping with private sector enterprises, and the need

for an information set that provides richer treatment of assets and liabilities,

governments have adopted accrual accounting as the basis for their prospective

(budgeting) and retrospective (historical) financial reporting.

Robinson (2002a) at page 3 argues that Australian governments modified cash

accounting:

“ … into a system of expenditure accounting and budgeting … wherein the

acquisition of goods and services was recognized and subjected to

budgetary constraints in the financial year when those goods or services

were received, regardless of whether they were paid for in that year or

payment was deferred and a liability … incurred.”

Use of the cash conceptual basis by the national and state governments was largely

(though not completely) abandoned in Australia at the end of the 1990s as part of the

accrual output budgeting (AOB) reforms (see Kelly and Wanna (2004) and Robinson

(2002a)). Bartos (2000) notes that Australia was the third country to introduce

accrual budgeting and at that time had the largest public sector accrual budget in the

world.

Though each format has over time been produced using both the cash and accruals

conceptual bases, as previously stated, the two frameworks now used are based on

accruals concepts8. The suitability of the accrual basis for producing public financial

reports (either ex ante or ex post and in accordance with the GFS or GPFR

8 Kelly and Wanna (2004) noted that the national government still provides some special purpose financial reports, produced on a cash accounting basis, as supplements for decision making purposes during budgetary activities.

37


framework) for any level of government is the subject of a spirited debate. Perceived

benefits of the accruals basis include:

(a) an improved capacity to monitor efficiency in the allocation and use of

government resources and thus, enhanced sustainability of fiscal policy (ABS

(2000)), though members of the debate are divided on the validity of using this

information in assessing government performance, on the basis that profitability

is not a valid concept to apply to core government entities (see Robinson (1998));

(b) the richer information set it provides relating to assets and liabilities (though

members of the debate are divided on the validity of using this information in

decision-making, on the basis that government asset valuation methods are

inappropriate and lead to distortions in assessing service costs, making pricing

decisions and project valuation (Moll, Humphrey and Chow (2004));

(c) improved microeconomic efficiency of government service provision as a result

of its provision of a managerial decision tool and usefulness in the monitoring of

managerial performance at agency level (not supported by Robinson (1998), but

supported by most AOB enthusiasts and argued in national and state government

policy documents); and

(d) its provision of superior measures of the sustainability and intergenerational

equity implications of fiscal policy (Robinson (1998)).

In addition, debate exists as to the characteristics of reports at different levels of

aggregation. For example, Moll, Humphrey and Chow (2004) describe the debate

surrounding the usefulness of whole-of-government consolidated GPFR.

The chronology of reform of retrospective financial reporting is as follows. Since

38


1983, Commonwealth statutory authorities have been required to prepare historical

GPFR and department business undertakings were expected to prepare these data on

an accrual basis. During the 1980s, the Australian Accounting Research Foundation

and standard settings boards (Australian Accounting Standards Board and Public

Sector Accounting Standards Board) promoted the idea of GPFR for both public and

private sector (Guthrie (1998)). Since 1 July 1992, local government bodies (the

lowest tier of government under Australia’s model of federalism) have produced

historical accrual financial GPFR, as a consequence of the introduction of Australian

Accounting Standard AAS27 Reporting by Local Governments (AAS27). Since 1997,

national and state government departments have been required to produce historical

GPFR following the introduction of Australian Accounting Standard AAS29

Reporting by Government Departments (AAS29). Commonwealth Government

departments had produced audited GPFR since 1994–95.

From 1 July 1999, the Commonwealth Government and the respective State and

Territory Governments (i.e. New South Wales, Queensland, South Australia,

Tasmania, Victoria, Western Australia, the Australian Capital Territory and the

Northern Territory) have been required to provide consolidated whole-of-government

historical GPFR as a consequence of the introduction of Australian Accounting

Standard AAS31 Financial Reporting by Governments (AAS31) (Moll, Humphrey

and Chow (2004)). Some had produced these reports prior to the introduction of

AAS31.

The chronology of prospective financial reporting (budgeting) reform in Australia is

as follows. Since 1991, the national and most state governments9, by agreement, have

produced budgetary financial reporting in a uniform format termed the Uniform

9 The Victorian Government is an exception in that it provides most of its publicly available budgetary financial reports in GPFR format.

39


Presentation Framework (UPF). The UPF is based on GFS reporting standards and

adopts a GFS format which is described in detail in Chapter 3.

GFS reports were produced on a cash conceptual basis until 1998-99. The UPF was

amended in 1997, and again in 2000, to accommodate the change of GFS to an

accruals basis. This reform was driven by the revision by the International Monetary

Fund of its Manual on Government Finance Statistics to an accruals basis and the

subsequent broad adoption of accrual-based budgeting by governments world-wide.

The financial reporting reform described above was also accompanied by broader

resource allocation reforms. Kelly and Wanna (2004) distinguish between the

adoption of the accrual conceptual basis for (retrospective and prospective) reporting

and adoption of new governance (outputs-outcomes) frameworks that were an

additional component of AOB. They conclude that the evaluation of accrual based

prospective financial reporting can only be fully comprehended in the context of the

larger set of reforms that have occurred simultaneously at the end of the 1990s.

Described as a revolution in government accounting by Boxall (1998)) and as

representing huge changes in the mechanism of central financial control by Robinson

(2002b)), these combined developments have comprised a central part of wider

ranging public sector changes collectively referred to as New Public Management

(NPM) reforms.

These reforms were described by Kelly and Wanna (2004) at page 96 as :

“ … the fundamental logic of public sector budgeting … (is shifted) … to

include the independent pricing of outputs, the use of market prices and

contestable prices, and the use of diverse output delivery agents through

40


tendering and outsourcing.”

The intent of the reforms is described by Kelly and Wanna (2004) at page 96 as to:

“… reduce the incidence of budget lock-in, lessen the influence of

incrementalism, and to better align the responsibility for decision making

with the accountability for performance.”

The extent to which these benefits have been achieved remains the topic of debate.

These reforms have also been accompanied by changes in Australian governments’

capital management approaches. Paradoxically, the amended focus away from inputs,

resulting from the amended resource frameworks adopted, has been accompanied by

an increased focus on capital inputs to the public production process i.e. the use of

public assets in the government production process. This development has been the

subject of some criticism. For example, Johnstone (1999) argues that this new focus

on comparing service prices between public and private potential providers

incorporates a cost comparison methodology which is mistaken in its treatment of the

costs to government agencies of owning and using capital assets, and that this tilts the

analysis against any in-house (i.e. public sector provider) bids.

This section outlined practices through which public resource allocation, and thus,

public capital formation, occurs. An understanding of this is justified in light of the

methodological contribution made by this thesis i.e. that institutional arrangements

such as fiscal policy rule adoption are found to be influential in determining public

investment levels and, in turn, potentially productivity and GDP growth outcomes.

41


2.5.2 Credit agencies’ public debt rating methodologies

A full coverage of the methodologies by which credit ratings agencies determine

ratings is beyond the scope of this study. However, a brief coverage is provided

below because of the significance that debt ratings have assumed in governments’

fiscal policy rules.

Credit ratings agencies, such as Moody’s Investors Service (Moody’s) and Standard

and Poor’s Ratings Services (S&P’s)10, facilitate the functioning of capital markets by

expanding the information set of investors. They do this by assigning to publicly-

traded debt instruments of private and public corporations a rating which signals the

default risk imbedded in those instruments.

According to Moody’s (1998) at page 3:

“A rating is essentially an opinion on the ability and willingness of an

issuer to pay its debt obligations on a full and timely basis.”

Investors use this information in determining the required return and thus, given a

fixed income stream, the price they are willing to pay for the financial asset

represented by that debt. Debt ratings, then, are significant to debt-issuing

governments for several reasons. The first arises from the previous discussion. That

is, a government’s interest expense is in part determined by the risk rating assigned to

its debt. Secondly, debt ratings are widely considered to signal the sustainability of a

government’s fiscal management practices and therefore acquire significance that is

not only financial but also political in nature. Thirdly, credit ratings agencies mention

in their ratings reports any accounting practices deemed questionable and therefore a

10 While a number of credit rating agencies exist, only three are recognized worldwide. The third of the three is Fitch Ratings (Ylagan (2005)).

42


view exists that agencies have accrued some additional significance as pseudo-

auditors of accounting methodologies (Glater (2003)). Debt ratings vary by term and

are assigned to categories of long, short and medium term. Table 2.1 presents ratings

assigned by Moody’s to long-term debt.

Table 2.1 Moody’s long-term credit ratings

Source: Moody’s (2004)

Countries are assigned a foreign currency country rating which provides a ceiling on

the foreign-currency denominated debt ratings of debt issuers domiciled in that

country, including both the national and sub-national governments. According to

Moody’s (1998) at page 5:

“ …the country rating is intended to capture the risk that, in a currency or

foreign exchange crisis mode, the sovereign government may choose to

limit foreign currency payments by the entities subject to its sovereignty.”

43


In assigning a foreign currency country ratings, factors considered include: (a) risk of

political regime change, which could lead to a general repudiation of debts, civil war

or invasion; (b) whether the country has a well-defined system of contract law, which

allows for legal procedures for collection of unpaid debts including seizure of

collateral; (c) whether the country has a deep financial system of proven effectiveness

in making payments and is technically secure against breakdown; (d) stability of the

regulatory/legal environment against malleability, corruption, etc; and (e) the extent

of any tendency toward hyperinflation (Moody’s (2004) page 57).

Certain factors relevant to assigning a foreign currency rating are also relevant in

assigning a local currency rating. Guidelines for rating debt issued in local currency

are higher than guidelines for rating debt issued in foreign currency since, according

to Moody’s (2004) page 58:

“ … even if a government reschedules its debt, this does not necessarily

lead to large-scale disruption of the local currency payments system.”

However, the default risk on local currency instruments is considered to be usually

lower than that on foreign currency instruments. Factors relevant in assigning a rating

on local currency instruments include the government’s ability to mobilize currency

to meet its domestic currency commitments. It is frequently the case that

governments are more able to mobilize domestic resources than foreign resources.

Additional factors include the structure of the economy and its capacity generally, to

mobilize resources, which might be affected by the period of establishment of the

government, and the size of the public sector relative to its usable resource base.

Factors influencing the rating of a national government are also likely to be relevant

to a sub-national government. Additionally, the extent to which a national

44


government would support a fiscally-constrained sub-national government is relevant

to the credit rating of the sub-national government. According to Standard & Poor’s

(2004), the Australian State governments’ relationship with the Commonwealth

Government is a supporting credit factor. In addition, there is a direct link between

the country ceiling and the credit rating of a sub-national government. However, the

ceiling does not apply to debt instruments denominated in domestic currency and so

the domestic currency debt rating of a sub-national government may exceed the

country ceiling of the national government.

Moody’s approach to rating sub-national governments can be grouped into five broad

categories: 1. institutional framework; 2. economic fundamentals; 3. government

finances; 4. debt profile and 5. government structure and political dynamics (Moody’s

(2004)). Of these, budgetary performance and debt profiles are the most impacted by

fiscal policy rules and thus are of most relevance to this study.

In assessing government finances, Moody’s considers the budgetary framework, that

is, the structure of its revenue and expenditure base, as well as its budgetary

performance. Critical components of the budgetary framework include the extent of

the government’s taxation powers (i.e. breadth of tax sources such as sales tax, real

estate, corporate, or income tax, stability of taxation sources, control over tax rates,

extent of dependence on fiscal transfers from upper-tier government) and functions

assigned to the government, or assumed by it and the degree of financial flexibility

arising (as an indicator of the capacity to adjust spending if faced with significant

budgetary imbalances). Key indicators of budgetary performance include the ability

of the government to fully cover its operating costs with its recurring revenue base,

thus protecting the timely payment of debt service, and its ability to generate a surplus

45


from its operating account to finance a portion of its capital program. Emphasis is

placed on the ability of assets to generate cash for use in meeting debt obligations

rather than with the value of assets as stated on the debt issuer’s balance sheet.

Debt profile analysis recognizes differences between direct debt and contingent

liabilities and also takes into account likely future borrowing needs and refinancing

requirements. Direct debt analysis focuses on the statutory framework for debt

issuance and specific characteristics of the debt. The statutory framework for debt

issuance comprises issues such as the legal structure through which a government is

allowed to issue and service debt, relative ranking of payment obligations with other

debt or financial obligations and whether any particular security is pledged to support

payment or whether the entire revenue base is pledged as a source of payment.

Specific characteristics of the government’s debt include its overall level, the relative

burden it imposes on the revenue base, structure of the debt, the extent to which any

foreign currency risk is managed, where debt is foreign-currency denominated, and

short term debt exposure resulting from market access and interest rate risks.

Analysis of the significance of contingent liabilities addresses such issues as the

extent to which a government may become liable for debt issued by other entities such

as through explicit guarantees, ownership or other factors.

Table 2.2 lists ratios used by Moody’s, and defines relevant financial terms. Perusal of

the table indicates a strong focus on budgetary balance and minimization of net debt.

As could be expected, the methodologies of Moody’s and S&P’s show significant

similarities. For example, S&P’s report for 2004 indicates that critical criteria used in

rating the State of Victoria included (a) balance sheet measures (ratio of non-financial

public sector net financial liabilities to operating revenue, net direct debt of the

46

Cha

pter

2: R

evie

w o

f The

oret

ical

and

Em

piric

al L

itera

ture

Tab

le 2

.2 M

oody

’s fi

nanc

ial t

erm

s, de

finiti

ons a

nd r

atio

s

Rat

ios

Fina

ncia

l ter

ms

Def

initi

on

Ow

n So

urce

of R

even

ue a

s % o

f To

tal R

even

ue

Ow

n-So

urce

Rev

enue

In

clud

es lo

cally

col

lect

ed re

venu

es le

vied

by

or o

n be

half

of th

e lo

cal/r

egio

nal a

utho

rity.

Incl

udes

loca

l tax

es, u

ser f

ees,

inve

stm

ent

earn

ings

and

tran

sfer

s fro

m st

ate-

owne

d en

terp

rises

. In

terg

over

nmen

tal R

even

ue a

s % o

f To

tal R

even

ue

Inte

rgov

ernm

enta

l Rev

enue

A

lso

know

n as

Inte

rgov

ernm

enta

l Tra

nsfe

rs. I

nclu

des a

mou

nts

rece

ived

from

hig

her l

evel

s of g

over

nmen

t, w

heth

er in

tend

ed a

s ge

nera

l fis

cal a

id o

r as c

ost-s

harin

g fo

r par

ticul

ar sp

endi

ng c

ateg

orie

s In

tere

st E

xpen

se a

s % o

f Tot

al

Rev

enue

To

tal R

even

ue

Incl

udes

taxe

s, in

terg

over

nmen

tal t

rans

fers

, tra

nsfe

rs fr

om st

ate-

owne

d en

terp

rises

, fee

s and

all

othe

r rev

enue

s. Fi

nanc

ing

Surp

lus/

(Req

uire

men

t) as

%

of T

otal

Rev

enue

Fi

nanc

ing

Surp

lus/

(Req

uire

men

t) ne

t of

cap

ital e

xpen

ditu

res,

as %

of

Tota

l Rev

enue

Fina

ncin

g Su

rplu

s/(R

equi

rem

ent)

Def

ined

as T

otal

Rev

enue

less

Tot

al E

xpen

se.

Net

Gov

ernm

ent D

ebt p

er C

apita

N

et D

ebt a

s % o

f GD

P N

et D

ebt a

s % o

f Tot

al R

even

ue

Net

Gov

ernm

ent D

ebt

= G

ross

deb

t les

s sel

f-su

ppor

ting

debt

and

sink

ing

fund

s.

Net

For

eign

Cur

renc

y D

ebt a

s % o

f G

ross

Deb

t N

et F

orei

gn C

urre

ncy

Deb

t =

Gro

ss fo

reig

n cu

rren

cy e

xpos

ure

net o

f cur

renc

y sw

aps.

Shor

t-ter

m D

ebt

Incl

udes

all

debt

out

stan

ding

with

a m

atur

ity o

f les

s tha

n on

e ye

ar,

incl

udin

g cu

rren

t mat

uriti

es (n

et o

f am

ount

s to

be p

aid

from

sink

ing

fund

s) o

f lon

g-te

rm d

ebt,

as w

ell a

s flo

atin

g ra

te d

ebt.

Shor

t-ter

m D

ebt a

s % o

f Gro

ss D

ebt

Gro

ss d

ebt

Incl

udes

dire

ct d

ebt o

f the

gov

ernm

ent a

nd c

ontin

gent

liab

ilitie

s suc

h as

gua

rant

eed

debt

. So

urce

: Ada

pted

from

Moo

dy’s

(199

8)

47

Chapter 2: Review of Theoretical and Empirical Literature

general government sector to operating revenue); (b) management capacity

(commitment to fiscal discipline); (c) financial performance (surpluses including after

capital spending); (d) riskiness of revenue stream (proportion of revenue dominated

by core government revenues and low risk government-owned trading enterprises); (e)

strength and diversity of the economic base; (f) fiscal flexibility (per cent of revenue

received in grants from the federal government and therefore not under the State’s

control) compared with expenditure responsibility and (g) liquidity (debt profiles,

holdings of cash and liquid assets) (Standard & Poor’s (2004)). Table 2.3 lists the

relevant ratios used by Standard & Poor’s in formulating their ratings, segmented by

government sector.

This section has outlined the methodologies of two of the most important credit

ratings agencies. All examined governments assigned particular significance to debt

ratings in determining their fiscal strategies and several targeted a specific debt rating

to be achieved for its general government debt. This is presumably in response to

awareness that credit agencies’ debt ratings are effective in determining interest rates

on public debt, as reported in Poterba (1996).

2.6 EMPIRICAL INVESTIGATIONS OF THE EFFECTIVENESS OF

FISCAL POLICY RULES

The type of fiscal policy rules covered in this thesis are to be distinguished from the

type of fiscal policy rules often addressed by writers of mainstream economic

textbooks such as Dornbusch et al (2006). That analysis addresses fiscal policy rules

targeting macroeconomic aggregates such as levels or rate of growth of Gross

Domestic Product in absolute or per capita terms. Dornbusch and the great majority

of studies dismiss that type of fiscal policy rules as being flawed in terms of

48

Cha

pter

2: R

evie

w o

f The

oret

ical

and

Em

piric

al L

itera

ture

Tab

le 2

.3 S

tand

ard

& P

oor’

s rat

ios

Sect

or

Rat

ios

Non

-fin

anci

al p

ublic

sect

or

Cor

e go

vern

men

t Fi

nanc

ial p

erfo

rman

ce

Net

lend

ing/

reve

nue

Cas

h su

rplu

s (de

ficit)

/rece

ipts

C

ash

surp

lus a

fter c

apita

l exp

endi

ture

Bal

ance

shee

t N

et d

ebt/r

even

ue

Net

fina

ncia

l lia

bilit

ies/

reve

nue

Dire

ct d

ebt/r

even

ue

Bud

geta

ry p

erfo

rman

ce

Acc

rual

ope

ratin

g ba

lanc

e/op

erat

ing

reve

nue

Net

lend

ing/

oper

atin

g re

venu

e C

ash

surp

lus (

defic

it)/re

ceip

ts

Net

inte

rest

exp

ense

/ope

ratin

g re

venu

e

Acc

rual

ope

ratin

g ba

lanc

e/op

erat

ing

reve

nue

Net

lend

ing/

oper

atin

g re

venu

e C

ash

surp

lus (

defic

it)/re

ceip

ts

Net

inte

rest

exp

ense

/ope

ratin

g re

venu

e D

ebt b

urde

n N

et d

ebt/o

pera

ting

reve

nue

Net

fina

ncia

l lia

bilit

ies/

oper

atin

g re

venu

e N

et d

ebt/o

pera

ting

reve

nue

Net

fina

ncia

l lia

bilit

ies/

oper

atin

g re

venu

e

Key

fina

ncia

l rat

ios

Net

ope

ratin

g ba

lanc

e/op

erat

ing

reve

nue

Net

cas

h flo

w fr

om o

pera

tions

/rece

ipts

N

et in

tere

st e

xpen

se/o

pera

ting

reve

nue

Net

inte

rest

cov

er

Cap

ital e

xpen

ditu

re g

row

th

Cap

ital e

xpen

ditu

re/(e

xpen

ses +

cap

ex)

D

ivid

ends

/ope

ratin

g in

com

e C

ash

surp

lus (

defic

it)/re

ceip

ts

Net

lend

ing/

oper

atin

g re

venu

e

Ope

ratin

g re

venu

e/G

SP

Ope

ratin

g ex

pens

es/G

SP

Cas

h op

erat

ing

bala

nce/

cash

rece

ipts

A

ccru

al o

pera

ting

bala

nce/

oper

atin

g re

venu

e N

et in

tere

st re

venu

e/op

erat

ing

reve

nue

Net

lend

ing/

oper

atin

g re

venu

e C

ash

surp

lus(

defic

it)/re

ceip

ts

Cap

ital e

xpen

ditu

re/to

tal p

aym

ents

Sour

ce: a

dapt

ed fr

om S

tand

ard

& P

oor’

s (20

04)

49


usefulness for economic management purposes. In contrast, the fiscal policy rules

surveyed in this study target fiscal aggregates appearing in governmental financial

statements, such as a measure of operating balance or of net assets, as described in

detail in Chapter 4. Thus, the coverage of the mainstream textbook literature is, with

few exceptions, of limited relevance to the primary focus of this study..

A comprehensive coverage of the characteristics of different types of fiscal policy

rules can be found in Kopits and Symansky (1998). They state that the ultimate

rationale for adoption of fiscal rules is sustained growth, through short or medium-

term macroeconomic stabilisation (including price stability), to be achieved by

attaining a government financial position that is sufficiently flexible to enable a

moderate countercyclical role through the operation of automatic stabilisers. For

example, flexibly-implemented balanced budget rules may be used to offset the deficit

bias inherent in the politically rational behaviour of democratically elected

governments, restoring sufficient fiscal capacity to enable a countercyclical role. The

OECD (2003) broadly agrees.

Debate continues as to whether this objective is better served by discretionary policy

or by design and implementation of sensible fiscal rules. The debate rests on the

observation that while rules such as a strict balanced-budget requirement could impair

the short-run stabilisation and tax smoothing roles of fiscal policy, lack of fiscal

discipline may also reduce the countercyclical role of fiscal policy and render it

procyclical.

Kopits and Symansky (1998) and the OECD (2003) agree that fiscal rules may also (a)

enhance the effectiveness of monetary policy (by reducing the burden on it), (b)

achieve longer-run fiscal sustainability, with a rule that requires accumulation of

financial assets at least equal to total contingent liabilities serving intergenerational

50


equity (by preventing potential future increases in public indebtedness associated with

e.g. accumulating employee entitlements or prospective ageing of the population), or

(c) prevent negative spillovers within a federation (when implemented at sub-national

levels of government). Rules-based approaches may provide time consistency that

may not result from discretionary approaches.

Kopits and Symansky (1998) and the OECD (2003) also acknowledge that rules based

approaches may reduce transparency of fiscal policy by encouraging one-off practices

or creative accounting to meet strict targets. Similarly to discretionary approaches,

the nature of the adjustment implemented under rules influences the macroeconomic

repercussions of the rule. Where one-off measures, reductions in investment, tax

increases or temporary wage freezes are used, maximum beneficial effects on

investment, employment and growth will not be gained.

Kopits and Symansky (1998) at page 12, state that:

“Binding fiscal policy rules are likely to influence the level and

composition of government expenditure and taxation ... For the most part,

economic performance under fiscal rules has been mixed.”

Where governments of advanced economies, such as CFA franc zone and European

Union countries, United States states, Indonesia, Canada (at both the national and sub-

national levels) and New Zealand, have adopted fiscal policy rules and made attempts

to comply with them, those economies have experienced declines in inflation and

interest rates, reduced crowding out effects on private investment and experienced

some alleviation of external imbalance (Kopits and Symansky (1998). However,

these gains have come at the price of tax increases or distortions in the composition of

government expenditures. In certain cases, reductions in public investment has been

used to achieve the reduced expenditure required or taxes have either been increased

51


or their administration distorted such as by advancing payments, sometimes leading

also to reduced transparency in financial reporting.

Kopits and Symansky (1998) assess the macroeconomic effects of balanced-budget

rules and rules limiting government deficits and debt by, firstly, drawing on the

literature to ascertain likely effects of compliance with these rules on interest rates

and growth, and, secondly, using a set of stochastic simulations of the effect of rules

adoption and compliance on output variability in the short term. Using the IMF’s

MULTIMOD model, simulations were run for Germany and the United States (US).

These simulations found a US multiplier of about 1 for spending decreases and 0.7 for

tax increases. For Germany, fiscal multipliers of approximately half those of the US

were found. Interest rates were found to be sensitive to a change in the public debt

ratio, with an increase of government debt equal to 25 per cent of GDP leading to an

increase in long run interest rates of between 125 and 500 basis points.

Overall, it was found that, where rules are sufficiently flexible to allow for the

operation of automatic stabilizers, fiscal rules contribute only marginally to short-run

output variability while sometimes providing a degree of fiscal discipline unlikely to

be attained through discretionary approaches. Thus rules are able to assist with

correcting the deficit bias and contribute to improved economic outcomes. However,

in developing economies, compliance with fiscal policy rules that constrain

government borrowing from domestic sources may lead to increases in indebtedness

to external sources.

A later study by Hemming, Kell and Mahfouz (2002) surveyed the emerging literature

52


regarding impacts of fiscal policy rules. They note that estimates of fiscal multipliers

may be derived from macroeconomic model simulations and reduced form equations,

and that the former may be carried out as empirically estimated macroeconomic

models or as calibrated general equilibrium models. With respect to simulations, they

report that (a) a range of multiplier estimates are typically reported; (b) in all cases of

simulation analysis excluding the impacts of credibility, short-term multipliers were

positive, ranging from 0.1 to 3.1 with most expenditure multipliers between 0.6 and

1.4 and most tax multipliers between 0.3 and 0.8; (c) expenditure multipliers vary

between countries; (d) most models confirm that short-term tax multipliers are smaller

than are short-term expenditure multipliers, and the difference varies between

countries; (e) alternative monetary regimes have little effect on short-term fiscal

multipliers; (f) analysis including credibility effects find that debt reduction has fiscal

multiplier effects that are generally positive but small, even when credibility is low,

but that can be negative if credibility is high; and (g) long-term multipliers are smaller

than short-term multipliers due to crowding out effects and some long-term

multipliers are negative.

Simulations of calibrated general equilibrium models have focused on steady-state

fiscal policy impacts, again on macroeconomic aggregates such as output. The main

conclusions drawn by Hemming, Kell and Mahfouz (2002) included (a) positive

multipliers, resulting from unanticipated increases in government spending, typically

in the range of 0.3 to 1.2; (b) differential impacts of permanent and temporary

changes in expenditure; (c) differential impacts on employment of unanticipated debt-

financed increases in government expenditure depending on assumptions regarding

the effect of public employment on private capital and labour productivity; and (d)

dependence of the output effect of a deficit-financed tax reduction on type of taxation,

53


with resulting multipliers ranging from positive to negative.

Studies generating reduced form equations estimates of fiscal multipliers usually but

not exclusively produce positive short-term as well as long-term expenditure

multipliers. The overwhelming majority of these studies use US data.

Poterba (1996) reports the US experience with fiscal policy rules prohibiting

budgetary deficits during the 1980s and 1990s. Following a failed attempt to pass a

balanced budget amendment in 1982, continued deficits led to the circumstance by

1985 where the national government was borrowing to meet interest commitments.

The Gramm-Rudman-Hollings bill was enacted in 1985 and required phased

reductions on the deficit culminating in a balanced budget in 1990-91. The bill was

prescriptive as to what expenditure reductions were to be made to achieve this target.

The Bill was declared unconstitutional in 1986 however new legislation was passed in

1987. A series of deferrals of the deficit targets were passed and accounting

manipulations used to achieve only some of the targeted expenditure reductions, with

the result that substantial deficits remained at the end of the 1980s. Poterba (1996)

cites Reischauer as having estimated that half the deficit reduction was due to these

factors. Poterba (1996) also cites Gramlich as stating that, to the extent that primary

deficits dropped at the time of enactment of the legislation, this was probably

coincidental. However, Poterba (1996) also reports Hahm at al as stating that the

fiscal policy rules had some effectiveness. Overall, Poterba (1996) finds that results

available to date are insufficiently precise to enable more specific judgements about

the effects of rules on fiscal outcomes. Further budgetary reform was enacted in 1990,

with increasingly specific requirements regarding the timing of reductions in

expenditure. Again, conclusions appear to have been hard to draw regarding the

54


efficacy of these arrangements.

Poterba (1996) states that examination of the effect of fiscal policy rules at the sub-

national level in the United States also does not lead to clear cut conclusions. Nearly

all states, excepting only Vermont, have balanced budget rules which exhibit

significant variation from state to state. For example, Poterba (1996) at page 24 states:

“ … in forty-four states, the governor must submit a balanced

budget…thirty-seven states impose a stricter standard, requiring that the

legislature enact a balanced budget … (which)…nonetheless allow for

actual revenues and expenditure to diverge from balance if realisations

differ from expectations … (and)… the actual budget may be in deficit and

the state can borrow to carry this deficit forward to future years … the

third and strictest type of balanced budget rule combines a requirement

that the legislature enact a balanced budget with a prohibition on deficit

carry-forward … in 24 of the 37 states that require the legislature to enact

a balanced budget.”

In addition, state fiscal policy rules frequently apply to only the general fund and do

not apply to special funds, capital spending funds and trust funds. Also, most states

have no formal mechanisms for enforcement of their fiscal policy rules. Poterba cites

a number of studies on state-level budgetary outcomes, for example, Abrams and

Dougan as finding that balanced budget fiscal policy rules have not had substantial

effects on US state fiscal outcomes and Bohn and Inman as finding that balanced

budget fiscal policy rules have a statistically significant effect in reducing deficits and

that this arises from lower expenditure levels rather than increased taxation. Related

55


studies of the effectiveness of net debt fiscal policy rules find that only stringent

requirements, such as that state debt must be approved by the entire electorate in

referendum, are highly negatively correlated with state indebtedness (Poterba (1996)).

A criticism of fiscal policy rules is that compliance has lead to distortions in the

composition of government expenditures (unwarranted cuts in public investment),

lack of transparency in the budget process (adoption of creative accounting and one-

off measures) or distortions in the tax structure or administration processes. Simes

(2003) considered the Australian government history of use of fiscal rules in the

context of the institutional arrangements under which fiscal policy is set. The study

did not extend, however, to empirical analysis of the impact of fiscal policy on capital

formation, productivity or growth.

2.7 SUMMARY

This chapter described mainstream economic theories of the effect of government

expenditure on the macroeconomy. While neoclassical theories focus on tax

smoothing, Keynsian theories of varying complexity typically focus on impacts on the

demand side of the economy, with most attention paid to the size and sign of fiscal

multipliers. Observations of a deficit bias by elected governments provide an

empirical counterpoint to these theories. The increasing focus on supply side impacts

provides justification for the focus of this study, namely, the experience of public

capital formation in the context of fiscal policy rules adoption by Australian

governments.

Attention is then paid to institutional arrangements, and Australian governments’

budgeting practices are described, including reforms during the past twenty years.

This provides an introduction to the more detailed treatments provided in Chapters 3

56


and 4. A particular set of institutional arrangements relevant to pricing of public debt,

i.e. credit ratings agencies’ methodologies, is also presented because of its potential

importance in reducing public capital formation.

Finally, a review of the empirical literature regarding the effectiveness of fiscal policy

rules in influencing macroeconomic or fiscal out comes is presented. Studies

focusing on the effect of fiscal policy rules on macroeconomic aggregates indicate

generally positive outcomes result from fiscal policy rules adoption, but that the

outcomes vary according to the level of advancement of the economy under scrutiny,

with generally less favourable outcomes found for less advanced economies. Studies

focusing on the effect of fiscal policy rules on fiscal aggregates have less clear cut

findings however generally find that effectiveness is limited at best. Significantly,

there has been no previous study published that focuses on an empirical determination

of the impact of fiscal policy rules, as defined in this thesis, on fiscal or

macroeconomic outcomes (including public capital formation or GDP/GSP) in the

Australian context. As mentioned in Chapter 1, this thesis seeks to contribute to

overcoming this gap.

57

Chapter 3: Financial Management Frameworks and Fiscal Policy Rules

CHAPTER 3

FINANCIAL MANAGEMENT FRAMEWORKS

AND FISCAL POLICY RULES

3.1 INTRODUCTION

Previous chapters have described fiscal policy rules and the results of empirical

exploration of the fiscal and other outcomes associated with their adoption. In order

to explore fiscal outcomes in the Australian context, it is necessary to understand how

those outcomes are measured.

The purpose of this chapter is fourfold. Firstly, it outlines the range of fiscal

outcomes that may follow adoption of fiscal policy rules. The discussion, which

occurs in Section 3.2, focuses on the possible role of factors not under the control of

government, in order to distinguish intentional from unintentional fiscal outcomes

(and thus intentional from unintentional compliance).

Secondly, it describes the governmental financial reporting frameworks employed in

the Australian context, including their evolution over time and similarities and

differences between them. This occurs in Section 3.3.

Thirdly, Section 3.4 specifies and interprets the intended effect of fiscal policy rules

on key measures of fiscal outcomes. The aim is to identify a link between fiscal

policy announcements and measures of fiscal outcomes. The latter may take the form

of either a line item, or the difference between two or more line items, in financial

reports. Reporting frameworks have changed over time and, accordingly, fiscal

measures used to record outcomes have also changed over time. The impact of these

58


changes is also traced.

Fourthly, Section 3.5 outlines interdependencies between fiscal policy rules that arise

as a result of the integrated nature of financial reporting frameworks. As a result of

these interdependencies, certain fiscal policy rules may imply constraints on fiscal

measures additional to the stated measure. Section 3.6 concludes. This chapter thus

provides a basis for identification and understanding of fiscal measures which may be

constrained by adoption of fiscal policy rules.

3.2 IMPACTS OF FISCAL POLICY RULES

This section addresses the issue of penalties and their role in motivating governments

to comply with constraints imposed on fiscal measures by fiscal policy rules. The

possible role of factors outside the control of government is also addressed in order to

distinguish intentional from unintentional fiscal outcomes. The ex post or ex ante

nature of fiscal outcomes, constrained by fiscal policy rules, is then addressed.

Finally, an approach is outlined by which to recognise the potential impact of fiscal

policy rules prior to their formalisation.

3.2.1 The role of penalties

Following adoption of fiscal policy rules, a range of possible outcomes exists. Much

of the literature on fiscal policy rules assumes that governmental fiscal behaviour, and

thus fiscal outcomes, can be expected to change after the adoption of fiscal policy

rules. The adoption of fiscal policy rules is thus assigned a determinative role in

generating behavioural change. For example, Kopits and Symansky (1998) at page 1

state:

“In many countries the deterioration (in fiscal performance) recently has

59


been reversed – in a number of cases, as part of the convergence toward meeting

fiscal rules.”

The OECD (2002) at page 125 states:

“Fiscal sustainability demands structural reforms … as well as effective

fiscal rules.”

Milesi-Ferretti (1997) at page 1 states:

“Numerical fiscal rules can play a role in enhancing fiscal responsibility

insofar as they are adopted in a framework for budgetary

reform …otherwise they are … likely to be less effective….”

Tanner (2004) at page 1 notes that:

“Fiscal rules ... are often thought to inhibit counter-cyclical public

borrowing ….”

However, whether or not adoption of fiscal policy rules leads to behavioural change

may depend on the existence of penalties for non-compliance. Kopits and Symansky

(1998) include among characteristics of a best practice fiscal rule the notion of

enforceability. They note that institutional arrangements, including penalties, vary

widely internationally and that it is unclear which arrangements are most effective.

This implies that constitutional or legal statutes are needed and should be

accompanied by penalties for non-compliance, supported by some form of authority

for enforcement of those penalties. Finally, they state that clear, agreed consequences

of non-compliance should be included and that these arrangements should suit

individual countries’ institutional arrangements.

However, it is observable that actual fiscal policy rules often do not meet these ideal

60


standards. Many fiscal policy rules have no formal penalties accruing to the adoptive

government. This is the case with all fiscal policy rules adopted by Australian

governments. Most governing legislation simply does not include any provision for

penalties for non-compliance. For example, the relevant New South Wales legislation,

the General Government Debt Elimination Act 1995 (‘the Act’), contains specific

provisions for absence of penalties. In Part 6, the Act specifies that nothing in it

places on any person an obligation enforceable in a court of law or administrative

review body. The Part further provides that failure to comply with the Act does not

affect the validity of any legislation, does not prevent introduction or passage of any

future legislation and does not affect the validity of actions taken by public officials or

agencies nor does it expose any person to civil or criminal liability. Thus, only

reputational penalties may possibly result from non-compliance with fiscal policy

rules by Australian governments.

