OECD COMPENDIUM OF PRODUCTIVITY INDICATORS · 2021. 4. 25. · 8 OECD Compendium of Productivity...

OECD COMPENDIUM OF PRODUCTIVITY INDICATORS

2008

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT

2 OECD Compendium of Productivity Indicators ©OECD 2008

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT

Pursuant to Article 1 of the Convention signed in Paris on 14th

December 1960, and which came into force on 30

th September 1961, the Organisation for Economic Co-operation and Development (OECD) shall promote

policies designed:

To achieve the highest sustainable economic growth and employment and a rising standard of living in member countries, while maintaining financial stability, and thus to contribute to the development of the world economy.

To contribute to sound economic expansion in member as well as non-member countries in the process of economic development; and

To contribute to the expansion of world trade on a multilateral, non-discriminatory basis in accordance with international obligations.

The original member countries of the OECD are Austria, Belgium, Canada, Denmark, France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States. The following countries became members subsequently through accession at the dates indicated hereafter: Japan (28

th April 1964), Finland (28

th January 1969), Australia (7

th June 1971),

New Zealand (29th

May 1973), Mexico (18th

May 1994), the Czech Republic (21st

December 1995), Hungary (7

th May 1996), Poland (22

nd November 1996), Korea (12

th December 1996) and the Slovak Republic

(14th

December 2000). The Commission of the European Communities takes part in the work of the OECD (Article 13 of the OECD Convention).

©OECD 2008

OECD Compendium of Productivity Indicators ©OECD 2008 3

FOREWORD

Over the past few years, productivity and economic growth have been an important focus of OECD work. This work has included both efforts to improve the measurement of productivity growth, as shown in the development of the OECD Productivity Manual, published in 2001, as well as work to enhance the understanding of the drivers of productivity performance. In the course of this work, questions about data choices and the measurement of productivity were examined at several occasions. At the same time, OECD was confronted with a growing interest in internationally comparable data on productivity growth.

The continued interest of many OECD member countries in productivity led to a decision to develop an OECD Productivity Database, based on data that were considered to be as comparable and consistent across countries as possible. This database and related information on methods and sources is available through the OECD Internet site and free of charge at:

www.oecd.org/statistics/productivity

In 2005, a large number of indicators on productivity were combined in one document for the first time to coincide with an OECD workshop on productivity measurement, held in Madrid. The present document constitutes the 2007-2008 update of the productivity compendium. It draws on the OECD Productivity Database, but also includes indicators drawn from other sources, such as the OECD STAN database, which enables productivity calculations for individual industries, and the OECD System of Unit Labour Costs and Related Indicators.

The compendium includes indicators as well as methodological notes and describes the measurement challenges and data choices that were made as well as the remaining measurement problems. Further details are available in a number of specific annexes.

Texts, tables and graphs presented in this compendium are available on-line at the following address:

www.oecd.org/statistics/productivity/compendium

The present document was prepared by the OECD Statistics Directorate (STD) and the Economic Analysis and Statistics Division (EAS) of OECD’s Directorate for Science, Technology, and Industry (STI). Agnès Cimper, Julien Dupont, Joaquim Oliveira-Martins and Paul Schreyer prepared the text, tables and graphs presented in the compendium. Additional contributions to text and tables from Nadim Ahmad, Benoît Arnaud, David Brackfield, Colin Webb and Pascal Marianna are gratefully acknowledged. Joseph Loux supervised the publication process.


TABLE OF CONTENTS

Highlights 7

Background: defining and measuring productivity 11

A) Economy-wide indicators of productivity growth 13

A.1. Growth in GDP per capita 14

A.2. Labour productivity growth, growth in GDP per hour worked 16

A.3. Alternative measures of labour productivity growth 18

A.4. Capital productivity 20

A.5. Growth accounts for OECD countries 22

A.6. The contribution of multi-factor productivity and ICT capital to GDP growth 24

A.7. The decomposition of labour productivity growth in MFP growth and capital deepening 26

B) Productivity levels 29

B.1. Income and productivity levels 30

B.2. Historical income and productivity levels, 1973-2006 32

B.3. Alternative measures of output 34

B.4. Heterogeneity of labour productivity by size class and industry 36

B.5. Heterogeneity of labour productivity by size class, total manufacturing 38

C) Productivity growth by industry 41

C.1. Contribution of key activities to aggregate productivity growth 42

C.2. Productivity growth in manufacturing 44

C.3. Productivity growth in services 46

D) Impact of labour productivity on unit labour costs 49

D.1. Unit labour costs and labour productivity - Total economy 50

D.2. Unit labour costs, labour productivity and labour compensation per unit labour input - Industry 52

D.3. Unit labour costs, labour productivity and labour compensation per unit labour input - Market services

54

Annex 1 – OECD productivity database 58

Annex 2 – OECD estimates of labour productivity levels 65

Annex 3 – OECD databases relevant to productivity analysis 70

Annex 4 – OECD system of Unit Labour Cost and Related Indicators 74

Annex 5 – Multi-factor productivity measures in OECD countries 81

References 92


HIGHLIGHTS

This third edition of the OECD Compendium of Productivity Indicators brings together the latest data and productivity indicators in four broad areas: a) economy-wide productivity growth and b) levels; c) productivity growth by industry; and, d) impact of labour productivity on unit labour costs. Concerning economy-wide indicators of productivity growth the main highlights are as follows:

Over the period 1970-2006, average annual growth in GDP per capita has been above 2% in most OECD countries, even exceeding 4% in Ireland and Korea. In the second half of the 1990s, Hungary, Ireland, Korea, Poland and the Slovak Republic experienced the highest rates of GDP per capita growth. More recently, however, several OECD countries have experienced a productivity growth slowdown compared to long-term trends, in particular Italy and Portugal. In parallel, since the beginning of the new millennium, the rate of labour utilisation has also decreased in many European countries. Both trends have induced a deceleration of GDP per capita growth. In contrast, the Japanese economy has experienced a recent pick-up in both labour utilisation and productivity growth.

Labour productivity growth (or GDP per hour worked) has varied considerably among OECD countries. For example, in the first half of the 2000s, labour productivity growth in Hungary, Korea and the Slovak Republic ranged from 4.3 to 5.2%, while Italy and Mexico experienced less than 0.5% growth. In a number of OECD countries, labour productivity growth had accelerated in the second half of the 1990s but slowed again in turn of the millennium. Between 1995-2000 and 2001-2006, Australia, Ireland, Mexico and Portugal display a particularly strong deceleration.

As an alternative measure of productivity, GDI (Gross Domestic Income) per hour worked displays approximately the same profile as GDP per hour worked over the past twenty years. Only in countries, such as Australia and Korea, that have experienced largest shifts in terms of trade and/or where foreign trade accounts for a large share of GDP, the differences between the two measures are more significant.

The user costs of capital relative to labour have declined. Reflecting the relative abundance of capital resources, capital productivity has decreased almost everywhere since 1985 (notably in Canada, Spain and the United Kingdom), with the exception of Finland where output per unit of capital input displays on average positive growth over the last decade.

In most OECD countries, GDP growth was mainly driven by capital and Multi-Factor Productivity (hereafter, MFP) growth. From 1985 to 2006, capital inputs accounted for around one third of GDP growth. Information and Communication Technologies (ICT) explains the bulk of capital’s contribution to GDP growth in Australia, Denmark, France, New Zealand, Sweden, the United Kingdom and the United States. Despite this fact, the contribution of ICT capital to GDP growth fell in most OECD countries between the periods 1985-2006 and 2001-2006 (notably in Austria, Portugal, Sweden, Switzerland and the United States).

Multi-factor productivity (MFP) was particularly important for overall growth performance in Belgium, Finland, Ireland and Japan. It also helped strengthen growth in Sweden, the United Kingdom and the United States in more recent years (2001-2006). But, in many other countries, MFP growth has slowed down significantly, in particular Italy, New Zealand, Portugal and Switzerland. Italy and Switzerland even experienced negative MFP growth between 2001 and 2006.

Still, growth in labour inputs contributed significantly to GDP growth in Australia, Canada, Ireland, the Netherlands, Spain and the United States, while fast ageing countries, such as Germany and Japan, experienced negative growth in labour inputs.


The second section presents productivity levels, as key indicators of economic convergence across countries. Main results are:

In 2006, Ireland, Luxembourg, Norway and the United States had the highest levels of per capita income. GDP per capita ranged from over USD 39 000 in Ireland, Luxembourg, Norway and the United States to less than USD 17 000 in Mexico, Poland and Turkey. As a general pattern, most OECD countries have higher levels of GDP per hour worked than GDP per capita relative to the United States. The differences in GDP per capita are therefore due to lower levels of labour utilisation than in the United States (though this could be partly due to disparities in measurement of working hours).

Since the 1970s, GDP per capita and labour productivity have broadly converged in the OECD area, with Ireland and Korea displaying the highest rates of catch-up in terms of GDP per capita. Economies that had relatively high income levels in the 1970s have had lower rates of catch-up. In terms of average labour productivity levels, several European countries have recently surpassed the United States, while Australia, Canada, Mexico and New Zealand are still below the United States.

Differences between GDP, Net Domestic Product (NDP) and Gross National Income (GNI) per hour worked are relatively small, as gross income inflows from abroad tend to be offset by gross outflows. This suggests that GDP per hour worked can be used as a relatively good proxy for other alternatives measures of output and productivity levels.

There is considerable heterogeneity in labour productivity figures across countries and industries. Comparisons of labour productivity by size class show that for most industries, particularly in the manufacturing sector, the larger the business the higher the labour productivity level. This, in part, reflects higher degrees of capital intensity in larger businesses and economies of scale.

The third section presents indicators of productivity growth by industry. It shows that:

For most OECD countries, manufacturing productivity growth has slowed down recently, but large cross-industry differences can be observed. High- and medium-high technology industries, such as electrical and optical equipment and transport equipment, have typically experienced relatively higher rates of productivity growth than low-technology manufacturing industries. In many countries, the highest aggregate labour productivity growth performances are still in the manufacturing sector. This was the case for example in the Czech Republic, Finland, Korea, the Slovak Republic and Sweden.

Overall Market Service labour productivity growth also decreased during the period 2000-2005 compared to 1995-2000, although service sectors that invest more in ICT and have more highly skilled workforces displayed higher productivity growth. These include post and telecommunications, finance and insurance, and computer services.

Reflecting the growing importance of the service sector, over the period 2000-2006, Market Services accounted for more than half of labour productivity growth in Greece, Luxembourg, New-Zealand, Norway, the United Kingdom and the United States. Between 1995-2000 and 2000-2006, the contribution of Market Services to labour productivity growth has also increased in Belgium, the Czech Republic, France, Luxembourg and New Zealand. This growing contribution of Market Services is sometimes linked to an increasing share in total value added, but can also reflect genuine faster labour productivity growth in services.

The last section presents the impact of labour productivity on an important indicator of competitiveness, the unit labour costs. Stronger growth in labour productivity than in average labour compensation will have a downward impact on growth in unit labour costs, though developments in average labour compensation also matter. Main results are as follows:


Reflecting a certain wage equalisation within economies, a similar development for annual growth rates of average labour compensation can be observed both for Industry and Market Services in most countries. Given the typically higher annual average growth rates for labour productivity in Industry, noted above, unit labour costs for Industry have tended to decrease relative to Market Services. This reflects to some extent the impact of globalisation on Industry, as well as a more intensive use of capital.

The gap between unit labour costs and productivity in Market Services was particularly marked in Poland, Hungary, Slovak Republic and Turkey. In these countries, long-run annual average growth rate (1986-2006) for unit labour costs in Market Services has been around or above 10%, while average annual labour productivity growth in Market Service sectors has been less than 3%.


DEFINING AND MEASURING PRODUCTIVITY

Productivity isn’t everything, but in the long run it is almost everything. A country’s ability to improve its standard of living over time depends almost entirely on its ability to raise its output per worker.

Paul Krugman, The Age of Diminishing Expectations (1994)

Productivity is commonly defined as a ratio between the output volume and the volume of inputs. In other words, it measures how efficiently production inputs, such as labour and capital, are being used in an economy to produce a given level of output. Productivity is considered a key source of economic growth and competitiveness and, as such, is basic statistical information for many international comparisons and country performance assessments. For example, productivity data are used to investigate the impact of product and labour market regulations on economic performance. Productivity growth constitutes an important element for modelling the productive capacity of economies. It also allows analysts to determine capacity utilisation, which in turn allows one to gauge the position of economies in the business cycle and to forecast economic growth. In addition, production capacity is used to assess demand and inflationary pressures.

There are different measures of productivity and the choice between them depends either on the purpose of the productivity measurement and/or data availability. One of the most widely used measures of productivity is Gross Domestic Product (GDP) per hour worked. This measure captures the use of labour inputs better than just output per employee. Generally, the default source for total hours worked is the OECD Annual National Accounts database, though for a number of countries other sources have to be used. Despite the progress and efforts in this area, the measurement of hours worked still suffers from a number of statistical problems. Namely, different concepts and basic statistical sources are used across countries, which can hinder international comparability. In principle, the measurement of labour inputs should also take into account differences in workers’ educational attainment, skills and experience. Accordingly, the OECD has started to develop adjusted labour input measures.

To take account of the role of capital inputs, an appropriate measure is the flow of productive services that can be drawn from the cumulative stock of past investments (such as machinery and equipment). These services are estimated by the OECD using the rate of change of the ‘productive capital stock’, which takes into account wear and tear, retirements and other sources of reduction in the productive capacity of fixed capital assets. The price of capital services per asset is measured as their rental price. In principle, the latter could be directly observed if markets existed for all capital services. In practice, however, rental prices have to be imputed for most assets, using the implicit rent that capital goods’ owners ‘pay’ to themselves (or the ‘user costs of capital’).

After computing the contributions of labour and capital to output, the so-called multi-factor productivity (MFP) can be derived. It measures the residual growth that cannot be explained by the rate of change in the services of labour, capital and intermediate outputs, and is often interpreted as the contribution to economic growth made by factors such as technical and organisational innovation.

Against this background, a broad overview of productivity indicators is presented in four areas. International comparisons of economy-wide indicators of productivity growth are first presented followed by international comparisons of income and productivity levels, including a measure of productivity heterogeneity by enterprise size classes. Thirdly, productivity growth indicators by industry and services are examined. Finally, the impact of labour productivity on unit labour costs is discussed.

Unless noted otherwise, GDP refers to the total economy.


A. ECONOMY-WIDE INDICATORS OF PRODUCTIVITY GROWTH


A.1. GROWTH IN GDP PER CAPITA

Gross Domestic Product (GDP) per capita measures economic activity or income per person and is one of the core indicators of economic performance. GDP per capita is a rough measure of average living standards or economic well-being; per capita GDP growth can be broken down into a part which is due to labour productivity growth (GDP per hour worked) and a part which is due to increased labour utilisation (hours worked per capita). Growing labour utilisation can have considerable impacts on the growth of GDP per capita. A slowing or declining rate of labour utilisation combined with high labour productivity growth can be indicative of a greater use of capital and/or of a decreasing employment of low-productivity workers.

Definition

The indicator hereafter is calculated using GDP and population estimates published in the OECD Annual National Accounts database. For zone aggregates, GDP estimates have been converted to constant US dollars, using 2000 constant Purchasing Power Parities (PPPs). Series on hours were mostly derived from the OECD Annual National Accounts; when this source was not available, the OECD Employment Outlook was used instead.

Comparability

All OECD countries follow the 1993 System of National Accounts, except for Turkey that is using the 1968 System of National Accounts. Hours worked correspond to actual hours worked, although methods to derive actual hours worked may vary somewhat between countries.

In the chart hereafter on this page, OECD aggregate does not include Czech Republic, Hungary, Poland and Slovak Republic while in the chart on the right page, OECD aggregate does not include Poland and Turkey.

Overview

The figure highlights the key role of productivity growth in determining the GDP per capita. Over the period 1970-2006, growth in GDP per capita has been above 2% in most OECD countries, but significantly more in some countries, notably Ireland and Korea for which the average growth rate went over 4%. In the second half of the 1990s, Hungary, Ireland, Korea, Poland and the Slovak Republic experienced high rates of growth in GDP per capita. More recently, many OECD countries have experienced a deceleration in their income growth relative to long-term trends, notably Italy and Portugal.

Since the beginning of the new millennium, many European countries have decreased in the rate of labour utilisation, which was also accompanied by a sharp decline in labour productivity growth. In contrast, compared with the second half of the nineties, the Czech Republic, Japan and the Slovak Republic experienced a pick-up in both labour utilisation and labour productivity growth. Noteworthy, the estimates

shown here are not adjusted for differences in the business cycle; cyclically adjusted estimates might show a somewhat different pattern.

Growth in GDP per capita

Total economy, percentage change at annual rate

1970-2006 2001-2006

0 2 4 6

Turkey

Slovak Republic

Hungary

Czech Republic

Korea

Poland

Greece

Ireland

Luxembourg

Sweden

Finland

Iceland

United Kingdom

New Zealand

Australia

United States

NAFTA

Norway

Mexico

Canada

Spain

OECD

Japan

Denmark

G7

Belgium

Austria

EU15

Netherlands

France

Switzerland

Germany

Portugal

Italy

%

Sources OECD Productivity Database:

www.oecd.org/statistics/productivity.

OECD Annual National Accounts Database.

For further reading

OECD (2001), Measuring Productivity – OECD Manual, OECD, Paris.

Pilat, D. and P. Schreyer (2004), “The OECD Productivity Database – An Overview”, International Productivity Monitor, No.8, Spring, OECD, Paris, pp. 59-65.


GROWTH IN GDP PER CAPITA A.1.

The contribution of labour productivity and labour utilisation to GDP per capita

Total economy, percentage change at annual rate, 2001-2006

-2 0 2 4 6

Turkey

Slovak Republic

Hungary

Czech Republic

Korea

Poland

Greece

Ireland

Luxembourg

Sweden

Finland

Iceland

United Kingdom

New Zealand

Australia

United States

NAFTA

Norway

Mexico

Canada

Spain

OECD

Japan

Denmark

G7

Belgium

Austria

EU15

Netherlands

France

Switzerland

Germany

Portugal

Italy

Growth in GDP per capita

-2 0 2 4 6

Growth in GDP per hour worked

=

-2 0 2 4 6

Turkey

Slovak Republic

Hungary

Czech Republic

Korea

Poland

Greece

Ireland

Luxembourg

Sweden

Finland

Iceland

United Kingdom

New Zealand

Australia

United States

NAFTA

Norway

Mexico

Canada

Spain

OECD

Japan

Denmark

G7

Belgium

Austria

EU15

Netherlands

France

Switzerland

Germany

Portugal

Italy

Growth in labour utilisation +


A.2. GROWTH IN GDP PER HOUR WORKED

Productivity growth is measured by relating changes in output to changes in one or more inputs to production. The most common productivity measure is labour productivity, which links changes in output to changes in labour input. It is a key economic indicator and it is closely associated with standards of living.