In such cases, no legislative or financial penalties are imposed on governments that do

not comply with their own rules. This contrasts with examples of fiscal policy rules

where financial or judicial penalties are, at least notionally, enforceable against

transgressing governments. Examples of the latter are highlighted by Kopits and

Symansky (1998) at page 3. They include judicial penalties for non-compliance with

yearly current balance rules applying to national and sub-national governments in

Germany; judicial penalties applying to European Union members for non-

compliance with a rule prohibiting borrowing from central banks and limiting gross

debt to 60 percent of Gross Domestic Product; and financial penalties applying to

CFA franc zone11 members for non-compliance with rules limiting borrowing from

11 The CFA Franc Zone is a common currency area spanning most of the former French colonies of West and Central Africa. CFA stands for ‘le franc des Colonies Francaises d’Afrique’

61


central banks to twenty per cent of the previous year’s revenue. Certain of these

penalties are supra-national and there is an extensive literature arguing that most have

proven to be largely ineffective.12 Thus, the international experience indicates that

penalties associated with fiscal policy rules may not be determinative of governmental

attempts to change fiscal behaviour in accordance with fiscal policy rules. A further

question also arises regarding the ability of governments to comply with fiscal policy

rules even when willing to do so.

3.2.2 Intentional and unintentional compliance

Table 3.1 provides a schematic format that summarises the possible levels and causes

of compliance and non-compliance with an example rule (in this case, a rule requiring

budgetary balance) together with an indication of the likely impact and attempts to

model outcomes. In column 1, the type of rule is presented as a starting point. An

indication of whether governments, having adopted fiscal policy rules, then comply

with them, may be gained by examination of the relevant fiscal measures over time to

reveal whether changes occur in those measures at the time of adoption. Column 2

presents the two possibilities regarding fiscal outcomes, i.e. fiscal outcomes do or do

not change at the time the fiscal policy rule is adopted.

However, a determination that fiscal outcomes changed when a fiscal policy rule was

adopted does not provide conclusive proof either of behavioural change or its absence.

This is because exogenous factors may intervene which are not under the control of

government (further discussed below). Thus, it is not possible to assign to fiscal

policy rules a causal role in changes in fiscal outcomes. For this reason, a distinction

is drawn regarding the intentionality of either compliance or non-

12 See, for example, Tanzi (2005) and Faini (2006).

62


compliance with constraints imposed on fiscal measures by adoption of a fiscal policy

rule. A circumstance of continued intentional compliance is included, reflecting a

possible motive for rules adoption of sustaining a positive trend. Column 3 lists these

‘levels of compliance’ that could occur for each possible outcome.

Column 4 then expounds the fiscal outcome that will logically follow from each level

of compliance, based on fiscal accounting frameworks. In column 5, possible

explanations for each level of compliance are presented. Column 6 then specifies the

resulting analytical outcome.

63

Cha

pter

3: F

inan

cial

Man

agem

ent F

ram

ewor

ks a

nd F

isca

l Pol

icy

Rul

es

Tab

le 3

.1 P

ossi

ble

fisca

l out

com

es fo

llow

ing

adop

tion

of fi

scal

pol

icy

rule

s

1. E

xam

ple

of

rule

2.

Pos

sibl

e ou

tcom

e 3.

Lev

els o

f co

mpl

ianc

e 4.

Fis

cal

outc

ome

5. E

xam

ple

of r

easo

n fo

r ou

tcom

e 6.

Ana

lytic

al o

utco

me

1. In

tent

iona

l co

mpl

ianc

e

1.

Bud

get i

s ba

lanc

ed.

1.G

over

nmen

ts c

ompl

ied

with

rule

s an

d ex

ogen

ous f

acto

rs d

id n

ot

inte

rven

e to

pre

vent

com

plia

nce.

1. A

naly

sis w

ill c

orre

ctly

de

tect

impa

ct.

Fisc

al

outc

omes

ch

ange

at

time

of

rule

s ad

optio

n

2. U

nint

entio

nal

com

plia

nce

2.

Bud

get i

s ba

lanc

ed.

2.G

over

nmen

ts d

id n

ot a

dopt

m

easu

res t

o en

sure

com

plia

nce

but

exog

enou

s fac

tors

inte

rven

ed to

cr

eate

com

plia

nce

(e.g

. un

antic

ipat

ed in

crea

ses i

n re

venu

e or

dec

reas

es in

exp

endi

ture

oc

curr

ed).

2. A

naly

sis w

ill in

corr

ectly

de

tect

impa

ct.

3. In

tent

iona

l non

-co

mpl

ianc

e

3.

Bud

get i

s no

t ba

lanc

ed.

3.G

over

nmen

t did

not

com

ply

with

ru

les a

nd e

xoge

nous

fact

ors d

id n

ot

inte

rven

e to

cre

ate

com

plia

nce.

3. A

naly

sis w

ill c

orre

ctly

de

tect

no

impa

ct.

4. U

nint

entio

nal

non-

com

plia

nce

4.

Bud

get i

s no

t ba

lanc

ed.

4.G

over

nmen

ts a

dopt

ed m

easu

res t

o en

sure

com

plia

nce

but e

xoge

nous

fa

ctor

s int

erve

ned.

4. A

naly

sis w

ill in

corr

ectly

de

tect

no

impa

ct.

Bud

get i

s to

be

bala

nced

Fi

scal

ou

tcom

es

do n

ot

chan

ge a

t tim

e of

ru

les

adop

tion

5. C

ontin

ued

inte

ntio

nal

com

plia

nce

5.

Bud

get i

s ba

lanc

ed.

5.G

over

nmen

t was

obs

ervi

ng im

plic

it ru

les b

ased

regi

me

prio

r to

adop

tion

of ru

les a

nd c

ontin

ued

to d

o so

.

5. A

naly

sis w

ill in

corr

ectly

de

tect

no

impa

ct.

64


The existence of automatic stabilisers (i.e. elements of expenditure and revenue not

under the control of governments and variable with the business cycle) is accepted in

mainstream economics (for example, see McTaggart, Findlay and Parkin (2006)).

This relates mainly to national governments to whom accrue most of the

responsibilities for business cycle stabilisation. However, the fiscal positions of

Australian State governments (and sub-national governments of many other nations)

are also dependent on the business cycle. In Australia, this is due to State

governments’ current reliance on: (a) cyclically-related revenue sources of federal

taxes on consumption, namely the Goods and Services Tax, which is now fully

distributed in accordance with fiscal federalism arrangements to State governments

after deduction of administration costs (A New Tax System (Commonwealth-State

Financial Arrangements) Act 1999 Schedule 2), and (b) State taxes and duties such as

stamp duties, which are strongly affected by property market conditions and which in

turn are business cycle dependent given the critical role in these markets of interest

rates and income levels (NSW Government (2002) at page ii).

The interdependence of fiscal measures is described in this chapter in Section 4.5.

This interdependence is such that exogeneity of revenue items may also affect

governmental expenditure and investment behaviour, either in the current fiscal year

(typically termed the ‘budget year’ in budget documents) or in future fiscal years.

The extent to which this may occur in the budgeted year depends on the flexibility

with which governments undertake budget monitoring practices following budget

formulation. For example, where unanticipated revenues are received, and this is

detected mid-year, a government may spend in excess of budgeted expenditure in the

knowledge that the increment to revenues will at least equal the increment to

expenditure. In these circumstances, a non-negative budgetary outcome will still be

attained and a budgetary balance rule will not be abrogated. In such cases, legislative

65


authority for expenditure, in excess of expenditure authorised by the original bill of

supply, may be obtained by passing an additional bill of unanticipated expenditure.

While a full political economy of budget determination is outside the scope of this

study, this tendency of jurisdictions to spend windfall revenues, rather than reducing

other revenue-raising avenues and maintaining expenditure unchanged, is known in

the political economy literature as the ‘flypaper effect’ (Hines and Thaler (1995),

Courant and Ribinfeld (1990) and Courant, Gramlich and Rubinfeld (1979)). To the

extent that those windfall revenues are not fully spent, better-than-budgeted financial

operating outcomes result.

Therefore, an examination of fiscal measures for evidence of change at the date of

adoption of fiscal policy rules may or may not show that changes occur in those

measures at the time of adoption of fiscal policy rules, even if governmental

behaviour changed in compliance with those rules following their adoption. At least

two possible reasons can be identified.

Firstly, if fiscal measures change at the time of adoption of fiscal policy rules, this

may be because governmental behaviour changed in intentional compliance with

those rules. Such an outcome assumes that behaviour prior to adoption of fiscal

policy rules would not produce financial outcomes in accordance with the rule, and

further, that exogenous factors do not intervene after the adoption of such rules.

Secondly, the state of revenues and expenditure, especially unplanned revenues, could

present an appearance that rules are being followed although governments have not

changed their behaviour to ensure compliance. In other words, exogenous factors

intervene. This has possibly occurred in the province of Alberta, Canada, where rules

relating to budgetary balance may have been met primarily due to receipt of

66


unexpectedly high oil revenues (Kneebone (2003)). This could be termed accidental

or unintentional compliance.

If fiscal measures did not change at the time of adoption of fiscal policy rules, three

conclusions are possible. Firstly, it is possible that governments did not observe the

rules after adoption i.e. governments did not change their behaviour and the fiscal

policy rules can be said to have had no effect on fiscal measures. This could be

described as intentional non-compliance.

Secondly, governments may have changed behaviour to comply with fiscal policy

rules, only to have exogenous factors intervene to prevent compliance. This could be

termed unintentional non-compliance.

Finally, it is possible that governments had been observing the rules prior to their

formal adoption and continued to do so after adoption. In other words, governments

had been obeying an implicit rules-based regime prior to their announcement (and,

usually, legislation) of those rules. Possible motives underlying this include testing,

where the government follows a rule for a period of time to observe its effects, and/or

their ability to obey the rule, prior to committing publicly to that rule. Behaviour

following adoption of fiscal policy rules could be characterised as (continued)

intentional compliance in this case. In such a case, a change in behaviour may be

traceable to a period prior to official adoption of fiscal policy rules13.

In addition, it is not necessarily the case that fiscal policy rules were ineffective in

13 As discussed further in Section 3.2.4, it is for this reason that a ‘bracketing’ approach is adopted. Under the bracketing approach, where budgetary or other documentation contains information sufficient to identify the commencement of such a regime, the period of adoption of fiscal policy rules is assumed to commence at that time.

67


influencing government behaviour, even if that behaviour was established prior to

formal adoption of those rules. The rules may have held government to a pattern of

behaviour from which it may otherwise have diverged. That is, governments may, in

the absence of fiscal policy rules, have chosen to reverse their previous positions.

This inability to ascribe causation, relating to fiscal behavioural change or non-change,

subsequent to adoption of fiscal policy rules, is in accordance with normal statistical

inference decision rules wherein the null hypothesis can be said to be unable to be

rejected rather than that the alternative hypothesis has been disproven. For example,

see Selvanathan, Selvanathan, Keller and Warrack (2004). Nevertheless, this inability

to specify causation between fiscal policy rules adoption and changed fiscal behaviour

complicates the analysis. For this reason, a number of analytical techniques are used,

with the advantage that multiple lenses are used to focus on the research question,

incorporating a number of different perspectives. Chapter 5 provides further

explanation.

3.2.3 What constitutes compliance – prospective or retrospective measurement

In addition, it is necessary to clarify what constitutes compliance. Kopits and

Symansky (1998) note that fiscal rules can be defined in terms of the degree of

stringency, precision and enforcement of the statutory instrument by which rules are

established and that a narrow definition would require both ex ante and ex post

compliance, subject to tangible penalties.

Compliance, whether intentional or unintentional, is necessary in order to gain the

credibility benefits, such as reduced risk premia on public debt, that are among the

usual motivations for adoption of fiscal policy rules. If an ‘announcement effect’ is

created (wherein adoption of fiscal policy rules leads to an increase in the credibility

68


of a government’s fiscal policy) prior to establishment of a record of compliance, this

beneficial effect will quickly dissipate if non-compliance is observed.

In the Australian experience, fiscal policy rules have been typically formulated as

requiring ex ante observance. That is, they are formulated in terms of budgetary

measures only. In this case, only budgeted fiscal measures need meet the constraints

imposed by fiscal policy rules and actual fiscal outcomes, measured retrospectively,

are deemed irrelevant. Hence, formulation of budgets, in which relevant fiscal

measures comply with constraints imposed by fiscal policy rules, is a necessary

condition to adherence to fiscal policy rules. A strict interpretation of fiscal policy

rules indicates that such budgetary adherence is both necessary and sufficient to

constitute adherence.

However, when fiscal rules are interpreted in the context of governmental financial

management practices (wherein governments monitor actual fiscal outcomes for

uniformity with budgeted fiscal outcomes) and in light of the argument in the previous

paragraph, it is clear that budgetary adherence alone is insufficient. A further

condition is necessary for adherence to fiscal policy rules. That condition is that

retrospective fiscal outcomes14, or the relationships between them, also comply with

those budgets. That is, fiscal policy outcomes, when measured retrospectively, must

also comply with budgeted (prospective) fiscal policy outcomes.

That all Australian governments intend compliance with their fiscal policy rules is

evidenced by the fact that each has reported annually on their ex post performance in

meeting the constraints imposed on fiscal measures by fiscal policy rules. As

previously discussed, this may relate to compliance levels 1, 2 or 5 of Table 3.1.

14 Historical fiscal outcomes are termed ‘actuals’ in budgetary documentation.

69


Intentions regarding compliance can be determined by examination of prospective

financial statements. However, it is not always possible to determine, by examination

of historical financial statements, whether the historical outcome was intentional or

unintentional. Where non-compliance has occurred, the intentionality or otherwise of

the outcome can to some extent be determined by comparison of prospective and

historical financial statements. For example, if a budgetary balance rule has not been

met, and comparison of prospective with historical financial statements reveals that

this was due to actual expenditures in excess of budgeted expenditures, of a type not

under the control of government, then it can be reasonably concluded that non-

compliance was unintentional. However, in some cases, it is not possible to

determine the intentionality of the outcome by such an examination due to the lack of

detail provided in financial statements.

Hence, in examining fiscal measures for evidence of compliance or otherwise, and

ignoring the question of intentionality, it is necessary to examine those fiscal

measures in their ‘ex post’ form. That is, the fiscal measures must be examined in

historical financial reports rather than in budgeted financial reports. Such measures

are examined in this study using Government Finance Statistics (GFS) data in Chapter

6.

Further, an awareness of when the effect of fiscal policy rules can be expected to

commence is necessary to an examination of compliance.

70


3.2.4 Bracketing

Section 3.2.2 presents a range of possible observable outcomes subsequent to

adoption of fiscal policy rules. These include intentional compliance, intentional non-

compliance, unintentional non-compliance and unintentional compliance. Where the

first occurs, and this represents a change in behaviour, a change in fiscal outcomes

may be discernible. However, where an implicit rule has been adopted and obeyed

for some time prior to its being made public, a change in fiscal outcomes may not be

discernible. In such a case, a change in behaviour may be traceable to a period prior

to official adoption of fiscal policy rules.

For example, suppose a government announced an intention to legislate a specific

fiscal policy rule in Year 1 and passed the legislation in the following year (‘Year 2’).

In such a case, it is reasonable to expect that the government’s fiscal strategy in Year

1 would reflect fiscal constraints imposed by the fiscal policy rule in that year. This is

despite the fact that legislation enacting the fiscal policy rule did not take effect until

Year 2. Such a delay could be explained by time taken to draft legislation, legal

‘sitting times’ whereby legislation must be tabled for a minimum period before the

parliamentary vote and the passage of time before the Governor gives assent.

For this reason, where budgetary documentation contains information sufficient to

identify the commencement of such a regime, for example, when announcing the

formulation of impending legislation, for purposes of analysis, the period of adoption

of fiscal policy rules can be argued to extend from this time to the first reporting

period (inclusive) for which the existence of fiscal policy rules can be identified.

This approach is termed ‘bracketing’ and is analogous to the event window adopted in

event studies methodology. The event window usually includes the event date as well

71


as a period before the event, in order to capture the effects of any pre-announcement

anticipation or information leakage (Alankar (2003) page 92).

3.3 AUSTRALIAN GOVERNMENTAL FINANCIAL REPORTING

FRAMEWORKS

Australian national and sub-national governments present financial reports in both

prospective (budgetary) and retrospective (historical) formats, and in accordance with

both generally accepted accounting principles used by the accounting profession and

national accounting principles used by economists and statisticians.

3.3.1 Budgetary Uniform Presentation Framework

The Commonwealth Government of Australia (2000) at page 1 states:

“The May 1991 Premiers’ Conference agreed to the introduction of the

Uniform Presentation Framework (UPF) in 1991. The primary objective

of the UPF is to ensure that Commonwealth, State and Territory

governments provide a common ‘core’ of financial information in their

budget papers … (and)… facilitate a better understanding of individual

governments’ budget papers and provide for more meaningful

comparisons of each government’s financial results and projections.

The format of the UPF is based on the reporting standards of the

Australian Bureau of Statistics (ABS) Government Finance Statistics (GFS)

framework. This ensures a high degree of consistency in the preparation

and presentation of financial data. In 1997 all jurisdictions agreed to a

revised Uniform Presentation Framework which simplified, rationalised

and enhanced reporting requirements, while maintaining consistency with

the ABS cash-based GFS. The revised framework also

72


previewed the shift from a cash to an accrual reporting framework.”

Australian national and sub-national governments thus have by agreement presented

prospective financial statements (contained in budgetary documentation) in

accordance with a Uniform Presentation Format (UPF) since 1991, including

prospective financial statements based on the reporting standards of the GFS (then

cash-based) framework. In 1997, governments agreed to a revised UPF containing

prospective financial statements which were, however, still formulated according to

cash-based GFS standards. In 2000, jurisdictions agreed to an amended UPF (to be

adopted by most governments from 2000-2001) containing prospective financial

statements formulated according to the newly-developed accrual-based GFS

framework.

3.3.2 Cash and accrual bases of accounting

The differences between cash and accrual-based accounting frameworks relate mainly

to temporal aspects of expense and revenue recognition. This is the subject of an

extensive literature and the interested reader may find further coverage in any first

year accounting textbook, for example, see Deegan (2005). Cash-based accounting

frameworks trace flows of cash only at the time of the flow. For example, in a cash-

based accounting framework, the purchase of an asset is recognised as an expense

equal to the total amount of the purchase price in the reporting period in which the

purchase took place. By comparison, in an accrual-based accounting framework, such

a transaction would result in recognition, during the reporting period in which the

purchase took place, of acquisition of an asset and of a depreciation (or amortisation)

expense equal only to the portion of the value of the asset consumed during the

reporting period in which the purchase took place. The remainder of the value of the

73


asset would be recognised as an expense during subsequent reporting periods as the

economic benefits encapsulated in the asset were consumed during those periods.

Similar differences apply to recognition of accruing revenues and expenses.

Adoption by Australian governments of differing bases for accounting frameworks

during the period examined in this study created problems for empirical testing

relating to the adoption of fiscal policy rules. These problems, and their resolution,

are discussed in Chapter 6.

3.3.3 Government Finance Statistics (GFS) reporting framework

GFS is a financial measurement framework under which all Australian governments

are required to provide historical financial reports to the Australian Bureau of

Statistics (ABS). It is based on the System of National Accounts 1993 and the

international statistical standard developed by the International Monetary Fund.

The ABS (2005a) at page 1 states:

“The term ‘government finance statistics’ refers to statistics that measure

the financial activities of governments and reflect the impact of those

activities on other sectors of the economy ...

The GFS system covers all activities of governments that can be measured

in money terms. The system focuses on monetary measures of transactions

and other economic flows that involve governments, and the money values

of assets and liabilities held by governments. Examples of activities

covered by GFS include government spending, lending, taxing, and

borrowing. As well, the statistics include information about the value of

government investments and debt ...

74


The Australian GFS system is based on two important international

statistical standards… the system of national accounts embodied in System

of National Accounts 1993 (SNA93), issued jointly by the United Nations

(UN), the International Monetary Fund (IMF), the Commission of

European Communities, the Organisation for Economic Co-operation and

Development (OECD) and the World Bank … (and)… the international

statistical standard for compiling GFS developed, in consultation with

member countries, by the IMF.”

Table 3.2 provides a schematic presentation of the conceptual framework underlying

the current accrual-based GFS framework which highlights the integrated nature of

stocks and flows.

75


Table 3.2 Integrated nature of GFS stocks and flows

Stocks Flows Stocks

Other Flows

Opening Balance Sheet

$

Transaction Flows

$

Revaluations $

Other Changes

in Volume

of Assets

$

Closing Balance Sheet

$

GFS Revenues (1) GFS Expenses (2)

GFS Net Worth (3) = (4 – 7 – 8)

GFS NOB = (1 – 2) =

Change in Net Worth due to transactions

Change in Net Worth

due to Revaluations

Changes in Net Worth due to Other

Changes in

Volume of

Assets

Assets (4) Non-financial assets (5)

GFS Net Lending (+)/Borrowing (-) = (1 – 2 – 5) = (6 – 7 – 8)

GFS Net Lending (+)/Borrowing (-) = (1 – 2 – 5) = (6 – 7 – 8)

Financial assets (6) Liabilities (7) Shares and other contributed capital* (8)

Note 1: Blank sectors indicate that items described in row headings are included in items described in column headings. Note 2: No items are logically classifiable to the shaded areas. Note 3: * This item is zero for the general government sector. Source: ABS (2000) page 6

GFS divides the above framework into four separate statements. Firstly, there is the

76


Operating Statement (which includes details of GFS revenues, GFS expenses15 and the net acquisition of non-financial assets). Secondly, there is the Statement of Stocks and Flows (which links the Operating Statement to the Balance Sheet, showing opening balances of assets and liabilities, flows affecting these elements during the reporting period, and their closing balances). Thirdly, there is the Balance Sheet (which shows assets, liabilities and GFS Net Worth for several years to provide an analytical time series). Fourthly and finally, there is the Cash Flow Statement (which identifies sources and applications of cash flows).

Tables 3.3 and 3.4 provide an outline of the GFS Operating Statement and Statement

of Stocks and Flows respectively. Important analytical balances derived in the

Operating Statement (Table 3.3) are GFS Net Operating Balance (equalling

transactions in revenues less transactions in expenses) and GFS Net

Lending/Borrowing (equalling Net Operating Balance less transactions in the net

acquisition of non-financial assets).

The Statement of Stocks and Flows (Table 3.4) provides the analytical balance of

GFS Net Worth (equalling assets less liabilities for the general government sector).

15 The ABS (2000) states at page 7 “In a broad conceptual sense, GFS revenues are defined as transactions that increase net worth and GFS expenses as transactions that decrease net worth”.

77


Table 3.3 Outline of GFS Operating Statement

Source: ABS (2000) page 7

Table 3.4 Outline of GFS Statement of Stocks and Flows

Note: * This item is zero for the general government sector. Source: ABS (2000) page 8

The GFS framework defines institutional sectors within government, in accordance

with the Standard Institutional Sector Classification of Australia, as comprising the

non-financial corporations sector, the financial corporations sector and the general

government sector (ABS (2005a) at page 11).

Transactions only (excludes Revaluation and Other Changes in

Volume of Assets) $

GFS Revenues Less GFS Expenses

Equals GFS Net Operating Balance Less Net acquisition of non-financial assets Equals GFS Net Lending(+)/Borrowing(-)

Stocks Flows Stocks Other Flows

Opening Balance Sheet

$

Transaction Flows

$

Revaluations $

Other Changes

in Volume of Assets

$

Closing Balance Sheet

$

Assets (1) Non-financial Assets Financial Assets Liabilities (2) Shares and other contributed capital * (3)

GFS Net Worth (1 – 2 – 3)

78


3.3.4 General purpose financial reporting

In addition, governments have provided historical general purpose financial reports

(GPFR) in accordance with generally accepted accounting principles, and with the

purpose of meeting the requirements of Australian Accounting Standard 27 (AAS 27):

Financial Reporting by Local Governments, Australian Accounting Standard 29 (AAS

29): Financial Reporting by Government Departments and Australian Accounting

Standard 31 (AAS 31): Financial Reporting by Governments, since inception of those

standards.

GPFR for governments typically comprise an income or operating statement (a

statement of flows of revenues and expenses culminating in an operating surplus or

deficit result for the reporting period), a statement of financial position or balance

sheet (a statement of stocks of assets and liabilities, with equity calculated as a residue

of assets after deduction of liabilities, and equal to net worth) and a cash flow

statement (a statement of flows of cash underlying the operating statement and

statement of financial position). Supplementary reports may also be provided.

In their simplest form16, linkages between accrual-based general purpose financial

statements are as follows. Flows of revenues and expenses reported in the operating

statement impact stocks of assets, liabilities and equity reported in the statement of

financial position. The overall impact of those flows (the Operating Result) attained

in the operating statement flows into the statement of financial position. The change

in net equity and net worth between one reporting period and the next equals the

operating result in the latter reporting period.

16 This is in the absence of abnormal or extraordinary items, which were shown ‘below the line’ (i.e. after calculation of the operating result, these items are added or deducted to attain an operating result after abnormal and extraordinary items) until 1 July 2000.

79


3.3.5 Differences between GFS and GPFR

GFS share many similarities with general purpose financial reporting in accordance

with generally accepted (accrual) accounting principles, being based on the same

information sources.

The ABS (2005a) at page 2 states:

“The main sources of information for compiling GFS in Australia are the

data systems that support the public accounts of the Commonwealth, and

each state and territory government. These accounts are largely geared

towards financial accountability and control and are in a format reflecting

accounting standards and legal and administrative imperatives in each

jurisdiction. The accounts that are the main sources of information for

compiling GFS in Australia generally comply with Australian accounting

standards for government entities.”

However, differences exist between the two approaches, arising from their different

roles and conceptual bases. The ABS (2005a) at page 2 further states:

“The accounting concepts established by these standards are generally

consistent with the statistical concepts employed in GFS. In developing

Australia’s GFS system, the ABS has worked closely with government

accountants and the Australian Accounting Standards Board and has

endeavoured to identify and document the small number of unavoidable

differences between the GFS system and accounting standards.”

The Commonwealth Government of Australia (2000) at page 5 notes:

80


“There are a number of conceptual differences between GFS and AAS31

that mainly reflect the economic focus of the former and the accounting

focus of the latter ... GFS data is broadly consistent with the data on

government transactions in the Australian National Accounts and

reflected in measures of economic activity.”

Consequently, the two approaches produce different headline measures. The

Commonwealth Government of Australia (2000) at page 5 further notes:

“… the GFS framework yields the net lending, or fiscal balance, measure,

which is important in assessing the impact of a government’s policies on

the economy. This concept is not found in AAS31 standards.”

The Commonwealth Government of Australia (2000) at page 13 states:

“The key measures in the GFS accrual framework are: GFS net operating

balance, GFS net lending (fiscal balance), cash surplus, net debt, net

worth, change in net worth, and net financial worth.”

The major difference between the two approaches lies in the treatment of asset

revaluation increments or decrements. The Queensland Government (2006) at page

179 states:

“The primary difference between GFS net operating balance and the

accounting surplus calculated reporting under Australian Accounting

Standards (AAS) is that valuation adjustments are excluded from the GFS

net operating balance.”

and

81


“The primary difference between GFS net operating balance and the

accounting surplus calculated reporting under Australian Accounting

Standards (AAS) is that valuation adjustments are excluded from the GFS

net operating balance.”

This accords with the Commonwealth Government of Australia (2000) at page 5,

statement that:

“The GFS excludes revaluation income and expenses because they are

outside the control of government.”

As a result, for example, because the impact of an asset revaluation decrement is not

included in a GFS operating result, the reporting entity may report a GFS operating

surplus, and yet also report diminution of net worth, in the same reporting period

under this measurement approach. Equally, when the impact of an asset revaluation

increment is not included in a GFS operating result, the reporting entity will report an

increase in net worth in excess of the GFS operating surplus.

According to the Commonwealth Government of Australia (2000) at page 18, the

GPFR Operating Result and GFS Net Operating Balance also reflect differing

treatment of provisions for doubtful debts, bad debts written off from provisions and

treated as capital transfers, abnormal items, distributions to owners (dividends) and

capitalised interest.

Table 3.5 shows reconciliation of GFS Net Operating Balance with the GPFR

Operating Result reported in accordance with AAS31.

82


Table 3.5 Reconciliation of GFS Net Operating Balance with Operating Surplus/Deficit

GFS Net Operating Balance minus Provisions for doubtful debts plus Bad debts written off from provisions and treated as capital

transfers plus/minus Gains/losses on assets, including derivatives plus/minus Adjustment for abnormal/extraordinary items (a) plus Distributions to owners (dividends) plus Capitalised interest Plus/minus Other adjustments (b) equals AAS 31 Operating Surplus/Deficit (a) Due to changes to Australian Accounting Standards, separate identification of abnormal items will not be required on the face of financial statements in respect of reporting periods beginning on or after 1 July 2000. (b) Calculated as a residual. May include adjustments for superannuation, coverage and unidentified differences. Source: ABS (2000) page 20

The Commonwealth Government of Australia (2000) at page 21 states that measures

of net worth also differ between GFS and GPFR:

“GFS Net Worth … (and) … AAS 31 Net Assets … are equivalent for the

large part. The GFS Net Worth measure represents total assets less

liabilities ... The AAS 31 Net Assets measure, in concept, represents Total

Net Assets (total assets less total liabilities) less outside equity interests,

though in the illustration in the standard outside equity interests are not

deducted from Total Net Assets – rather they are separately disclosed.”

Table 3.6 shows the reconciliation of GFS Net Worth with AAS 31 Net Assets.

83


Table 3.6 Reconciliation of GFS Net Worth with AAS 31 Net Assets

Source: Author’s own compilation based on ABS (2000) page 22

At the time of writing, efforts are underway to harmonise GPFR and GFS standards

(ABS (2005a) page V).

Australian governments have mostly, but not always, chosen to measure fiscal

outcomes in terms of line items, or the difference between two or more line items,

comprising part of financial reports produced in accordance with the GFS

frameworks. In a few cases, Australian governments have chosen measurement in

terms of GPFR items. This has complicated empirical investigation of fiscal

outcomes in the context of fiscal policy rules. The nature of these complications, and

action taken in response to them, is detailed in Chapter 6.

3.4 FISCAL MEASURES

The following sections proceed from the above general discussion to outline more

specific outcomes regarding fiscal policy rules, focussing in particular on outcome

measures relating to net debt, budgetary balance and net worth fiscal policy rules.

3.4.1 The net debt measure

Net debt is a measure derived from other measures presented in the GFS Statement of

GFS Net Worth

plus/minus

Adjustments (a)

equals

AAS 31 Net Assets

(a) Calculated as a residual. Includes adjustments for capitalised interest,

provision for bad debts, superannuation, coverage, valuation and unidentified differences.

84


Stocks and Flows and GFS Balance Sheet. It is provided separately in the GFS

Balance Sheet for reference purposes.

In general, net debt refers to the difference between debt owed and financial assets

owned by the government. According to the Commonwealth Government of

Australia (2000) at page 16:

“Net debt comprises the stock of selected gross financial liabilities less

financial assets. … The net debt measure is limited in that it does not

include accrued employee liabilities or outstanding claims associated

with insurance type activities, which can be substantial.”

Specifically, net debt is calculated as debt (equal to Borrowing + Deposits held +

Advances received) (for the general government sector) less financial assets (equals

Cash and deposits + Advances paid + Investments, loans and placements) (ABS

(2005a) page 54).

The measure is unchanged regardless of which reporting framework is adopted.

The Commonwealth Government of Australia (2000) at page 16 states that:

“The concept of net debt is the same under cash and accrual-based

financial reporting.”

Governments adopting net debt fiscal policy rules usually target reduced net

debt, often for each annual reporting cycle. This was the case for all of the

Australian governments examined.

3.4.2 Budgetary balance measures

Fiscal policy rules adopted by selected Australian governments are detailed in Chapter

5. Fiscal rules commonly relate to budgetary balance i.e. a requirement that flows of

85


revenue and expenditure items must equal in total. Most Australian governments

have committed to a balanced or surplus budget, however, this was sometimes subject

to the state of the business cycle. In addition, the expected level of surplus (e.g. zero

or positive), sectoral coverage (e.g. General Government or Total Public sector) and

method of measurement (e.g. GFS Net Operating Balance or GPFR Operating

Balance) varied between governments. Therefore, these budgetary balance rules must

be interpreted with an awareness of the meaning of the terms in which those rules are

formulated (the topic of the following sections of this chapter).

Basis of measurement – cash or accrual

When measured on a cash basis, budgetary balance implies that cash incomings equal

cash outgoings of all kinds (including on capital items). Such an outcome would

produce a zero result in a cash-based operating statement. When measured on an

accruals basis, budgetary balance implies that accrual revenues equal accrual

expenditures. However, due to differences between cash and accrual reporting

frameworks, the two outcomes are significantly different, with the difference relating

mainly to timing of recognition of revenues and expenses (see Section 3.3.1). In

addition, the meaning of budgetary balance in an accrual-based GFS framework

differs from the meaning of budgetary balance in an accrual-based GPFR framework,

due to different treatments of adjustments for revaluation of assets. See Section 3.3.4

for further discussion.

Basis of measurement - GPFR

The relevant fiscal measure to a budgetary balance fiscal rule, measured in GPFR

terms, is the Operating Result. That is, a government which formulated a budgetary

balance fiscal policy rule with reference to fiscal outcomes measured in GPFR terms

86


would use Operating Result as the targeted fiscal measure. For example, if that

government sought budgetary balance, the target would be expressed as a zero

Operating Result. More commonly, a positive Operating Result is sought. In these

cases, the target is expressed as ‘budgetary surplus’.

Basis of measurement – accrual-based GFS

When budgetary balance is measured in accordance with accrual-based GFS, as

previously described, two relevant fiscal measures exist, being GFS Net Operating

Balance and GFS Net Lending/Borrowing. Fiscal outcomes, relating to budgetary

balance and measured in accrual-based GFS terms, are usually measured in terms of

GFS Net Lending/Borrowing. When this line item is zero, the result is often termed

‘fiscal balance’. This is essentially a reference to cash inflows broadly equalling (‘in

balance with’) cash outflows.

Robinson (2002a) at page 294 states that fiscal balance is:

“… officially defined as equal to the operating balance minus a … ‘capital

adjustment’ ... (which)…equals general government capital expenditure minus both

depreciation and general government asset sales receipts; it is, in other words, net

investment…”

From 1999-2000 to the end of the period examined, the Australian government

committed to a target related to fiscal balance on average over the economic cycle.

Fiscal balance is described in the Commonwealth Government 1998-99 budgetary

documents as the accrual counterpart of the underlying cash balance, and consistent

with the Australian National Accounts concept of net lending. It equals the accruals

operating result less revaluations (changes in the value of assets and liabilities

87


resulting from non-transaction flows such as changes in price of foreign exchange),

plus depreciation and less capital expenditure. For this reason, fiscal balance is

described as implying general government saving equal to general government

investment in the Commonwealth Government’s 2000-01 budgetary documentation.

The stringency of a fiscal policy rule requiring fiscal balance defined in such terms

depends on the quantum of net investment. As Robinson (2000) explains, the

meaning of fiscal balance is such that where net investment is negative, this means

that operating results may be negative (to a level equal to, but not exceeding, the

negative level of net investment) without abrogating the target. However, where net

investment is positive, a fiscal balance requirement means that operating balance must

at least equal net investment. Such a requirement is thus more stringent than a

budgetary balance requirement measured in terms of GFS Net Operating Balance.

Period of measurement – annual or cyclical

The operating result may be measured in terms of the reporting period i.e. the relevant

financial year. However, there is some incidence of governments adopting a fiscal

policy rule requiring budgetary balance subject to the state of the business cycle. For

example, the Commonwealth Government in 1998-99 adopted a fiscal policy rule

targeting budgetary balance on average over the economic cycle. While the specific

meaning can vary between jurisdictions (and is often not specified), the general intent

of such a rule is usually to impose fiscal responsibility while enabling the government

to fulfil a stabilisation role.

The meaning of a rule requiring budgetary balance on average over a business cycle is

that while the rate of growth of the economy is projected to be positive (and, usually,

above some threshold level), the operating outcome (whether measured in terms of

88


GFS Net Operating Balance or GFS Net Lending/Borrowing) for the government

should be at least non-negative i.e. zero or positive. However, the rule also allows for

a negative operating result where trend growth rates are either negative or below some

threshold level, which may be the average of the known past rates of economic

growth plus anticipated rates over the remainder of the cycle. Calculation of such an

average is obviously complicated by difficulties of prediction.

Anticipating the discussion in following chapters, Table 3.7 provides rates of change

in chain volume measures of Gross Domestic Product and Gross State Product for the

period 1995-6 to 2003-04, for the Commonwealth, New South Wales, Victorian,

Queensland and Western Australian Governments. It shows that, with the exception

of Western Australia in 2001, rates of growth were positive during the entire period.

Table 3.7 Gross Domestic/State Product, Chain volume measures - percentage change from previous year

Source: ABS 2005 (b)

As growth rates have been constant during much if not all of the period examined, a

fiscal policy rule requiring budgetary balance on average over a business cycle may

be interpreted as intending that positive operating results should be achieved in each

year examined.

1995-

96 1996-

97 1997-

98 1998-

89 1999-2000

2000-01

2010-02

2002-03

Commonwealth (GDP) 4.1 3.9 4.5 5.2 4.0 1.9 3.8 3.2 New South Wales (GSP) 4.9 4.1 4.6 4.2 4.5 3.2 1.8 2.2 Victoria (GSP) 4.4 2.9 4.7 6.8 3.7 1.5 4.1 2.8 Queensland (GSP) 3.6 4.7 4.4 6.3 6.3 1.4 6.4 5.5 Western Australia (GSP) 5.0 2.5 6.0 4.0 3.0 -1.3 5.6 5.6

89


3.4.3 Net worth measures Basis of measurement – cash or accrual

Cash-based reporting frameworks provide only partial balance sheets detailing

selected financial assets and liabilities only (Commonwealth Government of Australia

(2000) page 13). Under such frameworks, the concept of net worth reduces to a

concept of net financial worth. This is equivalent to net debt when liabilities exceed

financial assets i.e. when net financial worth is negative.

Accrual-based reporting frameworks provide full balance sheets from which can be

derived a measure of net worth. However, as previously described, the concept of net

worth varies between GFS and GPFR frameworks, described in the following two

sub-sections.

Basis of measurement - GPFR

Sections 3.3.2 details the accrual-based GFS measure of GFS Net Worth and Section

3.3.4 specifies the differences between GFS Net Worth and the concept of net worth

reported within GPFR. A net worth fiscal policy rule formulated in GPFR terms will

typically target Net Assets, as was the case in the New South Wales and Victorian

jurisdictions.

Basis of measurement – accrual-based GFS

A government which formulates its net worth fiscal policy rule in GFS terms will

typically use GFS Net Worth as the relevant fiscal measure. It may thus establish a

net worth rule requiring that GFS Net Worth be at least maintained or increased, or

increased by some minimum amount. The Commonwealth, Queensland and Western

90


Australian Governments adopted the former approach.

Basis of measurement – annual or cyclical

Net worth fiscal policy rules may also target some measure of change on an average

basis over a business cycle. The preceding discussion regarding cyclical

measurement of a budgetary balance target also applies in such a case. That is, as

growth rates have been constant during most of the period examined, a fiscal policy

rule requiring increased net worth on average over a business cycle may be interpreted

as intending that increased net worth should be achieved in each year examined.

3.5 INTERDEPENDENCE OF FISCAL POLICY RULES

Interdependence of fiscal policy rules arises from the previously-described linkages

between fiscal measures and financial reports. Such linkages are most marked where

financial reports are accrual-based, as cash-based reporting frameworks do not have

the integrated nature of accrual-based frameworks.

However, interdependence of fiscal policy rules also depends on whether the fiscal

outcome, constrained by the rule, is measured in a GPFR or GFS framework, even

when the GFS methodology used is accrual-based, and hence, both are accrual-

based17. Further, when the operating outcome is measured in an accrual-based GFS

framework, whether interdependence exists between a budgetary balance rule and net

worth rule depends on the terms in which the budgetary balance outcome is measured

i.e. GFS Net Operating Balance or GFS Net Lending/ Borrowing.

This essentially reduces to consideration of whether the fiscal outcome is measured in

cash or accrual terms, regardless of the basis of the reporting framework used. This is

17 GPFR are, by definition, accrual-based.

91


because a target of positive net lending (reported under an accrual-based GFS

framework) is equivalent to a target of positive operating outcome (reported under a

cash-based GFS framework). Many budgetary balance rules are measured in cash

terms and typically target non-negative GFS Net Lending (and hence are substantively

net debt reduction rules). As previously described, when taken alone, such a

requirement may be met by a negative Net Operating Balance where Net Acquisition

of Non-Financial Assets is also negative (and at least as large as the negative Net

Operating Balance) (Robinson (2002a) page 294).

Examination of Table 3.2 indicates that such an outcome leads to diminution of net

worth in the absence of positive Changes in Net Worth due to Revaluations or

Changes in Net Worth due to Other Changes in Volume of Assets. However, if a net

worth rule requiring non-negative change to net worth also exists, this outcome is

prevented and a budgetary balance rule leads to the more intuitive outcome of non-

negative GFS Net Operating Balance. In addition, where a budgetary balance rule

(measured in accrual terms) is met and a positive GFS Net Lending outcome is

attained, as long as the combined effect of Changes in Net Worth due to Revaluations

and Changes in Net Worth due to Other Changes in Volume of Assets is non-negative

(or, if negative, does not exceed the GFS Net Lending outcome), GFS Net Worth will

increase and a net worth rule will be met. In addition, if the GFS Net Lending/

Borrowing surplus is retained partly or wholly in the form of financial assets or

applied to reduction of debt, a fiscal policy rule requiring diminution of net debt will

also be met.

92


CONCLUSION

Figures 3.1 to 3.3 inclusive diagrammatically summarise approaches whereby fiscal

outcomes may be measured by Australian governments which have adopted fiscal

policy rules. Figure 3.1 depicts alternative possible ways of measuring fiscal

outcomes constrained by adoption of net debt fiscal policy rules. Figure 3.2 addresses

budgetary balance fiscal policy rules and Figure 3.3 provides this information for net

worth fiscal policy rules.

This provides a basis for specifying fiscal measures affected by adoption of fiscal

policy rules. Chapter 4 outlines the history of fiscal regimes adopted by Australian

governments, including fiscal policy rules adopted during the 1980s and 1990s.

Adoption dates for rules-based regime are specified, including the measurement

approaches used i.e. the fiscal measures targeted. Chapter 5 outlines methods and

data used to determine, in Chapter 6, the extent to which Australian governments have

succeeded in complying with fiscal targets. In addition, Chapter 6 presents results of

investigation of whether adoption of a net worth fiscal policy rule by the Australian

national and five State governments has coincided with changed levels of public

investment.

93

Cha

pter

3: F

inan

cial

Man

agem

ent F

ram

ewor

ks a

nd F

isca

l Pol

icy

Rul

es

Figu

re 3

.1 M

easu

rem

ent o

f fis

cal o

utco

mes

in th

e co

ntex

t of a

net

deb

t fis

cal p

olic

y ru

le

T

ype

of r

ule

Fr

amew

ork

B

asis

of

mea

sure

men

t

Fi

scal

mea

sure

by

whi

ch m

easu

red

Pe

riod

of

mea

sure

men

t

Com

plia

nce

(whe

ther

in

tent

iona

l or

non-

inte

ntio

nal)

cons

titut

es

cash

acc

ount

ing

c

ash

a

nnua

l

N

et d

ebt

GPF

R

acc

rual

N

et

cas

h

D

ebt

GFS

cy

clic

al

acc

rual

no in

crea

se in

Net

Deb

t ove

r fin

anci

al y

ear

no in

crea

se in

Net

Deb

t ove

r cy

cle

(as d

efin

ed)

94

Cha

pter

3: F

inan

cial

Man

agem

ent F

ram

ewor

ks a

nd F

isca

l Pol

icy

Rul

es

Figu

re 3

.2 M

easu

rem

ent o

f fis

cal o

utco

mes

in th

e co

ntex

t of a

bud

geta

ry b

alan

ce fi

scal

pol

icy

rule

Typ

e of

rul

e

Fram

ewor

k

Bas

is o

f m

easu

rem

ent

Fi

scal

mea

sure

by

whi

ch m

easu

red

Pe

riod

of

mea

sure

men

t

Com

plia

nce

(whe

ther

in

tent

iona

l or

non-

inte

ntio

nal)

cons

titut

es

ann

ual

cas

h ac

coun

ting

c

ash

cyc

lical

B

udge

tary

GPF

R

acc

rual

c

yclic

al

bala

nce

annu

al

cas

h

c

yclic

al

GFS

a

nnua

l

a

ccru

al

Net

cas

h flo

w (i

n-

flow

s les

s out

lays

) po

sitiv

e ne

t cas

h flo

w o

ver

cycl

e (a

s def

ined

)

Ope

ratin

g Re

sult

posi

tive

Ope

ratin

g R

esul

t ov

er fi

nanc

ial y

ear

posi

tive

Ope

ratin

g Re

sult

over

cy

cle

(as d

efin

ed)

GFS

Def

icit/

Su

rplu

s

annu

al

GFS

Net

Ope

r-

atin

g Ba

lanc

e or

GFS

Net

Le

ndin

g/Bo

rr-

owin

g

posi

tive

GFS

Def

icit/

Su

rplu

s ove

r fin

anci

al y

ear

posi

tive

GFS

Def

icit/

Su

rplu

s ove

r cyc

le (a

s de

fined

) po

sitiv

e G

FS N

et O

pera

ting

Bala

nce

or G

FS N

et

Lend

ing/

Borr

owin

g ov

er

finan

cial

yea

r

posi

tive

GFS

Net

Ope

ratin

g Ba

lanc

e or

GFS

Net

Le

ndin

g/Bo

rrow

ing

over

cy

cle

(as d

efin

ed)

posi

tive

net c

ash

flow

ove

r fin

anci

al y

ear

cycl

ical

95

Cha

pter

3: F

inan

cial

Man

agem

ent F

ram

ewor

ks a

nd F

isca

l Pol

icy

Rul

es

Figu

re 3

.3 M

easu

rem

ent o

f fis

cal o

utco

mes

in th

e co

ntex

t of a

net

wor

th fi

scal

pol

icy

rule

Typ

e of

rul

e

Fram

ewor

k

Bas

is o

f m

easu

rem

ent

Fi

scal

mea

sure

by

whi

ch m

easu

red

Pe

riod

of

mea

sure

men

t

Com

plia

nce

(whe

ther

in

tent

iona

l or

non-

inte

ntio

nal)

cons

titut

es

an

nual

C

ash

acco

untin

g

c

ash

cy

clic

al

annu

al

Net

wor

th

G

PFR

accr

ual

N

et A

sset

s

cycl

ical

an

nual

cash

G

FS

c

yclic

al

ac

crua

l

GFS

Net

Wor

th

ann

ual

cy

clic

al

Net

fina

ncia

l ass

ets/

lia

bilit

ies (

if ne

gativ

e, e

qual

s Net

D

ebt)

posi

tive

chan

ge in

net

fin

anci

al a

sset

s/ li

abili

ties

over

fina

ncia

l yea

r

posi

tive

chan

ge in

Net

As

sets

ove

r fin

anci

al y

ear

posi

tive

chan

ge in

Net

As

sets

ove

r cyc

le (a

s de

fined

)

Net

fina

ncia

l ass

ets/

lia

bilit

ies (

if ne

gativ

e, e

qual

s Net

D

ebt)

posi

tive

chan

ge in

net

fin

anci

al a

sset

s/lia

bilit

ies

over

fina

ncia

l yea

r

posi

tive

chan

ge in

GFS

N

et W

orth

ove

r fin

anci

al

year

po

sitiv

e ch

ange

in G

FS

Net

Wor

th o

ver c

ycle

(as

defin

ed)

posi

tive

chan

ge in

net

fin

anci

al a

sset

s/lia

bilit

ies

over

cyc

le (a

s def

ined

)

posi

tive

chan

ge in

net

fin

anci

al a

sset

s/lia

bilit

ies

over

cyc

le (a

s def

ined

)

96

Chapter 4: Fiscal Policy Rule Adoption by Australian National and State Governments

CHAPTER 4

FISCAL POLICY RULE ADOPTION BY

AUSTRALIAN NATIONAL AND STATE GOVERNMENTS

4.1 INTRODUCTION

Fiscal policy rules are defined for this thesis as forms of agreement, either

between governments or between a government and its constituency, that are

enshrined in some lasting and authoritative form so as to remain effective beyond

the life of the adoptive government. For example, if a government adopts a fiscal

strategy targeting a particular fiscal outcome, but fails to put in place any

mechanism that causes that target to both bind and outlive itself, that

arrangement would not constitute a fiscal policy rule for purposes of this study.

Typically, fiscal policy rules are made binding on the initiating and succeeding

governments by legislation. This is in accordance with Kopits and Symansky

(1998) who state that an intention that a rule be applied on a permanent basis by

succeeding governments is essential to designation of a fiscal target, contained in

a fiscal strategy, as a fiscal policy rule (see Chapter 1).

The purpose of this chapter is to detail the fiscal strategies adopted by Australian

national and State jurisdictions and identify the dates on which they adopted

fiscal policy rules (as defined) relating to budgetary balance, debt reduction and

net worth. A general version of fiscal policy rules is interpreted (in Chapter 3) as

constraints on fiscal measures reported in accordance with either GPFR or GFS

accounting frameworks. This general application of fiscal policy rules is applied

to the specific fiscal policy rules adopted by Australian governments, allowing

identification of the precise targets adopted and the sector (general government

97


or total public sector18) to which they apply. This is the foundation for the

empirical investigation of compliance and public capital formation described in

Chapters 5 and 6.

4.2 AUSTRALIAN FISCAL POLICY RULES REGIMES

The background to Australian fiscal policy rules regimes is international,

stretching back to just after World War 2. Australian governments did not

operate in isolation, but relied on international precedent in determining their

fiscal regimes.

The Bretton Woods agreement comprised an early if indirect form of fiscal rule,

effecting more than the widely-acknowledged and evident commitment to

pegged exchange rates (Williamson (1985)). Williamson states that Bretton

Woods imposed three implicit rules on parties. The first of these is said to have

allowed (or required) adjustment of exchange rates where disequilibrium existed

(in order to achieve ‘external balance’ in the medium term). The second required

that monetary and fiscal policies be aimed at achieving full employment in the

short run (termed ‘internal balance’). The third related to maintenance of

reserves of foreign currency and/or gold as a buffer to enable the continual

pursuit of internal balance.

Kopits and Symansky (1998) trace the commencement of negative fiscal trends

in major groups of countries to the breakdown of the Bretton Woods system in

the late 1960s or 1970 and subsequent abandonment of the gold standard. While

various examples of fiscal policy rules can be found dating from the 1940s, the

1990s saw more widespread adoption of fiscal policy rules. This occurred as

numerous governments sought to correct deterioration in their fiscal positions 18 The total public sector comprises the general government and public corporations sectors.

98


during the preceding decades, having achieved only moderate success at

reversing this trend through discretionary policies. Kopits and Symansky (1998)

list the Netherlands, European Union members, New Zealand, Germany, Japan,

United States sub-national governments, Canada, Indonesia, Argentina, Chile,

Ecuador, Hungary, Peru, CFA franc zone members, Brazil, Egypt, Morocco,

Philippines and the Slovak Republic as having adopted fiscal policy rules of

various types. OECD (2002) also reports Switzerland as having adopted fiscal

policy rules. Listed by Kopits and Symansky (1998) as considering adoption of

fiscal policy rules were the United States national government19 and Costa Rica.

Also considering adoption of fiscal policy rules were Brazil, India and Nigeria

(IMF (2004)).

The Commonwealth and all State Australian Governments except the Tasmanian

Government adopted fiscal policy rules during the 1990s20. Most of these

governments adopted fiscal policy rules relating to budgetary balance, debt

reduction and net worth. However, the measurement approach, and thus targeted

fiscal measures, varied between jurisdictions. As noted in Chapter 6, outcomes

also varied between jurisdictions.

19 The proposal was defeated in 1997 (Milesi-Ferretti (1997). 20 Territory governments were excluded from the analysis due to their relatively small size in terms of population and economic activity.

99


4.2.1 Commonwealth Government

Table 4.1 presents the experience of fiscal policy rules adoption by the

Commonwealth Government.

Table 4.1 Commonwealth Government fiscal policy rules

Fiscal policy rule type

Measure adopted

Target Relevant Sector

Date of adoption

Bracketing span

Budgetary surplus

1998-99 cash-based GFS Net Operating Balance 1999-2000 to 2003-04 accrual-based GFS Net Lending/ Borrowing

Balance 1998-88 budget sector 1999-2000 to 2003-04 general government

1998-99 1996-97 to 1998-99

Debt reduction

Net Debt 1998-99 to 1999-2000 Reduction to 10 per cent GDP by 2000-01

General government

1998-99 1996-97 to 1998-99

Net worth GFS Net Worth

Improvement over the medium to long term

General government

1999-2000

N/A

The fiscal trilogy

In 1984, in response to steady growth in the size of government and a

deteriorating fiscal position, the Australian Government adopted a fiscal ‘trilogy’,

that: (a) taxes would not be increased; (b) government expenditure would not be

increased; and (c) the budget deficit would be reduced both in current terms and

as a proportion of GDP (National Commission of Audit (1996) page 3).

100


Each part of the trilogy was to apply in 1985-86 and over the life of the

Parliament, and as such, do not meet the requirement that fiscal policy rules

remain effective beyond the life of the initiating government. That is, no

effective mechanism was adopted to make the strategy binding on either the

initiating or succeeding governments.

The government continued this strategy of fiscal consolidation until the early

1990s when the recession induced recurring budgetary deficits. In 1993-94, a

medium term fiscal strategy was adopted with a target of budgetary deficit equal

to approximately one per cent of GDP by 1996-97 (National Commission of

Audit (1996) page 3). This strategy also fails to meet the criteria by which fiscal

policy rules are defined in this study, as it did not include an enforcement

mechanism to make the strategy binding on the initiating and succeeding

governments.

1996-97 Introduction of The Charter of Budget Honesty Bill and commencement

of bracketing

In 1996, the government introduced, for legislation in 1998, a Charter of Budget

Honesty Bill. The aim of the Charter was to provide for greater discipline,

transparency and accountability in fiscal policy (Commonwealth of Australia

(1996) page 3).

The Commonwealth of Australia (1996) states (at page 2) that:

“ … the adoption of the new fiscal framework through legislation

has the objective of applying the new arrangements to all future

governments by law…”.

101


For this reason, targets adopted under the Charter are considered to constitute

fiscal policy rules for purposes of this study.

The proposed Charter of Budget Honesty was included in the 1996-97 Budget

papers and it is for this reason that bracketing21 commences in this year. That is,

while fiscal policy rules were not considered to have been adopted until their

formalization in 1998-99, it is to be anticipated that changed government

behaviour, if any, could commence in this year. This is because targets adopted

for fiscal measures from this year onward are consistent with provisions of the

Charter of Budget Honesty. The Charter of Budget Honesty specifies principles

on which fiscal policy was to be based, rather than identifying fiscal measures

and specifying targets. Thus, minor changes have been made to measures and

targets during ensuing years. The exact measures and targets are described

below.

The Charter of Budget Honesty Act 1998 (in Sections 2 and 9) commits each

succeeding government to produce a medium-term fiscal strategy each year, to

specify shorter-term fiscal measures against which fiscal policy is set and

assessed and to state targets for these measures. The requirement that the

government’s fiscal strategy be based on principles of sound fiscal management

is at Section 2. These are specified (at Section 8) as including prudent risk

management (including with reference to general government debt and

contingent liability levels), ensuring that fiscal policy contributes to achieving

adequate national savings and stabilization, adopting spending and taxation

levels consistent with reasonable taxation stability and predictability, maintaining

taxation integrity and consideration of intergenerational effects.

21 Bracketing is described in Section 3.2.4.

102


In accordance with this proposed Charter of Budget Honesty, the short-term

fiscal strategy announced in the 1996-97 budget targeted reduced underlying

budget deficits as a prerequisite to a medium-term fiscal strategy of movement to

underlying budgetary balance22. The relative fiscal measure (GFS Net Operating

Balance) was stated in accordance with the cash-based GFS framework applying

at the time. The mechanism adopted to achieve this outcome was expenditure

constraint. A commitment was also made to revenue constraint, specifically, to

adopt no new taxes or increase any existing taxes (although no undertaking to

limit total taxation or other revenue was made). While the improvement in

underlying balance was expected to reduce general government net debt, aided

by application of proceeds of substantial equity asset sales, the ultimate aim of

the medium-term strategy was to increase national saving (Commonwealth of

Australia (1996) pages 1 to 5).

The objective of the medium-term fiscal strategy was unchanged in the 1997-98

budget, i.e. reduced underlying deficit in order to increase national saving and

reduce general government net debt. The target of reduced underlying budget

deficits every year was amended, to enable the stabilization role to be effectively

undertaken. The target became reduced underlying budget deficits on average

over the course of the economic cycle. Two more stringent targets were also

adopted, to achieve underlying balance by 1998-99 and maintain surplus (i.e.

positive underlying balance) beyond 1998-99 while economic growth was

22 “Underlying budget balance” is described in Commonwealth of Australia (1996) at page 2 as measuring directly the budget sector’s contribution to public sector net lending i.e. savings less investment, and as being negative if in deficit and positive if in surplus. The underlying budget outcome reflects a flow-of-funds concept.

103


‘solid’23. Commitments to expenditure and revenue constraints adopted in 1996-

97 were retained (Commonwealth of Australia (1996) pages 1 to 5).

1998-99 - Formalization of fiscal policy rules

The Charter of Budget Honesty Bill was passed into law in 1998. For this reason,

fiscal policy rules are deemed to have been adopted in 1998-99. Reflecting this,

in 1998-99, the primary objective of Commonwealth Government medium-term

fiscal strategy was to maintain underlying budgetary balance, on average, over

the course of the economic cycle. The relevant fiscal aggregate (GFS Net

Operating Balance) continued to be stated in accordance with the cash-based

GFS framework, as had been the case since 1996-97. This required an

underlying budgetary surplus in 1998-99 as well as surpluses in forward

estimates years for which estimated economic growth was sound.

A continuing objective of increasing national saving and, particularly, net debt

reduction is evidenced by the accompanying intention to halve net debt from its

level of 20 per cent of GDP in 1995-96 to 10 per cent in 2000-01. Existing

commitments to taxation and expenditure constraint were retained

(Commonwealth of Australia (1998) pages 1 to 5).

In 1999-2000, the objective of Commonwealth fiscal strategy was unchanged.

However, the primary target was redefined, under the newly-adopted accrual-

based GFS budgetary framework, as maintaining fiscal balance (Commonwealth

of Australia (1999) pages 1 to 5). As in Section 3.4.2, when the GFS Net

Lending/Borrowing measure equals zero, the result is often termed ‘fiscal

balance’. This is the accrual counterpart of the underlying cash balance, and is

23 No definition of the term ‘solid’ was provided.

104


consistent with the Australian National Accounts concept of net lending.

The focus of budgetary balance also moved from the budget sector to the general

government sector from this time. The Commonwealth of Australia (1999 b) at

page 2 stated:

“The Commonwealth budget sector consists of those departments

and agencies whose day-to-day transactions are recorded in the

Official Commonwealth Public Account. The general government

sector has a slightly wider definition than the budget sector (eg

including ABS, CSIRO) and includes resident public entities which

are mainly engaged in the production of goods and services outside

the normal market mechanism for consumption by governments and

the general public. Costs of production are mainly financed from

public tax revenues. Goods and services are provided free of charge

or at nominal charges well below costs of production.”

As in Chapter 3, the essential difference between cash and accrual bases of

accounting is the timing of recognition of expenses and revenues and the

resulting different financial reporting formats. Accrual-based accounting

frameworks provide the capacity to produce full balance sheets. The change in

accounting methodology thus provided an improved information set on the basis

of which were established five fiscal targets supplementary to the primary target.

Of these, four were unchanged. These were maintenance of budgetary surpluses

over the forward estimates period where economic growth prospects were

deemed sound, reduction of the ratio of general government net debt to GDP to

10 per cent by 2000-01, expenditure constraint and taxation constraint. The fifth

105


was improvement of the government’s net assets position over the medium- to

long-term (Commonwealth of Australia (1999) page 1). This constituted

adoption of a net worth fiscal policy rule.

In 2000-01, 2001-02, 2002-03 and 2003-04, the primary target of fiscal balance

on average over the course of the economic cycle were unchanged. The

Commonwealth of Australia ((2000) page 1, (2002) page 1 and (2003) pages 1-5

to 1-6) stated that supplementary objectives included:

“.. maintaining fiscal surpluses over the forward estimates period

while economic growth prospects remain sound, no increase in the

overall tax burden from its 1996-97 level and improving the

Commonwealth net worth position over the medium to longer term.”

The net debt and expenditure constraint aims were not included in the 2000-01

strategy statement, on the basis that each had been achieved. Nor were they

reintroduced in the remainder of the period under scrutiny (Commonwealth of

Australia (2001) page 1–3, Commonwealth of Australia (2002) page 1–5,

Commonwealth of Australia (2003) pages 1-5 to 1-8). This is a continuation of

the 1996-97 strategy, restated in accrual terms and with the addition of the net

assets target. From 2003-04, the net assets component was restated as net worth

(which is conceptually the same measure in accrual-based GFS terms).

Dating, interpretation and interdependencies of Australian Government fiscal

policy rules

As previously stated, adoption of fiscal policy rules relating to budgetary surplus

and debt reduction by the Commonwealth Government is dated to 1998-99 (the

106


first financial year following legislation of the Charter of Budget Honesty Act

1998 (Cw)). Bracketing is applied from 1996-97, when the policy was first

announced. From 1996-97 to 1998-99, the budgetary balance target (of reduced

annual underlying budget deficits) was measured in cash-based GFS terms. The

concept of net worth is not measurable in these terms. Adoption of a fiscal

policy rule relating to net worth is dated to 1999-2000, when accrual-based GFS

was adopted for reporting purposes. Bracketing is not applied to adoption of the

net worth fiscal policy rule.

From 1999-2000, the budgetary balance target was redefined as maintaining

fiscal balance, on average, over the course of the economic cycle. Robinson

(2000) argues that such a measure is inherently biased against capital

accumulation due to the interaction between GFS Net Operating Balance and Net

Investment previously described. However, as discussed in Section 3.4.4,

interdependences between fiscal measures are such that, as the Commonwealth

Government also adopted a secondary target of maintaining or increasing net

worth from 1999-2000, both requirements taken simultaneously mean that

positive budgetary balance (GFS Net Operating Balance) was also targeted from

1999-2000. Although this was expressed as an average over the economic cycle,

as pointed out in Section 3.4.2.3, consistent economic growth rates during the

period since adoption of this rule implies that the target may be interpreted on an

annual basis. The combined effect of these factors is that positive GFS Net

Operating Balance was targeted in each year.

4.2.2 New South Wales Government

Table 4.2 summarises the fiscal policy rules adopted by the New South Wales

107


Government.

Fiscal strategy 1988-89 to 1991-92

The NSW government in its 1988-89 Budget adopted a five-year, medium-term

financial strategy which remained in place until 1991-92 (National Commission

of Audit (1996) Appendix G page 8). The four goals of the strategy related to

the budget sector which comprises budget-dependent general government sector

Table 4.2 NSW Government fiscal policy rules


Measure adopted

Target Relevant sector

Date of adoption

Bracketing span

Budgetary surplus

1996-97 to 2000-01 cash-based GFS Net Operating Balance 2001-02 to 2003-04 accrual- based GFS Net Lending/ Borrowing

Surplus commen-cing 1998-99

General government

1996-97 1995-96

Debt reduction

Net Debt Reduct-ion to a ‘sustain-able level’ by 2004-05 and elimin-ation by 2020

General government

1996-97 1995-96

Net worth GPFR Net Assets

Positive rate of change (in real terms)

General government

1996-97 1995-96

108


agencies and is therefore a component of the general government sector24. The

goals were to balance the budget in order to constrain debt, expenditure

constraint (with a target of zero real growth in recurrent expenditure), increased

investment in infrastructure, and revenue reform (by maximization of non-

taxation revenue and reduction of taxation revenue) (New South Wales

Government (1990) page 14). The budgetary balance requirement was originally

expressed in terms of the NSW Government’s Consolidated Fund and changed in

1991-92 to a measure based on the GFS format which was cash-based at that

time (New South Wales Government (1991) page 13).

Fiscal strategy 1992-93

In 1992-93, the strategy was restated with an overall target of capping budget

sector net debt in real terms over the medium term. This was to be achieved by

expenditure restraint, capping of capital payments in real terms after an isolated

increase in 1992-93, taxation increases and accessing equity held in government

trading enterprises by dividend and sale (privatizations) (New South Wales

Government (1992) pages 1-6 to 1-9).


The 1993-94 budget financial strategy retains the target of capped budget sector

real net debt, but sets a target date of 1995-96 and adds a target of budgetary

24 NSW Government (2005a) at page 1-2 states “The general government sector includes all NSW Government agencies that receive Parliamentary appropriations or are regulatory in nature. Budget dependent general government sector agencies receive an appropriation from the Consolidated Fund. Non-budget dependent general government sector agencies are generally self-financing through the imposition of regulatory and user charges. Non-budget dependent agencies may also return a surplus to the Consolidated Fund.” Non-budget dependent general government sector agencies include the Registry of Births, Deaths & Marriages, Workers’ Compensation (Dust Diseases) Board and the WorkCover Authority.

109


balance thereafter (New South Wales Government (1993) page 1-5). The 1994-

95 budget restates objectives of debt containment and elimination of the

budgetary deficit (New South Wales Government (1994) page 1-3).

While the intended period of application exceeded the life of the initiating

government, no effective mechanism existed to support this and hence none of

the preceding strategies are considered to constitute fiscal policy rules as defined

for purposes of this study. Rather, adoption of rules is considered to have

occurred later, as described in the next section.

1995-96 – The General Government Debt Elimination Act (NSW) 1995 and

commencement of bracketing

The General Government Debt Elimination Bill (‘the Bill’) was tabled in June

1995 (New South Wales Government (1995) page 2). The 1995-96 budget

marked adoption by the (newly-elected) New South Wales Government of a new

financial strategy in accordance with the Bill. Following enactment in late 1995,

the Bill became the General Government Debt Elimination Act (NSW) 1995 (‘the

Act’. ) The Act at Part 2 includes short term fiscal targets of ‘sustainable’25

general government budgetary surpluses in (cash-based) GFS terms by 1998-99,

medium-term fiscal targets of reduction of general government sector net debt to

a ‘sustainable level’26 by 2004-05 and long-term fiscal targets of elimination of

general government net debt by 2020 and unfunded State sector superannuation

liabilities by 2030.

25 Sustainability with respect to this target was defined as “a level at which the budget can absorb the full impact of an economic cycle without the need for significant corrective action on the revenue or expenditure side”. 26 No definition of the term ‘sustainable’ with respect to this target was provided.

110


The Act also includes fiscal principles and states that the government should aim

to pursue policy in accordance with those principles, but that departure from

those principles is allowable if reasons are given for the departure and an

approach is outlined to enable a return to them. Fiscal principles stated at Part 3

included maintenance or increase of general government sector net worth,

taxation restraint, and expenditure restraint.

Net worth is defined in the Act at Part 4 as meaning net assets indicated in

audited financial statements annually tabled in Parliament. This is therefore a

GPFR not GFS measure. In confirmation of this, New South Wales Government

(1996) at page 1-4 measures progress in relation to this target in terms of

Consolidated Financial Statements and Public Accounts—which are both GPFR

financial statements (though only the Public Accounts are audited). These

reports differ only in their time coverage (New South Wales Government

(1997c), New South Wales Government (1997b) ). This net worth measure was

stated in real terms and is more stringent than a nominal measure.

From 1995-96 onward, fiscal policy has been set in accordance with these

principles (New South Wales Government (1996), New South Wales

Government (1997a) , New South Wales Government (1998), New South Wales

Government (1999), New South Wales Government (2000) ; New South Wales

Government (2001a)).

1996-97 – Enactment of the Bill and adoption of fiscal policy rules

The Bill received assent on 19 December 1995 (General Government Debt

Elimination Act 1995 No. 83 (NSW) page 1). For this reason, adoption of fiscal

policy rules by the New South Wales Government is dated from 1996-97, the

111


first full fiscal year following enactment of the Act. As previously stated, fiscal

policy from 1995-96 onward was set in accordance with the Act.

Until 1999-2000, the medium-term budgetary target was budgetary balance,

measured in cash-based GFS terms, on an annual basis. From 1999-2000, this

was measured over the course of a full business cycle (New South Wales

Government (2001a) page 1-20). A new focus on fully funding accruing non-

debt liabilities such as superannuation was also evident in this budget. Although

GFS had moved to an accrual basis in 1999-2000, the budgetary target was still

defined in cash terms in 1999-2000 and 2000-01. From 2001-02, the budgetary

surplus target was relabelled as net lending27 surplus. That is, the relevant fiscal

aggregate became accrual-based GFS Net Lending/ Borrowing.

2002-03 onwards - The Public Finance and Audit Amendment (Budgeting and

Financial Reporting) (NSW) Act 2002

In 2002, the NSW Government passed the Public Finance and Audit Amendment

(Budgeting and Financial Reporting) Act which amended the General

Government Debt Elimination Act 1995. The amendments aimed to update the

Act to reflect the shift from cash-based to accrual-based reporting, clarify fiscal

principles and distinguish between principles and ‘implementation instruments’28.

However specific targets remained unchanged in 2002-03 and 2003-04, as the

legislation merely reflected changes which had already been adopted in fiscal

strategies.

27 New South Wales Government (1999) at page 1-15 describes net lending as “the accrual measure most closely matching … changes in net financial liabilities”. New South Wales Government (2001a) at page 1-20 further describes net lending as “‘broadly equivalent to the former Government Finance Statistics cash result while being derived from the Cash Flow Statement.” 28 This term is used analogously to the term ‘targets’.

112


Dating, interpretation and interdependencies of New South Wales Government

fiscal policy rules

Adoption of fiscal policy rules in New South Wales, relating to budgetary

surplus, net debt and net worth, is therefore dated to 1996-97 (the first financial

year after enactment of the legislation). The targets were to be achieved

regarding budgetary surplus by 1998-99, regarding net debt by 2004-05 and

regarding net worth from 1996-97. However, changes in fiscal behaviour

(leading to achievement of these targets by those dates) in all cases were to be

undertaken from 1996-97. For this reason, adoption is dated to that year with

bracketing applied from 1995-96, when the legislation was first introduced.

However, as previously outlined, measurement of budgetary balance targets has

differed over time. From 1995-96 until 1998-99, the target was reduction in

budgetary deficits until annual positive GFS Net Operating Balance in cash-

based GFS terms could be achieved by 1998-99. (As previously discussed, GFS

Net Operating Balance in cash-based GFS terms is the broad equivalent of GFS

Net Lending/Borrowing in accrual-based GFS terms.) From 1999-2000 to 2000-

01, identical surpluses were targeted but these were now to be achieved on

average over the course of a full business cycle29. From 2001-02, positive GFS

Net Lending/Borrowing in accrual-based GFS terms was targeted on average

over the course of the business cycle.

As in the case of the Commonwealth Government fiscal policy rules, if Net

Investment is negative, GFS Net Operating Balance may also be negative

without abrogating the target of positive GFS Net Lending/Borrowing. However,

29 Again, as growth rates have been constant during the period examined, the fiscal policy rule may be interpreted as intending that positive budgetary balance should be achieved in each year.

113


as with the Commonwealth Government, this is mitigated by the New South

Wales Government’s adoption of a principle of net worth maintenance. Thus the

New South Wales Government fiscal policy rule target of positive GFS Net

Lending/Borrowing must also be interpreted as intending that positive GFS Net

Operating Balance should be achieved. Additionally, although the target was

stated as an average over the economic cycle, consistent economic growth rates

during the period implies that the target may be interpreted on an annual basis.

That is, positive GFS Net Operating Balance should be achieved in each year.

Throughout the period, net debt reduction was also targeted. Additionally, the

principle of net worth maintenance in real terms persisted during this time.

4.2.3 Victorian Government

Table 4.3 presents the history of fiscal policy rules adoption by the Victorian

Government.