Definition

The output measures used for calculations are Gross Domestic Product estimates from OECD Annual National Accounts database, based on the 1993 System of National Accounts. Labour input measures used are estimates of the hours actually worked. They reflect regular hours worked by full-time and part-time workers, paid and unpaid overtime, hours worked in additional jobs and time not worked because of public holidays, annual paid leaves, strikes and labour disputes, bad weather, economic conditions and other reasons.

Comparability

The OECD and National statisticians work together to ensure that the data on hours actually worked are as comparable as possible, though they are based on a range of different sources of varying reliability. In most countries, the data are taken from household labour force surveys, while the rest use establishment surveys, administrative sources or a combination of sources. One problem is that for several EU countries, the estimates are made by the OECD using results from the Spring European Labour Force Survey. The results reflect a single observation in the year, and the survey data have to be supplemented by information from other sources for hours not worked due to public holidays and annual paid leave. Annual working hours reported for the remaining countries are provided by national statistical offices and are estimated using the best available sources. In general, the data are best used for comparisons of trends over time rather than for inter-country comparisons of level of productivity.

Although the GDP estimates are based on common definitions, the methods used by most countries to estimate value added in government services assume that labour productivity growth is zero. This means that countries with large government sectors or with

government sectors that were growing during the period considered will, by assumption, have lower growth in GDP per hour worked than other countries. In the charts, OECD aggregate excludes Poland and Turkey.

Overview

Labour productivity growth varies considerably among OECD countries. For example, in the first half of the 2000s, labour productivity growth in Hungary, Korea and the Slovak Republic ranged from 4.3 to 5.2% to a growth rate of less than 0.5% in Italy and Mexico.

In a number of OECD countries, labour productivity growth accelerated in the second half of the 1990s but slowed again in the first half of the new millennium. Between 2001-2006 and 1995-2000, the Czech Republic and Hungary were the only countries to experience a significant acceleration of growth in GDP per hour worked while over the same period, Australia, Ireland, Mexico and Portugal saw a strong deceleration in labour productivity growth.

The rates shown here are not adjusted for differences in the business cycle; cyclically adjusted estimates might show a somewhat different pattern.

Sources

OECD Productivity Database: www.oecd.org/statistics/productivity


Further information

Ahmad, N., F. Lequiller, P. Marianna, D. Pilat, P. Schreyer and A. Wölfl (2003), “Comparing Labour Productivity Growth in the OECD Area: The Role of Measurement”, STI Working Papers 2003/14,

OECD (2004), “Clocking In (and Out): Several Facets of Working Time”, OECD Employment Outlook 2004, Chapter 1, OECD, Paris.

OECD, Paris.OECD (2001), Measuring Productivity – OECD Manual, OECD, Paris.

Pilat, D. and P. Schreyer (2004), “The OECD Productivity Database – An Overview”, International Productivity Monitor, No.8, Spring, OECD, Paris, pp. 59-65.


GROWTH IN GDP PER HOUR WORKED A.2.

GDP per hour worked

Average annual growth in percentage, 1995-2006

0

1

2

3

4

5

6

7%

Average annual growth in percentage, 1995-2000 and 2001-2006

0

1

2

3

4

5

6

7%

2001-2006 1995-2000


A.3. ALTERNATIVE MEASURES OF LABOUR PRODUCTIVITY GROWTH

GDP is the most frequently used measure of macro-economic activity but makes no allowance for the using up of capital goods during the production process. An allowance for depreciation of capital can then be subtracted from GDP to obtain the corresponding net concept of Net Domestic Product (NDP). It is also useful to compare GDP with Gross Domestic income (GDI) per hour worked over time. Real GDI measures the purchasing power of the total incomes generated by domestic production (including the impact on those incomes of changes in terms of trade).

Definition

GDP is a gross measure that does not account for capital used in production. The associated loss in value, depreciation, reduces the net value of production that is available as net income in any given year. The observation has often been made that a growing part of capital goods is short-lived (for example computers), and that this structural shift in the composition of assets leads to higher overall depreciation. A case could thus be made to measure productivity on the basis of NDP as well as GDP. Countries with a structure of fixed assets that is biased towards short-lived assets would exhibit a relatively lower NDP per hour worked than GDP per hour worked, reflecting relatively higher depreciation.

GDI is equal to gross domestic product at constant prices plus the trading gain (or less the trading loss) resulting from the changes in terms of trade. The terms-of-trade effect arises because real GDI is obtained by deflation with the price index for domestic final demand rather than the price index of GDP. It measures the rate at which exports can be traded against imports from the rest of the world. The difference between movements in GDP at constant prices and real GDI are not always small. If imports and exports are large relative to GDP, and if commodity composition of the goods and services which make up imports and exports are very different, the scope for potential trading gains or losses may be large. If the prices of a country’s exports rise faster (or fall more slowly) then the prices of its imports – that is if its terms of trade improve – less exports are needed to pay for a given volume of imports so that a given level of domestic production of goods and services can be reallocated from exports to consumption or capital formation. Thus, an improvement in terms of trade makes it possible for an increased volume of goods and services to be purchased by residents out of the incomes generated by a given level of domestic production.

Comparability

Net measures require reliable estimates of depreciation and the empirical basis for depreciation estimates is generally not well established. For similar types of assets, significant differences exist in the service lives and depreciation rates that are used by different countries. These rates are sometimes based on assumptions more than in-depth empirical studies or are based on evidence dating back a number of years. This reduces the quality of depreciation estimates as well as their

international comparability. The international GDP-NDP comparisons should thus be interpreted with caution.

Despite its analytical usefulness, it should be borne in mind that GDI is a measure of income and not a measure of production. However, as has been pointed out by some authors (Kohli 2004), terms of trade effects resemble technical change. Changes in constant prices GDP would only reflect technical changes whereas GDI is a measure that reflects both technological change and terms of trade. This makes GDI a meaningful measure even in the con text of production analysis.

Long-term trends

Real depreciation has grown somewhat faster than real GDP in the past years in many OECD countries, reflecting investment in new technologies and a shift in the structure of investment and capital stocks towards shorter-lived assets. As a consequence, NDP per hour worked has risen somewhat slower than GDP per hour worked. The gap turned out to be relatively large in the United States, and Denmark whereas in Australia, Czech Republic, Denmark, Iceland and Slovak Republic, NDP growth exceeded GDP growth. Similar to the growth rates comparisons of NDP and GDP per hour worked, levels comparisons of NDP and GDP per hour worked can be compared (see B.3).

The differences between GDP, NDP and GDI are reduced over the long term. Looking at the trend growth of real GDP and GDI per hour worked over the past twenty years shows that differences between the two measures are relatively small, except for a few countries (e.g. Australia and Korea). By definition, the difference between the two measures is most important in those countries that experienced the largest shifts in their terms of trade and/or where foreign trade accounts for a large share of GDP.

Sources



Further information

Boarini, Romina, Åsa Johansson and Marco Mira d’Ercole (2006), “Alternative Measures of Well-Being”, OECD Statistics Brief, n°11, May.

Commission of the European Communities, OECD, IMF, United Nations, World Bank (1993), System of National Accounts 1993, Brussels/Luxembourg, New York, Paris, Washington DC.

Kohli, Ulrich (2004), “Real GDP, Real Domestic Income and Terms of Trade Changes”, Journal of International Economics, 62.




ALTERNATIVE MEASURES OF LABOUR PRODUCTIVITY GROWTH A.3.

Growth in NDP per hour worked compared with growth in GDP per hour worked and growth in GDI per hour worked

Total economy, percentage change 2001-20061, annual rate

-1

0

1

2

3

4

5

6

% GDI per hour worked GDP per hour worked NDP per hour worked

Growth in NDP per hour worked compared with growth in GDP per hour worked and growth in GDI per hour worked

Total economy, percentage change 1985-20062, annual rate

-1

0

1

2

3

4

5

6

%GDI per hour worked GDP per hour worked NDP per hour worked


A.4. CAPITAL PRODUCTIVITY GROWTH

While labour productivity is the most common partial productivity measure, capital productivity provides another, supplementary piece of information about productivity growth.

Definition

Capital productivity is measured as the ratio between output and capital input. Capital productivity has to be distinguished from the rate of return to capital. The former is a physical, partial productivity measure; the latter is an income measure that relates capital income to the value of the capital stock.

From the viewpoint of the economic theory, the objective is to measure the flow of productive services that capital delivers in production. This measure relies on the assumption that capital services are a fixed proportion of the productive capital stock. In which case, the rate of change of capital services coincides with the rate of change of the capital stock. The latter is estimated by cumulating investment flows and correcting them for retirement, wear and tear and obsolescence. The aggregate flow of capital services is obtained by weighting the flow of capital services of each type of asset by its share in total capital income.

Comparability

Price indices are vital for in measuring volume investment, capital services and user costs. Accurate price indices should be constant quality deflators that reflect price changes for a given performance of ICT investment goods. Thus, observed price changes of ‘computer boxes’ have to be quality-adjusted for comparison of different vintages. There are differences how countries deal with quality adjustment with possible consequences for the international comparability of price and volume measures of ICT investment. In particular, those countries that employ hedonic methods to construct ICT deflators tend to register a larger drop in ICT prices than countries that do not. The OECD uses a set of ‘harmonised’ deflators to control for some of the differences in methodology and assumes that the ratios between ICT and non-ICT asset prices evolve in a similar manner across countries, using the United States as the benchmark. Although no claim is made that the ‘harmonised’ deflator is necessarily the correct price index for a given country, we feel that the possible error due to using a harmonised price index is smaller than the bias arising from comparing capital services based on national deflators. However, from an accounting perspective, adjusting the price index for investment goods for any country implies an adjustment of the volume index of output. In most cases, such an adjustment would increase the measured rate of volume output change. At the same time, effects on the economy-wide rate of GDP growth appear to be contained (see Schreyer (2001) for a discussion).

Long-term trends

Two important drivers shape capital productivity: overall efficiency or multi-factor productivity growth and the amount of labour input per unit of capital in production. The fewer hours worked are available per unit of capital, the lower capital productivity. Generally, the cost of using capital has declined relative to labour, so that the amount of labour input per capital input has declined as well, leading to the observed fall in capital productivity over the last twenty years. The fall in capital productivity since 1985 has been very pronounced in Canada, Spain and the United Kingdom but also in Australia, Denmark, Italy, Japan, New Zealand, Switzerland, Sweden and the United States. Notable exceptions to the decline in output per unit of capital input is Finland where capital productivity grew over most of the last decade. Over the recent years, capital productivity rose in Belgium, Finland, Japan and Sweden and the decline in capital productivity slowed significantly in New Zealand, the United Kingdom and the United States. Over the same period, Australia, Ireland, Italy, the Netherlands and Portugal experienced a sharp deceleration in capital productivity growth. Note, however, that like other productivity measures, capital productivity varies considerably with the business cycle as no adjustments have been made for variations in the rate of capacity utilization.

Source


Further information


Schreyer, P. (2003), “Capital Stocks, Capital Services and Multi-factor Productivity Measures”, OECD Economic Studies No. 37, 2003/2, OECD, Paris, pp. 163-184.

Schreyer, P. (2001), “Computer Price Indices and International Growth and Productivity Comparisons”, Economics of Innovation and New technology Vol. 10.

Schreyer, P., P.E. Bignon and J. Dupont (2003), “OECD Capital Services Estimates: Methodology and a First Set of Results”, OECD Statistics Working Paper 2003/6, OECD, Paris.



CAPITAL PRODUCTIVITY GROWTH A.4.

Growth in capital productivity, 2001-20061 compared with 1985-2006

2

Total economy, percentage change, annual rate

-3.0

-2.0

-1.0

0.0

1.0

2.0

3.0

%2001-2006 1985-2006

Growth in capital productivity, 2001-20061 compared with 1995-2000

Total economy, percentage change, annual rate

-3.0

-2.0

-1.0

0.0

1.0

2.0

3.0

% 2001-2006 1995-2000

1. 2001-2004 for Australia, Belgium, Japan and Switzerland, 2001-2005 for Austria, Denmark, Finland, Netherlands, Portugal and the United

Kingdom. 2. 1985-2004 for Australia, Belgium and Japan, 1985-2005 for Denmark, Finland, Netherlands and the United Kingdom, 1990-2006 for New

Zealand and Spain, 1991-2006 for Germany, 1995-2004 for Switzerland, 1995-2005 for Austria and Portugal.


A.5. GROWTH ACCOUNTS FOR OECD COUNTRIES

Economic growth can be increased by increasing the amount and types of labour and capital used in production, and by attaining greater overall efficiency in how these factors of production are used together, i.e. higher multi-factor productivity (MFP). Growth accounting involves breaking down growth of GDP into the contribution of labour input, capital input and MFP.

Definition

The growth accounting approach is based on the microeconomic theory of production and directly related to the calculation of MFP growth, measured by deducting from output growth the growth of labour and capital inputs. Turned around, the same relation can be used to explain output growth by the rates of change of labour and capital inputs and by MFP growth.

In these calculations, the growth rate of labour and capital inputs is weighted with their share in total costs. Thus, the contribution of labour to GDP growth is measured as the speed with which labour input grows, multiplied by the relative importance of labour captured by its share in total costs. The growth contributions of capital or of certain types of capital are measured in a similar way so that the growth contribution always reflects two effects, the growth rate of the input and its relative importance in production.

Comparability

The role of information and communication technology (ICT) for growth in GDP and MFP is analysed thanks to the differentiation between ICT and non ICT-capital. ICT related capital include hardware, communication and software. Non-ICT capital include transport equipment and non residential construction, products of agriculture, metal products and machinery other than hardware and communication equipment, and other products of non-residential gross fixed capital formation.

The appropriate measure for capital input with the growth accounting framework is the flow of productive services that can be drawn from the cumulative stock of past investments in capital assets. These services are estimated by the OECD using the rate of change of the ‘productive capital stock’. This measure takes into account wear and tear and retirements, i.e., reductions in the productive capacity of the fixed assets. The price of capital services for each type of asset is measured as their rental price. In principle, the latter could be directly observed if markets existed for capital services. In practice, however, rental prices have to be imputed for most assets, using the implicit rent that capital goods’ owners ‘pay’ themselves (or ‘user costs of capital’).

The measure of total hours worked is an incomplete measure of labour input because it does not account for changes in the skill composition of workers over time, such as educational attainment, and work experience. Adjustment for such attributes would provide a more accurate indication of the contribution of labour to production. In the absence of these adjustments, as is the case in the series shown here, more rapid output growth due to a rise in skills of the labour force are captured by the MFP residual, and not attributed to labour. This should be kept in mind when interpreting rates of MFP growth.

Long-term trends

From 1985 to 2006, GDP growth in most OECD countries was for a large part driven by growth in capital and MFP. In many countries, growth in capital accounted for around one third of GDP growth from 1985 to 2006. Over the same period, ICT capital services represented between 0.2 and 0.6 percentage point of growth in GDP. ICT accounts for a bulk of capital’s contribution to GDP growth in Australia, Denmark, France, New Zealand, Sweden, the United Kingdom and the United States; its contribution was more modest in Italy and even smaller in Austria and Ireland. From 1985 to 2006, MFP growth was also an important source of growth in Belgium, Finland, Ireland and Japan but its contribution was very small in Canada, Italy, New Zealand, Spain and Switzerland. Growth in labour input was also important for a few countries over 1985-2006, notably Australia, Canada, Ireland, the Netherlands, Spain and the United States.

Source


Further information





GROWTH ACCOUNTS FOR OECD COUNTRIES A.5.

Contributions to growth of GDP, 1985-20061 and 2001-2006

2

Percentage points

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

Italy Germany Japan France United Kingdom United States Canada

%

Labour input ICT capital Non-ICT capital Multi-factor productivity

2001-20061985-2006

G7 countries

-1.0

0.0

1.0

2.0

3.0

4.0

5.0

6.0

%

Labour input ICT capital Non-ICT capital Multi-factor productivity

1985-2006 2001-2006

Other OECD countries

1. 1985-2004 for Australia, Belgium and Japan, 1985-2005 for Denmark, Finland, Netherlands and the United Kingdom, 1990-2006 for New Zealand and Spain, 1991-2006 for Germany, 1995-2004 for Switzerland, 1995-2005 for Austria and Portugal. 2. 2001-2004 for Australia, Belgium, Japan and Switzerland, 2001-2005 for Austria, Denmark, Finland, Netherlands, Portugal and the United Kingdom.


A.6. GROWTH ACCOUNTS – THE CONTRIBUTION OF MULTI-FACTOR PRODUCTIVITY AND ICT CAPITAL TO GDP GROWTH

Growth accounting typically involves breaking down the growth of GDP into three components — the contribution of labour, the contribution of capital and multi-factor productivity (MFP).

Definition

MFP is the change in GDP that cannot be explained by changes in the quantities of capital and labour that are made available to generate GDP. MFP is sometimes described as disembodied technological progress, because it is the increase in GDP that is not embodied in the amounts of either labour or capital. MFP growth comes from more efficient use of labour and capital inputs, for example through improvements in the management of production processes, organisational change or more generally, innovation. Growth in MFP is a significant factor in explaining the long-term growth in real GDP.

MFP growth is measured by deducting from output growth the growth of labour and capital inputs. Turned around, the same relation can be used to explain output growth by the contribution of labour and capital inputs, and by MFP growth.

In these calculations, the growth rate of labour and capital inputs is weighted with their share in total costs. Thus, the contribution of labour to GDP growth is measured as the speed with which labour input grows, multiplied by the relative importance of labour captured by its share in total costs. The growth contributions of ICT capital are measured in a similar way so that the growth contribution always reflects two effects, the growth rate of the input and its relative importance in production.

Comparability

Correct measurement of investment in ICT in both nominal and volume terms is crucial for estimating its contribution to economic growth and performance. Data availability and measurement of investment in ICT based on national accounts (SNA93) vary considerably across OECD countries, especially as regards measurement of investment in software, deflators applied, breakdown by institutional sector and temporal coverage. In the national accounts, expenditure on ICT products is considered as investment only if the products can be physically isolated (i.e. ICT embodied in equipment is considered not as investment but as intermediate consumption). This means that investment in ICT may be underestimated and the order of magnitude of the underestimation may differ depending on how intermediate consumption and investment are treated in each country’s accounts.