114


Table 4.3 Victorian Government fiscal policy rules


Measure adopted


Date of adoption

Bracketing span

Budgetary surplus

GFPR Operating Result

Surplus of at least $100M per annum

General government

2000-01 N/A

Debt reduction

Net Debt Non-positive rate of change

General government

2000-01 N/A

Net worth GPFR Net Assets

Increase of $100M per annum

General government

2000-01 N/A


In 1991-92, in a context of mounting public debt, persistent budgetary deficits

and reduced infrastructure provision, the Victorian Government adopted a debt

management strategy targeting budgetary surplus and ultimately, public debt

reduction (National Commission of Audit (1996) page 8). This strategy was also

evident in the 1992-93 budget.


In October 1992, the (newly-elected) Victorian Government announced a three-

year medium-term fiscal strategy, based on expenditure reductions and revenue

increases, targeting budgetary balance by 1995-96 and reductions in public sector

debt. The strategy was formulated in cash-based GFS terms and applied to the

budget sector. As targets were achieved (such as budgetary balance in 1994-95),

the fiscal strategy was modified accordingly.

115


The National Commission of Audit (1996) at pages 8 to 9 states that fiscal targets

for 1995-96 to 1998-99 included:

“..achievement of, by no later than 1998-99, a sustainable current

account surplus sufficient to cover depreciation of the budget sector

capital stock without increasing the State’s overall tax level,

maintenance of budget sector investment on infrastructure … at an

average of around 1.25 per cent of GSP and application of net

proceeds of privatisation to debt reduction.”

The National Commission of Audit (1996) at page 8 further states that the

strategy also included:

“…longer term budget objectives to reduce State debt and debt

servicing ratios to levels consistent with the restoration of Victoria’s

former AAA credit rating and to bring Victoria’s overall tax effort

into closer alignment with the average of the Australian States.”

The Victorian Government adopted an accrual-based budgeting and reporting

framework in 1998-99 (Victorian Government (1998) at page 1). Accordingly,

the budget balance target was amended to maintenance of an operating surplus

sufficient to fully fund infrastructure investment (a more stringent measure)30. It

continued to apply to the budget sector.

By 1998-99, the fiscal strategy included a number of long-term budget targets,

30 Under accrual principles, this measure requires a budget surplus be attained after expenses that include additional non-cash outlays (such as accruing superannuation liabilities) not previously counted in the cash-based approach, as well as the depreciation non-cash outlay which was previously counted. The policy further required that this excess be at least equal to infrastructure investment.

116


including maintaining public sector debt at levels consistent with a triple-A credit

rating, maintaining a budget operating surplus sufficient to fully fund

infrastructure investment and bringing Victoria s tax rates into alignment with the

national average. It also included medium-term targets of reducing liabilities,

including net debt, by privatisation and maintaining budget sector investment at

around 1 per cent of GSP (Victorian Government (1998) page 1). The strategy

was stated in budget sector terms. In 1999-2000, the budget sector investment

target was reduced to a minimum 1 per cent of GSP (Victorian Government

(1999) page 1).

While some of the previously-discussed fiscal strategies included terms such as

long-term objectives and medium-term operational targets, indicating an

intention that the strategy apply beyond the life of the initiating government, no

authoritative mechanism or legislation was in place to achieve this. That is, the

strategies either were not binding on either the adoptive government or its

successors. Therefore, none of these strategies are considered to constitute fiscal

policy rules.

2000-01 - The Financial Management (Financial Responsibility) Act (Vic) 2000

and adoption of fiscal policy rules

In April 2000, the (newly-elected) Victorian Government passed the Financial

Management (Financial Responsibility) Act 2000 (the ‘Act’). Section 4 of the

Act provides that the Act binds the Crown as far as the legislative power of the

Parliament permits.

The Act amends the Financial Management Act (Vic) 1994 to include, at

Sections 23C, a requirement that the government operates in accordance with

117


several broadly worded financial management principles. These principles (at

Section 23D) include prudent management of risks (arising from general

government debt, public corporations ownership, changes in the structure of the

taxation base and, generally, management of government assets and liabilities),

pursuit of expenditure and revenue policies consistent with stable and predictable

levels of the tax burden, taxation integrity, and consideration of the inter-

generational effects of policy.

The Act further requires, at Sections 23E, 23F and 23G, that the government

provide a statement of its short and long-term financial strategy and long-term

and short-term objectives in each budget and budget update. These strategies

and objectives must be based on the previously-mentioned principles of financial

management. Further, the Act provides that specific key financial measures must

be specified and targets for those measures established. Sections 23H, I and J of

the Act provides that the government must prepare estimated financial statements,

which set out the projected financial results for the budget sector resulting from

the stated fiscal strategy, and that these statements must be based on generally

accepted accounting principles. In other words, the Victorian Government’s

fiscal policy rules are to be measured in accordance with GPFR.

In accordance with this legislation, in 2000-2001 the Victorian Government

amended the budget sector budgetary surplus target (sufficient to fully fund

budget sector infrastructure investment) to a general government sector31

operating surplus of at least $100 million in each year. (The budget surplus

measure was equivalent to the GFPR Result from Ordinary Activities (later

renamed (GPFR) Operating Surplus/Deficit), with the accrual-based GFS Net 31 While the Victorian Government (2000a) at page 14 states that the target relates to the budget sector, the Victorian Government (2002b) at page 1 states that the general government and budget sector are the same for the Victorian Government.

118


Operating Balance reported as an alternative measure (Victorian Government

(2000a) pages 15, 23 and 197). Further, the Victorian Government removed the

medium-term operational target of budget-sector investment of 1 per cent of

GSP. Instead, a $1 billion infrastructure reserve was established (funded by a

large surplus in the previous year and added to from surpluses in subsequent

years), from which to fund investment over subsequent years. However, no

specific level of investment per annum was targeted.

Finally, the Victorian Government adopted long-term taxation restraint and

general government net financial liabilities constraint policies, with a short-term

target being to maintain a triple-A credit rating (Victorian Government (2000a)

page 14). Net financial liabilities include net debt and net non-debt liabilities

such as net unfunded superannuation liabilities (Victorian Government (2000a)

page 22). As a result, this is not explicitly a net debt rule. However, the focus of

the credit rating agencies on net debt and the adoption of the AAA credit rating

as a short-term target are considered justification for interpretation of this as an

implicit net debt fiscal policy rule.

The budgetary surplus and liabilities constraint targets were unchanged in 2001-

02, 2002-03 and 2003-04 (Victorian Government (2001a), (2002a) and (2003a)).

However an additional measure by which to report budgetary surplus was added.

From 2003-04, a standardized operating surplus was also reported, equal to the

general government GPFR Operating Surplus/Deficit less the difference between

the actual and expected rates of return on financial assets. While the taxation

119


constraint objective changed to one of simplification reforms, targeting of

taxation reduction continued.

Dating of fiscal policy rule adoption by the Victorian Government

Thus adoption of fiscal policy rules relating to budgetary balance and debt

reduction by the Victorian government is dated to 2000-01 and no bracketing is

applied. Dating of adoption of fiscal policy rules relating to net worth or

investment, however, warrants further discussion.

While the Kopits and Symansky definition of a fiscal policy rule includes that a

rule is typically defined in terms of an indicator of overall fiscal performance,

inclusion of the term ‘typically’ indicates that this is not an essential component

of the definition. Further, as previously stated, between 2000-01 and 2002-03,

the Victorian Government fiscal policy rule targeted a general government sector

operating surplus of at least $100 million in each year and the operating surplus

measure adopted was the GPFR Operating Surplus/Deficit (though alternate

measures were also reported). It was previously established in Section 3.2.4 that

interdependence between fiscal measures is such that achievement of a positive

GPFR Operating Surplus/Deficit has the effect of increasing net worth in GPFR

terms. Therefore, the budgetary balance rule adopted is also substantively a net

worth rule. Thus adoption of a fiscal policy rule relating to net worth by the

Victorian government is also dated to 2000-01 (and bracketing is not applied).

4.2.4 Queensland Government

Table 4.4 details fiscal policy rules adopted by the Queensland Government.

120


Table 4.4 Queensland Government fiscal policy rules


Measure adopted


Date of adoption

Bracketing span

Budgetary surplus

Cash-based GFS Net Operating Balance 1997-98 Accrual-based GFS Net Operating Balance 1999-2000 – 2003-04

Positive General government

1999-2000

1997-98

Net debt Net Debt Negative General government

1999-2000

1996-97


Positive rate of change

Total public

1999-2000

1996-97

121


Fiscal strategy early to mid-1990s

As early as 1990-91, the Queensland Government adopted a fiscal strategy

comprising three main elements. The first element was fully funding accruing

long-term (e.g. superannuation and workers compensation) liabilities. The

second was borrowing restraint, where assets deemed ‘social’ (i.e. non-revenue

generating, such as schools and hospitals) were funded from recurrent revenues,

and borrowing was only undertaken to acquire assets capable of generating

revenue at least equal to the interest charges applicable to that debt. (This does

not equate to a budgetary balance target, even in the cash terms used at the time,

though a surplus budget in cash terms does not conflict with this requirement.)

The third element was revenue restraint, comprising a commitment not to

introduce new taxes or increase fees and charges on average at a rate exceeding

inflation (National Commission of Audit (1996) page 9, Queensland Government

(1991) pages 5 to 6).

In 1991-92, the policy of maintaining financial assets sufficient to fully meet

accruing superannuation liabilities was extended to include full actuarial funding

of motor vehicle third party insurance liabilities. The borrowing and revenue

restraints were maintained unchanged (Queensland Government (1991) pages 5

to 6). The 1992-93, 1993-94 and 1994-95 fiscal strategy showed no change from

1991-92 (Queensland Government (1992) page 4, Queensland Government

(1993) page 5, Queensland Government (1994) page 5). In 1995-96, the

objective of taxation restraint was restated as “maintaining Queensland as the

low tax State” with other strategy components unchanged (Queensland

Government (1995) pages 3 to 4).

122


1996-97- foreshadowing the Charter of Social and Fiscal Responsibility and


In 1996-97, the (newly elected) Queensland Government set its fiscal policy

parameters in accordance with a proposed charter of fiscal responsibility. The

proposed charter had objectives of maintaining competitive taxes, establishing an

efficient regulatory system, efficient service delivery and maintaining the State’s

net worth32. The net worth objective was adopted despite the usage of a cash

budgeting methodology at that time. As previously discussed in Section 3.4.3,

cash-based reporting methodologies provide insufficient information to either

estimate net worth or to estimate a budgetary outcome sufficiently inclusive of

all expenses to enable an estimate of changes in net worth resulting from

operations in a reporting period. Net worth, when measured on a cash basis,

essentially reduces to a measure of net financial worth. Thus it is considered that

this objective comprises an implicit objective of net debt reduction or

maintenance. This is supported by the borrowing constraints adopted in the

strategy.

In accordance with these objectives, the 1996-97 budget maintained ‘core fiscal

policy parameters’ of low taxes, fully funding contingent liabilities, not

borrowing for recurrent expenditure, and restricting borrowing to infrastructure

projects which could generate a revenue stream sufficient to service the debt.

Thus, the previous objectives of taxation restraint and funding liabilities were

retained unchanged. The borrowing restraint objective was changed in part. In

particular, the restriction of borrowing to acquire assets, capable of generating

revenue at least equal to the interest charges payable on that debt, was further

32 This appears to refer to the total public sector.

123


narrowed to infrastructure projects only. In addition, the ban on borrowing for

non-revenue generating assets was changed to prohibit borrowing for recurrent

expenditure (Queensland Government (1996) page 4).

These ‘core fiscal policy parameters’ were also adopted in 1997-98, and specific

fiscal strategies were added, including generalized restraints on revenue raising

(comprising both taxes and charges) and a positive budgetary outcome

(Queensland Government (1997) page 4). Thus, the budgetary target was cash-

based GFS Net Operating Balance.

In 1998-99, the ‘core fiscal policy parameters’ were also adopted. So too were

the additional elements of revenue restraint (comprising no taxation increases,

revenue concessions, a limit on increases in charges to not more than Consumer

Price Index increases), the requirement to attain budgetary surplus in cash-based

GFS terms, and maintenance or increase of total State net worth (Queensland

Government (1998) page 4). Budgetary documentation for this year contains a

commitment to develop a Charter of Social and Fiscal Responsibility. A

commitment was also given to develop the next budget on an accrual basis. Both

of these commitments were subsequently fulfilled (Queensland Government

(1998) page 4).

1999-2000 - The Charter of Social and Fiscal Responsibility and fiscal policy

rules adoption

Amendments were made to the Financial Administration and Audit Act (Qld)

1977 (the “Act”) requiring that the government prepare a Charter of Social and

Fiscal Responsibility (‘a Charter’) and table it in the Legislative Assembly

124


within 90 days. The amended Financial Administration and Audit Act (Qld)

1977 at Section 6B states that the purpose of the Charter was to:

“… state the broad social and fiscal objectives of the government

and establish a framework for assessing … performance in achieving

the objectives.”

The amended Financial Administration and Audit Act (Qld) 1977 at Section 6C

requires that the government is required to have regard, in preparing the Charter,

for fiscal policy principles including transparency, accountability and prudent

risk management. It is a further requirement of the revised Act, at Section 6D,

that the Charter include the government’s financial objectives including,

particularly, the prudent management of the State’s net worth. Amendment,

withdrawal and replacement of, the Charter is allowed however withdrawal of a

Charter is ineffective until a replacement is tabled in the Legislative Assembly

(Financial Administration and Audit Act (Qld) 1977 Section 6E ).

The first Charter, produced in August 1999, outlined the government’s fiscal

strategy elements. These were termed ‘principles’ and included: (a) taxation

restraint (maintenance of State taxation levels competitive with taxation levels

imposed by other states and territories), (b) maintenance of an overall general

government budgetary surplus, measured in (accrual-based) GFS terms, at least

equal to additional interest charges applicable to new borrowings for investment

in real assets, (c) restriction of borrowing for investment purposes only (and

where consistent with maintaining a AAA credit rating) and subject to the

requirement of maintaining an operating surplus (i.e. after inclusion of interest

expenses), (d) maintenance of financial assets sufficient to cover all accruing and

125


future liabilities of the general government sector, and (e) maintenance, and (if

possible) an increase of total public sector net worth (Queensland Government

(1999) pages 1 to 4).

Again, the strategy is basically unchanged (though with greater restriction on

debt usage). The strategy was adopted in the 1999-2000 budget which was

produced in September 1999, after legislation of the Charter (Queensland

Government (1999) page 1). For this reason, the Queensland Government’s

adoption of fiscal policy rules is dated to 1999-2000.

The fiscal strategy was unchanged in 2000-01 to 2003-04 inclusive despite

production of a new Charter in October 2001 (Queensland Government (2000)

pages 9 to 10, Queensland Government (2001) pages 4 to 5, Queensland

Government (2003) page 2, Charter of Social and Fiscal Responsibility

Queensland Government October 2001 page 5). The new Charter stated the

Queensland Government’s revised whole-of-government outcomes.

Dating of fiscal policy rules adoption by the Queensland Government

Adoption of fiscal policy rules by the Queensland government relating to

budgetary surplus, debt reduction and net worth maintenance are thus dated to

1999-2000, with bracketing applied from 1996-97 for debt reduction and net

worth fiscal policy rules (when principles comprising the Charter were first

announced and its formulation foreshadowed). Bracketing is applied from 1997-

98 for the budgetary balance fiscal policy rule.

4.2.5 South Australian Government

Table 4.5 summarises fiscal policy rules adopted by the South Australian

126


Government.

Table 4.5 South Australian government fiscal policy rules


Measure adopted


Date of adoption

Bracketing span

Budgetary surplus

Accrual-based GFS Net Lending/ Borrowing (excluding depreciation effects)

Positive (excluding depreciation)

General government

2002-03 N/A

Debt reduction

Net Debt Negative rate of growth

General government

2002-03 N/A

Net worth N/A N/A N/A N/A

1993-94 - Debt Management Strategies

The National Commission of Audit traces fiscal consolidation strategies adopted

by the South Australian Government to 1993-94. According to the National

Commission of Audit (1996) pages 9 to 10:

“In 1993, South Australia implemented a debt management strategy

(DMS) in agreement with the Commonwealth Government as part of

the Commonwealth’s special assistance package to assist in the bail-

out of the State Bank of South Australia. The main element of that

strategy was to reduce net debt to sustainable levels.”

First four year plan 1994-95 to 1997-98

This strategy was extended in 1994-95 to a medium-term strategy of four-year

plans. The first four-year plan, covering 1994-95 to 1997-98, targeted, firstly,

127


movement toward budgetary balance or surplus (through reduced expenditures

particularly salaries) at least equal to capital investment, by non-commercial

sector agencies, secondly, debt reduction in real terms and, lastly, progressive

accumulation of financial assets equal to superannuation liabilities (by 2024), in

the non-commercial sector (National Commission of Audit (1996) pages 9 to 10,

South Australian Government (1998) pages 3, 6 to 7). These targets were

measured in cash-based GFS terms (South Australian Government (1998) page

3).

Second four year plan 1998-99 to 2001-02

In 1998-99, the budget documentation included the provision of an accrual

budgeting presentation in addition to the cash-based GFS presentation and a

second four-year plan, to cover the period to 2001-02. This plan retained the

objectives for the non-commercial sector. The first of these was net debt

reduction (based on an assumption that electricity utilities sale proceeds would

be used to reduce debt and interest expenses) with the aim of achieving an AA

plus credit rating in the short term and an AAA credit rating in the medium term.

The second was elimination of unfunded superannuation liabilities (although the

target date for this was extended by ten years to 2034). The third was

maintenance of a competitive tax regime and the last was ongoing achievement

of budgetary balance (the goal of budgetary balance having been met in 1997-98)

(South Australian Government (1998) pages 6 to 8).

The budgetary balance measure continued to be measured in cash-based GFS

terms, i.e. in accrual terms equivalent to the cash-based GFS Deficit/ Surplus.

The measure was therefore GFS Net Lending/Borrowing. The South Australian

128


Government (1998) at page 7 states that the equivalent measure, in terms of

GPFR, requires:

“…the achievement of a non commercial sector operating surplus

before depreciation equal to or greater than the total capital

investment in the non commercial sector.”

and that:

“… this is equivalent to the previous Australian Bureau of Statistics

Government Finance Statistics measure adapted from a cash to an

accrual basis.”

Financial statements on a cash basis were included in budget documentation,

although accrual-based financial statements were also provided in order to meet

the government’s responsibilities under the Uniform Presentation Framework33.

Because expenditure increased in real terms, measures to increase taxation and

non-taxation revenue were introduced to achieve this aim (South Australian

Government (1998) page12). These objectives were retained in 1999-2000,

2000-01 and 2001-02 (the remainder of the four year period) (South Australian

Government (1999) pages 2.1 to 2.2, South Australian Government (2000) pages

2.2 to 2.3, South Australian Government (2001) page 2.10 to 2.11). An AA plus

credit rating was achieved in December 1999 (South Australian Government

(2000) page 2.2). In 2001-02, taxation reduction in accordance with national

taxation reform was undertaken.

All of the preceding strategies fail to meet the definition of fiscal policy rules

adopted for the purpose of this study. This is because, although the strategies

33 The Uniform Presentation Framework was described in Section 3.3.3.

129


were intended to apply beyond the term of the initiating government, they were

not formulated in terms sufficiently authoritative to ensure that outcome.

2002-03 - The Charter of Budget Honesty, Fiscal Responsibility Framework and

adoption of fiscal policy rules

In May 2002, the new South Australian Government introduced the Public

Finance and Audit (Honesty and Accountability in Government) Amendment Bill

(SA) 2002 (‘the Bill’). The Bill required preparation of a Charter of Budget

Honesty (‘the Charter’) to improve the transparency of the government’s fiscal

management by detailing fiscal principles and targets.

The 2002-03 budget was prepared in accordance with the Charter. It retained for

2002-03 the objectives of positive general government sector budgetary balance

in cash-based terms, negative growth in net debt (to reach, on average, zero net

borrowing in the long term) and accumulation of financial assets sufficient to

equal accrued superannuation liabilities by 2034. Revenue-increasing measures

were adopted to achieve these targets. In addition, a new period of measurement

was announced, targeting balanced budgets, on average, over any four-year term.

As budgetary presentation had moved to a GFS accrual basis for the general

government sector (although cash-based GFS statements were also provided for

the non commercial sector), the fiscal target for the general government sector

became accrual-based GFS Net Operating Balance at least equal to investment

(but less depreciation). The intent was to ensure zero net borrowing. This target

therefore broadly equalled accrual-based GFS Net Lending/Borrowing.

However, the operating result was defined as excluding depreciation and hence

130


remained similar to the previous cash-based target (though slightly more

stringent as accruing liabilities are included in the operating result). In addition,

for reasons discussed in Chapter 3, achievement of such a result did not

necessarily equate to maintenance of net worth.

The strategy also adopted a new focus on prudent management of financial risks

in order to attain or maintain at least an AA plus credit rating. This added risk

recognition, associated with the existence of contingent liabilities such as

guarantees, to the previous aims of net debt reduction and financial assets

accumulation. While this was not presented as an explicit net debt rule, the

South Australian Government (2002 at page 2.6) states that:

“… risks associated with State finances are managed firstly by the

measures taken as part of other fiscal principles that act to constrain

the growth in net debt and unfunded superannuation liabilities.”

Hence, it is considered that this constitutes an implicit net debt rule. Similarly to

the Queensland Government’s emphasis on credit rating retention, the focus by

the South Australian Government on maintaining an AA credit rating, and the

emphasis placed on net debt by credit ratings agencies, further supports this.

Finally, the South Australian Government (2003) at page 1.5, reports against net

debt, indicating that net debt reduction was an intended outcome of this target.

The strategy included restrictions on borrowings by public non-financial

corporations to where commercial returns could be demonstrated (South

Australian Government (2002) pages 2.3 to 2.6).

In 2003-04, the primary fiscal target was budgetary balance on average in the

general government sector, measured as positive net lending before revaluation

131


effects. That is, the target was still measured on a basis similar to cash.

Liabilities funding, risk management and restrictions on the borrowings of public

non-financial corporations continued unamended.

Dating of fiscal policy rules adoption by the South Australian government

Adoption of fiscal policy rules relating to budgetary balance and net debt

reduction by the South Australian Government is dated to 2002-03 (and no

bracketing is applied). It is not considered that the South Australian Government

adopted a fiscal policy rule relating to net worth.

4.2.6 Western Australian Government

Table 4.6 summarizes the West Australian Government’s adoption of fiscal

policy rules.

Table 4.6 West Australian Government fiscal policy rules


Measure adopted


Date of adoption

Bracketing span

Budgetary surplus

Accrual-based GFS Net Operating Balance

Positive

General government and total public sector

2000-01 1997-98

Debt reduction

N/A N/A N/A N/A N/A


Positive rate of change

Total public sector

2000-01 1997-98

Fiscal strategy 1992-93

A focus on net debt reduction by the Western Australian Government first

appeared in 1992-93. Key targets were (a) containing increases in total State net

132


debt to at least one per cent less per annum on average than economic growth, (b)

accelerated repayment of general government debt leading to reductions in real

terms in the general government net financing requirement over the following

four years, and (c) an ultimate objective of regaining an AAA credit rating in the

medium term (Western Australian Government (1992) page 16).


Following a change of government, the fiscal strategy was strengthened in 1993-

94. The strategy was extended to eliminate the overall deficit on the

Consolidated Fund by the end of the government’s term in office. (This can be

considered a partial budgetary balance requirement, as the Consolidated Fund is

typically the main but not the only fund from which monies are sourced to

finance capital acquisitions and recurrent expenditures.) This also implies a

more stringent requirement regarding growth of total State debt, from a reduction

in the rate of growth to a reduction in the level (Western Australian Government

(1993) page 3). Additionally, the Western Australian Government (1993) at

page 3 stated that it aimed:

“… over the medium term to eliminate borrowings for non-income

generating general government services.”

The aim of ultimately regaining an AAA credit rating was retained.

The 1994-95 strategy again targeted net debt reduction and partial budgetary

balance (a surplus in the Consolidated Fund), to be achieved through expenditure

restraint and despite taxation restraint (Western Australian Government (1994)

pages 1 to 2). The 1995-96 strategy retained its focus on net debt reduction and

133


taxation restraint. The strategy also extended the target of partial budgetary

balance to budgetary surplus for the total public sector (Western Australian

Government (1995) pages 1 to 4). This is the equivalent of an operating surplus

in cash-based GFS terms, for the total public sector.

The 1996-97 strategy incorporated all of the elements of the 1995-96 strategy

and extended revenue restraint, previously applied through taxation measures, to

non-taxation measures (Western Australian Government (1996) pages 1 to 4).

The Western Australian Government (1996 at page 1) states that by May 1996

(the date of delivery of the 1996-97 budget), Western Australian government

debt had been placed on a positive credit watch towards regaining AAA status.

Attaining AAA status remained a key target of the budget strategy.

Again, none of these strategies meets the definition of fiscal policy rules adopted

for the purpose of this study. This is because the strategies were not intended to

apply beyond the life of the initiating government, nor were they formulated in

terms sufficiently authoritative to ensure that outcome.

1997-98 - The Government Financial Responsibility Bill (WA) 1998 and


The 1997-98 budget papers announced the introduction of a fiscal targeting

framework to be specified in legislation. This framework was outlined in the

Government Financial Responsibility Bill (‘the Bill’). The Bill provided, at

Section 6, broad fiscal principles as a basis for formulation of future fiscal policy

including intergenerational equity, policy stability and predictability in relation to

134


spending and taxing policies and prudent risk management34. Section 10 of the

Bill provided that the government was to plan financially in a way consistent

with the principles, but that departure from the principles was allowable within

the proposed legislative framework conditional on provision of an explanation of

why the departure was necessary and a plan for reversion to compliance with the

principles.

Section 7 of the Bill specified financial elements to be calculated in accordance

with the principles of external reporting standards. Section 4 provided that both

GFS and GPFR comprise external reporting standards, however Section 8

specified that financial statements are to include matters usually addressed in

GPFR as indicated by AAS31 (in particular, any reported measure of operating

surplus or total equity).

Section 11 required that the Government release a financial strategy statement at

least annually, and that the statement set out medium-term targets in relation to

the financial elements. Similarly to other legislation, the targets were not

legislated but were to be specified and monitored in budget documentation.

For 1997-98, the targets included, firstly, maintenance or increase in net assets

(i.e. GPFR Net Assets). Secondly, debt management was targeted, measured as a

reduced ratio of net interest cost to gross own source revenue35 for the total

public sector. This was achievable either as a result of net debt reduction,

reduced interest rates on outstanding debt (including potentially increased

outstanding debt) or increased gross own source revenue. Thirdly, a target of

expenditure reduction (reduced real per capita expenditure for the Consolidated

34 Western Australian Government (1997) page 153. 35 This typically comprises fees and charges.

135


Fund on average over 1997-98 to 2000-01, to be measured on a cash basis

pending introduction of accrual based methodology) was adopted. A fourth

target was accrual operating surplus, before extraordinary items, for the whole of

government, on average over 1997-98 to 2000-01, based on generally accepted

accounting principles rather than on GFS principles (and thus measured as GPFR

Net Surplus/Deficit). A fifth target was general government operating surplus

after deduction of net advances. Termed “underlying surplus”, this was the

equivalent cash-based GFS measure to the accrual operating balance target and

therefore equal to accrual-based GFS Net Lending/Borrowing. Sixthly, risk

management was targeted, including an intention during the period 1998-99 to

2000-01 to begin to accumulate financial assets to offset against accrued

superannuation liabilities as well as a program of privatization of government–

owned businesses. The final target was operating revenue sufficient to achieve

other targets, particularly the operating and underlying surplus targets (Western

Australian Government (1997) pages 153 to 161).

The aim of achieving an AAA credit rating was maintained, noting that an AA

rating had been attained with Moodys and that the State’s credit rating had been

placed on positive watch by Standard and Poors’ for a possible upgrade to an

AAA rating (Western Australian Government (1997) page 157). The Bill

included an additional financial management principle, namely a general

requirement that employment and economic effects be considered when

formulating expenditure and taxation policies (Government Financial

Responsibility Act (WA) 2000 section 6(d)).

In 1998-99, the Western Australian Government adopted an accrual budgetary

136


framework. The 1997-98 position was reaffirmed in 1998-99 and 1999-2000,

with the exception that the expenditures target was converted to an accrual basis

from 1998-99 and was required to be met each year, rather than on average, from

1999-2000 (Western Australian Government (1999) pages 12 to 14). Each

change effected a more stringent approach.

2000-01 - The Government Financial Responsibility Act (WA) 2000 and adoption

of fiscal policy rules

The Bill received assent on 5 July 2000, becoming the Government Financial

Responsibility Act (WA) 2000 (the ‘Act’). The 2000-01 Western Australian

Government budget was formulated for the first time in accordance with accrual-

based GFS adopted for the purpose of uniform presentation36. (One result of this

was that the general government sector became the focus of discussion on

financial aggregates.) Accordingly, the budgetary balance target was restated, as

a positive accrual-based GFS Net Operating Balance for both the general

government and total public sectors. The targets of maintenance of net worth,

debt management (so as to achieve an AAA credit rating) and expenditure

restraint were maintained (Western Australian Government 2000 pages 10 to 12).

However, the relevant measure for the net worth rule became accrual-based GFS

Net Worth for the general government sector.

The 2001-02 budget retained financial targets of retaining an AAA credit rating,

maintenance or increase of net worth of the total public sector, and general

36 Refer to Section 3.3.3 for a description of the Uniform Presentation Framework.

137


government sector accrual-based GFS operating surplus37 (i.e. positive accrual-

based GFS Net Operating Balance). The objective of achieving an accrual-based

GFS operating surplus for the total public sector was abandoned. The credit

rating retention target was interpreted in terms of two financial aggregates.

These were maintaining a ratio of net debt to non-financial public sector38

revenue not above 45 percent (which still allows net debt to rise if revenue rises),

and zero or negative growth of real per capita expense for the general

government sector. In addition, a new target of maintaining tax competitiveness

(a form of taxation restraint) was adopted (Western Australian Government

(2001) pages 7 to 12). This replaced the residual approach to revenue

management apparent in the previous framework.

The strategy was unchanged in 2002-03 (Western Australian Government (2002)

pages 7 to 14) and 2003-04, with one exception. In 2003-04, the ratio of net debt

to non-financial sector revenue target rose to being not above 47 per cent (from

not above 45 per cent) (Western Australian Government (2003) pages 9 to 15).

Dating of fiscal policy rule adoption by the Western Australian government

Adoption of fiscal policy rules by the Western Australian Government relating to

net worth and budgetary balance is dated to 2000-01, and bracketing is applied

from 1997-98. In the absence of a target specifically requiring reduction of net

debt, it is not considered that the Western Australian Government adopted a

37 That this is substantially unchanged from the previous year, despite a change of governing political party in the interim, attests to the efficacy of financial framework legislation, when that legislation contains clearly enunciated targets or principles, as a means of achieving longevity of financial targets. This is the case even when those targets are not included in the legislation, and are left to be set in annual Budget documentation at the discretion of successive governments of potentially differing political persuasions. 38 The Western Australian Government (2002) at page 8 describes the non-financial public sector as comprising the general government and public non-financial corporations sectors.

138


fiscal policy rule requiring net debt reduction at any time during the period

studied.

4.2.7 Tasmanian Government

In contrast to other jurisdictions, the Tasmanian Government did not adopt fiscal

policy rules during the period examined. The reason for this conclusion is the

absence of any effective mechanism whereby fiscal strategy objectives could be

made to outlast the life of the initiating government. The various fiscal strategies

adopted during the period are described below.

First five-year fiscal strategy

The National Commission of Audit (1996) at page 10 states:

“In 1990-91, Tasmania introduced a five-year financial strategy

with the initial assistance of the Commonwealth. That strategy set a

consolidated fund NFR39 target of $40m in 1994-95.”

This strategy, to cover the period 1990-91 to 1994-95, was aimed primarily at

debt containment and stabilizing increasing debt servicing costs.

Second five-year fiscal strategy

A new five-year financial strategy was announced in the 1994-95 budget. This

included reducing general government net debt to not more than 10.5 per cent of

Gross State Product by June 2000. This was to be achieved by applying the

proceeds of asset sales and by limiting further Consolidated Fund deficits to $35

million. The strategy also included maintenance of net interest costs as a

39 NFR means net financing requirement and deficit.

139


proportion of total revenue at or below the 1994-95 level, and taxation restraint

(to a point where taxation levels were to fall below the average of all States)

(National Commission of Audit (1996) page 10). Successive Tasmanian

governments applied this framework to 1997-98 (National Commission of Audit

(1996) page 10, Tasmanian Government (1998) page 11).

Third five-year fiscal strategy

The (newly-elected) Tasmanian government introduced a new fiscal strategy in

1998-99. The 1998-99 strategy was intended to apply until and including 2003-

04 (a five-year strategy with 1998-99 representing a transitional year due to the

delivery of the budget five months after the start of the financial year in

accordance with the timing of the State election). The strategy targeted, firstly,

reduced net debt. Specifically, total State Government net debt was to be

reduced to below 20 per cent of GSP by 2003-04, general government net debt

was to be reduced to below 10 per cent of GSP by 2003-04, there was to be no

increase in net debt even if GSP increased and there was to be a reduction in the

net interest cost ratio to below 5 percent by 2003-04. Secondly, the strategy

targeted reduced net liabilities by accumulation of financial assets at least equal

to accruing superannuation liabilities from 1999-2000. Thirdly, the strategy

targeted a budgetary surplus for the total public sector (in cash-based GFS terms)

and Consolidated Fund by 1999-2000. A fourth target was use of major asset

sale proceeds to either reduce debt or acquire real assets (related to the budgetary

surplus measure). Fifthly, reduction in total real non-salary recurrent operating

costs (a partial expenditure restraint target) was to be attained. Finally, the

strategy required taxation restraint (specifically, no new taxes or increase in

140


taxation rates) (Tasmanian Government (1998) pages 10 to 15).

In 1999-2000, the operating surplus target was boosted by further specifying a

general government surplus not less than 2.5 per cent of general government

revenue on average between 1999-2000 and 2003-04. The taxation restraint

objective was expanded to include targeted tax reduction. The asset sale

proceeds, partial expenditure restraint and net debt reduction targets were

unchanged (Tasmanian Government (1999) pages 14 to 17).

The budgetary surplus, asset sales, partial expenditure restraint, taxation

reduction, net debt and liabilities reduction targets continued to apply in 2000-01

and 2001-02 (Tasmanian Government (2000) pages 17 to 22, Tasmanian

Government (2001) pages 12 to 16).

Six-year fiscal strategy

A new six-year fiscal strategy was announced in the 2002-03 Budget, on the

basis that all targets of the fiscal strategy had already been established. The new

strategy targeted, firstly, a general government sector budgetary surplus

sufficient to achieve net debt targets (on a cash-based GFS basis). Secondly,

debt and liability targets were adopted, specifically, general government sector

net debt below $450 million by June 2005, and zero by June 2008 (and general

government sector net interest costs of zero by June 2008), total public sector net

debt below $1 billion by June 2008 and accumulation of financial assets

equivalent to total public sector accrued superannuation liabilities by June 2018

and to general government total liabilities by June 2017. Thirdly, competitive

taxation levels and, finally, maintenance of investment in real terms were

targeted (Tasmanian Government (2002) pages 16 to 19).

141


These were repeated in 2003-04 (Tasmanian Government (2003) pages 11 to 15),

the year the Tasmanian Government adopted an accrual basis for budgeting40.

The cash-based operating surplus target was reinterpreted into equivalent accrual

terms as the fiscal surplus (Tasmanian Government (2003) page 17).

Dating of fiscal policy rule adoption by the Tasmanian government

Despite the announced intention that these fiscal strategies apply for a period

longer than the life of the initiating government, in the absence of any

authoritative mechanism or legislation by which to enforce this, these fiscal

strategies do not meet the definition of fiscal policy rules adopted for purposes of

this study.

4.3 SUMMARY

Table 4.7 summarizes fiscal policy rules adopted by the Australian national and

State governments.

40 The Tasmanian Government had adopted in 1996 a Financial Management Reform policy incorporating progressive adoption of accrual reporting, accrual accounting and accrual financial management. However it was not until 2003-04 that budgeting was carried out in accrual terms in accordance with the Uniform Presentation Framework, and fiscal targets accordingly established in accrual terms (Tasmanian Government December 2003).