In particular, expenditure on software has only very recently been treated as capital expenditure in the national accounts, and methodologies still vary considerably across countries. Difficulties for measuring software investment are also linked to the ways in which software can be acquired, e.g. via rental and licences or embedded in hardware. Moreover, software is often developed on own account.

To tackle the specific problems relating to software in the context of the SNA93 revision of the national accounts, a joint OECD-EU Task Force on the Measurement of Software in the National Accounts has developed recommendations concerning the capitalisation of software (Lequiller, et al., 2003; Ahmad, 2003). These are now being implemented by OECD member countries.

Long-term trends

Multi-factor productivity growth was one of the factors that helped strengthen growth in Belgium, Japan, Sweden, the United Kingdom and the United States in the recent years (2001-2006) compared with the longer period 1985-2006. In other countries, including Austria, Australia, Denmark, Finland, France, Germany, Ireland, Italy, the Netherlands, New Zealand, Portugal and Switzerland, MFP growth slowed down in the recent years (2001-2006) compared to the longer period 1985-2006, sometimes significantly as in the case of Ireland, Italy, New Zealand, Portugal and Switzerland. MFP growth was negative in the recent years (2001-2006) in Italy and Switzerland but positive in the long term period 1985-2006. The contribution of ICT capital to GDP growth decelerated in most of OECD countries comparing 1985-2006 and 2001-2006. The deceleration over the recent years was particularly significant in Austria, Portugal, Sweden and the United States, and smallest in Belgium, Canada, France, Germany, Italy, the Netherlands, Spain, Switzerland and the United Kingdom. The contribution of ICT capital to GDP growth accelerated in the recent years in Australia.

Source


Further information

Ahmad, N. (2003), “Measuring Investment in Software”, STI Working Paper, 2003/6, OECD, Paris.

Lequiller, F., N. Ahmad, S. Varjonen, W. Cave and K.H. Ahn (2003), “Report of the OECD Task Force on Software Measurement in the National Accounts”, OECD Statistics Working Paper 2003/1, OECD, Paris.






GROWTH ACCOUNTS – THE CONTRIBUTION OF MULTI-FACTOR PRODUCTIVITY AND ICT CAPITAL TO GDP GROWTH

A.6.

Multi-factor productivity growth, 1985-2006 and 2001-2006 (or closest year available)1

Percentage points

-1.0%

-0.5%

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

2001-2006 1985-2006

Contributions of ICT capital, 1985-2006 and 2001-2006 (or closest year available)1

Percentage points

-0.1%

0.0%

0.1%

0.2%

0.3%

0.4%

0.5%

0.6%

0.7%

0.8%

2001-2006 1985-2006

1. 1985-2004 for Australia, Belgium and Japan, 1985-2005 for Denmark, Finland, Ireland, the Netherlands and the United Kingdom, 1990-2006 for Spain, 1995-2004 for Switzerland, 1995-2005 for Austria and Portugal, 1991-2006 for Germany.


A.7. GROWTH ACCOUNTS – THE DECOMPOSITION OF LABOUR PRODUCTIVITY GROWTH IN MFP GROWTH AND CAPITAL DEEPENING

Assuming a standard Cobb-Douglas function and constant return to scale, the growth in labour productivity can be decomposed into the MFP growth and the contribution of capital deepening, defined as a weighted capital input relative to the weighted labour input.

Definition

A breakdown of labour productivity growth between multi-factor productivity growth and capital deepening can be described as follows:

)

)(

)(

ln(*)ln()ln(

1

111

t

t

t

t

K

t

t

t

t

H

K

H

K

sMFP

MFP

L

L

Where:

L: labour productivity,

MFP: multi-factor productivity,

sK: share of capital income in total income,

K: capital input,

H: total hours worked actually worked in the entire economy.

Growth in capital deepening refers to the growth in the aggregate flow of capital services minus the growth in aggregate hours worked. Growth in capital deepening has a positive effect on labour productivity because a larger amount of capital per worker should increase the output per worker.

Comparability

The measure of total hours worked is an incomplete measure of labour input because it does not account for changes in the skill composition of workers over time, such as educational attainment, and work experience. Adjustment for such attributes would provide a more accurate indication of the contribution of labour to production. In the absence of these adjustments, as is the case in the series shown here, more rapid output growth due to a rise in skills of the labour force are captured by the MFP residual, and not attributed to labour. This should be kept in mind when interpreting rates of MFP growth.

The contribution of capital deepening could be decomposed thanks to the differentiation between information and communication technology (ICT) and non-ICT capital.

Long-term trends

In most OECD countries, labour productivity grew more slowly in the recent years compared with the long period 1985-2006 with the exception of Belgium, Sweden and the United States and in a lesser extent the United Kingdom. Labour productivity growth over the long period relied more on capital deepening and less on improvements in MFP, whereas over the recent years, MFP growth was an important source of growth in labour productivity. Its contribution to labour productivity growth increased in Belgium, Sweden, the United Kingdom and the United States. In most OECD countries, contribution of capital deepening to labour productivity slowed over the recent years, significantly in Finland, Japan and the Netherlands.

Source


Further information

Ahmad, N. (2003), “Measuring Investment in Software”, STI Working Paper, 2003/6, OECD, Paris.

Lequiller, F., N. Ahmad, S. Varjonen, W. Cave and K.H. Ahn (2003), “Report of the OECD Task Force on Software Measurement in the National Accounts”, OECD Statistics Working Paper 2003/1, OECD, Paris.





GROWTH ACCOUNTS – THE DECOMPOSITION OF LABOUR PRODUCTIVITY GROWTH IN MFP GROWTH AND CAPITAL DEEPENING

A.7.

Decomposition of labour productivity growth into multi-factor productivity growth and capital deepening, 2001-2006 (or closest year available)

1

Average annual growth rates

-1.0%

-0.5%

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

contribution of capital deepening MFP growth

Labour productivity growth

Decomposition of labour productivity growth into multi-factor productivity growth and capital deepening, 1985-2006 (or closest year available)

2

Average annual growth rates

-1.0%

-0.5%

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

contribution of capital deepening MFP growth

Labour productivity growth

1. 2001-2004 for Australia, Austria, Belgium, Japan, Portugal and Switzerland, 2001-2005 for Denmark, Finland, the Netherlands and the United Kingdom. 2. 1985-2004 for Australia, Belgium and Japan, 1985-2005 for Denmark, Finland, the Netherlands and the United Kingdom, 1990-2006 for Spain, 1995-2004 for Switzerland, 1995-2005 for Austria and Portugal, 1991-2006 for Germany.


B. PRODUCTIVITY LEVELS


B.1. INCOME AND PRODUCTIVITY LEVELS

Together with the analysis of growth rates, the levels of GDP per capita and GDP per hour worked are essential to assess the state of the convergence or divergence of economic performances across countries.

Definition

The differences in income levels can be decomposed into differences in labour productivity levels, measured as GDP per hour worked, and differences in the extent of labour utilisation, measured as the number of hours worked per capita. In countries with low levels of GDP per capita, the gaps in labour productivity levels are typically the most significant factor in determining differences in income. The estimates shown here are based on official OECD GDP converted to a common currency using OECD Purchasing Power Parities (PPPs) for 2006.

Comparability

Comparisons of income and productivity levels across countries require several demanding conditions. First, they require comparable data on output. All OECD countries, except Turkey, have implemented the 1993 System of National Accounts. For this reason, the output level in Turkey is likely to be understated relative to other OECD countries. Other differences, such as the measurement of software investment, can also affect the comparability of GDP across countries, although these differences are usually quite small. Second, in a number of countries, employment data are derived from labour force surveys which may not be entirely consistent with the national accounts. This reduces the comparability of labour utilisation levels across countries. The measure of labour inputs also requires hours worked which are derived either from labour force surveys or from business surveys. Several OECD countries estimate hours worked from a combination of these sources or integrate these sources in a system of labour accounts, which is comparable to the national accounts. The OECD Productivity Database uses consistent estimates of employment and hours worked. Nonetheless, the cross-country comparability of hours worked remains somewhat limited, generating a margin of uncertainty in estimates of productivity levels. The third problem relates to the conversion of output from national currency into a common unit. Market exchange rates cannot be used directly, as they are volatile and reflect other factors, such as capital and trade flows. The preferred alternative is to use Purchasing Power Parities (PPPs), which measure the relative prices of the same basket of consumption goods in different countries.

In charts, France includes overseas departments; GDP for Turkey is based on the 1968 System of National Accounts.

Overview

In 2006, GDP per capita in OECD countries ranged from over USD 39 000 in Ireland, Luxembourg, Norway and the United States to less than USD 17 000 in Mexico, Poland and Turkey. On average, income levels were about 70% of that of the United States, Norway is a notable exception with its GDP per capita 14% above that of the United States. Relative to the United States, most OECD countries had higher levels of GDP per hour worked than GDP per capita because their levels of labour utilisation were substantially lower than in the United States. This owes to disparities in working hours but also, in several countries, to high unemployment and low participation of the working-age population in the labour market. The difference between income and productivity levels was largest in European countries. For example, in Belgium, Ireland and the Netherlands, while productivity levels in 2006 surpassed that of the United States, income levels were considerably lower. In several non-EU countries, such as Canada, Japan, New Zealand and Switzerland, labour utilisation in 2006 was higher than in the United States, notably in Iceland and Korea, mainly owing to relatively long working hours and high rates of labour force participation.

Sources


OECD Annual National Accounts and Labour Force Statistics Databases.

For further reading

OECD (2004), “Clocking In (and Out): Several Facets of Working Time”, OECD Employment Outlook 2004, Chapter 1, OECD, Paris



INCOME AND PRODUCTIVITY LEVELS B.1.

Percentage point differences with respect to the United States, 2006

-80 -60 -40 -20 0 20 40

Norway

Ireland

Switzerland

Canada

Netherlands

Iceland

Austria

Australia

Denmark

Sweden

Belgium

United Kingdom

Finland

Germany

Japan

France

OECD

Spain

EU19

Italy

Greece

New Zealand

Korea

Czech Republic

Portugal

Hungary

Slovak Republic

Poland

Mexico

Turkey

Percentage gap with respect to

US GDP per capita

-80

-80 -60 -40 -20 0 20 40

Effect of labour utilisation

-80 -60 -40 -20 0 20 40

Norway

Ireland

Switzerland

Canada

Netherlands

Iceland

Austria

Australia

Denmark

Sweden

Belgium

United Kingdom

Finland

Germany

Japan

France

OECD

Spain

EU19

Italy

Greece

New Zealand

Korea

Czech Republic

Portugal

Hungary

Slovak Republic

Poland

Mexico

Turkey

Percentage gap with respect to US

GDP per hour worked

Income and labour productivity (USA = 100) Hours per capita and labour productivity (USA = 100)

TurkeyMexico

PolandKorea

Czech Republic

HungaryPortugal

Slovak Republic New Zealand

GreeceJapan

IcelandOECDItaly

SpainEU19

Switzerland

FinlandCanada

United Kingdom

Australia

AustriaDenmarkSweden

GermanyFrance United States

NetherlandsIreland

Belgium

Norway

20

40

60

80

100

120

140

20 40 60 80 100 120 140

GDP per hour worked

GDP per capita

TurkeyMexico

Poland Korea

Czech RepublicHungary

Portugal

Slovak Republic New Zealand

GreeceJapan

IcelandOECDItalySpain

EU19 Switzerland

FinlandCanada

United Kingdom

Australia

AustriaDenmarkSweden

GermanyFrance

United States

Netherlands

IrelandBelgium

Norway

20

40

60

80

100

120

140

20 40 60 80 100 120 140

GDP per hour worked

Hours worked per capita


B.2. HISTORICAL INCOME AND PRODUCTIVITY LEVELS, 1973-2006

The process of “catch-up” in average income implies that less advanced economies should experience faster growth in output per capita, typically by adopting the practices of more advanced economies, notably as regards capital, technology and labour utilisation. While less developed countries may grow more rapidly at the beginning of the catching-up process, their economic growth rates are expected to decline over time as their income levels come closer to those of the more advanced countries.

Definition

For each country, the rate of “catch-up” vis-à-vis the United States is calculated as the difference between the average annual compounded growth rate of its GDP per capita level over the period and the average annual compounded growth rate of the United States’ GDP per capita level over the same period.

Comparability

Comparisons of income and productivity levels for a particular year are derived from the time series of Gross Domestic Product (GDP), population, employment and hours worked of the OECD Productivity Database. For some countries, GDP and population data were also derived from Angus Maddison (2001), The World Economy: A Millennial Perspective, OECD Development Centre, OECD, Paris.

Calculations are based on GDP measures converted from national currencies to US dollars using 2006 Purchasing Power Parities.

Overview

Since the 1970s, GDP per capita and labour productivity have broadly converged in the OECD area. Over the period 1973-2006, Ireland and Korea had the highest rates of catch-up in GDP per capita with 2.3% and 3.8% per year, respectively. More advanced economies that started with relatively high income levels in the 1970s have had lower rates of catch-up or even stagnated or recently have diverged vis-à-vis the United States; this was also the case for less advanced economies such as Eastern European countries, Mexico and Turkey. Estimates of levels of GDP per hour worked display slightly different patterns. Since the beginning of the new millennium, several European countries have surpassed the United States in terms of average labour productivity levels. Only Australia, Canada, Mexico and New Zealand did not catch-up vis-à-vis the United States’ productivity levels.

Sources



For further reading

Angus Maddison (2001), The World Economy: A Millennial Perspective, Development Centre Studies, OECD, Paris.



HISTORICAL INCOME AND PRODUCTIVITY LEVELS, 1973-2006 B.2.

Catch-up and convergence in OECD income levels, 1973-2006, relative to the United States

Australia

Austria

Belgium

CanadaCzech Rep.

Denmark

Finland

France

Germany

GreeceHungary

Iceland

Ireland

Italy

Japan

Korea

Mexico

Netherlands

New Zealand

Norway

Poland

Portugal

Slovak Rep.

Spain

Sweden

Switzerland

Turkey

United Kingdom

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

-80 -70 -60 -50 -40 -30 -20 -10 0 10 20

Gap in GDP per capita (%), 2006

Gap in average growth rate (%), 1973-2006

Levels of GDP per hour worked in the OECD area, relative to the United States

0

20

40

60

80

100

120

140

Uni

ted

Stat

es =

100

1973 1985 2006


B.3. ALTERNATIVE MEASURES OF OUTPUT

GDP is the most frequently used measure of macro-economic activity but makes no allowance for the using up of capital goods during the production process. An allowance for depreciation of capital can then be subtracted from GDP to obtain the corresponding net concept of Net Domestic Product (NDP). Another issue is that GDP also makes no allowance for international transfer payments such as profits received from abroad or remittances sent abroad. The Gross National Income (GNI) is measured as the GDP adjusted for “net income from abroad”.

Definition

GDP is a gross measure that does not account for capital used in production. The associated loss in value, depreciation, reduces the net value of production that is available as net income in any given year. The observation has often been made that a growing part of capital goods is short-lived (for example computers), and that this structural shift in the composition of assets leads to higher overall depreciation. A case could thus be made to measure productivity on the basis of NDP as well as GDP. Countries with a structure of fixed assets that is biased towards short-lived assets would exhibit a relatively lower NDP per hour worked than GDP per hour worked, reflecting relatively higher depreciation.

It is also useful to compare GDP with GNI. This comparison permits to take into account the payments received from or sent to the rest of world. GNI is equal to GDP less net taxes on production and imports, less compensation of employees and property income payable to the rest of the world plus the corresponding items receiving from the rest of the world. For example, when company profits are transferred abroad, this leads to GNI being lower than GDP. Conversely, when foreign affiliates or domestic firms or residents abroad transfer payment to the domestic economy, this will raise GNI relative to GDP.

Comparability

Net measures require reliable estimates of depreciation and the empirical basis for depreciation estimates is generally not well established. For similar types of assets, significant differences exist in the service lives and depreciation rates that are used by different countries. These rates are sometimes based on assumptions more than in-depth empirical studies or are based on evidence dating back a number of years. This reduces the quality of depreciation estimates as well as their international comparability. The international GDP-NDP comparisons should thus be interpreted with caution.

NDP and GNI presented here have been converted using the same Purchasing Power Parities (PPPs) as for GDP.

Long-term trends

When data in the national accounts are taken at face value, a comparison between GDP and NDP per hour worked hardly changes the ranking of countries with respect to relative levels of labour productivity. On average, NDP accounts for about 85% of GDP, although there is some variation across countries. Nonetheless, in Belgium, Luxembourg, the Netherlands and Norway depreciation is large in relation to GDP. By contrast, Korea, Mexico, Poland or Turkey, are countries where NDP is relatively high compared to GDP. Similar to the level comparisons of NDP and GDP per hour worked, the growth rates of NDP and GDP per hour worked can be compared (see A.3).

In most OECD countries, the difference between GDP, NDP and GNI per hour worked is small since gross income inflows from abroad tends to be offset by gross outflows. This indicates that GDP per hour worked is a useful proxy for other measures of output in productivity calculations, especially when data availability restricts international comparability. However, the difference between GDP per hour worked and GNI per hour worked is significant for a few countries, notably Ireland and Luxembourg which are examples for countries where GDP per hour worked is significantly higher than GNI per hour worked, whereas Switzerland is a case where the opposite holds.

Sources



Further information

Ahmad, N., F. Lequiller, P. Marianna, D. Pilat, P. Schreyer and A. Wölfl (2003), “Comparing Labour Productivity Growth in the OECD Area: The Role of Measurement”, STI Working Papers 2003/14, OECD, Paris.






ALTERNATIVE MEASURES OF OUTPUT B.3.

Net Domestic Product and Gross Domestic Product per hour worked, 20061

Total economy, USD, current prices, current PPPs

0

10

20

30

40

50

60

70

80

USD mlnGDP per hour worked NDP per hour worked

Gross National Income and Gross Domestic Product per hour worked, 20061

Total economy, USD, current prices, current PPPs

0

10

20

30

40

50

60

70

80

USD mlnGDP per hour worked GNI per hour worked

1. 2005 for Australia, Canada, Ireland, Japan, Luxembourg, New Zealand, Poland and Turkey; 2004 for Mexico.


B.4. HETEROGENEITY OF LABOUR PRODUCTIVITY BY SIZE CLASS AND INDUSTRY

Many examples of productivity analyses focus on relatively aggregated industries, typically at the 2, or sometimes, 3-digit industry level. The focus is typically on productivity growth and, so, these analyses generally assume that homogeneity exists within these pre-defined industry groups. In other words: that the productivity growth observed at the aggregate level is representative of the growth at the business level. Where this assumption does not hold it is possible that changes in 2-digit labour productivity, say, could be attributed to labour merely because of changes in the market share of heterogeneous businesses (as well as changes that might be driven by entries and exits into the population).

Comprehensive estimates of productivity are therefore dependent on business level data. Unfortunately internationally comparable databases of individual businesses covering whole economies are not currently available; usually because of restrictions that preserve the anonymity and confidentiality of businesses and their data. The OECD’s Structural and Demographic Business Statistics database however provides a useful compromise between the need for very detailed data on businesses and the need for international comparability. The database provides information on value-added and employment in most OECD countries at the 4 digit industry level and by size class. It therefore allows researchers to conduct their analysis of productivity levels at least at a more detailed industry and size class level; although information on prices and price change are also needed when the focus is on productivity growth. The issue of prices notwithstanding the database is able to provide, on its own, at least a diagnostic assessment of the robustness of the homogeneity assumption by industry and additionally by size class.

Overview

The table illustrates that there is considerable heterogeneity in labour productivity figures across countries and industries; in other words that labour productivity figures can differ significantly by businesses. Of course such a comparison can overstate the true scale of heterogeneity across a sector since businesses within a given size class may indeed be homogeneous and indeed those in the largest size class will contribute disproportionately to overall value-added and employment. But the point of the table is merely to illustrate that heterogeneity can be considerable.

Comparisons of labour productivity by size class (see Structural and Demographic Business Statistics, 1996-2003, OECD) also show that in many industries, particularly in the manufacturing sector, and in most countries the larger the business the higher the labour productivity. This will no doubt, at least in part, reflect higher degrees of capital investment in larger businesses but it may also indicate larger economies of scale in larger businesses. Either way the comparisons across countries are of interest in themselves.

Definition

The accompanying table to this section tries to do this by comparing the coefficient of variation at the 2 digit industry level for OECD countries. The coefficient is calculated as the standard deviation of labour productivity estimates across the 4 digit industries within the 2-digit sector and across 5 size classes normalised by the simple un-weighted average of the labour productivity in these 4 digit industries and 5 size classes.

Comparability

For the United States, the coefficients of variation reflect the heterogeneity of labour productivity across size classes only and so tend to underestimate their size compared to estimates that reflect coefficients across size class and sector.

Ideally the estimates of labour productivity would have been produced using full-time-equivalent measures or number of hours worked but this has not been possible and instead, the number of persons engaged were used as the denominator for labour productivity, and for New Zealand and the United States number of employees. Because small businesses will typically have a higher percentage of persons engaged, and not on the pay-roll, than larger businesses it is possible that the estimates of relative labour productivity for small businesses in the United States are biased upwards. This bias is possibly reinforced by the fact that the numerator for the United States uses turnover and not value-added, which is used case for other countries (value-added at factor costs in the EU countries and value added at basic prices for Japan, Korea, Mexico and Turkey).

Data in the table below are from 2005, except for Czech Republic, Japan and Norway (2004), Mexico (2003), the United States (2002) and Turkey (2001).

Source

OECD, Structural and Demographic Business Statistics (SDBS), OECD Database.

For Further reading

Ahmad, Nadim, Francois Lequiller, Pascal Marianna, Dirk Pilat, Paul Schreyer and Anita Wölfl (2003), “Comparing Labour Productivity Growth in the OECD Area: The Role of Measurement”, STI Working Paper 2003/14, OECD, Paris.

OECD (2006), Structural and Demographic Business Statistics 1996-2003, 2006 Edition, OECD, Paris.

OECD (2003), OECD Science, Technology and Industry Scoreboard section D and Annex 1. OECD, Paris

OECD (2001), OECD Science, Technology and Industry Scoreboard, D.4. OECD, Paris.

OECD (2001), Measuring Productivity – OECD Manual Measurement of aggregate and Industry-level Productivity Growth, OECD, Paris.


HETEROGENEITY OF LABOUR PRODUCTIVITY BY SIZE CLASS AND INDUSTRY B.4.

Austria

Belgium

Czech Republic

Denmark

Finland

France

Germany

Greece

Hungary

Ireland

Italy

Japan

Korea

Luxembourg

Mexico

Netherlands

Norway

Poland

Portugal

Slovak republic

Spain

Sweden

Turkey

United Kingdom

United States1

15F

ood

prod

ucts

and

bev

erag

es0.

450.

310.

410.

240.

380.

360.

470.

600.

400.

430.

571.

081.

110.

320.

470.

740.

541.

470.

440.

191.

650.

620.

36

16T

obac

co p

rodu

cts

1.73

0.94

3.41

0.78

0.93

0.84

17T

extil

es0.

450.

190.

340.

100.

340.

140.

340.

350.

480.

240.

280.

590.

600.

390.

160.

250.

200.

490.

250.

140.

420.

200.

23

18W

earin

g ap

pare

l and

fur

0.27

0.11

0.31

0.62

0.65

0.54

0.68

0.67

0.37

0.44

0.53

0.06

0.35

0.54

0.37

19Le

athe

r an

d ar

ticle

s of

leat

her;

foot

wea

r0.

560.

100.

240.

310.

460.

630.

390.

120.

390.

390.

670.

390.

380.

360.

190.

170.

050.

460.

160.

22

20W

ood

& p

rodu

cts

of w

ood

& c

ork,

exc

. fur

nitu

re0.

430.

210.

290.

110.

140.

160.

330.

560.

210.

360.

140.

470.

630.

270.

290.

860.

490.

880.

300.

291.

220.

200.

20

21P

aper

and

pap

er p

rodu

cts

0.09

0.66

0.36

0.19

0.60

0.51

0.26

0.49

0.58

1.03

0.33

0.24

0.67

0.46

0.98

0.59

0.57

0.48

0.32

0.30

22P

rintin

g, p

ublis

hing

and

rec

orde

d m

edia

0.03

0.19

0.37

0.90

0.51

0.41

0.49

0.68

0.52

0.59

0.56

3.20

0.26

0.32

23C

oke

and

petr

oleu

m p

rodu

cts;

nuc

lear

fuel

4.24

0.76

0.32

1.33

1.56

1.03

1.71

0.36

0.85

24C

hem

ical

s an

d ch

emic

al p

rodu

cts

0.42

0.41

0.52

0.92

0.35

0.44

0.45

0.89

0.72

0.34

0.58

0.93

0.80

0.35

0.68

0.52

0.44

0.50

0.44

0.44

1.06

0.32

0.43

25R

ubbe

r an

d pl

astic

pro

duct

s0.

110.

260.

340.

050.

010.

070.

180.

520.

220.

420.

570.

940.

170.

310.

200.

280.

200.

280.

400.

060.

18

26O

ther

non

-met

allic

min

eral

pro

duct

s0.

450.

260.

460.

210.

402.

950.

340.

700.

380.

430.

300.

961.

800.

320.

300.

840.

550.

360.

680.

330.

950.

260.

24

27B

asic

met

als

0.05

0.27

0.61

0.11

0.79

0.37

0.81

0.27

0.38

0.76

0.86

0.22

1.03

0.68

0.56

0.40

0.52

0.45

0.21

0.37

28F

abric

ated

met

al p

rodu

cts,

exc

. mac

hine

ry0.

210.

190.

230.

100.

140.

120.

220.

430.

560.

250.

220.

280.

190.

620.

210.

150.

300.

620.

320.

280.

270.

510.

25

29M

achi

nery

and

equ

ipm

ent,

nec

0.16

0.30

0.18

0.07

0.24

0.26

0.19

0.28

0.17

0.24

0.39

0.55

0.84

0.16

0.35

0.37

0.19

0.35

0.38

0.44

0.60

0.36

30O

ffice

, acc

ount

ing

& c

ompu

ting

mac

hine

ry0.

060.

240.

180.

150.

270.

890.

790.

400.

440.

870.

420.

610.

540.

940.

190.

590.

100.

43

31E

lect

rical

mac

hine

ry &

app

arat

us, n

ec0.

230.

260.

140.

180.

560.

200.

270.

740.

260.

320.

390.

500.

590.

440.

290.

410.

530.

840.

340.

380.

480.

17

32R

adio

, TV

, com

mun

icat

ion

equi

pmen

t0.

440.

030.

280.

191.

470.

290.

250.

630.

280.

700.

950.

510.

510.

530.

440.

430.

490.

240.

240.

740.

28

33M

edic

al, p

reci

sion

& o

ptic

al in

stru

m.;

wat

ch &

clo

cks

0.28

0.38

0.23

0.30

0.17

0.21

1.52

0.48

0.43

0.38

0.66

0.37

0.69

0.34

0.33

0.29

0.34

0.25

0.26

0.29

0.60

0.33

34M

otor

veh

icle

s, tr

aile

rs a

nd s

emi-t

raile

rs0.

530.

440.

130.

600.

301.

220.

270.

250.

380.

661.

250.

350.

360.

770.

550.

230.

301.

060.

470.

240.

62

35O

ther

tran

spor

t equ

ipm

ent

0.71

0.12

0.35

0.34

0.57

0.28

0.62

0.60

1.28

0.37

0.36

0.46

0.72

0.35

0.39

36F

urni

ture

; man

ufac

turin

g, n

ec0.

280.

420.

300.

240.

170.

260.

410.

440.

230.

330.

430.

450.

060.

470.

750.

320.

380.

390.

630.

280.

120.

720.

21

37R

ecyc

ling

0.20

0.29

0.03

0.28

0.77

0.27

0.01

0.43

0.23

0.20

0.43

0.52

0.24

0.46

0.40

0.32

0.21

40E

lect

ricity

, gas

, ste

am a

nd h

ot w

ater

sup

ply

3.96

2.22

0.89

0.79

1.17

0.78

1.05

0.06

0.60

1.94

0.24

1.05

0.68

0.36

41C

olle

ctio

n, p

urifi

catio

n an

d di

strib

utio

n of

wat

er0.

220.

280.

060.

170.

250.

190.

550.

410.

130.

37

45C

onst

ruct

ion

0.18

0.35

0.34

0.13

0.12

0.20

0.48

0.56

0.41

0.29

0.15

0.80

0.24

0.18

0.25

0.54

0.16

0.13

0.23

0.22

50S

ale,

mai

nt.,r

epai

r ve

hic/

cycl

e, p

erso

n., h

/hol

d gd

.0.

410.

640.

260.

290.

290.

370.

370.

480.

470.

640.

290.

370.

640.

230.

350.

440.

410.

861.

060.

290.

380.

65

51W

/sal

e. tr

ade

& c

omm

issi

on e

xc m

otor

veh

. & c

ycle

0.29

0.38

0.34

0.26

0.33

0.19

0.32

0.46

0.34

0.83

0.34

0.40

1.24

0.30

0.28

0.39

0.38

0.41

0.33

0.16

0.29

0.38

52R

etai

l tra

de e

xc. m

otor

veh

. & c

yc.;

repa

ir h/

hold

gds

.0.

350.

280.

350.

380.

150.

280.

440.

460.

580.

320.

420.

030.

720.

290.

380.

550.

440.

660.

270.

240.

390.

10

55H

otel

s an

d re

stau

rant

s0.

230.

440.

700.

340.

130.

160.

330.

630.

490.

330.

160.

410.

310.

230.

570.

350.

400.

310.

150.

450.

27

60La

nd tr

ansp

ort,

tran

spor

t via

pip

elin

es0.

140.

090.

120.

250.

080.

032.

660.

300.

610.

270.

030.

090.

470.

140.

410.

110.

580.

20

61W

ater

tran

spor

t1.

271.

310.

610.

820.

261.

150.

140.

601.

030.

130.

630.

661.

460.

38

62A

ir tr

ansp

ort

0.29

63S

uppo

rt. &

aux

. tra

nspo

rt a

ct.,

trav

el a

genc

ies

activ

.0.

700.

310.

300.

170.

250.

070.

280.

390.

410.

120.

710.

560.

120.

610.

201.

090.

150.

230.

09

64P

ost a

nd te

leco

mm

unic

atio

ns0.

661.

400.

890.

671.

180.

670.

811.

210.

600.

930.

800.

800.

460.

930.

371.

311.

150.

19

65F

inan

cial

inte

rmed

iatio

n, e

xcep

t ins

uran

ce a

nd p

ensi

on fu

ndin

g3.

090.

27

66In

sura

nce

and

pens

ion

fund

ing,

exc

ept c

ompu

lsor

y so

cial

sec

urity

0.37

67A

ctiv

ities

aux

iliar

y to

fina

ncia

l int

erm

edia

tion

0.38

70R

eal e

stat

e ac

tiviti

es0.

750.

490.

980.

22

71R

entin

g of

mac

h. w

ithou

t ope

rato

r; r

ent.

of p

ers.

gds

.0.

290.

270.

510.

110.

030.

090.

310.

200.

922.

170.

470.

340.

500.

330.

260.

17

72C

ompu

ter

and

rela

ted

activ

ities

0.50

0.83

0.84

0.22

73R

esea

rch

and

deve

lopm

ent

0.23

1.15

0.48

0.07

74O

ther

bus

ines

s ac

tiviti

es0.

440.

510.

790.

280.

320.

390.

500.

910.

280.

660.

430.

530.

660.

380.

830.

730.

740.

470.

351.

06

ISIC

Rev

isio

n 3

an

d c

od

e d

escr

ipti

on

Hete

rog

en

eit

y i

n L

ab

ou

r P

rod

ucti

vit

y -

Co

eff

icie

nt

of

Vari

ati

on

at

the 2

dig

it i

nd

ustr

y l

evel

1.

Fo

r th

e U

nite

d S

tate

s,

the

co

eff

icie

nts

of

va

ria

tio

n r

efle

ct

the

he

tero

ge

ne

ity o

f la

bo

ur

pro

du

ctivity a

cro

ss s

ize

cla

sse

s o

nly

an

d s

o t

en

d t

o u

nd

ere

stim

ate

th

eir

siz

e c

om

pa

red

to

estim

ate

s t

ha

t re

fle

ct

co

eff

icie

nts

acro

ss s

ize

cla

ss a

nd

se

cto

r. F

or

this

co

un

try,

turn

ove

r is

use

d a

s n

um

era

tor

inste

ad

of

va

lue

ad

de

d.

Th

e c

oe

ffic

ien

t o

f va

ria

tio

n m

ea

su

res t

he

sta

nd

ard

de

via

tio

n o

f la

bo

ur

pro

du

ctivity n

orm

alis

ed

by t

he

ave

rag

e o

f la

bo

ur

pro

du

ctivity b

y s

ize

cla

ss a

nd

4 d

igit in

du

str

ies w

ith

in e

ach

2 d

igit s

ecto

r. I

t is

im

po

rta

nt

to n

ote

th

at

the

de

no

min

ato

r is

a s

imp

le a

ve

rag

e o

f th

e s

ize

cla

sse

s a

nd

4 d

igit in

du

str

ies u

se

d a

nd

, so

, is

no

t w

eig

hte

d b

y t

he

re

lative

co

ntr

ibu

tio

ns o

f a

pa

rtic

ula

r siz

e c

lass o

r 4

dig

it in

du

str

y t

o v

alu

e-a

dd

ed

or

em

plo

ym

en

t.


B.5. HETEROGENEITY OF LABOUR PRODUCTIVITY BY SIZE CLASS, TOTAL MANUFACTURING

The previous section demonstrated the, at times significant, heterogeneity in labour productivity across size classes and 4-digit industry sectors within a given 2-digit industry sector, and, so, the care that is needed when drawing conclusions, especially at the international level, based solely on 2 digit industry sector averages.

The table and the chart below illustrate this heterogeneity in a simplified form by showing the normalised labour productivity in the manufacturing sector by employment size class.

Definition

The normalised labour productivity figures shown below are calculated as labour productivity in a given size class as a percentage of the average labour productivity across all size classes. The size class breakdown used provides for the best comparability across countries given the varying data collection practices across countries. But for some countries slightly different conventions are needed as described in the footnotes to the table.

Overview

For every country, the highest labour productivity is observed in the biggest enterprise size class, possibly indicating higher degrees of capital investment in larger businesses, but it may also indicate economies of scale.

For the majority of countries (about 75 %), labour productivity increases monotonically with size class. Interestingly in Denmark, the Slovak republic and, to a lesser extent, the United Kingdom and the United States, labour productivity figures across small and medium-size enterprises are significantly more homogeneous than in other countries but this in part reflects the result of averaging throughout the manufacturing sector. At the 2 digit level for example the picture is more heterogeneous.

Source

OECD Structural and Demographic Business Statistics (SDBS), OECD database.

Further information



OECD (2003), OECD Science, Technology and Industry Scoreboard section D and Annex 1. OECD, Paris

OECD (2001), OECD Science, Technology and Industry Scoreboard, D.4. OECD, Paris.

OECD (2001), Measuring Productivity – OECD Manual Measurement of aggregate and Industry-level Productivity Growth, OECD, Paris.


HETEROGENEITY OF LABOUR PRODUCTIVITY BY SIZE CLASS, TOTAL MANUFACTURING

B.5.

Normalised Labour productivity, manufacturing

By size class as a percentage of total average, 20051

1-9 2 10-19 20-49

3 50-249

4 250+

5

Australia 68.0 61.3 73.5 86.4 139.5 Austria 58.7 64.8 73.6 92.0 126.5

Belgium 47.2 60.1 72.5 90.9 132.0 Czech republic 56.1 66.2 77.6 90.5 130.6

Denmark 78.0 72.9 83.0 93.4 116.7 Finland 72.5 68.3 69.2 82.1 121.9

France 59.1 73.3 81.0 86.0 126.0 Germany 49.8 58.1 74.3 88.7 122.5

Hungary 29.7 47.3 55.2 73.6 156.3 Ireland 30.0 28.1 31.6 67.8 154.7

Italy 54.0 81.6 99.0 122.1 146.2 Japan 42.9 55.4 65.8 97.8 157.8

Korea 41.3 50.9 59.3 88.0 189.8 Luxembourg 68.0 61.2 65.0 90.9 113.2

Mexico6

21.9 47.4 58.2 89.0 141.7 Netherlands 46.2 73.7 76.2 94.0 146.5

Norway 64.9 75.1 84.7 98.7 123.9 Poland 35.0 58.2 60.6 75.9 162.1

Portugal 49.3 66.6 79.8 105.4 191.6 Slovak republic 90.0 72.7 71.9 76.7 117.6

Spain 53.4 67.7 77.6 101.4 165.5 Sweden 73.7 94.4 92.8 92.6 113.6

Turkey 32.7 46.7 69.9 130.6 United Kingdom 74.5 74.4 81.4 90.1 122.0

United States 7 54.1 46.8 53.8 68.3 129.8

1. 2001 for Turkey, 2002 for the United States, 2003 for Mexico and New Zealand, 2004 for the Czech Republic, Japan and Norway. 2. For Ireland, only enterprises with 3 or more persons engaged are reflected, while data for Japan, Korea and Turkey do not include establishments with fewer than 4, 5 and 10 persons engaged, respectively. 3. 20-99 for the United States. 4. 50-199 for Australia, Korea and Turkey, 100-499 for the United States. 5. 200+ for Australia and Turkey, 500+ for the United States. 6. The data for Mexico refer to the following size classes: 0-10, 11-20, 21-50, 51-250, 251+ 7. For the United States, turnover is used as numerator. 8. 2002 for the United States, 2004 for Japan. 9. 20-99 for the United States. 10. 100-499 for the United States 11. 500+ for the United States


C. PRODUCTIVITY GROWTH BY INDUSTRY


C.1. CONTRIBUTION OF KEY ACTIVITIES TO AGGREGATE PRODUCTIVITY GROWTH

A breakdown of productivity growth by economic activity can highlight industries that are particularly important for overall productivity performance.