142

Cha

pter

4: F

isca

l Pol

icy

Rul

e A

dopt

ion

by A

ustra

lian

Nat

iona

l and

Sta

te G

over

nmen

ts

Tab

le 4

.7 K

ey a

dopt

ion

and

brac

ketin

g da

tes

Net

wor

th r

ule

Bud

geta

ry b

alan

ce r

ule

Net

deb

t rul

e A

ccru

al a

ccou

ntin

g Ju

risd

ictio

n A

dopt

ion

date

B

rack

eted

fr

om

Ado

ptio

n da

te

Bra

cket

ed

from

A

dopt

ion

date

B

rack

eted

fr

om

Ado

ptio

n da

te

Bra

cket

ed

from

C

omm

onw

ealth

19

99-2

000

N/A

19

98-9

9 19

96-9

7 19

98-9

9 19

96-9

7 19

99-2

000

N/A

N

ew S

outh

Wal

es

1996

-97

1995

-96

1996

-97

1995

-96

1996

-97

1995

-96

2002

-03

1995

-96

Vic

tori

a 20

00-0

1 N

/A

2000

-01

N/A

20

00-0

1 N

/A

1998

-99

N/A

Q

ueen

slan

d 19

99-2

000

1996

-97

1999

-200

0 19

96-9

7 19

99-2

000

1996

-97

1999

-200

0 19

96-9

7 So

uth

Aus

tral

ia

N/A

N

/A

2002

-03

N/A

20

02-0

3 N

/A

N/A

N

/A

Wes

tern

A

ustr

alia

20

00-0

1 19

97-9

8 20

00-0

1 19

97-9

8 N

/A

N/A

19

98-9

9 19

97-9

8

Tas

man

ia

N/A

N

/A

N/A

N

/A

N/A

N

/A

2003

-04

N/A

143


With the exception of Tasmania, each of the jurisdictions adopted fiscal policy rules

and enshrined them in legislation. Most of these jurisdictions adopted fiscal policy

rules relating to net debt, budgetary balance and net worth, with the exception of

South Australia (net worth) and Western Australia (net debt).

There was a strong focus on cash bases of measurement of the budgetary balance

rules even after adoption of accrual-based financial reporting and budgeting

frameworks. Net worth rules were measured, necessarily, in accordance with accrual

bases of measurement though these varied between GFS and GPFR. The

measurement of net debt rules is unaffected by selection of either a cash or accrual

basis or by selection of GFS or GPFR financial reporting frameworks.

The next chapter (Chapter 5) presents methods and data to be used to analyse the

fiscal outcomes resulting from adoption of the fiscal policy rules by Australian

governments. Results are presented in Chapter 6.

144

Chapter 5: Methods and Data

CHAPTER 5

METHODS AND DATA

5.1 INTRODUCTION

The research question is “What happens to public investment when governments

adopt fiscal policy rules?” In addressing this question, both inductive and deductive

processes have been used. Most studies test an existing theoretical structure.

However, no such structure exists with respect to public capital formation. This is

due to the assumed exogeneity of government behaviour in most macro-models and

resulting lack of attention paid to this area in mainstream economics (described in

Chapter 2). Hence an inductive process is used to develop initial conclusions. These

are then subjected to rigorous examination using deductive processes including

techniques of statistical inference.

This chapter sets out the methods and data used to explore Australian governments’

fiscal outcomes in the context of fiscal policy rules. Methods used draw on

information presented in previous chapters and accepted methodologies for examining

relationships of this type.

Several techniques are used in order to strengthen the overall analysis by overlapping

examination. This mitigates the weaknesses of an individual test and permits different

perspectives.

145


5.2 FISCAL POLICY RULES AND MODIFICATION OF THE ASSUMED

EXOGENEITY OF GOVERNMENT

This section describes the process undertaken to arrive at the research question. The

process is explained with reference to inductive and deductive thought processes and

scientific method. The research question is amexamined in the context of existing

economic theory and in particular with relation to potential shortcomings of that

theory.

The significance of a finding that is consistent with identified patterns is explained. A

pattern, the likely increase in public investment as a result of adoption of a net worth

fiscal policy rule, is identified. The likely cause of that pattern is identified as

compliance with the constraint imposed on fiscal measures by adoption of fiscal

policy rules. This is based on an assumption of unchanged composition of net worth.

An initial general conclusion is developed, i.e. that governments, having adopted

fiscal policy rules, act in a constrained manner to meet the fiscal constraints self-

imposed by adoption of fiscal policy rules. If this is so, adoption of fiscal policy rules

can be expected to coincide with ex post fiscal outcomes that meet the constraint

imposed on the relevant fiscal measure. Specifically, adoption of a fiscal policy rule

that the budget will be balanced should lead to a greater incidence41 of positive

operating balance reported in historical financial statements after, compared with

before, adoption of the fiscal policy rule. Similarly, a higher incidence of reduction in

net debt should occur after adoption of a fiscal policy rule requiring that net debt will

be reduced. In the case of adoption of a fiscal policy rule requiring maintenance or

increase of net worth, and assuming unchanged composition of net worth, public

41 The inability of governments to control all factors leading to a specific fiscal outcome, as previously discussed, leads to a conclusion that total compliance, i.e. one hundred per cent incidence of positive operating balance reported in historical financial statements, is unlikely.

146


investment is more likely to be at least maintained or increased after rules adoption.

The significance of the assumption of unchanged composition of net worth is now

explained. Increased public net worth does not necessarily infer increased public

capital (real assets), as net worth is a residual measure of net assets remaining after

deduction of liabilities from gross assets. As gross assets include both real and

financial assets, net worth could increase as a result of increases in financial asset

holdings, even though no increase in real assets has occurred. Alternatively, net

worth could increase as a result of reduction in liabilities with gross assets unchanged.

However, the ceteris paribus assumption is invoked on which to base an expectation

that, where a fiscal policy rule has been adopted requiring maintenance or increase of

net worth, holdings of real assets will also be maintained or increased. That is, the

composition of net worth will remain unchanged. Thus it is considered that adoption

of a net worth fiscal policy rule indicates an intention to increase public capital. This

is considered reasonable on the basis of observed stability of government service

provision. Further, this assumption adopts a conservative position, in that most

jurisdictions adopted net debt rules as well as net worth rules, as outlined in Chapter 4.

If any change to net worth composition occurred as a result of rules adoption, the

adoption of net debt rules would tend to indicate that net worth composition might

change toward a greater proportion of financial assets, or a lesser proportion of

liabilities. If so, this would make detection of changes in public investment more,

rather than less, difficult to detect.

The proposition that public investment is likely to be at least maintained or increased

147


after net worth rules adoptionis implies an important qualification of the usual

macroeconomic assumption of exogeneity of government expenditures. It is generally

assumed that only certain elements of government expenditures are business cycle

dependent. These are related to automatic stabilisers (e.g. see McTaggart, Findlay

and Parkin (2006) page 534)). The remainder of government expenditure is therefore

considered to be independent of levels of economic activity, regardless of the level of

government considered i.e. a national or sub-national government. Automatic

stabilisers are generally seen to affect only national governments. This is because it is

generally the national level of government in a federal system that bears responsibility

for transfer payments and collects income-dependent taxes. However, in Australia,

there is a strong correlation between State government revenues and the business

cycle. Adoption of certain fiscal policy rules may in part act as a mechanism of

transmission between levels of economic activity and governmental outlays.

The supposed fiscal policy rule transmission mechanism is as follows. There are

typically long lead times in amending revenue-raising capabilities, such as by

changing taxes or tax rates, for all levels of government. Thus, at least temporarily,

the state of the business cycle is likely to exert significant positive influence on a

government’s total revenues. Suppose a government has self-imposed a constraint on

its budgetary balance by adoption of a fiscal policy rule requiring a positive operating

outcome. That government’s revenues will be determined by the state of the business

cycle. The interdependence of fiscal measures is such that other fiscal measures are

impacted. If the government seeks to obey the fiscal constraint imposed by its

budgetary balance fiscal policy rule, it must restrict its recurrent expenditures,

including transfer payments, to less than approximately the amount of its revenues42.

42 The exact relationship between revenues and allowable expenditures in the context of a budgetary balance fiscal policy rule is dependent on the financial reporting framework used. Chapter 3 provides

148


This may have caused Australian State governments’ expenditures to fluctuate with

the business cycle in a manner closer to that of the Commonwealth government than

might otherwise have been the case.

Further, if the government in question has also adopted a fiscal policy rule requiring

reduction of net debt, this restricts its investment capability by limiting financing

sources to the amount of its operating surplus. Finally, if that government has also

adopted a fiscal policy rule requiring increase in net worth, the upper limit to its

possible increase in net worth is the amount of its budgetary surplus.

However, as previously discussed, increases in net worth do not necessarily lead to

increases in public investment. Increases in net worth can be achieved by increasing

financial assets rather than by increasing capital formation. The focus of this study is

public capital formation and, specifically, on whether public capital formation

changed following adoption of a net worth fiscal policy rule for any Australian

government.

These tentative conclusions comprise a theory, parts of which can be tested using

deductive techniques. The deductive process is characterised as comprising three

steps; firstly, development of a hypothesis, secondly, testing of that hypothesis using

empirical data and, finally, conclusion regarding support for, or rejection of, the

hypothesis (Samuel and Gupta (1993)). Samuel and Gupta (1993) further state that

deductive thinking is closely allied with scientific method. Scientific method, often

also characterised as a hypothetico-deductive approach, involves seven steps. These

are defining a problem, planning a research design, planning the sampling process,

collecting data, analysing data, arriving at conclusions and reporting the results.

information on the specific financial elements used to estimate operating balance and net worth measures in the two financial reporting frameworks used by Australian governments.

149


The following sections describe methods of analysis and data used to investigate in

part the tentative theory presented above. Fiscal outcomes are examined using

techniques of statistical inference. This allows specific conclusions tempered by a

specified level of statistical uncertainty.

5.3 ANALYTICAL TECHNIQUES

5.3.1 Overview

Firstly, the relevant fiscal measures of operating balance, net worth and net debt are

examined to determine to what extent compliance with adopted fiscal policy rules,

whether intentional or otherwise, occurred. Compliance can be ascertained by

comparison of targets, identified in Chapters 3 and 4, with relevant ex post fiscal

outcomes. The usefulness of establishing compliance is that it provides a basis for

interpretation of findings of other analyses. For example, suppose a finding is made

that public capital formation increased after adoption of net worth fiscal policy rules.

Awareness of whether the government under examination had, or had not, complied

with the constraint imposed by adoption of the net worth fiscal policy rule enables a

richer interpretation of that finding than if the finding were made without that

awareness.

After establishing the extent of compliance, an examination is made of whether the

rate of public capital formation changed after adoption of net worth fiscal policy rules.

Public capital formation is chosen for examination in light of emerging evidence that

public capital stocks have important impacts on private production (see Gramlich

(1994), Aschauer (2000) and Otto and Voss (1994 and 1996)) and the subsequent

issue of whether fiscal policy rules have affected productivity and growth through

their effect on public assets accumulation.

150


A selection of statistical techniques is used, motivated by a desire to develop a

number of perspectives on the topic of the study. These multiple perspectives are

intended to avert a narrowness of focus and provide broad coverage of the subject of

interest, namely, public investment by governments which have adopted net worth

fiscal policy rules. In addition, each test has a range of strengths and weaknesses and

use of a number of techniques enables overlapping examination (or triangularisation)

as a means of strengthening the analysis.

An initial visual examination is carried out to determine patterns and trends. This is

then supplemented by more rigorous examination of public capital formation in the

context of net worth fiscal policy rules. Examination is carried out by use of

statistical inference and econometric techniques.

An initial scrutiny of the experience of all jurisdictions is carried out using two non-

parametric tests and cross-sectional time series data. The first technique is a test of

independence to ascertain whether occurrence of positive rates of growth of public

capital formation is influenced by adoption of net worth fiscal policy rules.

Essentially, if the test rejects the hypothesis of independence, this would be one

possible indication that adoption of fiscal policy rules was associated with some

change in the way public capital formation proceeded. The second technique is a test

of differences between population proportions. The purpose of this test is to assess

whether adoption of net worth fiscal policy rules was associated with a change in the

likelihood that public capital formation would increase in any given year. For

example, if fiscal policy rules encouraged governments to favour growth in public

investment more than before, this could result in the latter variable experiencing

positive growth more frequently than previously. Specifically, the test seeks to

determine whether there is a significant difference between the proportion of positive

151


instances of growth of public capital formation before and after adoption of a net

worth fiscal policy rule.

Regression analysis is used to ascertain the degree to which variations in economic

activity and fiscal policy rules adoption contribute to variations in public capital

formation. The data are then examined for trends that might indicate that increases in

public capital formation are explicable by either the passing of time or by

governments increasing nominal outlays over time. A finding that a time-based trend

is significant would indicate a potential lack of relevance of fiscal policy rules

adoption to changes in public capital formation. Alternatively, a finding that the

passing of time is not a significant predictor of changes in public capital formation

would tend to support findings in the previous analysis.

Finally, an event study methodology is employed to assess the impact of adoption of a

net worth fiscal policy rule on public capital formation. Use of event studies

methodology provides another perspective on the phenomenon under investigation,

with the advantage that its validity is based on different statistical assumptions.

Further, analysis can be carried out on each jurisdiction for each year using this

approach.

152


5.3.2 Model specification and rationale

Non-parametric techniques

A test of independence of attributes is presented to ascertain whether occurrence of

positive rates of growth of public capital43 formation is associated with adoption of a

net worth fiscal policy rule. This analysis provides an initial albeit simple lens by

which to view public capital formation over time. The analysis is carried out in real,

per capita terms in order to test the capacity of the relevant governments to maintain

service levels, based on observed rivalry of many government services and assumed

constant capital productivity and composition of service provision.

The only technical requirement for this simple, non-parametric test is that the

minimum number of observations in any cell be 5. Sufficient data exist to meet this

requirement when all jurisdictions are aggregated. Accordingly, pooled cross-

sectional time series data are used. This introduces a possible weakness of the

analysis in that clarity of performance between jurisdictions is lost but the earlier,

simpler analyses and the later analyses on individual jurisdictions mitigate this loss by

offering insights into such individual performances.

The null hypothesis is that the attributes are independent, i.e. occurrence of positive

growth of public GFCF is independent of existence of a net worth fiscal policy rule.

The alternative hypothesis is that the attributes are dependent, i.e. occurrence of

positive growth of public GFCF may depend on existence of a net worth fiscal policy

rule. The test statistic is Chi-Square distributed for 1 degree of freedom.

43 Rates of growth of public capital formation are calculated and a binary series constructed with a value of 1 where the rate of growth was non-negative and 0 where negative. This is undertaken using a Laspeyres transformation in preference to using a log transformation as the Laspeyres transformation produces a binary series, whereas logs do not.

153


A test of differences between populations is employed to ascertain whether a different

proportion of non-negative instances of growth of public capital formation occurs

before and after adoption of a net worth fiscal policy rule. Again, the data are cross-

sectional time series.

The two possible outcomes of each trial are that the rate of growth is negative or the

rate of growth is positive. The two populations compared are the recorded growth

measures of GFCF measured on a real per capita basis before and after adoption of net

worth fiscal policy rules. The population parameter of interest is the population

proportion, i.e. the proportion of trials for which the outcome is a success, where

success is defined as “the rate of growth is positive.”44 The sample proportion is the

sample statistic used as an estimator of the population proportion.

The null hypothesis is that the proportion of instances of positive rates of growth is

equal between before and after adoption of a fiscal policy rule for each jurisdiction.

The alternative hypothesis is that the proportion of instances of positive rates of

growth after adoption of a net worth fiscal policy rule is greater than the proportion of

instances of positive rates of growth prior to adoption of a net worth fiscal policy rule.

The test statistic is the standardised normal random variable z of which the critical

value is 1.96 given a level of significance of .05 and a one-tailed (right-tailed) test.

This investigation is similar in nature to the previous and it is to be expected that the

two will provide similar results. Similarly to the test of independence, this technique

views the phenomenon under consideration as qualitative data. This technique differs

from the previous technique in that its validity is based on an assumed normal

distribution of the population proportion. However, this test yields a result equivalent

44 Again, rates of growth of GFCF are calculated and a binary series constructed using the Laspeyres method with a value of 1 where the rate of growth is non-negative and 0 where it is negative.

154


to that obtained with the Chi-squared test of independence, and the square of the z

value obtained for this test is equal to the Chi-square value obtained for the latter test

(Sheskin (2004) page 511).

Use of two such similar methodologies is justified because using different tests

overcomes in part issues associated with the nature of the data. In addition, use of

non-parametric tests does not confront the issues of the stationarity of time series data

which can act to reduce the suitability of use of parametric tests. Equally, use of

subsequent parametric methodology provides a check and further substantiation of the

findings of the non-parametric tests.

Multivariate linear regressions

More sophisticated analysis employing conventional ordinary least squares

methodology is then carried out to ascertain the degree to which, firstly, variations in

economic activity and, secondly, passage of time, both coupled with fiscal policy

rules adoption, explains variations in public capital formation.

Investigation of the impact of variations in economic activity is undertaken due to the

close relationship between economic activity and government revenues and the

posited relationship between government revenues and government outlays. For the

Commonwealth Government, there is a broadly accepted correlation between the

government’s revenue base and economic activity, related to the concept of automatic

stabilisers (e.g. see McTaggart, Findlay and Parkin (2006) page 534). State

governments’ revenue bases in turn are largely determined by the size of

Commonwealth Government revenues as well as by the size of their own revenue

155


base. Most Commonwealth Government revenues were derived from income taxes in

the period prior to the adoption of the Goods and Service Tax on 1 July 2000, and

from a combination of income taxes and goods and services taxes in the period

thereafter. These revenues form the basis for calculation of grants to State

governments, which comprise a significant proportion of State governments’ revenues.

States’ own revenues include payroll taxes and stamp duties and a strong correlation

typically exists between the business cycle and these revenues (see New South Wales

Government (1994) page 1-5).

Additionally, for both Commonwealth and State governments, inclusion of this

explanatory variable can be justified in terms of a notional capital-output ratio, where

each level of output requires a certain amount of capital, some of which is public

capital.

The focus is on rates of growth of public capital formation (or investment) and thus a

log-linear model is used. Intercept and slope dummy variables are employed so as to

enable separate identification of changes, if any, in the level and rate of change of

growth in public capital formation at the time of adoption of net worth fiscal policy

rules. Each jurisdiction is investigated separately, for the first time enabling

observations to be made regarding the impact of fiscal policy rules on behaviour at the

jurisdictional level.

Investigation is then carried out using a trend variable to ascertain the extent, if any, to

which changes in the rate of public capital formation are attributable to the passage of

time. This is investigated in order to mitigate the possibility of spurious findings in

earlier analyses and to further investigate the general observation that the ratio of

GFCF to GDP/GSP declines over time for all jurisdictions.

156


The rationale is that government investment may include a component that is time

dependent due to political economy issues related to budgetary size (see Alesina and

Perotti (1995), Tabellini and Alesina (1991) and Drazen and Grilli (1993))). This is

based on an observation that governments frequently announce increased expenditures

as an indication of policy validity. This would indicate that public capital formation

in any year was more influenced by public capital formation in the preceding year

than by other factors such as existence of fiscal policy rules. Such occurrences could

be explainable by factors underlying the deficit bias described in Chapter 2 (assuming

constancy of outlays components). This analysis essentially provides a check on the

previous regression analysis and investigates whether significant relationships found

in those analyses are merely due to such issues of political economy.

Analytical deficiencies include problems of serial correlation in time series data.

More important, however, is the difficulty of assigning precise relationships to various

causal factors that may influence observed variations in government fiscal behaviour.

In the context of multiple possible causes of variations in public capital formation, the

analysis in this study cannot be said to indicate conclusively the impacts of fiscal

policy rules adoption. This is because, as previously detailed, various other influences

(such as adoption of fiscal policy rules relating to other fiscal aggregates and adoption

of accrual accounting methodologies) exist simultaneously. As a result, test results

may be (in part or whole) the result of these influences rather than (primarily or

exclusively) the result of adoption of a net worth fiscal policy rule.

Event study

Finally, an event study methodology45 is employed to measure the impact of adoption

45 MacKinlay (1997) reviews, summarises and provides a concise history of event study methodology.

157


of a net worth fiscal policy rule on public capital formation. Use of event studies

methodology provides another perspective on the phenomenon under investigation.

This additional perspective has the advantage that its validity is based on different

statistical assumptions. In addition, the technique is not dependent on use of dummy

variables enabling use of a greater part of the information set provided by quantitative

data.

The methodology borrows heavily from MacKinlay (1997). An event study measures

the impact of a specific event (e.g. adoption of a net worth fiscal policy rule, as

defined in Chapter 2) on outcomes (e.g. growth of public capital formation, as

described in Section 6.2). This provides a less aggregated means of analysis than

previous analyses, in that it uses all the information contained in quantitative data and

is carried out for individual jurisdictions and individual years. It also provides an

alternative lens in that it does not use qualitative variables and thus analysis over the

post-event window also is not limited in this manner. Thus a particular strength of the

event study methodology, when used with a non-interregnum data series (the concept

of an interregnum is defined at Section 5.4.4), is that the methodology enables

identification of individual years for which abnormal levels of GFCF may have

occurred.

The flow of analysis of an event study may be generalised to comprise the following

steps. Firstly, it is necessary to define the event of interest and outcome to be studied.

Secondly, the period over which the study will be carried out must be identified. The

third step is identification of the selection criteria for the inclusion of a given entity in

the study, noting any potential biases which may have been introduced through the

sample selection. Fourthly, an appraisal must be carried out of the event’s impact,

158


including selecting a method of measuring normal and abnormal outcomes46. Finally,

the testing framework for abnormal outcomes must be designed, including defining

the null hypothesis and determining any techniques to be used for aggregating

abnormal outcomes for the entity under consideration.

Event studies have been criticised by Alankar (2003). In the context of finance,

Alankar (2003) argues that multicollinearity exists between stock and time-specific

dummy variables that are components of an event study. This multicollinearity is

argued to prevent estimation of time-specific fixed effects, which measure abnormal

returns and represent investor responses to an event. Further, additional collinearity

between the explanatory variables and the time dummy variable is argued to prevent

identification of the betas of the pricing model used. Alankar (2003) indicates that

aggregation of time series into a single portfolio series, setting time effects to zero for

dates far from the event and estimation over these dates only, is the technique

typically used to overcome this problem. This portfolio approach is argued to have

become the default methodology in event studies due to its intuitive aspects and

simplicity. This estimation procedure is further criticised on the basis that it utilises

incomplete information. Further it is argued that the problems of this approach are

magnified as the number of non-zero time-effect coefficients increase. However, as

the event study undertaken in this study is carried out on a jurisdiction-by-jurisdiction

basis, and hence no aggregation is necessary, this criticism does not appear to apply to

this study.

Event studies were originally developed to test the hypothesis that the stock market

46 These terms are defined later in this section. Following MacKinlay (1997), the abnormal return is the actual ex post fiscal outcome over the event window minus the normal fiscal outcome over the event window.

159


was efficient (Bhagat and Romano (2002)). The methodology did this by evaluating

the extent to which publicly available information became embedded in stock prices

over time. Immediate, total absorption would mean that an investor was unable to

earn abnormal profits by trading on the information after its release. This would be

consistent with an efficient market hypothesis, when efficiency is defined in this

restrictive manner. However, a broadening of use of the methodology resulted as

evidence accumulated that the stock market was efficient. Event studies became used

for analysis of the impact of any event which created abnormal movement of a

designated variable, where a link between the event and the variable was theorised.

MacKinlay (1997) states that event studies can be applied to a range of firm specific

and economy wide events such as mergers and acquisitions, earnings announcements,

debt or equity issues and announcements of macro-economic variables such as a trade

deficit. In the context of such a general application of the methodology, the question

of statistical power is particularly relevant.

The statistical power of an event study diminishes as the sample size decreases

(Bhagat and Romano (2002)). In this study, where every jurisdiction is separately

tested, the sample size is one. As a result, the statistical power is considered to be low.

This is because the variability of abnormal returns with a sample size of one is higher

than otherwise. Additionally, the announcement-period return of a lone sample

component may plausibly be affected by other information unrelated to the event

under study. Difficulties may arise in determining the effects of this information

separately from the effects of the event.

Furthermore, the power of the methodology decreases as the length of the event

window increases (Bhagat and Romano (2002)). In this study, the length of the event

160


window varies between jurisdictions. Where the event window for one jurisdiction is

longer than for another, the statistical power of the methodology is accordingly

reduced. This is acknowledged with respect to results for the Queensland and Western

Australian jurisdictions.

Ways of increasing the statistical power of the methodology include (MacKinlay

(1997)):

”… increasing the sample size, shortening the event window or …

developing more specific predictions to test.”

However, in this study, it is considered that the advantages of obtaining insights for

individual jurisdictions outweigh the possible advantages attainable by aggregation.

In addition, information obtained from publicly available data clearly indicate the

length of the event window. Finally, the specific nature of constraints, imposed on

fiscal measures by adoption of fiscal policy rules, limits capacity for developing more

specific predictions to test.

5.4 THE DATA

This section specifies data sources and indicates techniques used to address

difficulties of data availability. Crucially, these techniques include recasting cash-

based data series in equivalent accrual terms. The appropriate use of nominal or real

values and absolute or per capita levels is examined. The case for an extension of the

bracketing period is examined.

5.4.1 Fiscal outcomes

GFS Net Operating Balance, GFS Net Worth and GFS Net Debt data are sourced from

ABS Catalogue 5512.0 Government Finance Statistics 2004-05. GPFR Operating

161


Surplus/Deficit and Net Assets data for the Victorian Government are sourced from

Financial Reports for the State of Victoria for each of years 2000-01 to 2003-04.

GPFR Net Assets data for the New South Wales Government are sourced from New

South Wales Budget Sector Consolidated Financial Statements July-December 1996,

Public Accounts of the New South Wales Public Sector 1996-97, New South Wales

Financial Statements for General Government for the twelve months ended 30 June

1999 and New South Wales Report on State Finances 2000-01 to 2003-04. The data

are expressed on an historical cost basis which is consistent with nominal terms.

GPFR Net Assets data for the New South Wales Government have been deflated by

application of an implicit price deflator obtained from the ABS Catalogue 5206.0

Australian National Accounts: National Income, Expenditure and Product (Table 53

State and Local, General Government GFCF). Use of this deflator is based on an

implicit assumption that the rate of price change for State government assets is

equivalent to the rate of price change for local government GFCF.

Findings based on this analysis have to be interpreted in light of the extensive range of

methodological difficulties that attach to valuation and deflation of most elements of

balance sheets, particularly public sector balance sheets. In addition, the deflation of

balance sheets is complicated by the use of differing asset revaluation practices

between jurisdictions.

As discussed in Chapter 3, GFS are derived from jurisdictional accrual accounting

data. These data are based on historical cost. However, fixed (mostly physical) assets

are revalued periodically in accordance with accounting standards in an attempt to

align asset values with current prices and thereby address this issue. Therefore,

reported GFS values for public capital suffer from the same difficulties of

162


measurement as do GPFR measures of public assets.

5.4.2 Gross fixed capital formation

Public gross fixed capital formation (GFCF) data for the general government sector of

the Commonwealth Government is sourced from the ABS Catalogue 5206.0

Australian National Accounts: National Income, Expenditure and Product, Dec 2005,

for 1984-85 to 2003-04 inclusive. These original data are chain volume measures.

Public GFCF data for the general government sector of the New South Wales,

Victoria, Queensland and Western Australian governments are sourced from the ABS

Catalogue 5220.0 Australian National Accounts: State Accounts, for 1984-85 to 2003-

04 inclusive. The data are original and presented in nominal terms.

Fiscal policy rules are typically formulated in nominal terms only. An example is that

net worth should increase. A literal interpretation of this is that the nominal value of

net worth should increase in each year reported47. In other words, the relevant fiscal

measure is not targeted to increase in per capita or real terms48. However,

measurement in real terms has the obvious advantage that deflation of nominal

measures removes the effects of price changes and allows a closer focus on the level

of economic services, encapsulated in public assets, provided by governments over

time.

Similarly, fiscal outcomes reported in response to fiscal policy rules typically are

reported in absolute terms only. However, if one adopts an assumption of constant

47 This is unless the rule is formulated in terms of averages over a business cycle. However, as argued in Section 4.4, the consistent growth experience during the period over which rules have applied renders such a formulation equivalent to a requirement that nominal net worth should increase in each year. 48 There is one exception to this. The New South Wales Government has adopted target of maintaining and, if possible, increasing its total net worth in real terms.

163


levels of capital productivity49, and if the intent of those rules was to achieve

sustainability of service provision, maintenance of net worth per capita in real terms is

necessary to sustain service provision over time. Use of a per capita measure has its

basis in the observable fact that many core government services, such as education

and health care, are rivalrous in nature. This means that, if public capital productivity

is assumed to be stable, no economies of scale and thus constant returns to scale occur

and thus capital input requirements for these services increase proportionately with

population. (This analysis, of course, is less appropriate when the composition of

service provision changes, as may occur for example as a result of significant

demographic change.)

Therefore, a fiscal policy rule requiring that net worth be maintained may reasonably

be interpreted as meaning that net worth will be maintained in constant, per capita

terms. For this reason, public capital formation in constant, per capita terms is

investigated.

GFCF State government data are also converted into constant terms by application of

implicit price deflators obtained from the ABS Catalogue 5206.0 Australian National

Accounts: National Income, Expenditure and Product (Table 53 State and Local,

General Government GFCF). This again assumes that the rate of price change for

State government GFCF is equivalent to the rate of price change for local government

GFCF. These data are also calculated in per capita terms using estimated resident

population data from the ABS Catalogue 3105.0 Australian Historical Population

49 Relaxation of this assumption to account for increased productivity of technologically more advanced capital inputs is beyond the scope of this paper.

164


Statistics 2006.

General government sector and government corporations together comprise the total

public sector hence use of general government sector data excludes capital formation

by government corporations. This approach has the disadvantage that it fails to

recognise when governments use dividends from government corporations to achieve

desired operating outcomes for the general government sector. Dividends from

government owned corporations are reported as revenue in general government sector

operating statements. Thus a general government sector operating surplus may be

attained by payment of a dividend from a government owned corporation to a general

government sector entity. A general government sector operating surplus can occur at

the same time as a reduction of total public sector net worth by this mechanism.

Such a reduction of total public sector net worth may reduce the ability of government

to produce services overall. For example, electricity and water services are

commonly provided by government owned corporations – with important

implications for private production. These services by government owned

corporations are generally provided on a user-pays basis. However, it is in terms of

the general government sector that the majority of fiscal policy rules are formulated.

Therefore, exploration of general government data most closely accords with the

intended coverage of fiscal policy rules. In addition, this approach is adopted to more

closely consider investment for provision of public or merit goods, which are mostly

provided by the general government sector.

The jurisdictions included are the Commonwealth, New South Wales, Victorian,

Queensland and Western Australian Governments on the basis that these governments

adopted net worth fiscal policy rules. The South Australian and Tasmanian

165


Governments are not included as it has been determined that these governments did

not adopt net worth fiscal policy rules (see Section 4.3). State jurisdictions are

considered due to the preponderance of public capital formation by State governments

compared with the Commonwealth Government. The Commonwealth jurisdiction is

also considered in order to determine whether behavioural effects, if any, differ

between levels of government within the Australian federal system (see Section 3.2).

GFCF data for the period between 1984-85 and 1997-98 were reported on a cash basis

and so excluded Assets acquired below fair value, Assets donated and Assets acquired

under finance lease), which are included in the accruals-based data reported for 1998-

99 to 2003-04. With the exception of Assets acquired under finance lease in New

South Wales, these components were subtracted from the latter series in order to

provide a single series covering the entire period on a consistent basis for use in this

study. Assets acquired under finance lease were retained in the New South Wales

data as this component is highly significant, comprising over twenty per cent of the

total. In no other jurisdiction are Assets acquired under finance lease material to the

total.

The data therefore should not be interpreted as indicative of levels but do provide a

totally consistent basis for calculating rates of growth, in the assumed absence of a

change in rates of usage of finance leases and donations as financing sources.

5.4.3 Gross Domestic Product/Gross State Product

Annual GDP data are sourced from the ABS Catalogue 5204 Australian National

Accounts: National Income, Expenditure and Product 2005.

Annual GSP data from 1999-2000 to 2003-04 inclusive are sourced from the ABS

166


Catalogue 5220.0 Australian National Accounts: State Accounts 2004-05 (Table 1).

Both data sets are original and in chain volume terms. Annual GSP data from 1984-

85 to 1999-2000 inclusive are backwards extrapolations of the chain volume terms,

calculated in accordance with unpublished original data sourced from the Australian

Bureau of Statistics (at average 1998-90 prices), in accordance with the ABS’ advised

splicing methodology50. These data are also calculated in per capita terms using

estimated resident population data from the ABS Catalogue 3105.0 Australian

Historical Population Statistics 2006.

The following section describes adjustment of the data series to take account of the

effect of a transitional period during which governmental behaviour may have been

influenced by the impending adoption of fiscal policy rules.

5.4.34 Issues relating to the bracketed period

Section 3.2.4 introduced the term ‘bracketing’ as a means of identifying a transition

period during which governmental fiscal behaviour may have been influenced by the

relevant policy rules before their formal adoption. This transitional period was

identified from public documents in the manner described in Chapter 4. Using this

concept, the following analyses are carried out on data series that omit data points for

either the bracketed period (including the date of adoption of rules) or an extension of

the bracketed period, termed an ‘interregnum’. The intent is to remove ‘noise’ from

the data and obtain a clearer picture of fiscal outcomes before and after the

transitional period.

Table 5.1 summarises the relevant dates. In each case, the bracketing approach is

adopted for the interregnum and used in subsequent analysis.

50 Verbal communication with relevant ABS officers.

167


Table 5.1 Dates of adoption of fiscal policy rules and interregnum

The ratio of GFCF (in chain volume measures and on a per capita basis) to GDP/GSP

(also real per capita) is plotted for each jurisdiction. The ratios are visually examined

for pattern emergence around the time of adoption of fiscal policy rules relating to net

worth, in Figures 5.1 to 5.5 inclusive. This ratio is useful as an indication of public

investment intensity which is used as a proxy for public capital stock productive

capacity. Any change in the ratio is indicative of changes in the relative emphasis

adopted by a government in its decision-making related to outlays.

Information thus gained is used to either confirm use of the bracketing period as the

interregnum or to provide an alternate expanded period for use as the interregnum.

This approach is taken because of the possibility that governments may have adopted

changed behaviour at times different from times at which information (relating to

adoption of fiscal policy rules) emerged in documents in the public domain such as

budget papers. Exercise of judgement in data selection in this way is analogous to the

approach adopted in event analyses methodology, where an event window is

determined on a similar basis.

Jurisdiction FPR adoption date

Bracketing span

Interregnum span

Commonwealth 1999-2000 1999-2000 1999-2000 NSW 1996-97 1995-96 to

1996-97 1995-96 to 1996-97

VIC 2000-01 2000-01 2000-01 QLD 1999-2000 1996-97 to

1999-2000 1996-97 to 1999-2000

WA 2000-01 1997-98 to 2000-01

1997-98 to 2000-01

168


Figure 5.1 Ratio of GFCF to GDP, Commonwealth Government

0.0025

0.0030

0.0035

0.0040

0.0045

0.0050

0.0055

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

Source: ABS Catalogue 5206.0 National Income, Expenditure and Product Dec 2005 Note: GFCF and GDP/GSP are expressed in real per capita terms.

Figure 5.1 suggests a break in the ratio of GFCF to GDP for the Commonwealth

government at the date of adoption of fiscal policy rules in 1999-2000 and supports

the application of bracketing. The interregnum thus comprises 1999-2000 and one

data point is omitted from the series, reflecting a short transition period during which

governmental behaviour underwent change with respect to public capital formation.

169


Figure 5.2 Ratio of GFCF to GSP, New South Wales Government

0

0.002

0.004

0.006

0.008

0.01

0.012

0.014

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03


In Chapter 4, it was determined that, while the New South Wales Government

formally adopted fiscal policy rules from 1996-97, the transition period may have

started as early as 1995-96.

While the visual analysis of Figure 5.2 suggests a break at this point, this may be

spurious due to the large negative value for 1995-96. This is due to a large negative

value of nominal GFCF in 1995-96 of -$5,236M which further investigation reveals

resulted from transfer of regional road assets to local councils of $6,179M and asset

devaluations totalling $2,742M which were partly offset by acquisitions of $3,685M

(NSW Government (1997) page 1-6).

This potentially leads to a problem for analysis of the New South Wales

Government’s experience (see Section 5.7). In the absence of additional information,

the bracketing approach is adopted and the interregnum covers the period 1995-96 to

170


1996-97. Two data points are therefore omitted from the series.

Figure 5.3 Ratio of GFCF to GSP, Victorian Government

0.007

0.008

0.009

0.010

0.011

0.012

0.013

1984-85

1986-87

1988-89

1990-91

1992-93

1994-95

1996-97

1998-99

2000-01

2002-03


In Section 4.3.3, it was determined that the Victorian Government adopted a fiscal

policy rule relating to net worth in 2000-01 and no evidence for bracketing was found.