Definition

Labour productivity growth can be calculated as the difference between the rate of growth of output or value added and the rate of growth of labour input. Calculating a sector’s contribution to aggregate productivity growth requires a number of simple steps, as explained in the OECD Productivity Manual. First, the aggregate rate of change in value added is a share-weighted average of the industry-specific rates of change in value added, with weights reflecting the current price share of each industry in current price value added. On the input side, aggregation of industry-level labour input is achieved by weighting the growth rates of total employment (National Accounts detailed series on hours worked by industry are not available for many OECD countries) with each industry’s share in total labour compensation. Aggregate labour productivity growth can then be calculated as the difference between the aggregate growth in value added and the aggregate growth in labour input. An industry’s contribution to aggregate labour productivity growth is therefore the difference between its contribution to total value added and total labour input. If value added and labour shares are the same, total labour productivity growth is a simple weighted average of industry-specific labour productivity growth.

Similar approaches can be followed when production, instead of value added, is used as the output measure.

“Market services” refers to ISIC Rev.3 service activities 50 to 74. Further details are available in the chapter on productivity growth in services.

Comparability

For the graphs, the contributions have been scaled so that the sum of the absolute contributions equals 100. Therefore, irrespective of countries’ actual total labour productivity growth, the relative contributions of the different sectors can be compared. Difficulties in measuring output and productivity in services sectors should also be taken into consideration when interpreting the results.

In charts, data for Japan do not refer to 1995-2000 but to 1996-2000; data do not refer to 2000-2006 but to 2000-2002 for New Zealand; 2000-2003 for Australia; 2000-2004 for Portugal and Sweden; 2000-2005 for Canada, France, Hungary, Spain and the United States.

Overview

Over the period 2000-2006, “market services” accounted for the bulk of labour productivity growth in many OECD countries. Namely, in Greece, Luxembourg, New-Zealand, Norway, the United Kingdom and the United States, ‘market services’ accounted for over 55% of aggregate labour productivity growth. However, the highest aggregate labour productivity growth performances can still be attributed to the manufacturing sector. This was the case in the Czech Republic, Finland, Korea, the Slovak Republic and Sweden.

The contribution of “market services” to labour productivity growth has increased between 1995-2000 and 2000-2006 in Belgium, the Czech Republic, France, Luxembourg and New Zealand. This growing contribution of market services is sometimes linked to an increasing share in total value added, but in the Czech Republic, Japan and New Zealand, for example, it also reflects faster labour productivity growth in the market service sector. However, in several other countries, labour productivity growth in market services has slowed down in the most recent years.

Sources


For further reading

OECD (2004), “Clocking In (and Out): Several Facets of Working Time”, OECD Employment Outlook 2004, Chapter 1, OECD, Paris.


Pilat, D. and P. Schreyer (2004), “The OECD Productivity Database – An Overview”, International Productivity Monitor, No. 8, Spring, OECD, Paris, pp. 59-65.


CONTRIBUTION OF KEY ACTIVITIES TO AGGREGATE PRODUCTIVITY GROWTH C.1.

Contribution of key activities to growth of value added per person employed

Percentage points

2000-2006 (or latest year available)

4.5 4.0 3.8

3.1

2.82.0

1.7 1.71.7

1.7 1.61.5 1.3

1.21.2

1.1

1.11.0

1.0

1.0

0.70.7

-0.1

-0.5-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

Market services Manufacturing Other industries Other services Total economy labour productivity growth

1995-2000 (or earliest year available)

4.13.6

2.72.7

2.5

2.4 2.3 2.2

2.11.9

1.9 1.7

1.6

1.4 1.4

1.41.3

1.2

0.8

0.8

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

Market services Manufacturing Other industries Other services Total economy labour productivity growth


C.2. PRODUCTIVITY GROWTH IN MANUFACTURING

The manufacturing sector has historically been the main driver of aggregate productivity growth in OECD countries. While its contribution to aggregate productivity growth has become less important in recent years, particularly in some OECD countries, it still shows strong performance in many industries.

Definition

In the indicator hereafter, for each manufacturing industry, labour productivity growth is calculated as the difference between the rate of growth of the industry’s value added and the rate of growth of the industry’s total employment (number of persons engaged).

Comparability

Due to rapid technological progress, measuring productivity growth in some industries like electrical and optical equipment is particularly difficult. Some countries use so-called hedonic deflators to address the rapid changes in quality and characteristics of the ICT goods produced within this industry. Other countries do not, which implies that the international comparability of productivity growth rates in this industry may be somewhat limited.

Examining the role of ICT-producing sectors in economic growth is heavily influenced by measurement problems, both regarding outputs and inputs. The key measurement problem for the manufacturing of ICT goods on both the output and input side concerns prices, in particular how to statistically capture significant quality improvements associated with technological advances in goods such as computers and semi-conductors. The use of hedonic deflators is generally considered as the best way to address these problems. Several countries currently use hedonic methods to deflate output in the computer industry (e.g. Canada, Denmark, France, Sweden and the United States), however, these countries do not use exactly the same method. Some countries, such as the United States, apply their own hedonic deflator, others apply the United States hedonic deflator adjusted for exchange rates, and yet other OECD countries apply conventional methods to account for quality change when deriving deflators.

Adjusting for these methodological differences in computer deflators for the purpose of a cross-country comparison is difficult, since there are considerable cross-country differences in industrial specialisation. Only few OECD countries produce computers, where price falls have been very rapid; many only produce peripheral equipment, such as computer terminals. Similar differences in industry composition exist in Radio, Television and Communication Equipment (ISIC 32), which includes the semi-conductor industry. The differences in the composition of output are typically larger than in computer investment, where standardised approaches have been applied (e.g. Schreyer et al. 2003).

In charts, data for Japan refer to 1996-2000 and not to 1995-2000; data for Canada, Portugal and Sweden refer to 2000-2004; 2001-2005 for Poland; and not to 2000-2005.

Overview

For most OECD countries, manufacturing productivity growth was slower during 2000-2005 than in the period 1995-2000 with the exception of a few countries such as Japan, Norway, the Slovak Republic and the United Kingdom. Notable reductions in the growth of manufacturing productivity in recent years have occurred in Austria, Canada, Italy and Korea, possibly reflecting a strong structural shift in the manufacturing sector in these countries. Within manufacturing, large differences can be observed. High- and medium-high technology industries, such as electrical and optical equipment and transport equipment, have typically experienced relatively high rates of productivity growth while low-technology manufacturing industries, such as textiles, have tended to generate slightly lower rates of productivity growth. However, growth rates for the textiles industry remained quite high in 2000-2005 for some OECD countries, including Czech Republic, France, Norway, the United Kingdom and the United States – important in the face of increasing imports of low cost textiles from developing countries.

Manufacture of electrical and optical equipment is one of the industries with the highest rates of productivity growth, despite some slowing down since late 1990s. During the period 2000-2005 some OECD countries sustained annual productivity growth in this sector of over 10%, including Czech Republic, Finland, Hungary, Japan, Sweden and the United States.

Sources


OECD STAN Database (forthcoming 2008 edition).

For further reading

Pilat, D., A. Cimper, K. Olsen and C. Webb (2006), “The Changing Nature of Manufacturing in OECD Economies”, STI Working Paper 2006/9, OECD, Paris.

Triplett, Jack (2004), “Handbook on Hedonic Indexes and Quality Adjustment in Price Indexes: Special Application to Information Technology Products”, STI Working Papers 2004/9, OECD, Paris


PRODUCTIVITY GROWTH IN MANUFACTURING C.2.

Value added per person employed, percentage change at annual rate

2000-2005 and 1995-2000 Total manufacturing

-2

0

2

4

6

8

10

12

%

Textiles and textile products (ISIC 17 to 18)

-8

-6

-4

-2

0

2

4

6

8

10

12%

Basic metals and metal products (ISIC 27 to 28)

-8

-6

-4

-2

0

2

4

6

8

10

12%

Electrical optical equipment (ISIC 30 to 33)

-5

0

5

10

15

20

25

30

35%

Transport equipment (ISIC 34 to 35)

-5

0

5

10

15

20%


C.3. PRODUCTIVITY GROWTH IN SERVICES

Measuring productivity in services tends to be more challenging than measuring productivity in goods production. However, as the service sector now accounts for over 70% of OECD aggregate GDP and employment, and continues to grow, it is increasingly important to understand the impact of different services sectors on aggregate productivity.

Definition

In the indicator hereafter, for each service sector, labour productivity growth is calculated as the difference between the rate of growth of the service sector’s value added and the rate of growth of the service sector’s total employment (number of persons engaged).

Comparability

Measuring output and productivity growth in many services is not straightforward. What exactly does a lawyer or economist produce? How can the changing pricing schemes of telecommunications providers be compared over time? And how should one measure the ‘quantity’ of health services provided by hospitals? These and similar questions arise when statisticians attempt to measure the output of service industries and the difficulty of this task is hard to overstate.

Generally, measuring volumes in the National Accounts requires current price values of flows of goods and services that can be divided into volume and price components. Typically, this is more difficult for services than for goods. Characteristics of goods can normally be identified and changes in quantities and qualities are, in principle, measurable. However for services, even quantitative changes are often hard to measure, let alone quality change. Omitting qualitative changes does not necessarily mean that volume of services will be underestimated. For example, technical progress may improve the speed with which letters are delivered (a quality improvement) but may also lead to a decline in quality of service when self service in post offices leads to longer queuing times or increased distances to the nearest post office all of which may increase the burden on customers.

‘Market services’ here refers to ISIC Rev.3 service activities 50 to 74. In other words, it is an activity based proxy and excludes those industries that tend to be dominated by non-market production such as health, education and community and social services. Note that no adjustments have been made to remove estimates of household rentals (actual and imputed), which has no labour input associated, from value added (part of ISIC 70) - current practice when calculating labour productivity by major economic activity in OECD’s System of Unit Labour Costs (ULC) Indicators. Also, this data set uses hours worked, rather than persons engaged, as a measure of labour input wherever possible. Therefore, estimates of productivity growth of market services here may differ from those presented in the section focussing on ULCs.

In charts, total market services data for Japan do not include hotels and restaurants (ISIC 55) and do not refer to 1995-2000 but to 1996-2000; data do not refer to 2000-2005 but to 2000-2004 for Canada, Portugal and Sweden; 2001-2005 for Poland.

Overview

Several OECD countries experienced reduced growth in total market service labour productivity during the period 2000-05 compared to 1995-2000. The most notable falls occurred in Mexico, Portugal, the Slovak Republic and Switzerland. Countries that experienced a marked increase included Belgium, the Czech Republic, Hungary, Ireland and Japan. Productivity growth in market services continued to be very small in Italy and Spain. The variation across services sectors and across countries is considerable. The services sectors with the highest rate of productivity growth tend to be those that invest more in ICT and have more highly skilled workforces. Sometimes labelled ‘Knowledge-intensive services’, these include industries such as post and telecommunications (ISIC 64); finance and insurance; and certain other business services such as computer services (ISIC 72). Labour productivity in the hotels and restaurants sector seems to have declined considerably across OECD countries with few exceptions. However, the steep falls are partly due to the effect of using persons employed as a measure of labour input rather than hours worked - there are significant numbers of part-time workers in this sector. For most of those countries where hours worked data are available, estimated productivity still declines but to a lesser degree. It is also worth noting that any increase in the quality of output may not be captured in such services.

Sources


OECD STAN Database (forthcoming 2008 edition).

For further reading


OECD (2005), Enhancing Services Sector Performance, OECD, Paris.

Wölfl, A. (2005), “The Service Economy in OECD Countries”, STI Working Paper 2005/3, OECD, Paris.

Wölfl, A. (2003), “Productivity Growth in Service Industries: An Assessment of Recent Patterns and the Role of Measurement”, STI Working Paper 2003/7, OECD, Paris.


PRODUCTIVITY GROWTH IN SERVICES C.3.

Value added per person employed, percentage change at annual rate

2000-2005 and 1995-2000 Total market services

-3

-2

-1

0

1

2

3

4

%

Wholesale and Retail trade (ISIC 50 to 52)

-6

-4

-2

0

2

4

6

8

10%

Hotels and restaurants (ISIC 55)

-5

-4

-3

-2

-1

0

1

2

3

4%

Transport, storage, communication (ISIC 60 to 64)

-1

0

1

2

3

4

5

6

7

8%

Finance and insurance (ISIC 65 to 67)

-4

-2

0

2

4

6

8

10

12

14

16%


D. IMPACT OF LABOUR PRODUCTIVITY ON UNIT LABOUR COSTS


D.1. UNIT LABOUR COSTS AND LABOUR PRODUCTIVITY – TOTAL ECONOMY

Unit labour costs measure the average cost of labour per unit of output. As such, a unit labour cost (ULC) represents a link between productivity and the cost of labour in producing output, and can be an indicator of inflationary pressure on producer prices.

Definition

Unit labour costs are calculated as the ratio of total labour costs to real output, or equivalently, as the ratio of total labour costs per hour (or employee numbers if hours data is not available) to output per hour, (or total employment if hours data is not available). Therefore, a ULC indicator can be expressed as the following equation:

Comparability

The OECD provides users with two measures of Labour Productivity at the Total Economy level. The OECD Productivity Database (PD): www.oecd.org/statistics/productivity and the OECD System of Unit Labour Cost and Related Indicators database (ULCRI): http://stats.oecd.org/mei/default.asp?rev=3. The PD is concerned with providing the best possible internationally comparable annual estimates of labour productivity and multi-factor productivity (MFP) at the Total Economy level and providing underlying data for capital services by asset type. Whereas labour productivity data in the ULCRI database is produced as a by-product of calculating unit labour cost indicators.

The major difference between methodologies used for compiling labour productivity statistics in the two databases is the use of Hours data. The DP only uses Hours data (total employment multiplied by average hours worked) as the labour input whereas the ULCRI database uses Employment in Persons data if Hours data is not available in the OECD Annual National Accounts database (ANA). In addition the OECD ULCRI database uses Gross Value Added for the output measure whereas the PD uses Total Gross Domestic Product (GDP).

This variation in input data leads to differences (assessed by Labour Productivity annual growth rate correlation) for some countries, the most acutely affected being: Australia, New Zealand, Poland, Portugal, Sweden, United Kingdom and the United States. The major cause of differences in these countries is having different labour input measures. These differences in values and input data are fully explained and outlined in metadata and methodology papers compiled for both databases. At the Total Economy level the user is best advised to use the PD estimates for cross country comparisons. The user should understand that

the OECD ULCRI provides a number of competitiveness measures across eight economic activities for all OECD member countries (where possible). To ensure internal consistency in the system, a coherent methodology has been established for the system as a whole using the ANA database as the sole data source of annual data.

Long-term trends

Stronger growth in labour productivity than in average labour compensation will have a downward impact on growth in unit labour costs as shown by the formula on this page. However, the relationship between changes in unit labour costs and labour productivity is of course dependent on developments in average labour compensation which is generally related to the rate of inflation within a country.

In 1986-2006, annual average growth in unit labour costs in the Total Economy for the following nine OECD member countries was above 5 percent: Slovak Republic, Portugal, Czech Republic, Iceland, Greece, Hungary, Poland, Mexico and Turkey. Only Mexico and Turkey have maintained this annual average growth rate above 5 percent in the 2001-06 period. The remaining seven countries all saw their annual average unit labour cost growth rate fall to below this mark. While a number of these countries experienced strong annual average labour productivity growth in the 2001-06 period, their reduced unit labour cost growth is more associated with lower average labour compensation annual growth rates reflecting a lower inflation environment.

Japan stands out as having negative annual average unit labour cost growth for the whole 1986-2006 period. Japanese annual average labour productivity growth was 2.7 percent for this same period, only slightly above the OECD average, implying that annual average labour compensation growth in Japan for this period was very low relative to other countries.

Sources

OECD System of Unit Labour Costs and Related Indicators: www.stats.oecd.org/mei/default.asp?rev=3



For further reading

OECD Main Economic Indicators March 2007; Introduction of Unit Labour Cost Indicators: www.oecd.org/dataoecd/30/38/38219407.pdf Working Party on Statistics (SWIC), 17-18 November 2005; Development of comparable quarterly unit labour cost indexes for OECD countries. Launch of the OECD System of Unit Labour Cost and Related Indicators: www.oecd.org/dataoecd/0/19/38492896.pdf Working Group on Short-Term Economic Prospects May 2007; Introduction of new Competitiveness Indicators: www.oecd.org/dataoecd/37/12/38736937.pdf

http://www.oecd.org/statistics/productivity

http://stats.oecd.org/mei/default.asp?rev=3

http://www.stats.oecd.org/mei/default.asp?rev=3


http://www.oecd.org/dataoecd/30/38/38219407.pdf




UNIT LABOUR COSTS AND LABOUR PRODUCTIVITY – TOTAL ECONOMY D.1.

2001-2006, Unit Labour Costs compared with Labour Productivity Total Economy, percentage change at annual rate

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

10

% UNIT LABOUR COSTS LABOUR PRODUCTIVITY

12.9

1986-2006, Unit Labour Costs compared with Labour Productivity

Total Economy, percentage change at annual rate

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

10

% UNIT LABOUR COSTS LABOUR PRODUCTIVITY

10.4

24.5

52.6

Note: Please see Unit Labour Cost notes for details on country data coverage


D.2. UNIT LABOUR COSTS, LABOUR PRODUCTIVITY AND LABOUR COMPENSATION PER UNIT LABOUR INPUT– INDUSTRY

Definition

Annual labour productivity is defined as constant price gross value added divided by total labour input. In the case of total labour input, this is total hours worked by those in employment for Australia, Austria, Canada, Denmark, Finland, France, Germany, Greece, Hungary, Italy, Korea, Norway, Slovak Republic, Spain and Sweden. For all other countries total employment in persons is used as the total labour input measure.

Annual labour compensation per unit of labour input (average labour compensation) is defined as compensation of employees divided by total hours worked by employees or total employees (using the country split outlined in the above paragraph). In all cases the OECD System of Unit Labour Cost and Related Indicators database sources country data via the OECD’s Annual National Accounts database.

Comparability

The graphs in sections D2 and D3 concentrate on the economic activities Industry (the sum of ISIC divisions C, D and E) and Market Services (activity proxy based on the sum of ISIC divisions G, H, I, J and K). These economic activities have been highlighted as it is these activities where economy wide changes can best be seen, with most economies having seen a move in economic resource allocation from Industry to Market Services over the last 10-15 years.

From comparing the two sections, one can notice a similar development for annual growth rates for average labour compensation between Industry and Market Services for most countries. However, higher annual average growth rates for labour productivity are much more evident in Industry, resulting in lower overall growth in unit labour costs for Industry relative to Market Services. This effect may reflect to some extent the impact of globalisation on Industry (in particular Manufacturing), requiring countries to raise productivity in order for their businesses to be more competitive and survive in the face of global competition. Growth in labour productivity can also arise from more intensive use of capital, which may be more evident within at least some aspects of Industry (e.g. mining, equipment manufacturing) than some service industries which are more labour intensive (e.g. legal services). However, as explained in section C3 of this compendium, measuring productivity in service industries is particularly difficult due to the problem of measuring the volume (i.e. output) of service activities. Consequently, the quality of measuring the outputs of Market Services can differ across countries, thus affecting the quality of labour productivity measures and ultimately unit labour costs. One concern is where countries continue to use input measures (e.g. hours worked or employment) as a proxy

for output in some service activities which implies zero labour productivity growth (although an aggregate level adjustments for ‘estimated’ labour productivity growth may be made). It is therefore possible that productivity growth in Market Services for those countries measuring services output in this way may have been understated over the periods shown in the graphs.

Long-term trends

For the economic activity ‘Industry’ there has been a sustained push by multinational corporations to move production (normally, manufacturing activities) to countries that offer lower average labour compensation levels. To offset this relocation of production processes, OECD member countries have three options to remain competitive: keep labour compensation growth lower than other countries; increase labour productivity faster than other countries; and/or devalue their currency. If the first two actions occur, then the result leads to lower or negative unit labour cost growth. This combination of strong labour productivity growth and subdued growth in average labour compensation has been most evident in Japan and Ireland over the longer term (1986-2006). More recently, many countries have achieved strong labour productivity growth whilst keeping growth in average labour compensation subdued leading to declines in unit labour costs. In particular, as shown in the graph on the following page, for the period 2001-06; Japan, Poland, Sweden, Ireland, Finland and the United States all had annual average labour productivity growth above 5% with annual average labour compensation growth rates below 5%. An increase in unit labour costs over time may create pressure on producer prices. OECD member countries with long-term annual average unit labour cost growth in excess of 5 percent are: Hungary, Poland, Mexico and Turkey. All these countries have recorded high producer price growth at some time between 1986-2006.

Sources

OECD System of Unit Labour Costs and Related Indicators: www.stats.oecd.org/mei/default.asp?rev=3 OECD Annual National Accounts Database. OECD Productivity Database: www.oecd.org/statistics/productivity

For further reading OECD Main Economic Indicators March 2007; Introduction of Unit Labour Cost Indicators: www.oecd.org/dataoecd/30/38/38219407.pdf Working Party on Statistics (SWIC), 17-18 November 2005; Development of comparable quarterly unit labour cost indexes for OECD countries. Launch of the OECD System of Unit Labour Cost Indicators: www.oecd.org/dataoecd/0/19/38492896.pdf Working Group on Short-Term Economic Prospects May 2007; Introduction of new Competitiveness Indicators: www.oecd.org/dataoecd/37/12/38736937.pdf







UNIT LABOUR COSTS, LABOUR PRODUCTIVITY AND LABOUR COMPENSATION PER UNIT LABOUR INPUT– INDUSTRY

D.2.

2001-2006, Unit Labour Costs, Labour Productivity and labour Compensation per Unit of Labour Input Industry, percentage change at annual rate

-5

-4

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

10

11

12

13

%UNIT LABOUR COSTS LABOUR COMPENSATION LABOUR PRODUCTIVITY

16.9

1986-2006, Unit Labour Costs, Labour Productivity and labour Compensation per Unit of Labour Input

Industry, percentage change at annual rate

-5

-4

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

10

11

12

13


16.1 52.6/55.0



D.3. UNIT LABOUR COSTS, LABOUR PRODUCTIVITY AND LABOUR COMPENSATION PER UNIT LABOUR INPUT– MARKET SERVICES

The OECD has developed a unique new statistical product referred to as the OECD System of Unit Labour Cost and Related Indicators. The System outputs, which are updated at the end of each quarter, consist of long time series of annual and quarterly unit labour cost and related indicators. The related indicators include annual time series for: Exchange Rate Adjusted Unit Labour Costs; Labour Income Share ratios; Labour Productivity (per Unit Labour Input); and, Labour Compensation per Unit Labour Input (in both national currency and USD PPP adjusted). Data are available for all OECD member countries (and zones) for eight economic activities (sectors), namely: Total Economy; Manufacturing; Industry; Construction; Trade, Transport and Communication; Financial and Business Services; Market Services; and Business Sector.

Definition

The economic activity Market Services (ISIC G_K) is calculated as the addition of the activities: Trade, transport, and communication (ISIC G_I); and, Financial and business services (ISIC JK). Where the components are not strictly additive across these activities, i.e. gross value added in constant prices, annual chain-linking is undertaken using gross value added in current prices as the weights. For a majority of OECD member countries (except France, Germany, Ireland, Luxembourg, the United Kingdom, and the United States) the proportion of G_I in Market Services is over half. Over the last 10 years virtually all OECD member countries have recorded faster growth in Financial and business services (JK) than for Trade, transport, and communication services (G_I).

Comparability

Achieving comparability across countries and sectors is a major challenge, particularly for series compiled on a quarterly basis. This stems largely from a lack of uniformity in labour cost data available on an infra-annual basis across different economic activities within and across countries. In addition, coherence with quarterly indicators of real output may often be poor leading to large volatility in a derived statistic such as a unit labour cost. The OECD has attempted to overcome these problems by developing a stepwise framework for choosing suitable quarterly indicator data, that is then benchmarked to more reliable annual data to form a consistent set of temporally disaggregated quarterly time series of total labour costs and real output. These series are then used to compile quarterly unit labour cost indexes. The OECD is currently investigating the possibility of producing unit labour costs for the eight EU member states that are not part of the OECD (e.g. Bulgaria, Romania, Slovenia, etc) and the BRICS countries, namely: Brazil, China, India, Russian Federation, and South Africa.

Long-term trends

Looking at the graph of long-term (1986-2006) annual average growth rates for unit labour costs in Market Services for OECD member countries, the most striking feature is the four countries with unit labour cost (ULC) growth around or above 10 percent: Poland, Hungary, Slovak Republic and Turkey. These four OECD member countries also recorded moderate average annual labour productivity growth (less than 3%), indicating that growth in average labour compensation was driving the large growth in ULC. At the other extreme for this period, Japan has a negative annual average growth rate of 1.9 percent for unit labour costs. This result is clearly being driven by strong labour productivity growth (2.9%) combined with the lowest annual average growth rate of average labour compensation (0.9%) among the OECD member countries.

The short-term (2001-06) graph shows a somewhat compressed picture with only Hungary, Slovak Republic and Turkey recording annual unit labour cost growth above 5 percent: 6.7, 7.6 and 15.7 percent respectively. Japan continues to have negative unit labour cost growth (-2.6%) and also recorded a negative annual average growth rate (-0.9%) for average labour compensation over the last five years. An interesting feature of this graph is the negative annual average labour productivity growth recorded in the last five years for four countries: Portugal, Italy, Mexico, and Slovak Republic. This situation has resulted in these countries having their ULC growing faster than their average annual labour compensation, in other words workers are being paid more for less output in Market Services for these countries.

Sources

OECD System of Unit Labour Costs and Related Indicators: www.stats.oecd.org/mei/default.asp?rev=3

OECD Annual National Accounts Database. OECD Productivity Database: www.oecd.org/statistics/productivity

For further reading OECD Main Economic Indicators March 2007; Introduction of Unit Labour Cost Indicators: www.oecd.org/dataoecd/30/38/38219407.pdf Working Party on Statistics (SWIC), 17-18 November 2005; Development of comparable quarterly unit labour cost indexes for OECD countries. Launch of the OECD System of Unit Labour Cost Indicators: www.oecd.org/dataoecd/0/19/38492896.pdf Working Group on Short-Term Economic Prospects May 2007; Introduction of new Competitiveness Indicators: www.oecd.org/dataoecd/37/12/38736937.pdf







UNIT LABOUR COSTS, LABOUR PRODUCTIVITY AND LABOUR COMPENSATION PER UNIT LABOUR INPUT– MARKET SERVICES

D.3.

2001-2006, Unit Labour Costs, Labour Productivity and labour Compensation per Unit of Labour Input Market Services, percentage change at annual rate

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

10

11

12

13

14


15.7/18.1

1986-2006, Unit Labour Costs, Labour Productivity and labour Compensation per Unit of Labour Input

Market Services, percentage change at annual rate

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

% UNIT LABOUR COSTS LABOUR COMPENSATION LABOUR PRODUCTIVITY

16.5/55.3/54.1



ANNEXES


ANNEX 1: OECD PRODUCTIVITY DATABASE

Over several years, the analysis of productivity and economic growth has been an important focus of OECD work. To carry out this work, OECD needs comparable data on productivity growth and its underlying drivers. There has also been significant interest in productivity data by researchers and analysts outside the OECD and this led to the development of the OECD Productivity Database, designed to be as comparable and consistent across countries as is currently possible. The database and related information on methods and sources were made publicly available through the OECD Internet site on 15 March 2004.

Data coverage

The OECD Productivity Database currently provides annual estimates of:

Labour productivity growth, measured as Gross Domestic Product per hour worked, and all components data, for 29 OECD countries and some economic / geographical zones and for the period 1970-2006;

Capital services and multi-factor productivity for 19 OECD countries and for the period 1985-2006 or nearest year;

Labour productivity levels for 2006, for all OECD countries as well as a few economic / geographical zones.

All of the above data relate to the economy as a whole. In addition, the Productivity Database features industry-level measures of labour productivity growth, drawn from other OECD data sets such as the STAN Database for Structural Analysis, see www.oecd.org/sti/stan and the System of Unit Labour Costs Indicators, see http://webnet4.oecd.org/wbos/default.aspx?DatasetCode=ULC_ANN.

Series of labour and multi-factor productivity growth are presented as indices and as percentage changes at annual rates (i.e. rates of change).

Year-to-year rates of change are calculated as ln (VAR t / VAR t-1), where ln is the natural logarithm, VAR is the observed measure and t is a point in time;

Average annual growth rates over a period are calculated as [(VAR year final / VAR year initial) (1 / number of years

] – 1

The OECD Productivity Database is updated on a regular basis as new data become available. It is accessible through the OECD Internet site, at: www.oecd.org/statistics/productivity.

Data sources

The OECD productivity database combines – to the extent possible – a consistent set of data on Gross Domestic Product, labour input (measured as total hours worked) and capital services. Detailed on data sources used in the productivity database are provided as follows:

http://www.oecd.org/sti/stan

http://webnet4.oecd.org/wbos/default.aspx?DatasetCode=ULC_ANN

http://www.oecd.org/statistics/productivityo


Gross Domestic Product

Estimates for GDP are derived from the OECD’s Annual National Accounts (ANA) which is based on the 1993 System of National Accounts (SNA). Member countries officially submit data for inclusion in ANA via a questionnaire; the data resulting from this questionnaire differ somewhat from national sources and are more comparable across countries than those derived from OECD’s Quarterly National Accounts Database (QNA), thanks to several small methodological adjustments that are made. However, the differences with other OECD sources, such as the Quarterly National Accounts and the Economic Outlook Database, are minor for most countries.

Labour input

The estimates of labour productivity included in the database refer to GDP per hour worked, which is the reference for measuring labour productivity. GDP per hour worked requires estimates of trends in total hours worked that are consistent across countries. In the productivity database, the default source for total hours worked is generally the OECD’s Annual National Accounts. However, for a number of countries, the national accounts do not provide hours worked and other sources have to be invoked. Consistency of data is achieved by matching the hours worked that are collected by the OECD for its annual OECD Employment Outlook with the conceptually appropriate measure of employment for each individual country, i.e. the measure of employment for that country that is consistent with the measure of hours worked collected by the OECD.

Estimates of average hours actually worked per year per person in employment are currently available on an annual basis for 29 OECD countries (See OECD Employment Outlook, Statistical Annex Table F). The OECD Productivity Database includes, in addition, hours of work per employee for Hungary and Korea. These estimates are available from National Statistical Offices for 22 countries, 16 of which are consistent with National Accounts concepts and coverage (Austria, Canada, Denmark, Finland, France, Germany, Greece, Hungary, Italy, Korea, Norway, Slovak Republic, Sweden, Switzerland, Turkey and the United States).

To develop these estimates, countries use the best available data sources for different categories of workers, industries and components of variation from usual or normal working time (e.g. public holidays, annual leave, overtime, absences from work due to illness and to maternity, etc.). For example, in two countries (Japan and United States) actual hours are derived from establishment surveys for regular or production/non-supervisory workers in employee jobs and from labour force surveys (LFS) for non-regular or managers/non-supervisory employees, self-employed, farm workers and employees in the public sector. In three other countries (France, Germany and Switzerland), the measurement of annual working time relies on a component method based on standard working hours minus hours not worked due to absences plus hours worked overtime. Standard working hours are derived from an establishment survey (hours offered), an administrative source (contractual hours) and the labour force survey (normal hours), respectively. The coverage of workers is extended using standard hours reported in labour force surveys or other sources as hours worked overtime. Vacation time is either derived from establishment-survey data on paid leave or the number of days of statutory leave entitlements. Hours lost due to sickness are estimated from the number of days not worked from social security registers and/or health surveys.

On the other hand, the national estimates for several other countries (i.e. Australia, Canada, Czech Republic, Finland, Iceland, Mexico, New Zealand, Slovak Republic, Spain, Sweden and United Kingdom) rely mainly on labour force survey results. Annual working hours are derived using a direct method annualising actual weekly hours worked, which cover all weeks of the year in the case of continuous surveys. But, for labour force surveys with fixed monthly reference weeks, this method results in averaging hours worked during 12 weeks in the year and, therefore, necessitates adjustments for special events, such as public holidays falling outside the reference week (i.e. Canada and Finland). Finally, estimates of annual working time for four other EU member states are derived by the OECD Secretariat by applying a variant of the component method to the results of the Spring European Labour Force Survey (ELFS). A summary of the various measures available in December 2007 is shown in Table 1 below.


Two other considerations should be kept in mind. First, annual working-time measures are reported either on a job or on a worker basis. To harmonise the presentation, annual hours worked measures can be converted between the two measurement units by using the share of multiple job holders in total employment, which is available in labour force surveys, albeit no further distinction is possible between second and more jobs.

1

1. For example, the BLS-Office of Productivity and Technology (OPT) estimates of annual hours of work for the United States are reported on a (per) job basis and are later converted by the OECD Secretariat to a per worker basis by multiplying the job-based annual hours of work by (1 + CPS based share of multiple jobholders in total employment).


Table 1. Sources for labour input measures included in the OECD Productivity Database

Source for annual actual working hours Source for employment

Australia Australian Bureau of Statistics Annual National Accounts Austria Annual National Accounts Annual National Accounts Belgium OECD Employment Outlook Annual National Accounts Canada Annual National Accounts Annual National Accounts Czech Republic OECD Employment Outlook Annual National Accounts Denmark Annual National Accounts Annual National Accounts Finland Annual National Accounts Annual National Accounts France Annual National Accounts Annual National Accounts Germany Annual National Accounts Annual National Accounts Greece Annual National Accounts Annual National Accounts Hungary Annual National Accounts Annual National Accounts Iceland OECD Employment Outlook Annual National Accounts Ireland OECD Employment Outlook Annual National Accounts Italy Annual National Accounts Annual National Accounts Japan OECD Employment Outlook Annual National Accounts Korea Annual National Accounts Annual National Accounts Luxembourg OECD Employment Outlook Annual National Accounts Mexico OECD Employment Outlook Annual National Accounts Netherlands OECD Employment Outlook Annual National Accounts New Zealand OECD Employment Outlook Labour Force Statistics Norway Annual National Accounts Annual National Accounts Poland OECD Employment Outlook Annual National Accounts Portugal OECD Employment Outlook Annual National Accounts Slovak Republic Annual National Accounts Annual National Accounts Spain Annual National Accounts Annual National Accounts Sweden Annual National Accounts Annual National Accounts Switzerland Annual National Accounts Annual National Accounts Turkey No data Labour Force Statistics United Kingdom OECD Employment Outlook UK Office for National Statistics United States US Bureau of Labour Statistics US Bureau of Labour Statistics

Source: OECD Productivity Database (December 2007).

Second, given the variety of data sources, of hours worked concepts retained in data sources, and of measurement methodologies (direct measures or component methods

2) to produce estimates of annual working

time, the quality and comparability of annual hours worked estimates are constantly questioned, and are subject to at least two probing issues:

Labour force survey-based estimates are suspected of over-reporting hours worked compared to work hours reported in time-use surveys, in particular for those working long hours, like managers and professionals.

Employer survey-based estimates do not account for unpaid overtime hours and are sometimes suspected of under-reporting hours worked, with consequences on productivity levels and growth.

2 . However, both methods can be summarised by the following identity: Annual hours per worker = Standard weekly hours worked x Number of weeks actually worked over the year = Weekly hours actually worked x 52 weeks, considering weekly reference period for reporting hours worked.


The comparability of measures of hours worked across OECD countries thus remains an issue, and work is currently underway, notably through the Paris Group, a UN city group on Labour and Compensation, to further improve the available measures of hours worked.