Visual analysis of the Victorian data (in Figure 5.3) indicates a general downward

trend to 1992-93 followed by a fairly regular shark-tooth pattern prior to adoption of a

net worth fiscal policy rule. The figure does not provide further information to that

obtained from public domain documents. In the absence of conflicting information

regarding the commencement of bracketing, the interregnum thus comprises 2000-01

only and one data point is omitted from the series. Similarly to the case of the

Commonwealth Government, this suggests a short transitional period to adoption of

fiscal policy rules.

171


Figure 5.4 Ratio of GFCF to GSP, Queensland Government

0.011

0.013

0.015

0.017

0.019

0.021

0.023

0.025

1984-85

1986-87

1988-89

1990-91

1992-93

1994-95

1996-97

1998-99

2000-01

2002-03


In Section 4.3.4, the Queensland Government’s adoption of a net worth fiscal policy

rule was dated to 1999-2000 and bracketed from 1996-97. Visual analysis of Figure

5.4 indicates a break at 1999-2000 and, in the absence of conflicting information

regarding the commencement of bracketing, the interregnum thus comprises 1996-97

to 1999-2000 inclusive. Four data points are omitted from the series. The transitional

period is therefore longer than that for the Commonwealth, New South Wales and

Victorian Governments and equal longest with the Western Australian Government.

172


Figure 5.5 Ratio of GFCF to GSP, Western Australian Government

0.008

0.010

0.012

0.014

0.016

0.018

0.020

1984-85

1986-87

1988-89

1990-91

1992-93

1994-95

1996-97

1998-99

2000-01

2002-03


As described in Section 4.3.6, the Western Australian Government adopted a net

worth fiscal policy rule in 2000-01 and bracketing was applied from 1997-98. Visual

analysis of Figure 5.5 does not conflict with adoption of these dates for the

interregnum. Four data points are omitted from the series. The transition period for

Western Australia is thus equal longest, with Queensland, of all jurisdictions.

The bracketing period, identified from public pronouncements, is adopted as the

interregnum in all cases. Data points for the interregnum period are removed in

interregnum models. The intent is to remove observations that reflect partial effects

of fiscal policy rules adoption during this transitional period. This should enable

clearer identification, using the techniques described in the previous section, of any

differences in behaviour both before and after fiscal policy rules adoption.

173


5.5 SUMMARY

This chapter explains the methodologies by which the fundamental research question

of this study is investigated. The research question is defined as “What happens to

public investment when governments adopt fiscal policy rules?” and a suite of

empirical methods for testing associated hypotheses are constructed and described.

The extent of compliance with constraints on fiscal measures, imposed by fiscal

policy rules relating to budgetary balance, net debt and net worth, is to be determined.

Statistical techniques, by which estimates are to be determined of relationships

between public investment, economic activity, adoption of fiscal policy rules and the

passing of time, were then described. These include non-parametric tests of

independence and of differences between population proportions, and parametric

estimation using multivariate linear regressions and an event study. Use of the

selected techniques is rationalised on the basis that this allows development of broad

perspectives and triangularised examination of public investment by governments

which have adopted net worth fiscal policy rules.

Key data sets of public gross fixed capital formation and economic activity were

described, including transformations that have been applied to overcome issues of

suitability of the data for use in the techniques employed. Selection of data points for

omission in accordance with the concept of bracketing and an interregnum period,

indicative of a transitional period, were discussed.

174

Chapter 6: Compliance and Public Capital Formation

CHAPTER 6

COMPLIANCE AND PUBLIC CAPITAL FORMATION

6.1 INTRODUCTION

This chapter explores the experience of Australian governments’ fiscal outcomes in

the context of fiscal policy rules, using methods and data described in Chapter 5. This

enables scrutiny of whether governments complied with their own fiscal policy rules

and if fiscal policy rules have affected governmental fiscal behaviour.

This analysis is an attempt to correct the lack of theoretical and empirical coverage of

government’s role in the economy and to aid in the development of future studies.

The non-parametric methods adopted explore two issues. The first is whether

adoption of fiscal policy rules was associated with some change in the way public

capital formation proceeded. The second is whether adoption of net worth fiscal

policy rules was associated with a change in the likelihood that public capital

formation would increase in any given year. The parametric methods used broadly

enable findings from three perspectives. The link between economic activity and

public capital formation is investigated. A check is performed to determine whether

this is attributable merely to the passing of time. Finally, the impact of adoption of a

net worth fiscal policy rule on public capital formation is explored. Findings in regard

to these aspects of public investment provide insights, that are not possible given the

current lack of economic treatment, regarding government outlays.

In particular, findings indicating that public investment has increased since adoption

of net worth fiscal policy rules would represent an important modification of the usual

assumption of exogeneity of governmental outlays other than automatic stabiliser

175


expenditures. This may well be applicable to all governments, not just the Australian

governments examined here. The extent of external usefulness of this study depends

primarily on the reporting methodologies adopted by different governments, but its

validity rests on a methodology for marshalling complex sets of government data.

External validity will be greatest with respect to governments which adopt similar

reporting methodologies to GFS or GPFS.

The remainder of the chapter is organised as follows. The initial step is to determine

the extent to which governments were successful in meeting constraints imposed by

adoption of fiscal policy rules, by comparing fiscal targets with ex-post fiscal

outcomes (in Section 6.2). Operating balance, net worth and net debt data are

presented in a simple format. However, as discussed in Section 3.2.2, due to

uncontrollable factors impacting governmental fiscal outcomes, this does not

necessarily provide an indication of the intention, or otherwise, of governments to

meet constraints.

Attention is then focussed on whether an association exists between adoption of fiscal

policy rules and changes in the public capital formation process in any of the

jurisdictions (in Sections 6.3 to 6.7). Public capital formation is examined both before

and after adoption of a net worth fiscal policy rule (by the national and four State

governments) using techniques of analysis that are as simple and as robust as possible

given the complexities of the data. This analysis does not attempt to answer all of the

questions that might arise. In particular, it does not explore questions of political

economy or of the desirability of fiscal policy rules.

Section 6.3 presents the relevant time series and discusses the general trends and

patterns they exhibit. This is followed by a statistical and econometric analysis of the

176


data. A test of independence of attributes is presented in Section 6.4 to ascertain

whether occurrence of positive rates of growth of public capital formation is

influenced by adoption of net worth fiscal policy rules. In Section 6.5, another non-

parametric test is employed. This is a test of differences between population

proportions. It aims to assess whether adoption of net worth fiscal policy rules was

associated with a change in the likelihood that public capital formation would increase

in any given year.

Ordinary least squares methods are then carried out in Section 6.6. This ascertains the

degree to which variations in economic activity and fiscal policy rules adoption

contribute to an explanation of variations in public capital formation. The model is

linear in the coefficients. Intercept and slope dummy variables are employed so as to

enable separate identification of changes, if any, in the level of public capital

formation as well as in its relationship with economic activity level at the time of

adoption of net worth fiscal policy rules. Each jurisdiction is investigated separately,

for the first time enabling identification of behaviour at the jurisdictional level. In the

same section, we shall also consider an alternative model, where public capital

formation is postulated to depend on a time trend, rather than on economic activity.

The results from the two models turn out to be quite similar.

Finally, an event study methodology is employed in Section 6.7 to assess the impact

of adoption of a net worth fiscal policy rule on public capital formation. Use of event

studies methodology provides another perspective on the phenomenon under

investigation, with the advantage that its validity is based on different statistical

assumptions.

Section 6.8 addresses the research questions of whether or not governments complied

177


with their own fiscal policy rules and whether these fiscal policy rules appear to have

affected fiscal behaviour, especially public capital formation.

6.2 COMPLIANCE WITH FISCAL POLICY RULES

This section investigates how frequently governments have complied with constraints

self-imposed by adoption of fiscal policy rules. Australian governments typically

formulated fiscal policy rules as requiring ex ante observance, i.e. in budgetary terms.

However, to establish compliance, it is necessary to scrutinise ex-post fiscal measures

(see Section 3.2.3).

6.2.1 Budgetary balance rules

The measure adopted, nature of the constraint, sector to which the measure relates and

adoption dates are presented in Table 6.1. This information is presented for the

Commonwealth, New South Wales, Victorian, Queensland, South Australian and

Western Australian Governments. It is considered that the Tasmanian Government

did not adopt fiscal policy rules (see Section 4.3). This information provides a basis

for comparison of ex-post fiscal measures with targets in order to establish the extent

of compliance, though not intentionality (as previously discussed).

Table 6.2 provides the relevant ex-post fiscal outcomes specified in Table 6.1. These

are provided for the period from adoption of fiscal policy rules to 2003-04. The

exception is New South Wales, for which accrual-based GFS data are unavailable for

1996-97 and 1997-98. The data are presented in nominal terms, in accordance with

the terms in which all budgetary balance fiscal policy rules were formulated.

178

Cha

pter

6: C

ompl

ianc

e an

d Pu

blic

Cap

ital F

orm

atio

n

Tab

le 6

.1 M

easu

rem

ent o

f bud

geta

ry b

alan

ce fi

scal

pol

icy

rule

s

Gov

ernm

ent

Mea

sure

ado

pted

T

arge

t R

elev

ant S

ecto

r

Dat

e of

ado

ptio

n C

omm

onw

ealth

1998

-99

ca

sh-b

ased

GFS

Net

Ope

ratin

g Ba

lanc

e (e

quiv

alen

t to

accr

ual-b

ased

GFS

Net

Le

ndin

g/Bo

rrow

ing)

19

99-2

000

to 2

003-

04 a

ccru

al-b

ased

GFS

Net

Le

ndin

g/ B

orro

win

g

Bal

ance

19

98-8

8 bu

dget

sect

or

1999

-200

0 to

200

3-04

ge

nera

l gov

ernm

ent

1998

-99

NSW

19

96-9

7 to

199

9-20

00 c

ash-

base

d G

FS N

et

Ope

ratin

g Ba

lanc

e (e

quiv

alen

t to

accr

ual-b

ased

G

FS N

et L

endi

ng/B

orro

win

g)

2001

-02

to 2

003-

04 a

ccru

al- b

ased

GFS

Net

Le

ndin

g/ B

orro

win

g

Surp

lus f

rom

199

8-99

Gen

eral

gov

ernm

ent

1996

-97

Vic

toria

G

FPR

Ope

ratin

g Re

sult

Surp

lus o

f at l

east

$1

00M

per

ann

um

Gen

eral

gov

ernm

ent

2000

-01

Que

ensl

and

Acc

rual

-bas

ed G

FS N

et O

pera

ting

Bala

nce

Bal

ance

or s

urpl

us

Gen

eral

gov

ernm

ent

1999

-200

0 SA

A

ccru

al-b

ased

GFS

Net

Len

ding

/ Bor

row

ing

(exc

ludi

ng d

epre

ciat

ion

effe

cts)

B

alan

ce o

r sur

plus

(e

xclu

ding

de

prec

iatio

n)

Gen

eral

gov

ernm

ent

2002

-03

WA

A

ccru

al-b

ased

G

FS N

et O

pera

ting

Bala

nce

Surp

lus

Gen

eral

gov

ernm

ent

and

tota

l pub

lic

2000

-01

179

Cha

pter

6: C

ompl

ianc

e an

d Pu

blic

Cap

ital F

orm

atio

n

Tab

le 6

.2 B

udge

tary

bal

ance

fisc

al o

utco

mes

(gen

eral

gov

ernm

ent s

ecto

r)

1998

-99

1999

-200

0 20

00-0

1 20

01-0

2 20

02-0

3 20

03-0

4 Ju

risd

ictio

n A

ccru

al-b

ased

GF

S N

et L

endi

ng(+

)/ B

orro

win

g(-)

($m

) C

omm

onw

ealth

4,

717

11,8

145,

836

-3,5

156,

391

6,40

6 N

SW

-13

1,43

166

7 56

963

86

SA (e

xclu

des

depr

ecia

tion)

814

860

A

ccru

al-b

ased

GF

S N

et O

pera

ting

Bal

ance

Q

ueen

slan

d

1,

062

-856

-8

9517

3,33

7 W

A

200

229

289

832

G

PFR

Ope

ratin

g Su

rplu

s (+)

/Def

icit(

-)

VIC

12

17.1

27

3.4

235.

999

0.1

Sour

ces:

AB

S C

atal

ogue

551

2.0

Gov

ernm

ent F

inan

ce S

tatis

tics 2

004-

05, V

icto

rian

Gov

ernm

ent (

2001

b), (

2002

b), (

2003

b) a

nd (2

004)

N

ote:

Sha

ded

area

s ind

icat

e th

at th

is ti

me

perio

d is

prio

r to

adop

tion

of fi

scal

pol

icy

rule

s.

180


The Commonwealth Government met its budgetary balance target of positive GFS

Net Lending/Borrowing in each year except 2001-02. The New South Wales

Government met its budgetary balance target of positive GFS Net Lending/Borrowing

from 1999-2000, failing to meet the target only in 1998-99. The Victorian

Government met its target of a GPFR Operating Surplus of at least $100 million per

annum from 2000-01 in every year. The Queensland Government attained its target

of positive GFS Net Operating Balance in 1999-2000, 2002-03 and 2003-04. It did

not attain positive GFS Net Operating Balance in 2000-01 and 2001-02. The South

Australian Government attained its target in 2002-03 and 2003-04 of positive GFS

Net Lending (excluding depreciation). The Western Australian Government met its

target of positive GFS Net Operating Balance from 2000-01 onwards.

Thus the Commonwealth and New South Wales Governments complied with their

budgetary balance fiscal policy rules in five of six years, both failing to comply once.

The Victorian and Western Australian Governments complied in each of four years.

The Queensland Government complied in three of five years, failing twice. The

South Australian Government complied in each of the two years examined.

The Commonwealth and New South Wales Governments each failed to meet their

targets in only one year and the Queensland Government failed to meet its target in

two consecutive years. This is a total of four failures, compared with 23 years of

being able to comply. The success rate of all governments is 23 of 27 or 85 per cent.

The success rate of State governments is 18 of 21 or 86 per cent. On balance,

governments have mostly, though not always, met the constraints self-imposed by

adoption of budgetary balance fiscal policy rules.

As stated in Section 3.2.1, no formal penalties attach to non-compliance with fiscal

181


policy rules adopted by Australian governments although persistent failures to comply

would have the potential to generate criticism of the relevant government and could

contribute to a downgrading of their credit rating.

6.2.2 Net worth rules

Information contained in Chapters 3 and 4 is reproduced below in summary form,

since this forms a basis for comparison of ex-post fiscal outcomes in order to

determine the extent of compliance with net worth fiscal policy rules.

Table 6.3 summarises the measures used by governments to report their fiscal

outcomes relevant to net worth fiscal policy rules. The information is presented for

the Commonwealth, New South Wales, Victorian, Queensland and Western

Australian Governments. It is considered that the South Australian and Tasmanian

Governments did not adopt net worth fiscal policy rules, as defined for purposes of

this study (see Section 4.3), as neither adopted sufficiently authoritative means of

ensuring their longevity.

Table 6.4 provides the relevant fiscal outcomes. Again, the data are presented in

nominal terms, in accordance with terms in which the fiscal policy rules were

formulated. The exception is New South Wales, where the target was defined in real

terms (see Section 4.3.2). Accordingly, the New South Wales data have been deflated

by application of an implicit price deflator. The data again are accrual-based as only

accrual-based reporting frameworks can provide a comprehensive measure of net

worth (as described in Section 3.4.3).

182

Cha

pter

6: C

ompl

ianc

e an

d Pu

blic

Cap

ital F

orm

atio

n

Tab

le 6

.3 M

easu

rem

ent o

f net

wor

th fi

scal

pol

icy

rule

s

Juri

sdic

tion

Mea

sure

ado

pted

T

arge

t R

elev

ant S

ecto

r

Dat

e of

ado

ptio

n C

omm

onw

ealth

GFS

Net

Wor

th

Posi

tive

rate

of c

hang

e ov

er th

e m

ediu

m

to lo

ng te

rm

Gen

eral

gov

ernm

ent

1999

-200

0

NSW

G

PFR

Net

Ass

ets

Posi

tive

rate

of c

hang

e (in

real

term

s)

Gen

eral

gov

ernm

ent

1996

-97

Vic

toria

G

PFR

Net

Ass

ets

Incr

ease

of $

100M

per

ann

um

Gen

eral

gov

ernm

ent

2000

-01

Que

ensl

and

GFS

Net

Wor

th

Posi

tive

rate

of c

hang

e To

tal p

ublic

19

99-2

000

WA

G

FS N

et W

orth

Po

sitiv

e ra

te o

f cha

nge

Tota

l pub

lic

2000

-01

Tab

le 6

.4 N

et w

orth

fisc

al o

utco

mes

1996

-97

1997

-98

1998

-99

1999

-200

0 20

00-0

1 20

01-0

2 20

02-0

3 20

03-0

4 Ju

risd

ictio

n G

FS

Net

Wor

th (g

ener

al g

over

nmen

t) ($

m)

Com

mon

wea

lth

-31,

348

-40,

776

-42,

664

-44,

369

-47,

162

-30,

322

G

FS

Net

Wor

th (t

otal

pub

lic) (

$m)

Que

ensl

and

58,4

68

57,7

7557

,623

58,0

9264

,891

77,7

24W

A

31,6

9132

,197

35,9

9638

,016

43,7

47

GPF

R N

et A

sset

s (ge

nera

l gov

ernm

ent)

($m

) N

SW (n

omin

al)

22,9

5331

,288

34,8

9641

,204

43,8

2850

,337

54,6

5756

,601

NSW

(rea

l) 27

,390

36,8

0939

,121

45,7

8247

,484

53,2

1056

,059

56,6

01V

icto

ria

14

,619

16,5

7021

,845

23,6

0826

,279

Sour

ces:

AB

S C

atal

ogue

551

2.0

Gov

ernm

ent F

inan

ce S

tatis

tics 2

004-

05, V

icto

rian

Gov

ernm

ent (

2000

b), (

2001

b), (

2002

b), (

2003

b) a

nd (2

004)

, N

ew S

outh

Wal

es T

reas

ury

(199

9), N

ew S

outh

Wal

es G

over

nmen

t (19

97b)

, (19

97c)

, (1

999)

, (20

01b)

, (20

02b)

, (20

03) a

nd (2

004)

N

ote:

As t

arge

ts w

ere

defin

ed in

term

s of c

hang

e fr

om a

pre

viou

s per

iod,

whe

re a

vaila

ble,

the

rele

vant

mea

sure

for t

he p

erio

d im

med

iate

ly p

rior

to a

dopt

ion

of fi

scal

pol

icy

rule

s is m

ade

avai

labl

e (in

shad

ed se

ctio

ns) f

or c

ompa

rison

pur

pose

s. S

hade

d ar

eas t

hat d

o no

t con

tain

dat

a in

dica

te

that

this

tim

e pe

riod

prec

edes

the

rele

vant

per

iod.

183


The data indicate that the Commonwealth Government attained its target of improving

GFS Net Worth, in the short to medium term, only in the last year of this period.

During the entire period, the Commonwealth Government’s net worth position was

negative. However, the absolute size of the negative net worth attained in the last

year was smaller than at the commencement of the period. The New South Wales

Government attained its target of improving GFS Net Worth in real terms throughout

the entire period. The Victorian Government attained its goal of increasing GPFR Net

Assets by at least $100M per annum for each year of the period. The Queensland

Government attained its goal of increasing GFS Net Worth only in the last three years

of the period studied. It did not attain this goal in 1999-2000 and 2000-01. The

Western Australian Government attained its target of maintenance or increase of GFS

Net Worth during each of the years since a net worth fiscal policy was adopted.

Thus the Commonwealth Government succeeded only in one of the five years

examined. The New South Government succeeded in each of the seven years for

which comparison is possible. The Victorian Government succeeded in each of four

years. The Queensland Government succeeded in three of five years. The Western

Australian Government succeeded in each of four years. In total, Australian

governments succeeded in meeting constraints imposed by net worth fiscal policy

rules in 19 of 25 instances, constituting a 76 per cent success rate. State governments

succeeded in 18 of 20 instances, constituting a 90 per cent success rate.

Overall, Australian governments enjoyed partial success in meeting net worth fiscal

policy rule targets. However, State governments on average enjoyed greater success.

184


6.2.3 Net debt rules

Again, information provided in Chapters 3 and 4 is reproduced below as a basis for

comparison with ex-post fiscal outcomes. Table 6.5 summarises the measures used,

targets adopted, relevant sector and adoption dates for each government. The

Commonwealth, New South Wales, Victorian, Queensland and South Australian

Governments are included. It is not considered that the Western Australian or

Tasmanian Governments adopted a net debt fiscal policy rule (as described in Section

4.3). Table 6.6 provides the relevant fiscal outcomes.

185

Cha

pter

6: C

ompl

ianc

e an

d Pu

blic

Cap

ital F

orm

atio

n

Tab

le 6

.5 M

easu

rem

ent o

f net

deb

t fis

cal p

olic

y ru

les

Juri

sdic

tion

Mea

sure

ad

opte

d T

arge

t R

elev

ant S

ecto

r

Dat

e of

ado

ptio

n

Com

mon

wea

lthN

et D

ebt

1998

-99

to 1

999-

2000

red

uctio

n to

10

per c

ent G

DP

by 2

000-

01

Gen

eral

gov

ernm

ent

1998

-99

NSW

N

et D

ebt

Red

uctio

n

to a

‘sus

tain

able

leve

l’ by

200

4-05

and

elim

inat

ion

by 2

020

Gen

eral

gov

ernm

ent

1996

-97

Vic

toria

N

et D

ebt

Red

uctio

n

Gen

eral

gov

ernm

ent

2000

-01

Que

ensl

and

Net

Deb

t M

aint

enan

ce o

f a n

egat

ive

leve

l G

ener

al g

over

nmen

t 19

99-2

000

SA

Net

Deb

t R

educ

tion

G

ener

al g

over

nmen

t 20

02-0

3 T

able

6.6

Net

deb

t out

com

es (g

ener

al g

over

nmen

t)

Net

deb

t ($m

) 19

95-9

6 19

96-9

7 19

97-9

8 19

98-9

9 19

99-2

000

2000

-01

2001

-02

2002

-03

2003

-04

Com

mon

wea

lth

82,9

3571

,928

54,4

43

41,1

8935

,983

27,2

6519

,610

NSW

10

,522

10

,834

10

,264

13,3

1111

,703

7,

702

5,05

61,

587

-209

Vic

toria

3,

076

2,16

21,

024

1,29

71,

399

Que

ensl

and

-1

0,12

3 -1

0,67

2-1

1,60

9-1

1,84

3-1

4,85

1SA

1,30

366

622

4 So

urce

: A

BS

Cat

alog

ue 5

512.

0 G

over

nmen

t Fin

ance

Sta

tistic

s 200

4-05

N

ote:

As t

arge

ts w

ere

defin

ed in

term

s of c

hang

e fr

om a

pre

viou

s per

iod,

whe

re a

vaila

ble,

the

rele

vant

mea

sure

for t

he p

erio

d im

med

iate

ly p

rior

to a

dopt

ion

of fi

scal

pol

icy

rule

s is m

ade

avai

labl

e (in

shad

ed se

ctio

ns) f

or c

ompa

rison

pur

pose

s. S

hade

d ar

eas t

hat d

o no

t con

tain

dat

a in

dica

te

that

this

tim

e pe

riod

prec

edes

the

rele

vant

per

iod.

186


The Commonwealth Government achieved the targeted reduction of the ratio of

general government Net Debt to GDP ($752,434 in nominal terms in 2000-01 (ABS

Catalogue 5204 Australian System of National Accounts 2004-05)) to 10 per cent by

2000-01. It has continued to reduce Net Debt in each year since. While sustaining an

increase in Net Debt in two years (1996-97 and 1998-99) after adoption of a target of

reduction and eventual elimination, the New South Wales Government achieved its

target of reduced general government Net Debt over the entire period under scrutiny.

Further, it met its target of elimination of Net Debt by 2020 some 16 years early in

2003-04. The Victorian Government achieved the targeted reduction of general

government Net Debt in the first two years (2000-01 and 2001-02). Though it did not

reduce Net Debt thereafter, it met its target of maintaining a AAA credit rating

throughout the period. The Queensland Government achieved in every year its target

of non-positive Net Debt levels in the general government sector since adoption of

that target in 1999-2000. The South Australian Government achieved its target of

reducing Net Debt in each year of the relatively short period since adoption of that

target.

Thus the Commonwealth Government can be said to have succeeded in meeting its

net debt fiscal policy rule in each of the six years scrutinised. The New South Wales

Government succeeded in six of the eight years examined. The Victorian

Government attained two successes of four. The Queensland and South Australian

Governments succeeded in each of five and two years respectively.

Overall, Australian governments succeeded in 21 of 25 instances. This is an 84 per

cent success rate on an aggregated basis. State governments succeeded 15 of 19 times,

constituting a 79 per cent success rate.

187


Australian governments have enjoyed a slightly greater level of success in meeting net

debt fiscal policy rule targets than net worth fiscal policy rule targets. This may be

related to adoption of accrual accounting and its provision of an improved information

set which has enabled recognition of the extent of accruing net debt liabilities such as

employee entitlements including superannuation. Alternatively, greater emphasis on

debt as a result of its importance in determination of credit ratings could be an

explanatory factor. Either of these factors could have lead to accumulation of

financial assets at the expense of public capital formation.

6.2.4 Summary

This section set out to determine the extent of apparent compliance by governments

with constraints self-imposed by their adoption of fiscal policy rules. Targets,

identified in Chapters 3 and 4, were compared with ex-post fiscal outcomes to

determine the extent of compliance. As discussed in Section 3.2.2, this should not be

taken as an indication of intentionality.

It was found that there is some occurrence of non-compliance with constraints

imposed by adoption of fiscal policy rules. This is probably related to the absence of

penalties for non-compliance with fiscal policy rules adopted by Australian

governments (as described in Section 3.2.1). However, the incidence of compliance

was found to be much greater than the incidence of non-compliance.

The degree of compliance varied with the type of fiscal policy rule. Australian

governments enjoyed an 85 per cent success rate with respect to budgetary balance

fiscal policy rules, a 76 per cent success rate with respect to net worth fiscal policy

rules and an 84 per cent success rate with respect to net debt fiscal policy rules. State

governments enjoyed an 86 per cent success rate in meeting budgetary balance fiscal

188


policy rule targets, a 90 per cent success rate in meeting net worth fiscal policy rule

targets and a 79 per cent success rate in meeting net debt fiscal policy rule targets.

Hence, Australian governments were more successful in meeting budgetary balance

and net debt targets than net worth targets. Possible explanations include greater

awareness of the need to accrue financial assets to meet accruing non-debt liabilities

as a result of adoption of accrual-based financial reporting frameworks and a focus on

credit ratings.

However, State governments were most successful in meeting net worth targets,

almost as successful with meeting budgetary balance targets and somewhat less

successful in meeting net debt targets. This may reflect the starting net debt levels of

State governments, which were generally lower than those of the Commonwealth

government.

6.3 TRENDS IN PUBLIC CAPITAL FORMATION

The ratios of GFCF to GDP/GSP over the study period are displayed in Figure 6.1.

189


Figure 6.1 Trends in the ratio of GFCF to GDP/GSP, by jurisdiction

0.00200.00250.00300.00350.00400.00450.00500.0055

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

CW Linear (CW)

0.005

0.007

0.009

0.011

0.013

0.015

1985

-86

1987

-88

1989

-90

1991

-92

1993

-94

1995

-96

1997

-98

1999

-00

2001

-02

NSW Linear (NSW )

0.006

0.007

0.008

0.009

0.010

0.011

0.012

0.013

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

V IC Linear (VIC)

0.0100.0120.0140.0160.0180.0200.0220.0240.0260.028

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

QLD Linear (QLD)

0.0050.0070.0090.0110.0130.0150.0170.0190.021

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

WA Linear (WA)


190


It can be seen that the ratio generally fell over the period 1984-85 to 2003-04 for all

jurisdictions. This indicates that public capital formation has not increased as quickly

as GDP/GSP. As discussed in Section 3.5, under certain assumptions, this may

indicate that government service levels have decreased over time.

The data series included a structural break at the time of the shift from a cash-based to

an accrual-based financial reporting methodology. (Techniques used to overcome this

difficulty were described in Section 5.3.1.) While comparison of levels across

jurisdictions is not strictly valid, the ratio of public capital formation to economic

activity is sufficiently lower for the Commonwealth Government than for each of the

four State governments to justify a conclusion that this is not an artefact of these

techniques. Further, the lower level of public capital formation by the

Commonwealth, compared with the States, is in accordance with the generally lower

level of public capital formation by the Commonwealth Government than State

governments in Australia’s model of federalism (discussed in Section 5.3.1).

For completeness, each part of the ratio is further examined below. Figure 6.2 plots

the levels of real GFCF per capita over time. It indicates that the Victorian and

Queensland Governments have increased public capital formation in real terms per

capita over time while the Commonwealth, New South Wales and Western Australian

Governments have generally reduced public capital formation in real terms per capita.

Figure 6.3 presents levels of real GDP/GSP per capita. As can be expected, real

GDP/GSP per capita in all jurisdictions tended to rise during this period.

191


Figure 6.2 GFCF over time, by jurisdiction

100.00

110.00

120.00

130.00

140.00

150.00

160.00

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

$

CW Linear (CW)

250.00

300.00

350.00

400.00

450.00

500.00

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

$

NSW Linear (NSW)

200.00

250.00

300.00

350.00

400.00

450.0019

84-8

5

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03V IC Linear (VIC)

300.00

400.00

500.00

600.00

700.00

800.00

900.00

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

QLD Linear (QLD)

300.00

350.00

400.00

450.00

500.00

550.00

600.00

650.00

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

WA Linear (WA)

Source: ABS Catalogue 5206.0 National Income, Expenditure and Product Dec 2005, ABS Catalogue 3105.0 Australian Historical Population Statistics 2005 Note: GFCF is expressed in real per capita terms.

192


Figure 6.3 Levels of GDP/GSP, by jurisdiction

25000.00

30000.00

35000.00

40000.00

45000.00

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

CW

25000.00

30000.00

35000.00

40000.00

45000.00

1984

-85

1987

-88

1990

-91

1993

-94

1996

-97

1999

-00

2002

-03

$

NSW

25000.00

30000.00

35000.00

40000.00

45000.00

1984

-85

1987

-88

1990

-91

1993

-94

1996

-97

1999

-00

2002

-03

$

VIC

22000.00

27000.00

32000.00

37000.00

42000.00

1984

-85

1987

-88

1990

-91

1993

-94

1996

-97

1999

-00

2002

-03

$

QLD

25000.00

30000.00

35000.00

40000.00

45000.00

50000.00

1984

-85

1986

-87

1988

-89

1990

-91

1992

-93

1994

-95

1996

-97

1998

-99

2000

-01

2002

-03

$

WA

Source: ABS Catalogue 5206.0 National Income, Expenditure and Product Dec 2005 Note: GDP/GSP is expressed in real per capita terms.

193


A closer examination reveals that rates of growth were similar for all jurisdictions

with the exception of Western Australia which outperformed the other jurisdictions.

GSP levels for Queensland were below the national level. Victorian levels were close

to the national level. Levels of GSP for New South Wales and Western Australia

were above the national level.

Analysis with a finer temporal focus is carried out in the following sections, focusing

on whether the adoption of net worth fiscal policy rules coincides with changes in the

rate of growth of public capital formation, or in its relationship with economic activity.

6.4 NON-DEPENDENCE OF POSITIVE GROWTH IN PUBLIC CAPITAL

FORMATION AND NET WORTH FISCAL POLICY RULES

The Chi-square method is used to investigate whether the following two qualitative

variables are statistically related : (1) occurrence of positive rates of growth of GFCF;

and (2) existence of a fiscal policy rule requiring maintenance or increase of net worth.

The technique can also be interpreted as indicating whether differences exist between

the two populations of observations (before and after the adoption of rules). Rejection

of the hypothesis of independence would suggest that adoption of the net worth fiscal

policy rules may have been associated with some changes in the way public capital

formation grew over time.

The test is non-parametric. The only technical requirement is that the minimum value

in any cell be 5. Sufficient data exist to meet this requirement when observations

pertaining to all jurisdictions are aggregated. To achieve aggregation, timelines are

normalised where:

194


tk is the year of adoption of fiscal policy rules; and

tk-1 is the year before adoption of fiscal policy rules etc.

Observations for years comprising the interregnum were removed from the series.

This means that, for different jurisdictions, different numbers of observations were

dropped from the analysis.

The critical value of Chi-Square at the five per cent level of significance is 3.84.

Testing State jurisdictions only, the calculated value of the test statistic is 0.72;

therefore one cannot reject the null hypothesis of independence. That is, occurrence

of a positive rate of growth of GFCF is independent of the existence of a net worth

fiscal policy rule. In other words, no evidence has been found of an association

between (a) adoption of a net worth fiscal policy rule by State governments in

Australia and (b) whether public investment by State governments was likely to

increase each year.

Testing all jurisdictions (including the Commonwealth Government) produces a

calculated value of 0.576 and a similar conclusion. That is, based on this analysis, the

conclusion is that positive rates of growth of GFCF are not dependent on adoption of

a net worth rule by the Commonwealth and State governments in Australia. Use of

Yates’ correction for continuity yields lower calculated values and no change in the

basic conclusions (Sheskin (2004) page 502)

As a check on the effect of the interregnum approach, a non-interregnum model was

constructed where tk denotes the date of adoption of fiscal policy rules and no data

points are omitted, for both the State jurisdictions only and for all jurisdictions. The

calculated values of the test statistic are 0.639 for State jurisdictions only and 0.304

195


for all jurisdictions. Consequently, tests of both models produce conclusions

consistent with those based on interregnum models.

The above conclusions have to be considered in light of limitations associated with

this aggregated approach. First, the fact that the analysis is carried out on all

jurisdictions simultaneously means that it can only provide an aggregated finding.

For example, if one jurisdiction had a very weak or non-existent relationship between

adoption of fiscal policy rules and public capital formation, this could be sufficient to

influence the overall finding for the aggregated analysis, even if a significant

relationship existed for the majority of jurisdictions. In particular, the New South

Wales Government does not appear to have increased the rate of public capital

formation after adopting a net worth fiscal policy rule. In this case, the limitation is

quite significant because of the small number of jurisdictions involved.

Perhaps more importantly, the technique focuses only on the conditional probability

that a positive rate of growth of GFCF will occur and does not take into account the

magnitude of such growth. Consider, for example, the hypothetical case where a

jurisdiction had been undertaking increases in public capital formation in every year,

then adopted a fiscal policy rule and proceeded to increase public capital formation

even faster than before. Under those circumstances, the current method of analysis

will not capture the essence of the change, because it can only focus on the fact that

public investment grew in every year before the adoption of policy rules, as well as in

every year after that event. More discriminating techniques are required to distinguish

between a positive-but-lower rate of growth from a positive-and-higher rate. Such

methods are implemented in Section 6.6 below.

196


6.5 PROBABILITY OF INCREASED PUBLIC CAPITAL FORMATION

AFTER ADOPTION OF FISCAL POLICY RULES

A test of differences between population proportions is then performed to ascertain

whether a different proportion of positive instances of growth of GFCF occurred

before and after adoption of a net worth fiscal policy rule. Required conditions, on

which to base the assumption of normal distribution of the population proportion, are

that there be a large number of trials and that these trials be independent. A large

number of trials is again achieved by combining data for all jurisdictions. The clear

distinction between reporting periods again provides for independence between trials.

Again, due to differing adoption dates between jurisdictions, timelines are

standardised.

Using an interregnum model and examining State jurisdictions only, the calculated

value of the test statistic equals -0.86. Thus, one cannot reject the null hypothesis that

the proportion of “successes” after adoption of a net worth fiscal policy rule is the

same as the proportion of “successes” prior to adoption of this rule. That is, positive

rates of growth of GFCF occurred with essentially the same frequency before and

after existence of a fiscal policy rule requiring maintenance or increase of net worth.

Testing all jurisdictions produces a calculated test statistic of -0.759 and a similar

conclusion.

These results accord with the finding of the Chi-square test of independence carried

out in Section 6.4. The implication of these results is that, using these criteria, one

cannot find evidence that adoption of fiscal policy rules affected the likelihood that

public capital formation increased each year.