Capital input

The appropriate measure for capital input within the growth accounting framework is the flow of productive services that can be drawn from the cumulative stock of past investments in capital assets (see OECD, 2001a). These services are approximated by the rate of change of the ‘productive capital stock’ – a measure that takes account of wear and tear, retirements and other sources of reduction of the productive capacity of fixed assets. Flows of productive services of an office building, for instance, are the protection against rain or the comfort and storage services that the building provides to personnel during a given period (Schreyer et al., 2003). The price of capital services per asset is measured as their rental price. If there are markets for capital services, as is the case for office buildings, for instance, rental prices could be directly observed. For most assets, however, rental prices have to be imputed. The implicit rent that capital good owners ‘pay’ themselves gives rise to the terminology ‘user costs of capital’.

Capital input (S) is measured as the volume of capital services, assumed to be in a fixed proportion to the productive capital stock (see Schreyer, et al., 2003 for a more extensive explanation and for details of the computation of capital services). The productivity database publishes capital services data with calculations based on the perpetual inventory method (PIM) The PIM calculations are carried out by OECD, using service lives for different assets that are common across countries and correcting for differences in deflators for information and communication technology assets. Sources for the investment series by type of asset underlying the capital services series are national statistical offices

3 and the Groningen Growth and Development Centre Total Economy

Growth Accounting Database4 (http://www.ggdc.net).

Measures of multi-factor productivity in the OECD Productivity Database

The following methodology has been applied for the computation of multi-factor productivity (MFP) measures:

Rates of change of output

Output (Q) is measured as GDP at constant prices for the entire economy. Year-to-year changes are

computed as logarithmic differences: )Q

Qln(

1t

t

Rates of change of labour input

Labour input (L) is measured as total hours actually worked in the entire economy. Data on total hours has been specifically developed for the present purpose, see above. Year-to-year changes are computed as logarithmic

differences: )L

Lln(

1t

t

.

3. For Australia, Canada, France, Japan, Ireland, Italy, Germany, Spain, Sweden, Switzerland and the United States. 4. For Austria, Belgium, Denmark, Finland, Netherlands, Portugal, Sweden and the United Kingdom.

http://www.ggdc.net/


Rates of change of capital input

Capital services are computed for six different types of assets ( 7,...2,1iSi

t ) and aggregated to an overall

rate of change of capital services by means of a Törnqvist index:

7

1i i

1t

i

ti

1t

i

t21

1t

t

S

Slnvv

S

Sln with

7

1i

i

t

i

t

i

t

i

ti

t

Su

Suv

where i

tv is the share of each asset in the total value of capital services

7

1i

i

t

i

tSu . In this expression, the

value of capital services for each asset is measured by i

t

i

tSu where i

tu is the user cost price per unit of capital

services and i

tS is the quantity of capital services in year t.

Cost shares of inputs

The total cost of inputs is the sum of the remuneration for labour input and the remuneration for capital services. Remuneration for labour input has been computed as the average remuneration per employee multiplied by the total number of persons employed. This adjustment was necessary to correct for self-employed persons whose income is not part of the compensation of employees as registered in the national accounts. The source for data on compensation of employees and for the number of employees as well as the number of self employed is the OECD Annual National Accounts.

t

t

t

tt EEE

COMPLw

where

ttLw : remuneration for labour input in period t

tCOMP : compensation of employees in period t

tEE : number of employees in period t

tE : total number employed (employees plus self-employed) in period t.

Total cost of inputs is then given by:

7

1i

i

t

i

tttt SuLwC and the corresponding cost shares are

t

ttL

tC

Lws for labour input and

t

6

1i

i

t

i

tS

tC

Sus

for capital input.


Note that under perfect competition and constant returns to scale, the observed Solow residual can be viewed as an unbiased estimate of MFP growth. In this case, the shares of capital and labour in output valued at marginal costs measure the elasticity of output with respect to inputs. However, this is no longer the case under imperfect competition, when significant mark-ups generate difference between Solow residual and TFP growth (see Oliveira Martins, Scarpetta and Pilat, 1996 for a discussion and estimates). As showed in Hall (1990), a way of overcoming this problem is to calculate a Solow residual using cost rather than revenue shares. As the calculation of capital services enables to compute cost shares, the MFP estimates presented in this Compendium follow the latter approach.

Total inputs

The rate of change of total inputs is a weighted average of the rate of change of labour and capital input with the respective cost shares as weights. Aggregation is by way of a Törnqvist index number formula:

1t

tS

1t

S

t21

1t

tL

1t

L

t21

1t

t

S

Slnss

L

Llnss

X

Xln

Multi-factor productivity

Multi-factor productivity is measured as the difference between output and input change, or as “apparent multi-factor productivity”:

1t

t

1t

t

1t

t

X

Xln

Q

Qln

MFP

MFPln .


ANNEX 2: OECD ESTIMATES OF LABOUR PRODUCTIVITY LEVELS

Introduction

International comparisons of productivity growth can give useful insights in the growth process, but should ideally be complemented with international comparisons of income and productivity levels. An examination of income and productivity levels may give insights into the possible scope for further gains, and also places a country’s growth experience in the perspective of its current level of income and productivity. OECD has published estimates of labour productivity levels in various studies (e.g. Scarpetta, et al., 2000; OECD, 2003), and has released estimates of productivity on the OECD Internet site, at the following address: www.oecd.org/statistics/productivity.

Since the release of OECD estimates of productivity growth in the OECD Productivity Database in March 2004 (see Annex 1), more attention has turned to the measurement of productivity levels, since these are essential to assess the state of the convergence or divergence of economic performances across countries. Several statistical agencies and international organisations, including Eurostat, the UK Office for National Statistics (ONS), the US Bureau of Labour Statistics, and the International Labour Organisation, now release estimates of labour productivity levels, as do some academic institutions, such as the Groningen Growth and Development Centre, and some private institutions, such as the Conference Board. In several instances, notably in the case of Eurostat and the ONS, estimates of labour productivity levels serve as official yardsticks of economic performance and are used to measure progress with regards to explicit policy targets.

Given the importance attached to labour productivity levels, it is unfortunate that there is still considerable variation in the currently available estimates. Primarily, this seems due to differences in the choice of basic data. Indeed, much of the differences can be brought back to how different organisations select and combine information on the three components of labour productivity levels at the economy-wide level. These components are gross domestic product, labour input and a conversion factor for total GDP, typically a purchasing power parity (or PPP) that is needed to translate output in national currency units to a common currency.

This annex briefly discusses some of the main measurement issues for these components, as well as the different data choices that can be made. It focuses on the current OECD approach to measuring labour productivity levels, but also refers to other possible approaches, where appropriate. The discussion focuses on comparisons of labour productivity at the economy-wide level; the estimation of productivity levels for individual industries raises additional measurement issues that go beyond the scope of this annex.

Output: comparability and data choices

Most comparisons of labour productivity levels focus on GDP as the measure of output. Other measures of aggregate output, such as GNP or national income, have also been used in a few studies, but are not considered here. The measurement and definition of economic output is treated systematically across countries in the 1993 System of National Accounts (SNA). All countries in the OECD area have now implemented the 1993 SNA, except Turkey, which implies that its level of GDP is likely to be somewhat understated relative to other OECD countries. However, despite the harmonisation of GDP estimates through the 1993 SNA, there are some differences in estimation methods across countries (Ahmad, et al., 2003). These typically have only a small effect on growth



rates, but may be substantially more important for comparisons of output and productivity levels. Some of the main differences that are known to affect GDP levels are the following (Ahmad, et al., 2003):

Expenditure on military equipment. The coverage of government investment in the US National Income

and Product Accounts (NIPA) is more extensive than that recommended by the SNA, since it includes expenditures on military equipment (aircraft, ships, missiles) that are not considered assets by the SNA. The national accounts in most other OECD countries strictly follow the SNA in this matter. As the amount of public investment affects GDP, this results in a statistical difference in the measurement of GDP. Convergence on this issue is expected in the 2008 edition of the SNA. In the meantime, the OECD publishes data in its Annual National Accounts Database for the United States which adjust for this difference.

Financial Intermediation Services. Most banking services are not explicitly charged. Thus, in the SNA, the

implicit production of banks is estimated using the difference between interests received and paid. All OECD member countries have estimated this part of bank production, known as “Financial Intermediation Service Indirectly Measured” or “FISIM”. While it is relatively straightforward to recognise and estimate FISIM, the key problem is breaking it down between final consumers (households) and intermediate consumers (business and government). Only the first part has an overall impact on GDP. In the United States, Canada and Australia, such a breakdown has been estimated in the national accounts for some time, in accordance with the SNA. A breakdown between final and intermediate consumers has been implemented in most European countries, although the number of years of historical data that has also been revised varies; nevertheless, there are still a few countries for which the allocation of FISIM has not yet been implemented.

Software investment. Another significant issue in the comparability of GDP concerns the measurement of

software. The 1993 SNA recommended that software expenditures be treated as investment as long as the acquisition satisfied conventional asset requirements. This change added nearly 2% to GDP for the United States, around 0.7% for Italy and France, and about 0.5% for the United Kingdom. Doubts on the comparability of these data were raised when comparing “investment ratios”, which are defined as the share of software expenditures that are recorded as investment to total expenditures in software. These ratios range from under 4% in the United Kingdom to over 70% in Spain (Lequiller, et al., 2003; Ahmad, 2003). A priori, one would expect that these are roughly the same across OECD countries. An OECD-Eurostat Task Force confirmed that differences in estimation procedures contributed significantly to the differences in software capitalisation rates, and a set of recommendations describing a harmonised method for estimating software were formulated (Lequiller, et al, 2003; Ahmad, 2003). Many of these recommendations have begun to be implemented by some countries but differences in software measurement will nevertheless continue to have an impact on the international comparability of GDP levels for some time to come.

The informal economy. Another factor that may influence the comparability of GDP across countries is size

of the non-observed economy. In principle, GDP estimates in the national accounts take account of this part of the economy. In practice, questions can be raised about the extent to which official estimates have full coverage of economic activities that are included in GDP according to the SNA, or to which extent there some under-reporting is involved. Large differences in coverage could substantially affect comparisons of productivity levels.


It is not clear, a priori, how large the impact of these, and possible other, differences is on GDP levels. What is clear, however, is that there is a margin of uncertainty associated with the comparability of levels of GDP across countries. Consequently, there is also a range of uncertainty associated with estimates of productivity levels; small differences between countries (of a few percentage points) will obviously fall within this range of uncertainty. This is important in interpreting estimates of productivity levels; countries within a small range of income and productivity levels may not have income and productivity differences that are statistically or economically significant (Schreyer and Koechlin, 2002).

The data choices for GDP are fairly uniform across different sources. In the OECD estimates of productivity levels, data on GDP are derived from OECD's Annual National Accounts (ANA). The data from ANA are based on the OECD’s annual national accounts questionnaire to OECD member countries. The data resulting from this questionnaire may differ somewhat from national sources and are more comparable across countries than those derived from OECD’s Quarterly National Accounts (or the OECD Economic Outlook database), thanks to some small methodological adjustments that are made. For example, the US GDP estimates are adjusted for expenditure on military equipment, as discussed above. However, the differences with other OECD sources, such as the Quarterly National Accounts and the Economic Outlook database, are minor for most countries.

For two countries, Australia and New Zealand, the OECD’s Annual National Accounts provides GDP estimates for fiscal years. This creates an inconsistency with other countries, since comparisons of productivity levels ideally should correspond to the same (calendar) year. However, this problem is considered relatively limited and no adjustment is made.

Labour input

Comparable measures of labour input are of great importance for international comparisons of productivity levels. The OECD estimates of labour productivity levels are typically based on the same data choices as the OECD estimates of labour productivity growth.

5 Annex 1 therefore provides a more detailed discussion of the measures

of labour input used in the calculations of productivity levels.

Purchasing power parities for international comparisons

The comparison of income and productivity across countries also requires Purchasing Power Parity (PPP) data for GDP. Exchange rates are not suitable for the conversion of GDP to a common currency, since they do not reflect international price differences, and are heavily influenced by short-term fluctuations. The estimates used by the OECD are derived from its joint programme with Eurostat and refer to current-price PPPs (Schreyer and Koechlin, 2002). For the current set of comparisons, the most recent PPP benchmark comparison is used as the basis for the estimates.

The OECD does not recommend the use of PPP-adjusted estimates of GDP in time series, because of the difficulty to obtain PPPs that are consistent over time. This is why only one year of productivity level comparisons is included in the OECD Productivity Database. Users interested in adding a time dimension to this one year level comparison should use the corresponding database on productivity growth, which gives appropriate indices of productivity growth for individual OECD countries over a long time period.

OECD estimates of labour productivity levels for 2006, as of December 2007

Clearly, data for international comparisons of income and productivity are not perfect and some choices between different sources have to be made. In the OECD approach, GDP is derived from the OECD ANA database,

5. A number of OECD countries include estimates of hours worked in their national accounts. To the extent possible, these estimates are incorporated in the OECD estimates of productivity levels, as presented in Annex Table 2. They are fully incorporated in the OECD estimates of productivity growth and in the data choices for those calculations, as shown in Annex Table 1.


which incorporates the latest comparative information on GDP from OECD member countries. Data on employment for most countries are also from the OECD national accounts as these should have a better correspondence to the estimates of GDP. For a limited number of countries, no appropriate employment estimates are currently available from the national accounts, in which case employment is derived from the OECD Labour Force Statistics. Estimates of hours worked are either from the national accounts, or from the OECD Employment Outlook, as shown in Annex 1. To convert GDP to a common currency, the OECD uses current PPPs, which are developed in the OECD-Eurostat PPP programme.

Annex Table 2 presents the data choices and the resulting productivity level estimates for 2006. These estimates still require further work in the following ways:

1. For several OECD countries, the estimates of annual hours worked per person are not yet consistent with the national accounts. Currently, the OECD collects series on hours worked through two data collections, the OECD Employment Outlook and the OECD Annual National Accounts. On one hand, all OECD countries but one provide data on hours worked for the annual publication OECD Employment Outlook and few of these countries supply estimates of annual hours worked that are consistent with the national accounts concepts and coverage. On the other hand, a large number of OECD countries also provide estimates of total hours worked in the framework of the national accounts for inclusion in the OECD’s Annual National Accounts. Further investigation of these estimates of hours worked is needed.

2. The employment estimates that are currently incorporated in the national accounts are not necessarily consistent across countries or with the corresponding estimate of GDP. Addressing this problem will require further statistical work.

For analytical purposes, it is important that estimates of GDP per hour worked are combined with estimates of GDP per capita and estimates of GDP per person in the labour force and GDP per person of working age. The national accounts currently often do not include the necessary information on working-age population and labour force, and such data have commonly been derived from labour force statistics. The OECD’s change in method towards the national accounts as the main source of employment information requires that the link between labour force statistics (i.e. national concepts) and national accounts estimates of productivity (i.e. domestic concepts) is addressed.


Table 2. OECD estimates of labour productivity levels for 2006 (as of December 2007)

GDP, million

national

currency units,

based on ANA

PPP for

total GDP,

2006

GDP, million

USD

Employment

(1000 persons) 1

OECD source for

employment 2

Annual average

hours worked,

corresponding

to employment

estimates 3

OECD source

for average

hours worked 2

Total hours

worked (million

hours)

GDP per

hour

worked,

USD

GDP per

hour

worked,

USA=100

(1) (2) (3) (4) (5) (6) (7) (8)

Australia 4 1,038,652 1.41 735,330 10,226 ANA / LFS 1,728 ABS 17,675 41.6 83Austria 257,897 0.87 295,624 4,228 ANA 1,659 ANA 7,014 42.1 84Belgium 316,622 0.90 353,479 4,278 ANA 1,571 EMO 6,721 52.6 104Canada 1,442,463 1.20 1,197,776 16,758 ANA / LFS 1,736 ANA / EMO 29,094 41.2 82Czech Republic 3,231,576 14.30 225,958 5,082 ANA 1,997 EMO 10,148 22.3 44Denmark 1,642,215 8.58 191,474 2,822 ANA 1,584 ANA 4,471 42.8 85Finland 167,062 0.97 172,399 2,441 ANA 1,716 ANA 4,189 41.2 82France 1,791,953 0.91 1,962,072 25,278 ANA 1,555 ANA / EMO 39,297 49.9 99Germany 2,322,200 0.88 2,631,598 39,088 ANA 1,433 ANA 56,001 47.0 93Greece 213,985 0.70 303,605 4,700 ANA / LFS 2,052 ANA / EMO 9,645 31.5 62Hungary 23,757,230 129.94 182,834 3,905 ANA 1,989 ANA 7,768 23.5 47Iceland 1,141,747 104.94 10,880 170 ANA 1,794 EMO / EO 304 35.8 71Ireland 174,343 1.01 172,820 2,042 ANA 1,640 EMO 3,350 51.6 102Italy 1,475,401 0.87 1,699,152 24,754 ANA 1,800 ANA 44,568 38.1 76Japan 507,754,500 124.46 4,079,511 64,179 ANA / LFS 1,784 EMO 114,520 35.6 71Korea 847,876,400 762.02 1,112,668 23,131 ANA 2,357 ANA 54,522 20.4 41Luxembourg 33,852 0.92 36,936 319 ANA 1,604 EMO 512 72.2 143Mexico 9,149,911 7.22 1,267,894 42,198 ANA 1,883 EMO 79,467 16.0 32Netherlands 534,324 0.89 597,232 8,383 ANA 1,391 EMO 11,661 51.2 102

New Zealand 4 163,416 1.52 107,318 2,126 LFS 1,787 EMO 3,798 28.3 56Norway 2,147,986 8.89 241,714 2,419 ANA 1,407 ANA 3,403 71.0 141Poland 1,057,855 1.89 558,298 14,594 ANA 1,985 EMO 28,969 19.3 38Portugal 155,131 0.70 220,514 5,154 ANA / LFS 1,758 EMO 9,060 24.3 48Slovak Republic 1,636,263 17.26 94,797 2,132 ANA 1,749 ANA 3,729 25.4 50Spain 980,954 0.76 1,294,828 19,848 ANA 1,656 ANA / EO 32,869 39.4 78Sweden 2,899,653 9.16 316,657 4,423 ANA 1,601 ANA 7,083 44.7 89Switzerland 486,178 1.70 285,280 4,291 ANA 1,651 ANA / EO 7,084 40.3 80

Turkey 5 576,322 0.90 639,693 22,830 LFS 1,918 GGDC 43,788 14.6 29United Kingdom 1,301,914 0.65 1,996,983 28,960 ONS 1,669 EMO 48,326 41.3 82United States 13,132,900 1.00 13,132,900 152,621 BLS 1,708 BLS 260,631 50.4 100OECD 36,118,223 543,380 1748 949,664 38.0 75G7 26,699,992 351,638 1685 592,437 45.1 89North America 15,598,570 211,577 1745 369,191 42.3 84

OECD-Europe 6 13,845,133 209,311 1654 346,169 40.0 79EU15 12,245,374 176,719 1611 284,765 43.0 85

EU19 7 13,307,260 202,431 1657 335,378 39.7 79

Euro-zone 8 9,740,259 140,514 1600 224,885 43.3 86

6. Does not include Turkey.

7. All EU members that are also OECD member countries.

8. Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain.