197


However, the results must again be qualified by reference to the possibility that the

finding reached using aggregated data may not apply to some (or possibly many)

individual jurisdictions. The analysis in the next section will be carried out on a

jurisdiction-by-jurisdiction basis, providing findings for each government in turn and

a view of Commonwealth/ State differences. Another caveat is that the analytical

techniques used in this section view public capital formation and fiscal policy rules as

qualitative variables whose values can only be either Yes or No. Such an analysis

loses much of the information contained in available quantitative data, which can be

captured by the more sophisticated techniques to be employed next.

6.6 PARAMETRIC ESTIMATION OF FISCAL POLICY RULES

6.6.1 Fiscal policy rules and the relationship between public capital formation

and economic activity

Overview

To complement the qualitative-variable approach adopted in Sections 6.4 and 6.5, in

this section we turn to an approach that emphasises the numerical aspect of the

growth in public investment. Specifically, ordinary least squares methodology will be

carried out to ascertain the degree to which variations in economic activity and the

adoption of net worth fiscal policy rules explain variations in public capital formation.

This technique can potentially capture more details than the previous non-parametric

techniques. For example, suppose a government had increased public capital

formation by $1 million per annum prior to adoption of fiscal policy rules and by $1

billion per annum after adoption. While neither non-parametric technique would

recognise the difference, the regression technique described in the current section may

be able to do so.

198


The current method involves regressing the natural log of GFCF on the natural log of

Gross Domestic/State Product, an intercept dummy variable and a slope dummy

variable. The economic variables are logged, as the investigation focuses on rates of

change. Transformation by taking natural logs provides a convenient measure of

these. Further, the relevant coefficient (β2) can be interpreted as an elasticity.

The underlying model is:

Y = A.Xβ1 (1a)

or

y = β0 + β1x (1b)

where:

y = the natural log of GFCF real per capita and

x = the natural log of real GDP/GSP per capita.

In regression (log-linear) format, and allowing for a structural break, this may be

expressed as follows:

y = β0 + β1x + β2FPR + β3M + ε (2)

where:

FPR = an intercept dummy variable, with a value of 1 when a fiscal policy

rule is in place and a value of 0 otherwise and

M = a slope dummy variable, being the product of FPR and x, and having

a value of 0 when no fiscal policy rule exists and the natural log of

real GDP/GSP per capita when a fiscal policy rule exists.

Τhus, β0 is the original intercept term, while β1 measures the elasticity of public

capital formation with respect to economic activity. Further, β2 measures the

proportional impact on the level of public capital formation of the imposition of a net

199


worth fiscal policy rule. It indicates whether the level of GFCF increases at date of

adoption of rules and, if so, by how much. By contrast, β3 measures whether the

elasticity of public capital formation with respect to economic activity increases at

this date.

When a fiscal policy rule exists, Equation 2 becomes:

y = βo + β1x + β2 + β3x + ε = (βo + β2) + (β1 + β3)x + ε

in which the intercept term equals (β0 + β2) and the slope equals (β1 + β3).

When no fiscal policy rule exists, Equation 2 is simply:

y = β0 + β1x + ε

in which the intercept term equals β0 and the slope equals β1.

Following Pindyck and Rubinfeld (1992) and Seddighi et al (2000), such a

specification allows detection of changes in both the slope and intercept of the

regression line following adoption of net worth fiscal policy rules. In practice,

however, it is not useful to include all three regressors in one calculation due to the

presence of multicollinearity. High levels of correlation exist between the

multiplicative dummy and GDP/GSP (correlation coefficients in the order of .99 are

common). This can lead to spurious results of exceedingly high coefficients of

determination but low and unreliable t-statistics.

To address this problem, two separate regressions are carried out, thereby reducing

the severity of multicollinearity. The two models are:

y = (β0 + β2FPR) + β1x + ε (3)

y = β0 + (β1 + β3FPR)x + e (4)

200


The key hypothesis to be tested using the first model is that β2 ≠ 0, that is, the level of

GFCF changes at adoption of a net worth fiscal policy rule. The hypothesis to be

tested using the second model is that β3 ≠ 0, that is, the elasticity of GFCF with

respect to GDP/GSP changes at adoption of a net worth fiscal policy rule. Tables 6.7

to 6.11 inclusive provide results for each jurisdiction.

Commonwealth Government results

Table 6.7 Commonwealth Government results

According to standard statistical tests, explanatory power of the models for the

Commonwealth Government is satisfactory (adjusted R2 is 0.79 for each) and the

models are deemed to be useful (p-value of the F-statistic is very close to zero for

each). Each independent variable is significant at the five per cent level of

significance (p-values are 0.00 and 0.01 for x, 0.04 for the level dummy variable and

0.04 for the slope dummy variable).

The estimate of β1 indicates that real public capital formation per capita increased by

about 1.4% when the level of economic activity increased by 1%. The coefficients of

both dummy variables are positive. It is not possible to determine whether it is more

accurate to say that the intercept or slope of the regression line increases. Instead, the

two separate regressions indicate that there is a significant and positive relationship

Independent variables Constant ln GDP

rpc FPR (level)

dummy Multiplicative (slope) dummy

Eqn

Adj R2

Prob (F-stat) Coef Prob Coef Prob Coef Prob Coef Prob

3 0.79 0.00 -9.90 0.04 1.39 0.00 0.30 0.04 4 0.79 0.00 -9.86 0.04 1.39 0.01 0.03 0.04

201


between either the level or the output elasticity (or both) of public capital formation

and adoption of a net worth fiscal policy rule.

New South Wales Government results

Table 6.8 New South Wales Government results

For New South Wales, the explanatory power of the models is very poor (adjusted R2

equals 0.13 for each) and the models are not useful (p-value of the F-statistic is 0.14

for each). None of the independent variables are significant, even at the ten per cent

level of significance (p-values for x are 0.48 and 0.49, 0.99 for the level dummy

variable and 0.97 for the slope dummy variable).

As the results are of an interregnum model, the data series omits the anomalous

observation represented by the large negative value of GFCF for 1996-97. Thus the

data anomaly does not explain these results. Nor does the earlier analysis of

compliance provide an explanation. The New South Wales Government complied

with its net worth fiscal policy rule in each of the seven years examined, with its net

debt fiscal policy rule in six of eight years and with its budgetary balance fiscal policy

rule in five of six years. This compliance rate is not strikingly different from that of

any other jurisdiction. However, earlier discussion of net worth established that

increasing net worth does not necessarily imply increasing real assets. It is possible

that the New South Wales Government accumulated financial assets rather than real

assets. Such an outcome could explain both the New South Wales Government’s

Independent variables Constant ln GSP rpc FPR (level)


Eqn

Adj R2


3 0.13 0.14 3.17 0.42 0.27 0.48 0.00 0.99 4 0.13 0.15 3.23 0.42 0.26 0.49 0.00 0.97

202


success in compliance as well as these results.

Victorian Government results

Table 6.9 Victorian Government results



Eqn

Adj R2

Prob (F-stat)

Coef Prob Coef Prob Coef Prob Coef Prob

3 0.41 0.01 1.53 0.73 0.39 0.36 0.26 0.09 4 0.41 0.01 1.53 0.73 0.39 0.36 0.02 0.09

While explanatory power is not very high (adjusted R2 is 0.41 for each), the models

for the Victorian Government are considered sufficiently useful (p-value of the F-

statistic is 0.01 for each). The coefficients of both dummy variables are positive. At

the ten per cent level of significance, both the level and slope dummy variables are

significant with p-values of 0.09. These results support a conclusion that there is a

significant and positive relationship between either the level or the output elasticity

(or both) of public capital formation and adoption of a net worth fiscal policy rule by

the Victorian Government.

It does not appear, however, that there was a significant relationship between levels of

economic activity and public capital formation in Victoria. A possible explanation is

the financial arrangements adopted in Victoria where an infrastructure reserve was

established in 2000-01 as a source of funding for future capital acquisitions. This

would have negated the link between revenue-raising capacity arising from increased

economic activity and funding availability for capital acquisitions. Another possible

explanation is that GDP/GSP is merely standing in for a time trend, a possibility we

shall pursue in the next subsection.

Overall, the analysis supports a conclusion that real public capital formation per capita

203


either increased in level, or began to exhibit a positive relationship with GSP, or both,

when the Victorian Government adopted a net worth fiscal policy rule.

Queensland Government results

Table 6.10 Queensland Government results

Although the explanatory power of the models is not high (adjusted R2 is 0.36 for

both), the models are considered sufficiently useful (p-values of the F-statistic are

0.02 each). All independent variables are significant at the five per cent level of

significance (p-values are 0.02 and 0.03 for x, 0.01 for the level dummy variable and

0.01 for the slope dummy variable).

Surprisingly, the coefficient for x is negative (-1.37), suggesting that public capital

formation decreased when levels of economic activity increased. This unexpected

result warrants further investigation.

One possible explanation is a tendency by the Queensland Government to increase

accumulation of financial assets rather than real assets. Such a tendency could be

explained in terms of a decision to accumulate financial assets by which to meet

accruing non-debt liabilities such as employee entitlement (including superannuation)

liabilities. Both entitlement liabilities and GSP experienced rising trends during this

period. Thus, the negative coefficient might reflect a shift from real assets to

financial assets, at a time when the latter were growing rapidly to match growing



Eqn

Adj R2


3 0.36 0.02 20.09 0.00 -1.37 0.02 0.57 0.01 4 0.36 0.02 20.07 0.00 -1.37 0.03 0.05 0.01

204


liabilities. The unusually high level of negative net debt of the Queensland

Government (reported in Section 5.2.3) provides support for this line of thinking.

Coefficients for both dummy variables are positive. Again, the two regressions

indicate that there is a significant and positive relationship between either the level or

the output elasticity (or both) of public capital formation and adoption of a net worth

fiscal policy rule by the Queensland Government.

Western Australian Government results

Table 6.11 Western Australian Government results

Explanatory power is high (adjusted R2 is 0.65 and 0.64 respectively) and both models

for the Western Australian Government are useful (p-values of the F statistic are 0.00

in each case). All independent variables are strongly significant at the five per cent

level (p-values are 0.00 in all cases).

Similarly to the case of Queensland, the coefficient for x is negative (-1.32) indicating

that public capital formation and economic activity were inversely related in Western

Australia. A possible explanation for this, similarly to Queensland, is accumulation of

financial assets. While the Western Australian Government did not adopt a net debt

rule, it is possible that a focus on accumulation of financial assets to offset accruing

liabilities may have contributed to the negative relationship between public capital



Eqn

Adj R2


3 0.65 0.00 19.74 0.00 -1.32 0.00 0.48 0.00 4 0.64 0.00 19.75 0.00 -1.32 0.00 0.05 0.00

205


formation and economic activity. However, investigation of this is outside the scope

of this study.

The coefficients for both dummy variables are positive (0.48 for the level dummy and

0.05 for the slope dummy). Again, the two regressions indicate that there is a

significant, positive relationship between either the level or the output elasticity (or

both) of public capital formation and adoption of a net worth fiscal policy rule by the

Western Australian Government.

Summary of results and their significance

Most jurisdictions show a positive relationship between public capital formation and

adoption of net worth fiscal policy rule. Table 6.12 summarises these results.

Table 6.12 Summary of results

Relationship between ln GFCF rpc and Jurisdiction Constant Ln GDP/GSP Level dummy

variable Slope dummy variable

Commonwealth - + + + NSW Victoria + + Queensland + - + + WA + - + + Note: - indicates a significant, negative relationship, + indicates a significant, positive relationship, and a blank space indicates no significant relationship.

The models adopted are useful for all jurisdictions except New South Wales and were

particularly useful for the Commonwealth, Queensland and Western Australian

Governments. The investigated relationships between public capital formation and

adoption of net worth fiscal policy rules were significant at the five per cent level for

the Commonwealth, Queensland and Western Australian Governments and at the 10%

level for the Victorian Government. The only jurisdiction where this association was

206


not significant was New South Wales.

The relationship between economic growth and public capital formation was

significant and positive only for the Commonwealth Government. While significant,

the relationship was negative for the Queensland and Western Australian

Governments. The relationship was insignificant for the New South Wales and

Victorian Governments.

Overall, these results are consistent with the premise that adoption of fiscal policy

rules has coincided with changed governmental fiscal behaviour with respect to public

capital formation by a majority of the Australian governments examined. In particular,

public capital formation tended to increase, or became more responsive to changes in

GDP/GSP, after adoption of net worth fiscal policy rules.

These results could also be interpreted as providing support for the broader, implicit

assumption adopted in much of the literature (as described in Section 3.2.1) that fiscal

policy rules generally are effective in changing governmental behaviour. However, as

outlined in Section 3.2.2, intentionality or otherwise of compliance, where compliance

exists, with constraints imposed by fiscal policy rules, is not determinable in this

study.

Similar outcomes are found for the national and three sub-national jurisdictions

indicating that, when fiscal policy rules are adopted, level of government is not a good

predictor of outcomes. This implies that fiscal policy rules may be useful tools of

fiscal sustainability for all levels of government, including the local government level,

in the Australian and international contexts. However, the greater explanatory power

evident for the national level may indicate that a slightly different relationship exists

between public capital formation and levels of economic activity than exists at the

207


state level.

In addition, high levels of constant growth of public capital formation are found for

Queensland and Western Australia which are also the highest growth States. The

literature surveyed in Chapter 2, in particular Aschauer (1988) and (1989) and Otto

and Voss (1994), indicated correlation between high rates of public capital formation

and (lagged) economic growth.

These results must be considered in light of a number of qualifications. First, it is

recognised that the number of observations available for analysis is very small, with

consequences for both the degrees of freedom underlying the various statistical tests

and for the power of the tests. Second, the specifications assume constant variance of

the error term in the prior and post adoption periods. This is a fundamental

qualification, applying to all of these regressions. A priori, it could be expected that

the variance of the error term would decrease in the period following adoption of a net

worth fiscal policy rule, when the incidence of negative public investment is less

likely to occur. However, as previously discussed with respect to net worth, negative

investment in real assets may be offset by positive investment in financial assets

resulting in an overall increase in net worth. Thus, the assumption is not unacceptable.

Further, the validity of ordinary least squares regression methods relies on correctness

of the classical assumptions. Following Hill, Griffiths and Judge (2001), these are (a)

the expected value of the random error term e is zero (stationarity of the economic

variables); (b) the variance of e is constant (no heteroskedasticity); (c) uncorrelated

values of the dependent variable y, which has zero covariance, based on statistical

independence between values of the dependent variables (no serial correlation); and

208


(d) the independent variable x is not random and takes at least two different values.

Under these assumptions, least squares estimators b1 and b2 have the smallest variance

of all linear and unbiased estimators of β1 and β2. An additional option assumption,

that e is normally distributed, is sometimes made.

As previously discussed, collinearity has been minimised, though not fully removed,

by model specification. Heteroskedasticity can be diagnosed by visual inspection of a

scatter diagram of the error term plotted against the explanatory variable. Some

evidence of heteroskedasticity exists for both equations for the Commonwealth

Government. This suggests that the model will be estimated more precisely for lower

levels of economic activity than for higher levels. Evidence of heteroskedasticity

exists for the Victorian Government, suggesting that the model will be estimated less

precisely for lower levels of economic activity than for higher levels (as GFCF varies

more at low levels of GSP). No evidence of heteroskedasticity exists for the New

South Wales, Queensland or Western Australian Governments.

Visual inspection of graphed errors indicates the presence of positive autocorrelation

for the Commonwealth, New South Wales, Victorian and Queensland Governments.

Autocorrelation of the errors was not evident for the Western Australian Government.

Similar consequences for the least squares estimators as for heteroskedasticity apply.

Additionally, the question of non-stationarity, common in time series data, must be

addressed. There is a strong likelihood that both the main economic variables (real

GDP/GSP per capita and real GFCF per capita), in either level or log forms, are non-

stationary i.e. exhibit deterministic trends. The fact that real GDP/GSP per capita

grew over time, and real GFCF per capita exhibited a downward trend over time,

lends weight to this expectation. Following Gujarati (2006), popular methods by

209


which to diagnose non-stationarity are the use of either (augmented or unaugmented)

Dickey-Fuller unit root tests or the Box-Jenkins approach, based on visual inspection

of the correlogram. Should non-stationarity be detected, a unit root test is usually

applied to the residuals of an ordinary least squares regression carried out using the

non-stationary variables. If the residuals of that regression are found to be stationary

i.e. integrated of order 0, this indicates the existence of cointegration between the

economic variables. The significance of this is that the coefficients estimated by the

regression are said to represent long term equilibrium relationships between the

dependent and explanatory economic variables. Where cointegration is not found,

first differences of the economic variables may be used to detrend the series and

relationships estimated. However, the brevity of the time series used in this study

indicates that unit root tests are not suitable.

6.6.2 Fiscal policy rules and the relationship between public capital formation

and the passage of time

Overview

In this section, attention is focussed on the effects of the passing of time. Ordinary

least squares methodology is again used and GFCF is regressed on time, an intercept

dummy variable (which has a value of 1 when a fiscal policy rule is in place and a

value of 0 otherwise) and a slope dummy variable, which is the product of the fiscal

policy rules dummy variable and time. In such an analysis, a non-interregnum

approach is conceptually most appropriate as such an approach uses uninterrupted

time series data, although results obtained using an interregnum approach are also

presented for completeness.

As previously, the underlying model is initially considered in nonlinear form:

210


Y = A.eβ1.t (5a)

or

y = β0 + β1t (5b)

where:

y = the natural log of GFCF real per capita and

t = years from 1984-85 to 2003-04, designated by sequential

numbers 1 to 20.

In regression (log-linear) format, and again allowing for a structural break, this may

be expressed:

y = β0 + β1t + β2FPR + β3M2 + ε (6)

where:

FPR = an intercept dummy variable, with a value of 1 when a fiscal policy

rule is in place and 0 otherwise and

M2 = a slope dummy variable, being the product of FPR and t, and therefore

having a value of 0 when no fiscal policy rule is in place and the

sequential number of the year when a fiscal policy rule is in place.

Again, β0 is the original intercept term. β 1 measures the growth rate of public capital

formation before the adoption of fiscal rules. β 2 measures whether the level of public

capital formation increases at date of adoption of rules. β 3 measures whether the rate

of increase of public capital formation increases at that date.

When a fiscal policy rule is in place, Equation 6 is:

y = β0 + β1t + β2 + β3t + ε = (β0 + β2) + (β1 + β3)t + ε

in which the intercept term equals (β0 + β2 ) and (β1 + β3) measures the slope.

211


When no fiscal policy rule is in place, Equation 6 is simply:

y = β0 + β1t + ε

where β0 is the intercept term and β1 is the slope parameter.

For reasons analogous to those applying to the analyses carried out in Section 6.6.1,

two models are used. The models are specified (in log-linear format) as follows:

y = (β0 + β2FPR) + β1t + ε (7)

y = β0 + (β1 + β3FPR)t + e (8)

The key hypothesis to be tested using the first model is that β2 ≠ 0, that is, the level of

GFCF changes at the date of adoption of a net worth fiscal policy rule. The key

hypothesis to be tested using the second model is that β3 ≠ 0, that is, the rate of

increase of GFCF changes at that date.

Results

Table 6.13 provides results for all jurisdictions. As in Section 6.6.1, results for New

South Wales differ from those of the other jurisdictions. Both models are useful at the

five per cent level of significance for all jurisdictions except New South Wales, for

which the second model is useful at the ten per cent level of significance.

212

Cha

pter

6: C

ompl

ianc

e an

d Pu

blic

Cap

ital F

orm

atio

n

Tab

le 6

.13

Res

ults

of t

rend

ana

lysi

s

Inde

pend

ent v

aria

bles

C

onst

ant

Tre

nd

FPR

M

2 Ju

risd

ictio

n

Eqn

A

dj

R2

Prob

(F

st

at)

Coe

f Pr

ob

Coe

f Pr

ob

Coe

f Pr

ob

Coe

f Pr

ob

7 0.

85

0.00

4.

26

0.00

0

.03

.00

0.27

0.

01

Com

mon

wea

lth

8 0.

86

0.00

4.

27

0.00

0

.03

.00

0.02

0.

00

7 0.

12

0.16

5.

93

0.00

0

.00

.57

0.04

0.

59

NSW

8

0.19

0.

07

5.96

0.

00

-0.0

0 .8

9

0.

01

0.26

7

0.25

0.

04

5.62

0.

00

0.0

0 .9

4 0.

27

0.06

V

icto

ria

8 0.

27

0.03

5.

62

0.00

-0

.00

.97

0.02

0.

04

7 0.

39

0.01

6.

29

0.00

-0

.02

.10

0.45

0.

00

Que

ensl

and

8 0.

31

0.02

6.

29

0.00

-0

.02

.14

0.03

0.

01

7 0.

29

0.02

6.

18

0.00

-0

.02

.01

0.30

0.

01

WA

8

0.51

0.

00

6.27

0.

00

-0.0

3 .0

0

0.

03

0.00

213


For the Commonwealth and Western Australian jurisdictions, explanatory power of

the model is high, with adjusted R2 above 0.8 for the Commonwealth and 0.51

(second model) for Western Australia. The first model is less useful for the Western

Australia Government, with an adjusted R2 of 0.29. Explanatory power for the New

South Wales, Victorian and Queensland Governments is lower, varying between 0.12

and 0.39.

There is a significant positive relationship between public investment and the passing

of time at the five per cent level of significance for the Commonwealth Government.

No significant relationship exists between public investment and the passing of time

for the New South Wales or Victorian Governments. There is a significant negative

relationship between public investment and the passing of time at the five per cent

level of significance for the Western Australian Government and at the ten per cent

level of significance for the Queensland Government.

Again, it is not possible to determine whether the intercept or the slope of the

regression line increases. Instead, the two separate regressions indicate that there is a

significant and positive relationship between either the level or rate of growth (or both)

of public capital formation and adoption of a net worth fiscal policy rule for the

Commonwealth, Queensland and Western Australian Governments (five per cent

level of significance) and the Victorian Government (ten per cent level of

significance). Again, no such significant relationship exists for the New South Wales

Government. These findings generally are in accordance with findings of Section

6.6.1. This was to be expected given the correlation between the passing of time and

increases in GDP/GSP51.

51 An interregnum model was also used with similar results. The passing of time is significant to growth of public capital formation for the Commonwealth, Queensland and Western Australian Governments. In addition, growth of public capital formation increases at adoption of net worth fiscal

214


Comparison of relationships between public capital formation, fiscal policy rules,

economic activity and passage of time

Adjusted R2 are compared in order to determine whether economic activity or

passing of time, with fiscal policy rules adoption, has greater explanatory power.

For the Commonwealth Government, the time trend model discussed in this section

provides larger adjusted R2 (0.85 and 0.86) than those (0.79 for each) for the

economic activity model, as discussed in Section 6.5. The opposite is found for the

Victorian and Western Australian Governments. Both models provide similar results

for the New South Wales and Queensland Governments. These results are not

surprising, given the fairly constant growth rates of real economic activity per capita

in all jurisdictions over time and the likely resulting strong correlation between the

trend and GDP/GSP.

6.7 VARIATION IN PUBLIC CAPITAL FORMATION AND FISCAL

POLICY RULES ADOPTION

This section discusses the flow of analysis of an event study, as described in Chapter

5, in the context of this study. Results are then presented.

6.7.1 Defining the event and outcome of interest

The event of interest is defined as the adoption of fiscal policy rules. The outcome

measured in this study is the fiscal measure specifically affected by adoption of a

policy rules in these jurisdictions. Passage of time is not significant for the New South Wales and Victorian Governments. Nor does growth of public capital formation increase at adoption of a net worth fiscal policy rule by the New South Wales Government. However, growth of public capital formation increases at adoption of a net worth fiscal policy rule by the Victorian Government.

215


fiscal policy rule. Hence, this event study measures the impact of adoption of a net

worth fiscal policy rule on public capital formation.

6.7.2 Identifying the period of examination

The period of examination is the period for which data is available. This is from

1984-85 to 2003-04. The period of examination includes examination of periods

surrounding the event and thus comprises three stages. These are the estimation

window, the event window and the post-event window. The estimation window and

post-event window are positioned in time by reference to the event window.

Following MacKinlay (1997), it is customary to define the event window to be larger

than the specific period in which fiscal policy rules were adopted. As discussed in

Section 5.3.2, the interregnum represents a transitional period between when

governmental behaviour may first have changed and the formal adoption of fiscal

policy rules. Thus the interregnum periods previously described are conceptually

consistent with the event window and are adopted for that purpose. Then the event

window is defined as extending from the beginning of the interregnum up to and

including the year of formal adoption of a net worth fiscal policy rule by each of the

jurisdictions examined. The estimation window is defined as the period prior to the

event window and is used to calculate normal outcomes. The post-event window

follows the event window and is used to calculate abnormal outcomes. Figure 6.4

diagrammatically presents the above timelines.

216


Figure 6.4 Event study timelines

(estimation window) (event window ) (post-event window)

T0 T1 T2 T3

Outcomes are indexed in event time using τ, with τ = T0 to τ = T1 indicating the

estimation window in normalised timelines, τ = T1 + 1 to τ = T2 representing the event

window and τ = T2 + 1 to τ = T3 describing the post-event window. While T0 and T3 are

the same for each jurisdiction, and represent 1984-85 and 2003-04 respectively, T1

and T2 vary between jurisdictions, in accordance with different commencement dates

and lengths of interregnum. The length of the estimation window is defined as L1 =

T1 – T0, the length of the event window is L2 = T2 – T1 and the length of the post-event

window is L3 = T3 – T2. Again, these lengths vary between jurisdictions.

6.7.3 Determining the selection criteria for inclusion in the study

This step is relevant to a study where measurement of abnormal outcomes relies on a

market return-type model and judgements have to be made as to the measure to be

used for market returns. As this study does not use this method of measuring

abnormal outcomes, the step is not necessary for this study.

6.7.4 Selecting a method for measuring normal and abnormal outcomes

Central to an event study is the measurement of normal and abnormal outcomes.

Again following MacKinlay, the normal outcome is defined as the expected outcome

without conditioning on the event taking place. In the context of an event study

measuring the impact on the price of a given security of an earnings announcement,

MacKinlay (1997) at page 15 states that:

217


“There are two common choices for modelling the normal return – the

constant mean return model where Xτ is a constant, and the market model

where Xτ is the market return. The constant mean return model … assumes

that the mean return of a given security is constant through time. The

market model assumes a stable linear relation between the market return

and the security return.”

An analogous constant mean model would assume that mean GFCF is constant over

time. However, this is not supported by perusal of the data series presented in

Section 6.3. While another analogue would be the rate of change in GFCF, this also

is not constant over time (see Section 6.3).

An analogous model to the market model is one in which a stable linear relationship is

assumed between public capital formation of a jurisdiction and the level of GDP/GSP.

This is also not supported by the empirical observation (presented in Section 6.3) that

public capital formation has comprised a decreasing proportion of GDP/GSP over

time. Nonetheless, a market-type model offers potential improvements over a

constant mean outcomes model by removing the portion of variation in GFCF related

to variation in levels of economic activity.

As a consequence, the model employed in this study is:

ln GFCF iτ = αi + β i ln GDP/GSPit + εit

E(εit) = 0

Var (εit) = σ2εi

where:

ln GFCFiτ is the period-t natural log of GFCF real per capita for jurisdiction i

218


ln GD/SPiτ is the period-t natural log of GDP/GSP real per capita for

jurisdiction i

αi is the intercept term for jurisdiction i

βi is the slope parameter for jurisdiction i and

εit is the time period t disturbance term for jurisdiction i, with an expectation

of zero and variance σ2εi.

Under general conditions, ordinary least squares is used to produce consistent and

efficient estimators for the model parameters. This is based on assumptions that fiscal

outcomes are independent and normally distributed. MacKinlay (1997) notes that

while such an assumption is strong, it is empirically reasonable and therefore

generally does not lead to problems in practice, and that, in any case, the normal

outcome models are robust to deviations from the assumption. Estimation of the

market-type model is therefore based on ordinary least squares estimation of

parameters, for the ith jurisdiction in event time.

The abnormal return is the actual ex-post fiscal outcome (measured as the natural log

of real GFCF per capita) over the post-event window minus the normal or expected

outcome over the post-event window, estimated on an out-of-sample basis.

For jurisdiction i and event date τ, the abnormal return is:

ARiτ = Riτ – E(ln GFCF rpc iτ|Xτ)

where:

ARiτ is the abnormal return for time period τ for jurisdiction i,

Riτ is the actual ex-post return for time period τ for jurisdiction i,

E(ln GFCF rpc iτ|Xτ) is the expected, normal return for time period τ for

219


jurisdiction i, and

Xτ is the conditioning information (adoption of net worth fiscal policy rules)

for the normal return model.

6.7.5 Designing the testing framework for abnormal outcomes

The abnormal returns were tested using a framework of a null hypothesis that the

event has no impact on the distribution of fiscal outcomes and an alternative

hypothesis that this is untrue. Each abnormal outcome is thus a disturbance term of

the market model calculated on an out-of-sample basis. Given the assumptions, these

abnormal outcomes will be jointly normally distributed, with zero conditional mean

and conditional variance σ2(ARiτ).

This variance will constitute only disturbance variance and will not contain the usual

additional component of variance due to sampling error in αi and βi. Under H0, the

distribution of the sample abnormal outcome of a given observation in the event

window is ARiτ ~ (N(0, σ2(ARiτ))).

The abnormal outcome observations are considered separately. They are also

aggregated across time and jurisdictions in order to accommodate a multiple period

event window. Consideration on this basis enables overall inferences, regarding the

event of interest, to be drawn. Inferences about the cumulative abnormal outcomes

can also be drawn to test the null hypothesis. The sample variance measure from the

regression is an appropriate choice of estimator for the variance of the abnormal

returns. The null hypothesis is that abnormal returns are zero and the alternative

hypothesis is that this is not true. A finding that abnormal returns are zero would

indicate no change in growth of public capital formation following adoption of net

220


worth fiscal policy rules.

Following MacKinlay (1997) and Patell (1976), the common approach is taken where

each abnormal return is standardized using an estimator of its standard deviation.

When standardised by division by the sample standard deviation, the resulting statistic

is approximately t-distributed, for degrees of freedom equal to the number of years in

the estimation period minus two. Due to the different dates of adoption of fiscal

policy rules and resulting different lengths of time series in the estimation window,

degrees of freedom differ between jurisdictions and for the individual and cumulative

tests, and thus are calculated separately for each jurisdiction and for each test. The

null hypothesis thus can be tested in the usual way, using a one-tailed (right-tailed)

test and decision rule to reject the null hypothesis when the calculated value of the test

statistic exceeds its critical value.

6.7.6 Results

Tables 6.14 and 6.15 respectively present calculated and critical values of the test

statistic, for both the interregnum and non-interregnum models.

221


Table 6.14 Calculated values of the test statistic used in event analyses

Table 6.15 Critical values of the test statistic used in event analyses

Interregnum model CW NSW VIC QLD WA

1997-98 68.36 1998-99 196.08

1999-2000 96.31 2000-01 3.53 103.43 13.61 8.00 2001-02 6.23 141.47 20.42 12.37 5.47 2002-03 10.77 180.61 25.82 5.54 4.31 2003-04 13.65 149.00 24.12 9.63 6.12

Cumulative value 17.09 353.49 40.62 20.57 11.95 Non-interregnum model

CW NSW VIC QLD WA 1994-95 42.06 1995-96 -11.12 1996-97 88.22 3.56 1997-98 37.10 5.59 3.68 1998-99 106.42 6.63 4.06

1999-2000 5.39 52.27 10.38 5.53 2000-01 3.34 56.13 -4.16 7.87 8.99 2001-02 5.90 76.78 16.51 7.15 6.15 2002-03 10.20 98.03 20.88 3.20 4.84 2003-04 12.92 80.87 19.50 5.57 6.88

Cumulative value 16.88 198.20 26.37 17.66 15.17

Interregnum model Item CW NSW VIC QLD WA

Degrees of freedom 13 8 14 10 11Individual abnormal returns Critical value of t 1.771 1.86 1.761 1.812 1.796

Degrees of freedom 73 98 62 94 89Cumulative abnormal returns Critical value of t 1.667 1.662 1.671 1.662 1.664

Non-interregnum model Item CW NSW VIC QLD WA

Degrees of freedom 13 10 14 13 13Individual abnormal returns Critical value of t 1.771 1.812 1.761 1.771 1.771

Degrees of freedom 73 118 62 118 103Cumulative abnormal returns Critical value of t 1.667 1.66 1.671 1.66 1.66

222


For the interregnum model, calculated values of the test statistic are higher than the

critical values for all cumulative abnormal returns and for all except two individual

abnormal returns. Thus, the interregnum model produces a decision to reject the null

hypothesis, implying that after adoption of fiscal policy rules the growth of public

capital formation was altered for all jurisdictions, both when measured by cumulative

returns and by individual returns.

For the non-interregnum cumulative returns model, the null hypothesis is rejected for

all jurisdictions. For the non-interregnum individual returns model, the null

hypothesis is rejected for all jurisdictions in all years except New South Wales in

1995-96 and Victoria in 2000-01.

These results illustrate a particular strength of the event study methodology (non-

interregnum model) in that the methodology enables identification of individual years

for which abnormal levels of GFCF occurred. These years were 1995-96 for the New

South Wales Government and 2000-01 for the Victorian Government. The New

South Wales Government in 1995-96 incurred a large negative value for GFCF due to

transfers of assets to local government authorities and devaluations, as discussed in

Section 4.3. Similarly, in 2000-01, the Victorian Government established a $1 billion

infrastructure reserve with the proceeds of a large budgetary surplus in the preceding

year and adopted fiscal policy rules prescribing the use of those funds. Each of these

occurrences could be expected to effect rates of public capital formation in an

abnormal manner. However, linear regressions carried out on an interregnum model

exclude these data points.

Overall, results of the event study methodology provide confirmation of the linear

regression results. These were that the regression models used were useful in

223


explaining variations in growth of public capital formation by all jurisdictions except

the New South Wales Government and were of mixed usefulness in explaining

variations in growth of public capital formation by the Victorian Government.

Further, the event study methodology provides possible reasons for those findings.

The results imply that adoption of net worth fiscal policy rules has been associated

with a change in the growth of public capital formation by all jurisdictions in all years

except New South Wales in 1995-96 and Victoria in 2000-01.

6.8 CONCLUSION

6.8.1 Did governments comply with rules?

Perusal of fiscal aggregates indicates that governments have generally though not

always complied with fiscal constraints imposed on adoption of fiscal policy rules.

However, greater incidence of compliance was found than non-compliance.

Occurrence of non-compliance with constraints imposed by adoption of fiscal policy

rules is probably related to the absence of penalties for non-compliance with fiscal

policy rules adopted by Australian governments (as described in Section 3.2.1).

The degree of compliance varied with the type of fiscal policy rule, with constraints

imposed by net debt rules met more often than net worth and, in turn, budgetary

balance rules. Possible explanations for this include greater awareness of the need to

accrue financial assets to meet accruing non-debt liabilities after adoption of accrual-

based financial reporting frameworks. In the context of multiple influences on

governmental fiscal outcomes, intentionality is not ascertainable in this study.

6.8.2 Did public capital formation vary after adoption of rules?

Preliminary analyses in Sections 6.4 and 6.5 using cross-sectional time series data

224


indicate, firstly, that growth in public capital formation is independent of the existence

of a net worth fiscal policy rule. Secondly, analysis indicates that positive rates of

growth of GFCF occurred with the same frequency before and after existence of a net

worth fiscal policy rule.

More detailed analyses in Sections 6.6 and 6.7 provide an indication of individual

jurisdictional experiences, as well as numerical differences underlying the relevant

relationships. Regression analyses indicate that, at the time of adoption of fiscal

policy rules, there was an increase in the level, or growth rate, or output elasticity of

public capital formation (or a combination of these) in the Commonwealth, Victorian,

Queensland and Western Australian jurisdictions, but not in the New South Wales

jurisdiction.

For the Commonwealth Government, regression estimates indicate that real public

capital formation per capita increased by about 1.4% when the level of economic

activity increased by 1%. While it is not possible to determine whether it is more

accurate to say that the intercept or slope of the regression line increases, estimates

indicate that there is a significant and positive relationship between either the level or

the output elasticity (or both) of public capital formation and adoption of a net worth

fiscal policy rule.

These regression results contrast with those for the New South Wales Government.

For this jurisdiction, the explanatory power of the models is very poor and the models

are not useful, nore are any of the independent variables significant. This suggests

misspecification of some type, possibly because the New South Wales Government

accumulated financial rather than real assets during the period studied.

Victorian Government results support a conclusion that there is a significant and

225


positive relationship between either the level or the output elasticity (or both) of

public capital formation and adoption of a net worth fiscal policy rule. However,

results do not indicate a significant relationship between levels of economic activity

and public capital formation in Victoria. This may be due to differing infrastructure

financing arrangements adopted in Victoria.

For the Queensland Government, regression analysis indicates that there is a

significant and positive relationship between either the level or the output elasticity

(or both) of public capital formation and adoption of a net worth fiscal policy rule.

However, results indicate a significant, negative relationship between public capital

formation and economic activity, possibly because of a tendency by the Queensland

Government to increase accumulation of financial assets rather than real assets

(similarly to the case of the New South Wales Government).

Similarly to the case of Queensland, for the Western Australian Government, results

indicate that public capital formation and economic activity were inversely related,

again possibly because of greater accumulation of financial assets than real assets.

Results also indicate a significant, positive relationship between either the level or the

output elasticity (or both) of public capital formation and adoption of a net worth

fiscal policy rule by the Western Australian Government.

Comparison of relationships between public capital formation, fiscal policy rules,

economic activity and passage of time indicated that economic activity, with fiscal

policy rules adoption, has greater explanatory power or passing of time for the

Commonwealth Government. The opposite is found for the Victorian and Western

Australian Governments. Both models provide similar results for the New South

Wales and Queensland Governments. These results are not surprising in light of

226


observed fairly constant growth of real economic activity per capita in all jurisdictions

over time.

Event analysis confirms findings of the regression analysis, and further specifies the

years in which the New South Wales Government’s investment led to these results.

Event analysis also indicates why results for the Victorian Government require

acceptance of a higher level of significance for type II error than do other

governments, specifying a year in which results differed from all other years.

The finding of analyses at an individual jurisdictional level is that adoption of net

worth fiscal policy rules has coincided with changed rates of public capital formation.

The implication is that adoption of fiscal policy rules may have influenced

governmental decision-making pertaining to public capital formation. This is

significant when considered in the context of private production elasticities related to

public capital stocks and the resulting changes in levels of economic activity.

This chapter set out to ascertain what happened to public capital formation when

Australian governments adopted fiscal policy rules requiring maintenance or increase

of net worth. Scrutiny of fiscal measures provided conclusions that most

governments complied most of the time with the constraints imposed by those fiscal

policy rules. Application of inter-linking statistical inference and econometric

techniques provided further detail. Initial analyses indicated that when all

jurisdictions are considered together, there is limited evidence of changes in the rate

of public capital formation coinciding with adoption of net worth fiscal policy rules.

However, analysis of individual jurisdictions indicates that a main contributor to this

outcome may be the experience of the New South Wales, and to a lesser extent,

Victorian Governments. The New South Wales Government had a strikingly different

227


experience of public capital formation to that of the other governments. This was in

part due to its transfer of fixed capital to other levels of government and devaluation

of fixed capital. The Victorian Government had a slightly different experience, in

part due to arrangements adopted for financing infrastructure acquisitions. In general,

more sophisticated analysis indicated that rates of public capital formation increased

with adoption of net worth fiscal policy rules.

The findings differ by level and are generally stronger for the national jurisdiction

than for the state jurisdictions. It is possible that a more useful specification for state

governments would take into account the composition of those governments’ revenue

bases, specifically the extent to which state government revenues are sourced from

economic activity within the state and from the Commonwealth Government, revenue

for which in turn is linked with economic activity across all states.

These findings have some significance for the current state of understanding of the

usefulness of fiscal policy rules. Of further significance is the potential of these

findings for contribution to a new, more detailed treatment of government in

macroeconomic theory. An overarching theory of the effects of fiscal policy rules on

governmental fiscal behaviour is unavailable as mainstream macroeconomics does not

investigate determinants of government outlays in a detailed manner, the convention

being to treat them as exogenous. Analysis of public capital formation after adoption

of net worth fiscal policy rules therefore cannot draw upon an existing body of theory

and can neither validate nor invalidate such a body of theory.

This study attempts to mitigate the current treatment of government expenditures by

making findings regarding variations in public capital formation in the context of

fiscal policy rules adoption, without claiming causality between fiscal policy rules

228


adoption and those variations. The importance of pursuing such a line of

investigation can be inferred by reference to the developing literature on private

production elasticities to public capital formation, which is described in Chapter 2.

It is interesting to note that those jurisdictions experiencing increased growth of

public capital formation subsequent to adoption of fiscal policy rules also have

experienced higher growth rates than New South Wales. (While there is an obvious

problem of causation here, the very small (and declining) proportion of GDP/GSP

represented by GFCF means it is likely that this holds even with GFCF ‘stripped out’

of the GDP/GSP measures, though such analysis is beyond the scope of this study).

This is consistent with the findings of Aschauer (1988 and 1989) and Otto and Voss

(1994), described in Chapter 2, that public capital formation has a positive impact on

private sector production.

229

Chapter 7: Conclusion

CHAPTER 7

CONCLUSION

7.1 INTRODUCTION

This study investigated whether public capital formation by five Australian

governments changed at the time of their adoption of net worth fiscal policy rules. In

order to investigate this, firstly the degree to which governments complied with the

constraints imposed on fiscal measures by adoption of fiscal policy rules was

ascertained (Chapter 6). Due to the simultaneous existence of multiple identifiable

influences on governmental fiscal outcomes, intentionality of compliance where

observed cannot be ascertained (see Chapter 3).

It was found that the examined governments experienced a high level of compliance

with fiscal constraints imposed by adoption of fiscal policy rules. Perusal of the

institutional arrangements surrounding rules adoption indicated that the absence of

penalties for non-compliance may have contributed to the occasional occurrences of

non-compliance.

The degree of compliance varied with the type of fiscal policy rule. Net debt rules

were more frequently met than were net worth rules and, in turn, budgetary balance

rules. This may be due to the significantly enhanced information set available to

governments after adoption of accrual-based financial reporting networks (described

in Chapter 4), allowing greater awareness of asset and liability aggregates at the time.

Another probable contributing factor was the significance attributed by governments

to their credit ratings and, in turn, the focus on net debt by credit ratings agencies in

assessing those ratings.

230


Attention was then focused on the experience of public capital formation and whether

it changed at the date of adoption of net worth fiscal policy rules. Preliminary non-

parametric analyses indicated that this was not so and that positive rates of growth of

public investment were not dependent on existence of a net worth fiscal policy rule.

Further, the frequency of occurrence of positive rates of growth of public investment

was unchanged in the periods before and after governments adopted a net worth fiscal

policy rule.

A variety of techniques, described in Chapter 5, were used to investigate the research

question. This strengthened the overall analysis by overlapping examination and

mitigated the weaknesses of individual tests. It also permitted different perspectives

to be brought to bear on the issue.

In Chapter 6, parametric techniques were used to provide a more detailed view,

providing an indication of individual jurisdictional experiences and estimating

numerical relationships. The Commonwealth, Victorian, Queensland and Western

Australian Governments increased the level, growth rate or output elasticity of their

investment when they adopted a fiscal policy rule requiring, at a minimum, that they

maintain their net worth. The New South Wales Government did not. Its investment

simply cannot be modeled in this way. This is in part because of its history of transfer

of fixed capital to other levels of government and of capital devaluations. The

Victorian Government’s results require acceptance of a higher level of significance

for type II error than do other governments. The Victorian Government’s investment

experience, while successfully modeled in this way, showed the effects of unique

infrastructure financing arrangements which weakened the relationship between

economic activity and investment. These findings are confirmed by event analysis

231


which also specifies the years in which the New South Wales and Victorian

Governments’ investment experiences led to these results.

Importantly, these findings imply that the usual macroeconomic assumption of

exogeneity of government expenditures may be too strong in circumstances where

governments have adopted such fiscal policy rules. Specifically, it becomes

necessary to review the general assumption, described in Chapter 2, that only certain

elements of government expenditures, those that are related to automatic stabilisers,

are business cycle dependent. The remainder of government expenditure, usually

considered to be independent of levels of economic activity, can no longer be

considered to be so when certain institutional arrangements, such as fiscal policy rules,

exist. Instead, constraints imposed by adoption of fiscal policy rules assume the

position of determining upper or lower bounds on certain fiscal measures. However,

Chapter 3 establishes that it is a complicating feature of reporting frameworks that

increases in net worth do not necessarily lead to increases in public investment but

can instead be achieved by increasing financial assets.

This is despite the absence of penalties of a financial or judicial nature. The examined

governments possibly incurred reputational penalties only. It is unclear whether these

penalties were sufficient to motivate the high rates of compliance that were observed.

This is of some significance to the ongoing exploration of the desirability of

discretionary versus rules-based fiscal policy regimes, since one of the assumed

detractions of fiscal policy regimes is their usual lack of enforceability. The findings

of this study provide an indication of the strength of reputational penalties alone.

These may not be limited to electoral effects but may also include credit market

effects via impacts on fiscal policy validity.

232


The level of government under consideration is also significant in the context of these

findings. The findings hold regardless of the level of government considered i.e.

whether it is a national or sub-national government under consideration. As

automatic stabilisers are generally seen to affect only national governments, because it

is generally the national level of government in a federal system that bears

responsibility for transfer payments and collects income-dependent taxes, it has

usually been considered that sub-national governments’ expenditures are less affected

by business cycle stages than are national governments. However, in Australia, there

is a strong correlation between State government revenues and the business cycle.

Thus it is shown that adoption of certain fiscal policy rules may in part generate a

transmission mechanism between levels of economic activity and levels of

governmental outlays. Further, this appears to be unrelated to the stringency of any

penalties incurred by non-compliance with adopted fiscal policy rules.

This is significant because it indicates the potential of institutional arrangements to

foster sustainable fiscal policymaking. It is possible that the potency of the potential

sanctions imposable by the credit markets in the form of risk premia on public debt, as

well as the potential electoral impact, are the key causes of the findings of this study.

Therefore, consideration of the totality of institutional conditions should be made

before any decision is taken to adopt fiscal policy rules. In particular, when

international agencies such as the International Monetary Fund recommend fiscal

policy rules adoption to governments who are potential recipients of financial aid,

consideration of such issues should be a part of any assessment of the likely

enforceability and thus efficacy of fiscal policy rules.

Further, consistent with the literature on supply-side effects of public capital

233


formation, the jurisdictions experiencing increased growth of public capital formation

subsequent to adoption of fiscal policy rules are those which have experienced higher

growth rates than other jurisdictions. While this study does not extend to an

exploration of the existence and nature of a relationship between public capital

formation and economic growth, such further research is recommended in the

following section. This is because of the potential significance particularly for

governments of developing nations, which typically face continual constraints on

expenditure, placing a premium on optimising expenditure decisions as they attempt

to improve the living standards of their populations.

7.2 RESEARCH DIRECTIONS

This study suffers from data deficiencies. Due to the relatively recent adoption of

accrual-based reporting methodologies, it was necessary to transform accrual data into

cash-based equivalencies and splice the two datasets to create a series sufficiently

long to enable econometric analysis. As a result, the results must be interpreted with

caution. In particular, findings cannot be considered to represent e.g. levels of public

capital formation although trends and change are ascertainable. In addition, the

transformation introduces unavoidable difficulties, being reliant for its validity on the

proposition that certain elements of public capital formation, recognizable under an

accrual-based methodology, comprised a similarly negligible proportion of total

public capital formation before and after the dates of adoption of accrual-based

methodologies. A longer accrual-based data series would provide a greater level of

comfort with respect both to the transformation and for econometric analysis

generally. Hence, a study completed after the passage of some time would provide

greater analytical reliability than is possible to achieve in this study.

234


This study does not extend to an exploration of the impacts of increased public capital

formation on economic growth. However, the study identified that the high-growth

states (Queensland and Western Australia) experienced increases in the level or rate

of growth of public investment, while the low-growth state of New South Wales did

not, and the low-growth state of Victoria did but to a lesser extent. This is an

accordance with Aschauer (1988 and 1989) and Otto and Voss (1994) who

determined that public capital formation has a positive impact on private sector

production. In particular, Otto and Voss (1994) used Australian data, albeit at the

national level. No study has yet been carried out at the state level in Australia. An

obvious direction for subsequent research is to carry out such an analysis at the state

level for the above states.

In Chapter 6, it was pointed out that, due to the simultaneous existence of multiple

identifiable influences on governmental fiscal outcomes, it is not possible to

determine with certainty the cause of the observed changes in fiscal behaviour. For

example, an information effect may exist i.e. improved information made available by

adoption of accrual-based reporting methodologies may have influenced

governments’ decision-making. Exploration of this issue may produce useful

information by which to inform decisions of governments still using cash-based

reporting methodologies.

Therefore, it would be useful to examine the impact of fiscal policy rules adoption in

jurisdictions, having similar institutional arrangements and fiscal circumstances as

Australia, such as New Zealand and Canada. Such studies could cast further light on

the likely causes of the Australian experience as well as providing a basis for

international comparison. For example, the level of compliance experienced could be

235


considered in the light of any penal arrangements in order to provide an indication of

the effectiveness of reputational or other penalties. If public capital formation were

found to increase at the time of adoption of accrual-based financial reporting

frameworks rather than at the time of adoption of fiscal policy rules that focused on

net worth, it could be concluded that the information effect may outweigh the

reputational and credit market effects. This would substantially extend the current

level of awareness of the effects of fiscal policy rules adoption on public fiscal

outcomes.

Further, the integrated nature of the financial reporting methodologies (by which are

measured the financial aggregates upon which fiscal policy rules impose constraints)

means that increases in net worth can be achieved by increasing financial assets rather

than by increasing capital formation. The high rate of compliance with net debt rules

and the emphasis that the examined jurisdictions placed upon achieving favourable

credit ratings indicates that a relationship may exist between adoption of net debt

rules and the rate of financial assets accumulation or of debt retirement. A fruitful

line of enquiry could include exploration of the existence and nature of this

relationship.

236


LIST OF REFERENCES

Abelson, P. (2003), Public Economics Principles and Practice, Applied Economics Australia

Alankar, A.G. (2003), ‘Essays in Latent Variable and Event Study Econometrics’, PhD thesis, University of California, Berkeley

Alesina, A. and Perotti, R. (1995), ‘The Political Economy of Budget Deficits’, IMF Staff Papers, Vol. 42, No. 1

Andres, J. and Domenech, R. (2003), ‘Automatic Stabilizers, Fiscal Rules and Macroeconomic Stability’, Documento de Trabajo, n. 0314, Banco de Espana

Aschauer, D. (1988), ‘The Equilibrium Approach to Fiscal Policy’, Journal of Money, Credit and Banking, Vol. 20, pp 41-62

Aschauer, D.A. (1989), ‘Is Public Expenditure Productive?’, Journal of Monetary Economics, Vol. 23, pp 177-200

Aschauer, D. (1997), ‘Do States Optimise? Public Capital and Economic Growth’, Working Paper No. 189, The Jerome Levy Economics Institute, Annandale-on-Hudson, NY

Aschauer, D.A. (2001), ‘Output and Employment Effects of Public Capita’, Public Finance & Management, Vol. 2, No. 2, pp 135-160

Auerbach, A. and Kotlikoff, L. (1987), Dynamic fiscal policy, Cambridge University Press, New York, NY Australian Bureau of Statistics (2000), ‘Information Paper: Accruals-based Government Finance Statistics’, Cat. No. 5517.0, AGPS, Canberra

Australian Bureau of Statistics (2005a), Government Finance Statistics, Cat. No. 5512.0, AGPS, Canberra

Australian Bureau of Statistics (2005b), Australian National Accounts: National Income, Expenditure and Product, Cat. No. 5206.0, AGPS, Canberra

Baker, H.K. and Mansi, S.A. (2002), ‘Assessing Credit Rating Agencies by Bond Issuers and Institutional Investors’, Journal of Business Finance & Accounting, Vol. 29, Nos. 9 and 10, pp 1367-1398

Barro, R.J. (1984), Macroeconomics, John Wiley & Sons, New York, NY

Barsky, R.B, Mankiw, N.G. and Zeldes, S.P. (1986), ‘Ricardian Consumers with Keynesian Propensities’, American Economic Review, Vol. 76, No. 4, pp 676-691

Barton, A. (2003 ), ‘The Department of Defence – Australia’s most profitable business?’, Australian Accounting Review, Vol. 13, No. 2, pp 35-40

237


Bartos, S (2000), ‘The ‘whole’ truth: The DoFA view’, Australian CPA, Vol. 70, Iss. 3, pp 67-70

Baunsgaard, T. (2001), ‘Fiscal Policy in Nigeria: Any Role For Rules?’, IMF Working Paper, WP/03/155

Besley, T. and Ghatak, M. (2001), ‘Government Versus Private Ownership on Public Goods’, The Quarterly Journal of Economics, pp 1343-1372

Bhagat, S. and Romano, R. (2002 ), ‘Event Studies and the Law: Part I: Technique and Corporate Litigation’, American Law and Economics Review, Vol. 4, No. 1, pp 141-167

Bohn, B. (1992), ‘Budget deficits and government accounting’, Carnegie-Rochester Conference Series on Public Policy, Vol. 37, pp 1-84

Bosshardt, D.I. (2003), ‘Capital structure, investment unanimity, and public goods: the case for social responsibility’, The Quarterly Review of Economics and Finance, Vol. 43, pp 239-260

Boxall, P. (1998), ‘The Revolution in Government Accounting’, Australian CPA, Vol. 68, No. 3, pp 18-20

Brunila, A. (2002), ‘Fiscal policy: Coordination, discipline and stabilization’, Bank of Finland Discussion Papers, 7

Buiter, W.H. (1997), ‘Generational Accounts, Aggregate Savings and Intergenerational Distribution’, Economica, No. 64, pp 605-626

Buiter, W.H. (2001), ‘Notes on ‘A Code For Fiscal Stability’’, Oxford Economic Papers, 53, pp 1-19

Carne, K. (2004), ‘Towards strategic policy rules for public finance’, Proceedings of the Inaugural Conference of the Academy of World Business, Marketing and Development, July 2004.

Churchill, M. (1992), ‘Accrual Accounting in the Public Sector’, Australian Accountant, June

Commonwealth Government of Australia (2000), Accrual Uniform Presentation Framework, Canprint Communications, Canberra

Commonwealth of Australia (1996), Budget Statement 1, Commonwealth Budget Papers 1996-97, AGPS, Canberra

Commonwealth of Australia (1997), Budget Paper No. 1, Commonwealth Budget Papers 1997-98, AGPS, Canberra

Commonwealth of Australia (1998), Budget Paper No. 1, Commonwealth Budget Papers 1998-99, Canprint Communications, Canberra

238


Commonwealth of Australia (1999), Budget Paper 1, Commonwealth Budget Papers 1999-00, Canprint Communications, Canberra





Commonwealth of Australia (1996), Charter of Budget Honesty, Commonwealth Budget Papers 1996-97, AGPS, Canberra

Courant, P.N., Gramlich, E.M. and Rubinfeld, D.L. (1979), ‘The Stimulative Effects of Intergovernmental Grants: Or Why Money Sticks Where it Hits’, ‘Fiscal Federalism and Grants-in-Aid’, Coupe Papers on Public Economics, The Urban Institute, Washington

Courant, P.N. and Rubinfeld, D.L. (1990), ‘Tax Reform: Implications for the State-Local Public Sector’, Readings in Public Sector Economics, D C Heath and Company

Creel, J. (003), ‘Ranking Fiscal Policy Rules: the Golden Rule of Public Finance vs. the Stability and Growth Pact’, OFCE, No. 2003-04

Dadgostar, B. and Mirabelli, F. (1998),‘Does Government Public Capital Expenditure Matter? Evidence for Canada’, Economic Analysis & Policy, Vol. 28, No. 2, pp 199-211

Davis, M. (2006), ‘Accounting debate divides profession’, The Australian Financial Review, 4 April, p10

Davis, M. (2006), ‘Treasury chiefs revolt over budget rules’, The Australian Financial Review, 3 April, p3

Deegan, C. (2005), Australian Financial Accounting 4th edition, Mc Graw-Hill

Department of Treasury and Finance, Tasmanian Government (2003), ‘Accrual Budget and Financial Management Realisation Plan – Phase 4 – Version 3.0’ accessed 6 December 2005 at http://www.treasury.tas.gov.au/domino/dtf/dtf.nsf/LookupFiles/ORP_Phase_4_version_3.0_December_2003.pdf/$file/ORP_Phase_4_version_3.0_December_2003.pdf

Dornbusch, R., Bodman, P., Crosby, M., Fischer, S. and Startz, R. (2006), Macroeconomics 2nd Edition, McGraw Hill

Faini, R. (2006), ‘Fiscal policy and interest rates in Europe’, Economic Policy Cambridge, Vol. 21, Issue 47, pp443-451

239

http://www.treasury.tas.gov.au/domino/dtf/dtf.nsf/LookupFiles/ORP_Phase_4_version_3.0_December_2003.pdf/$file/ORP_Phase_4_version_3.0_December_2003.pdf�

http://www.treasury.tas.gov.au/domino/dtf/dtf.nsf/LookupFiles/ORP_Phase_4_version_3.0_December_2003.pdf/$file/ORP_Phase_4_version_3.0_December_2003.pdf�


Financial Administration and Audit Act (Qld) (1977)

Financial Management (Financial Responsibility) Act (Vic) (2000)

Gans, J., King, S., Stonecash, R. and Mankiw, N.G. (2003), Principles of Economics 2nd Pacific Rim Edition, Harcourt

General Government Debt Elimination Act No. 83 (NSW) (1995)

Glater, J.D. (2003), ‘Top Debt Rating Agencies Take a Look at Accounting’, New York Times, Jan 11, (Late Edition) (East Coast), New York, p C1

Government Financial Responsibility Act (WA) (2000)

Gramlich, E.M. (1994), ‘Infrastructure Investment: A Review Essay’, Journal of Economic Literature, Vol. XXXII, pp 1176-1196

Gujarati, D.N. (2006), Essentials of Econometrics 3e, McGraw-Hill/Irwin, New York

Guthrie, J. (1998), ‘Application of Accrual Accounting in the Australian Public Sector – Rhetoric or Reality’, Financial Accountability & Management, 14 (1), pp 1-19

Hakes, D.R. and McCormick, P. (1996), ‘Did Smith Address Ricardian Equivalence Before Ricardo?’, Eastern Economic Journal Bloomsburg, Vol. 22, Iss. 1, pp 57-61

Harris, T. (2004), ‘No progress without debt’, Australian Financial Review, 4 January, p 46

Hemming, R., Kell, M. and Mahfouz, S. (2002), ‘The effectiveness of fiscal policy in stimulating economic activity – A review of the literature’, IMF Working Paper, WP/02/208, December

Hill, R.C., Griffith, W.E. and Judge, G.G. (2001), Undergraduate Econometrics 2nd Edition, John Wiley and Sons Inc, NewYork

Hines, J.R. and Thaler, R. (1995), ‘Anomalies: The Flypaper Effect’, The Journal of Economic Perspectives, Vol. 9, No. 4, pp 217-226

H.M. Treasury (1998), The Code for Fiscal Stability - Pre-Budget Report Publications, London

Holtz-Eakin, D. (1993), ‘State-specific estimates of state and local government capital’, Regional Science and Urban Economics, Vol. 23, pp 185-209

Huber, B. (1992), ‘Time consistency of government financial policy in a small open economy’, European Economic Review, Vol. 36, pp 1545-1555

Inman, R.P. and Rubinfeld, D.L. (1996), ‘Designing tax policy in federalist economies: An overview’, Journal of Public Economics, Vol. 60, pp 307-334

International Monetary Fund (2001), Manual on Fiscal Transparency accessed on 21 January 2007 at http://www.imf.org/external/np/fad/trans/manual/sec03a.htm

240


International Monetary Fund (2004), Transcript of an IMF Book Forum Washington DC accessed on 17 January 2005 at http://www.internationalmonetaryfund.org/external/np/tr/2004/tr041006.htm

Johnstone, D. (1999), ‘Public-Sector Outsourcing and Charging In-house Bidders for Their use of Capital’, Australian Accounting Review, Vol. 9, No. 1, pp 33-41

Kaufmann, B. (2002), ‘Harmonisation for government reporting’, Australian CPA, Vol. 72, Iss. 8

Kelly, J. and Wanna, J. (2004), ‘Crashing Through with Accrual-Output Price Budgeting in Australia - Technical Adjustment or a New Way of Doing Business?’, American Review of Pubic Administration, Vol. 34, No. 1, pp 94-111

Khalid, A.M. (1996), ‘Ricardian equivalence: Empirical evidence from developing economies’, Journal of Development Economics, Vol. 51, pp 413-432

King, S. (1997), ‘National Competition Policy’, The Economic Record, Vol. 73, No. 222, pp 270-284

King, S. and Pitchford, R. (1998), ‘Privatisation in Australia: Understanding the Incentives in Public and Private Firms’, The Australian Economic Review, Vol. 31, No. 4, pp 313-28

Kneebone, R.D. (2003), ‘Past, Current, and Future Fiscal Policy Issues in Canada’, Kohsaka, A. (ed) Fiscal Policy Issues in the Pacific Region, Routledge, London

Kopits, G. and Symansky, S. (1998), ‘Fiscal Policy Rules’, International Monetary Fund Occasional Paper 162

Kotlikoff, L.J., Razin, A. and Rosenthal, R.W. (1990), ‘A Strategic Altruism Model in Which Ricardian Equivalence Does Not Hold’, The Economic Journal, Vol. 100, pp 1261-1268

Levtchenkova, S. and Petchey, J. (2000), ‘Regional Capital Stock Data for Australia’ in Australian Economic Review, Vol. 33, No. 2, pp 193-197

Lougheed, A.L. (2001), ‘A Negative Fiscal Multiplier?’, University of Queensland Discussion Paper No. 293

Loukoianova, E.,Vahey, S.P. and Wakerly, E.C. (2003), ‘A Real Time Tax Smoothing Based Fiscal Policy Rule’, Cambridge Working Papers in Economics, 0235, Faculty of Economics (formerly DAE), University of Cambridge

McCallum, B.T. (1993), ‘Unit Roots in Macroeconomic Time Series: Some Critical Issues’, Federal Reserve Bank of Richmond Economic Quarterly, Volume 79/2, pp 13-41

MacKinlay, A.C. (1997), ‘Event Studies in Economics and Finance’, Journal of Economic Literature, Vol. XXXV, pp 13-39

McTaggart, D., Findlay, C. and Parkin, M. (2006), Economics 4th Edition, Pearson

241

http://ideas.repec.org/s/cam/camdae.html�


McWilliams, A. and Siegel, D. (June 1997), ‘Event Studies in Management Research: Theoretical and Empirical Issues’, Academy of Management Journal, Vol. 30, Issue 3, pp 626-656

Mikhailitchenko, S., Nguyen, D.T. and Smith, C. (2005), ‘Estimates of Capital Stocks for the States and Territories of Australia, 1985-2004’, Economic Analysis and Policy, Volume 35, Nos. 1 & 2, pp 1-22

Milesi-Ferretti, G.M. (1997), ‘Fiscal Rules and the Budget Process’, Centre for Economic Policy Research Discussion Paper, No. 1664

Moll, .J, Humphrey, C. and Chow, D. (2004), ‘Questions of Value: An Analysis of Experiences with Whole of Government Accounting in the UK, New Zealand and Australia’, Paper presented at the APIRA 2004 Conference

Monadjemi, M.S. (1991), ‘The interest rate neutrality of fiscal deficits: testing for Ricardian equivalence and capital inflow’, Journal of International Money and Finance, Vol. 10, pp 541-551

Moody’s Investors Service (2004), Moody’s Ratings System accessed on 9 October 2006 at http://www.moodys.com

Moody’s Investors Service (1998), ‘Subnational Governments: A Rating Agency Perspective – Rating Methodology’ accessed on 9 October 2006 at http://www.moodys.com

National Commission of Audit (1996), ‘Report to the Commonwealth Government (Australia)’ accessed on 9 February 2005 at http://www.finance.gov.au/pubs/ncoa/appg.htm

New South Wales Government (1990), Budget Paper 2 - Budget Statement 1990-91, Government Printer, Sydney






New South Wales Government (1997a), Budget Paper 2 - Budget Statement 1997-98, Government Printer, Sydney

242

http://www.moodys.com/�

http://www.moodys.com/�

http://www.finance.gov.au/pubs/ncoa/appg.htm on 9 February 2005�








New South Wales Government (1997b), NSW Budget Sector Consolidated Financial Statements July-December 1996 accessed on 21 June 2006 at http://www.treasury.nsw.gov.au/pubs/treasure/finstat.htm

New South Wales Government (1995), NSW Treasurer’s Budget Speech 1995-96 accessed on 28 November 2005 at http://www.treasury.nsw.gov.au/bp95/budspch.htm

New South Wales Government (1997c), Public Accounts of the NSW Public Sector 1996-97 accessed on 21 June 2006 at http://www.treasury.nsw.gov.au/pubaccts.htm

New South Wales Government (2001b), New South Wales Report on State Finances 2000-01, Government Printer, Sydney

New South Wales Government (2002b), New South Wales Report on State Finances 2001-02, Government Printer, Sydney

New South Wales Government (2003), New South Wales Report on State Finances 2002-03, Government Printer, Sydney

New South Wales Government (2004), New South Wales Report on State Finances 2003-04, Government Printer, Sydney

New South Wales Treasury (1999), General Government Financial Statement for the Twelve Months ended 30 June 1999, Government Printer, Sydney

Nourzad, F. (2000), ‘The productivity effect of government capital in developing and industrialized countries’, Applied Economics, Vol. 32, pp 1181-1187

Organisation for Economic Cooperation and Development (2002), Economic Outlook No. 72, Paris

243

http://www.treasury.nsw.gov.au/pubs/treasure/finstat.htm�

http://www.treasury.nsw.gov.au/bp95/budspch.htm�

http://www.treasury.nsw.gov.au/pubaccts.htm�


Organisation for Economic Cooperation and Development (2003), Economic Outlook No. 74, Paris

Otto, G.D. and Voss, G.M. (1996), ‘Public Capital and Private Production in Australia’, Southern Economic Journal, Vol. 62, No. 3pp 723-738

Otto G and Voss G M (1994), ‘Public Capital and Private Sector Productivity’, Economic Record, Vol. 70, No. 209, pp 121-132

Patell, J.M. (1976), ‘Corporate Forecasts of Earnings Per Share and Stock Price Behaviour: Empirical Test’, Journal of Accounting Research, Vol. 14, No. 2, pp 246-276

Pindyck, R.S. and Rubinfeld, D. L. (1992), Microeconomics, Second Edition, MacMillan

Poon, W.P.H. and Firth, M. (2005), ‘Are Unsolicited Credit Ratings Lower? International Evidence From Bank Ratings’, Journal of Business Finance & Accounting, Vol. 32, Nos. 9 and 10, pp 1741-1771

Poterba, J. (1996), ‘Do Budget Rules Work?’, Burch Working Paper, No. B96-12

Queensland Government (1991), Budget Overview Budget Paper No. 2 1991-92, Goprint, Queensland










244




Queensland Government (2003), Budget Strategy and Outlook, Budget Paper No. 2 State Budget 2003-04, Goprint, Queensland

Queensland Government (2006), Budget Strategy and Outlook, Budget Paper No. 2 State Budget 2006-07, Goprint, Queensland

Queensland Government (2001), Charter of Social and Fiscal Responsibility, Goprint, Queensland

Queensland Government (2003), Whole-of-Government Approach for Establishing Asset Information Requirements for Asset Financial and Asset Management Reporting and Asset Strategic Planning, Goprint, Queensland

Quiggin, J. ( 1995), ‘Whole-of-Government Accounting and Pubic Sector Net Worth – Submission to Joint Committee on Public Accounts, Parliament of Australia’ accessed on 15 October 2005 at https://mail01.domino.gu.edu.au/mail101/s788195.nsf

Robinson, M. (2002a), ‘Accrual Accounting and Australian Fiscal Policy’, Fiscal Studies, Vol. 23, No. 2, pp 287-300

Robinson, M. (1998), ‘Accrual Accounting and the Efficiency of the Core Public Sector’, Financial Accountability & Management, Vol. 14, No. 1, pp 21-37

Robinson, M. (2002b), ‘Financial Control in Australian Government Budgeting’, Public Budgeting & Finance, pp 80-93

Samuel, S.N. and Gupta, D.B. (1993), Issues in Applied Economics: An Australian Text, MacMillan Education Australia Pty Ltd

Seddighi, H.R., Lawler, K.A. and Katos, A.V. (2000), Econometrics A Practical Approach, Routledge

Selvanathan, A., Selvanathan, S., Keller, G. and Warrack, B. (2004), Australian Business Statistics 3rd Edition, Nelson Australia

Sheskin, D.J. (2004), Handbook of Parametric and Nonparametric Statistical Procedures Third Edition, Chapman & Hall

South Australian Government (1998), Budget Statement No. 2 1998-99, Government Printer, Adelaide


245

https://mail01.domino.gu.edu.au/mail101/s788195.nsf�





Standard & Poors (2004), Credit Ratings – Credit Reports Victoria (State of)

Standard & Poors (2003), Victoria’s ‘AAA’ Rating : A Comparative Study of Financial and Economic Performance McGraw-Hill

Stiglitz, J.E. (1988), Economics of the Public Sector Second Edition, W W Norton & Company

Stiglitz, J.E. (1997), Economics, Second Edition, W W Norton & Company

Tanner, E. (2004), ‘Fiscal rules and countercyclical policy: Frank Ramsey meets Gramm-Rudman-Hollings’, Journal of Policy Modelling, Vol. 26, pp 719-731

Tanzi, V. (2005), ‘Fiscal Policy and Fiscal Rules in the European Union’, CESifo Forum Munchen, Vol. 6,Iss.3, pp57-64

Tasmanian Government (1998), Budget Paper No. 1 Budget Overview 1998-99, Government Printer’s Office, Hobart





Tasmanian Government (2003), Budget Paper No. 1 Budget Overview2003-04, Government Printer’s Office, Hobart

Tridimas, G. (2001), ‘The economics and politics of the structure of public expenditure’, Public Choice, Vol. 106, pp 299-316

Victorian Government (1998), Budget Statement 1998-99 accessed on 21 September 2004 at http://www.dpc.vic.gov.au/domino/web_notes/budgets/budget98.nsf/9bcce3ea1

246

http://www.dpc.vic.gov.au/domino/web_notes/budgets/budget98.nsf/9bcce3ea1�


Victorian Government (1999), Budget Statement 1999-2000, Government Printer for the State of Victoria

Victorian Government (2000a), Budget Statement 2000-01, Government Printer for the State of Victoria

Victorian Government (2001a), Budget Statement 2001-02, Government Printer for the State of Victoria

Victorian Government (2002a), Budget Statement 2002-03, Victorian Government Printer

Victorian Government (2003a), Budget Statement 2003-04, Victorian Government Printer

Victorian Government (2000b), Financial Report for the State of Victoria 1999-2000, No. 39, Session 2000, Government Printer for the State of Victoria

Victorian Government (2001b), Financial Report for the State of Victoria 2000-2001, No. 90, Session 2001, Government Printer for the State of Victoria

Victorian Government (2002b),Financial Report for the State of Victoria 2001-2002, No. 182, Session 1999-2002, Victorian Government Printer

Victorian Government (2003b), Financial Report for the State of Victoria 2002-2003, No. 28, Session 2003, Victorian Government Printer

Victorian Government (2004), Financial Report for the State of Victoria 2003-2004, No. 91, Session 2003-2004, Victorian Government Printer

Western Australian Government (2000), 2000-01 Budget Economic and Fiscal Outlook, Government Printer, Perth




Western Australian Government (1997), Budget Paper No. 2 1997-98, Government Printer, Perth

Western Australian Government (1998), Budget Paper No. 2 1998-99, Government Printer, Perth

Western Australian Government (1992), Budget Speech 1992-93, Government Printer, Perth

247






Western Australian Government (1999), 1999-2000 Budget Summary and Financial Strategy, Government Printer, Perth

Williams, R. (2005), ‘Fiscal Federalism: Aims, Instruments and Outcomes The 2005 Giblin Lecture given on behalf of ANZAAS Presented at the Conference of Economists, University of Melbourne, 27 September 2005’

Williamson, J. (1985), ‘On the System in Bretton Woods’, American Economic Review, Vol. 75, No. 2 , Papers and Proceedings of the Ninety-Seventh Annual Meeting of the American Economic Association, pp 74-79

Ylagan, A.H.C. (2005), ‘Corporate Watch’, BusinessWorld Manila, 1 Aug, p1

248

FISCAL POLICY RULES AND PUBLIC CAPITAL FORMATION IN AUSTRALIA · PDF file ·...

Documents

Transcript of FISCAL POLICY RULES AND PUBLIC CAPITAL FORMATION IN AUSTRALIA · PDF file ·...