Source : OECD Productivity Database, December 2007.

1. The employment estimates for Austria, Canada, Greece, Japan, United Kingdom and United States refer to jobs.

2. ABS = Australian Bureau of Statistics; ANA = OECD Annual National Accounts; LFS = OECD Labour Force Statistics; EMO = OECD Employment Outlook; ONS = UK Office for National

Statistics; GGDC = Groningen Growth and Development Centre; BLS = US Bureau of Labor Statistics.

3. The estimates of annual hours worked for Austria, Canada, Greece, Japan, United Kingdom and United States refer to hours worked per job. Data for France, Greece, Italy and

Switzerland are estimates.

4. GDP estimates refer to fiscal years.

5. GDP for Turkey is based on the 1968 System of National Accounts.


ANNEX 3: OECD DATABASES RELEVANT TO PRODUCTIVITY ANALYSIS

Annual National Accounts Database (ANA): The OECD’s Annual National Accounts presents a consistent

set of data mainly compiled on the basis of the 1993 System of National Accounts (SNA). It contains data from 1970 whenever possible for OECD member countries. The database contains a wide selection of accounts: main aggregates (GDP by expenditure, GDP by kind of activity, GDP by income and disposable income, saving and net lending), detailed breakdown by kind of activity for gross value added (at current and constant prices), components of value added, gross fixed capital formation, employment and hours worked. It also includes final consumption expenditure of households by purpose and simplified accounts for general government. Detailed accounts by institutional sector are also available. The database also provides comparative tables based on exchange rates and purchasing power parities.

Publications: OECD, National Accounts of OECD Countries: Volume I: Main Aggregates, 1995-2006, 2008 Edition; OECD, National Accounts of OECD Countries: Volume II: Detailed Tables 1994-2005, 2007 Edition. Also available on line on SourceOECD (www.sourceoecd.org).

Labour Force Statistics (LFS): The OECD’s Labour Force Statistics provides detailed statistics on population,

labour force, employment and unemployment, broken down by gender, as well as unemployment duration, employment status, employment by sector of activity and part-time employment. It also contains participation and unemployment rates by gender and detailed age groups, as well as comparative tables for the main components of the labour force. Data are available for each OECD member country and for OECD-Total, Euro zone and EU15. The time series cover 20 years for most countries. It also provides information on the sources and definitions used by Member countries in the compilation of statistics.

Publication: OECD, Labour Force Statistics: 1985-2005, 2006 Edition. Available on SourceOECD (www.sourceoecd.org).

Economic Outlook (EO): The OECD Economic Outlook Database is a comprehensive and consistent

macroeconomic database of the OECD economies, covering expenditures, foreign trade, output, employment and unemployment, interest rates and exchange rates, the balance of payments, outlays and revenues of government and of households, and government debt. For the non-OECD regions, foreign trade and current account series are available. The database contains yearly and quarterly information, both for the historical period and for the projection period. The historical data is based on data from national statistical agencies and OECD sources such as the Quarterly National Accounts, the Annual National Accounts, the Quarterly Labour Force Statistics, the Annual Labour Force Statistics and the Main Economic Indicators.

Publication: OECD, Economic Outlook, twice yearly. The database is available on CD-ROM: http://www.oecd.org/dataoecd/47/40/32108765.pdf A statistical annex to the OECD Economic Outlook is available at: http://www.oecd.org/document/61/0,2340,en_2649_34109_2483901_1_1_1_1,00.html

Productivity Database: The OECD Productivity Database presents productivity at the aggregate level and is

designed to be as comparable and consistent across countries as is currently possible (see Annex 1). The database and related information on methods and sources were first made publicly available through the OECD Internet site on 15 March 2004. The OECD Productivity Database currently provides annual estimates of:

http://www.sourceoecd.org/



http://www.oecd.org/document/61/0,2340,en_2649_34109_2483901_1_1_1_1,00.html


Labour productivity growth, measured as GDP per hour worked, for 29 OECD countries and the period 1970-2006;

Capital services and multi-factor productivity for 19 OECD countries for the period 1985-2006 or nearest year;

Labour productivity levels for 2006.

The OECD Productivity Database is updated on a regular basis as new data become available. It is accessible through the OECD Internet site, at: www.oecd.org/statistics/productivity.

Structural Analysis Database (STAN): The STAN database for Industrial Analysis includes annual measures

of output, labour input, investment and international trade by economic activity that allow users to construct a wide range of indicators focused on areas such as productivity growth, competitiveness and general structural change. The industry list based on the International Standard Industrial Classification (ISIC) Rev. 3 provides sufficient detail to enable users to highlight high-technology sectors and is compatible with those lists used in related OECD databases in the ‘STAN’ family (see below). STAN-Industry is primarily based on member countries’ annual National Accounts by activity tables and uses data from other sources, such as national industrial surveys/censuses, to estimate any missing detail. Since many of the data points in STAN are estimated, they do not represent the official member country submissions. See: www.oecd.org/sti/stan

Publication: STAN-industry is available on line via SourceOECD (www.sourceoecd.org) with the latest tables showing data up to 2003. STAN-industry is also available on CDROM together with the latest versions of STAN–R&D (ANBERD), STAN–Bilateral Trade Database (BTD) and a set of derived STAN Indicators. See www.oecd.org/sti/stan/indicators.

System of Unit Labour Costs and Related Indicators (ULCRI): The ULCRI, which is updated at the end of

each quarter, consist of long time series of annual and quarterly Unit Labour Cost and related indicators compiled using a specific methodology to maximise comparability across countries. The related indicators include annual time series for: Exchange Rate Adjusted Unit Labour Costs; Labour Income Share ratios; Labour Productivity levels and indices and; Labour Compensation per unit labour input levels and indices. Data are available for all OECD member countries and the Euro area for a wide range of sectors including Total Economy, Manufacturing & Industry, Market Services and the Business Sector.

Publication: All time series together with detailed methodological information and country data sources are freely available through the OECD System of Unit Labour Cost and Related Indicators web portal at: http://stats.oecd.org/mei/default.asp?rev=3. A quarterly news release containing the latest quarterly updates together with a discussion of main features for the economic activities of Industry and Market Services is released around the middle of January, April, July and October. Quarterly time series for the unit labour cost indexes for all economic activities are also published in the monthly Main Economic Indicators publication. Elements of the OECD System of Unit Labour Cost and Related indicators also appear in a number of other OECD publications, such as the OECD Factbook and the OECD Economic Outlook.

Structural and Demographic Business Statistics Database (SDBS): The OECD’s Structural and

Demographic Business Statistics Database provides annual information on the industrial structure of OECD economies. It contains data, from 1995 whenever possible, for OECD member countries on turnover, value added, production, operating surplus, number of employees, wages and salaries, investment, number of enterprises, and other variables. This information is available at a very detailed sectoral level (International Standard of Industrial Classifications, Revision 3, 4-digit level) and is broken down by employment size classes.

The database also provides business demography and entrepreneurship Statistics such as business birth, death and survival rates which, in combination with the size class dimension, provides a tool to inform policy areas such as entrepreneurship. The OECD is currently expanding the collection of data in this field.

http://www.oecd.org/statistics/productivityo

http://www.oecd.org/sti/stan


http://www.oecd.org/sti/stan/indicators



Publication: OECD, Structural and Demographic Business Statistics, 1996-2003, 2006 Edition. Detailed information on SDBS is also available at: http://www.oecd.org/document/17/0,3343,en_2649_34233_36938705_1_1_1_1,00.html.

Other databases of interest are Activities of Foreign Affiliates (AFA) and Activities of Foreign Affiliates in the Services sector (FATS):

The former presents detailed data on the performance of foreign affiliates in the manufacturing industry of OECD countries (inward and outward investment). The data indicate the increasing importance of foreign affiliates in the economies of host countries, particularly in production, employment, value added, research and development, exports, wages and salaries. In AFA, the measures are broken down by country of origin and by industrial sector (based on ISIC Rev. 3).

The latter presents the activities of foreign affiliates in the services sector of OECD countries (inward and outward investment). The data indicate the increasing importance of foreign affiliates in the economies of host countries and of affiliates of national firms implanted abroad. In FATS, there are five measures (production, employment, value added, imports and exports) broken down by country of origin (inward investments) or implantation (outward investments) and by industrial sector (based on ISIC Rev. 3).

Publication: OECD, Measuring Globalisation: Economic Globalisation Indicators. Available on SourceOECD (www.sourceoecd.org).

http://www.oecd.org/document/17/0,3343,en_2649_34233_36938705_1_1_1_1,00.html



Table 3. Current country coverage of main databases used in this publication

General databases Productivity Database Industry-level

ANA LFS EO PROD CAP PRODL STAN ULCRI SDBS

Australia X X X X X X X X X Austria X X X X X X X X X Belgium X X X X X X X X X Canada X X X X X X X X X Czech Republic X X X X X X X X Denmark X X X X X X X X X Finland X X X X X X X X X France X X X X X X X X X Germany X X X X X X X X X Greece X X X X X X X X X Hungary X X X X X X X X Iceland X X X X X X X X Ireland X X X X X X X X X Italy X X X X X X X X X Japan X X X X X X X X X Korea X X X X X X X X Luxembourg X X X X X X X X Mexico X X X X X X X X Netherlands X X X X X X X X X New Zealand X X X X X X X X X Norway X X X X X X X X Poland X X X X X X X X Portugal X X X X X X X X X Slovak Republic X X X X X X X X Spain X X X X X X X X X Sweden X X X X X X X X X Switzerland X X X X X X X X Turkey X X X X X X X United Kingdom X X X X X X X X X United States X X X X X X X X X OECD Non-member countries

X X


ANNEX 4: OECD SYSTEM OF UNIT LABOUR COST AND RELATED INDICATORS

The following metadata tables outline those countries that, for a specific economic activity, could not provide the full data series required for the graphs published in sections D1, D2 and D3. The dates required for the economy activities – Total Economy, Industry Market Services – were 1986-2006.

All data were sourced from the OECD’s System of Unit Labour Cost and Related Indicators database, with the exception of Labour Productivity – Total Economy which was sourced via the OECD Productivity Database. If a country is not listed, or the cell is blank, then this country has all the required data for the listed economic activity.

Table 1. Annual Unit Labour Costs

Country Total Economy Industry Market Services

Czech Republic 1992-2006 1995-2006 1995-2006

Greece 1995-2006 1995-2006

Hungary 1992-2006 1995-2006 1995-2006

Iceland 1986-2005 1986-2005 1986-2005

Mexico 1986-2004 1995-2004 1995-2004

Poland 1992-2006 1992-2006 1992-2006

Portugal 1986-2004

Slovak Republic 1993-2006 1995-2006 1995-2006

Switzerland 1990-2006 NA NA

Table 2. Annual Labour Productivity

Country Total Economy Industry Market Services

Austria 1995-2006

Canada 1986-2005 1986-2005

Czech Republic 1993-2006 1995-2006 1995-2006

France 1986-2005 1986-2005

Greece 1995-2006 1995-2006

Hungary 1992-2006 1995-2006 1995-2006

Iceland NA NA

Mexico 1991-2006 1995-2004 1995-2004

New Zealand 1989-2005 1989-2005

Poland 1992-2006 1992-2006 1992-2006

Portugal 1986-2004 1986-2004

Slovak Republic 1994-2006 1995-2006 1995-2006

Switzerland NA NA

United States 1986-2005 1986-2005


Table 3. Annual Labour Compensation per Unit Labour Input

Country Industry Market Services

Canada 1986-2003 1986-2003

Czech Republic 1995-2006 1995-2006

France 1986-2005 1986-2005

Greece 1995-2006 1995-2006

Hungary 1995-2005 1995-2005

Iceland NA NA

Ireland (1995) 1986-2006 (1995) 1986-2006

Mexico 1995-2004 1995-2004

New Zealand 1989-2003 1989-2003

Poland 1992-2005 1992-2005

Portugal (1995) 1986-2004 (1995) 1986-2004

Slovak Republic 1995-2006 1995-2006

Sweden (1993) 1986-2006 (1993) 1986-2006

Switzerland NA NA

Turkey 1988-2006 1988-2006

Notes:

1. Labour Productivity - Total Economy (section D1 of the publication); data for the Euro area and Poland are sourced via the OECD System of Unit Labour Costs and Related Indicators database not the OECD Productivity Database.

2. Table 3 has three countries (Ireland, Portugal and Sweden) with dates in brackets. In these cases the dates in brackets are the actual dates that data provided by the countries allows these series to be calculated. However, using the identity that: “Labour Compensation per Unit of Labour Input is equivalent to Labour Productivity per Unit Labour Input multiplied by Unit Labour Costs”, the Labour Compensation per Unit Labour Input series has been extended to the new date indicated.

Symbols:

NA – In these cases, the countries lack either ‘Total Labour Costs’ or ‘Employment’ variables. The derived variable cannot be compiled, or if compiled the derived variable would not be comparable with other OECD member countries.

Growth Rates:

For section D, annual average growths rates are calculated as follows, using Labour Productivity (Real Output/Hours) as the example:

Annual average growth rates for Real Output per hour worked are computed as compound growth rates,

i.e. annual average growth rate per period = 1/Re

/Re

k

t

kt

HoursOutputal

HoursOutputal

where t and t+k are two points in time and k is the number of years between them.


Variable descriptions

Below are more detailed descriptions of the compilation methodology for the annual variables available through the OECD System of Unit Labour Costs and Related Indicators online database, accessible through the user interface at: http://stats.oecd.org/mei/default.asp?rev=3

Annual Unit Labour Costs

Key statistical concept

Annual unit labour costs are calculated as the quotient of total labour costs and real output. Time series are presented in both level and index form where the base year of real output is 2000. Unit labour costs (ULC) measure the average cost of labour per unit of output. They are calculated as the ratio of total labour costs to real output. The OECD System of Unit Labour Cost and Related Indicators database produces annual and quarterly ULC and related indicators according to a specific methodology to ensure data are comparable across OECD countries. For detail on country data sources, see: http://stats.oecd.org/mei/default.asp?lang=e&subject=19

Recommended uses and limitations

Every effort has been made to ensure that data are comparable across countries. Therefore cross country comparisons of unit labour cost levels (in ratio form) can be used for static analysis (i.e. comparison of unit labour cost levels across countries or economic activities at a point in time) together with indexes which show comparable development in unit labour costs over time. However, for some countries unit labour cost levels are not presented due to a lack of data to make an adjustment for the self-employed. For these countries only unit labour cost indexes are made available for analysis. Furthermore, the adjustment for the self employed assumes that labour compensation per hour or per person is equivalent for the self employed and employees of businesses. This assumption may be more or less valid across different countries and economic activities thus affecting the comparability of unit labour cost level data.

Sector: Sector coverage

The economic activities listed are derived from the International Standard Industrial Classification (ISIC Rev. 3): http://www.ilo.org/public/english/bureau/stat/class/isic.htm

Annual Total Labour Costs


The target variable for annual total labour costs is compensation of employees (COE) compiled according to the System of National Accounts 1993, adjusted for the self employed by multiplying COE by the ratio of total hours worked by all persons in employment to total hours worked by all employees of businesses. This target variable covers a significant part of total labour costs such as wages and salaries; bonuses; payments in kind related to labour services (e.g. food, fuel, housing, etc); severance and termination pay and; employers' contributions to pension schemes, casualty and life insurance and workers compensation.

However, COE excludes some relevant items of total labour cost such as the cost of employee training, welfare amenities and recruitment; taxes on employment (e.g. payroll tax) and; fringe benefits tax. Furthermore, the adjustment for the self employed assumes that labour compensation per hour or per person is equivalent for the self employed and employees of businesses. This assumption may be more or less valid across different countries and economic activities.

In the interest of producing the longest possible time series for annual total labour costs, current series are often linked to related historical time series provided to the OECD sometime in the past. As a result time series extend back to 1970 for most OECD member countries. The variables and their sources used for each country and economic activity are noted in the country metadata under the heading ‘Sources'.


http://stats.oecd.org/mei/default.asp?lang=e&subject=19

http://www.ilo.org/public/english/bureau/stat/class/isic.htm


Aggregation and consolidation

Total labour costs for economic activities G_K (Market Services) and C_K (Business Sector) are compiled by summing their respective activity components.

Annual Real Output


The target variable for annual real output is constant price value added compiled according to the System of National Accounts 1993. In the interest of producing the longest possible time series for annual real output, current series are often linked to related historical time series provided to the OECD sometime in the past. As a result time series extend back to 1970 for most OECD member countries. The variables and their sources used for each country and economic activity are noted in the country metadata under the heading 'Sources'.


All volume series of real output are re-referenced such that the national currency series are expressed in prices of the prevailing OECD base year. Series for activity aggregates G_K and C_K are compiled through annual chain linking of their respective components, using current price value added data as weights.

Other manipulations

The real output of activity J_K is adjusted to remove the (estimated) component attributed to the services provided by a dwelling to its occupants as this activity has no associated labour input. For a detailed explanation of this issue and the methodology used to perform the adjustment, see http://www.oecd.org/dataoecd/37/31/37664867.pdf. Consequently the published time series of real output for activities, J_K, G_K and C_K will differ from related national source data.

Annual Labour Productivity


Labour Productivity is defined in the OECD System of Unit Labour Cost and Related Indicators database as real output divided by total labour input. The labour input measure used is hours worked by those in employment for Australia, Austria, Canada, Denmark, France, Germany, Greece, Hungary, Italy, Korea, Netherlands, Norway, Slovak Republic, Spain, Sweden and Switzerland. For all other countries total employment in persons is used as the labour input measure.


The main purpose of the Labour Productivity measure compiled through the OECD System of Unit Labour Cost and Related Indicators is to enable users to decompose movements in the annual Unit Labour Cost into a numerator which shows Labour Compensation per Unit Labour Input and a denominator which shows Labour Productivity.

Estimates of Labour Productivity are very sensitive to the quality of data used for the labour input measure. This issue is explained in depth in the OECD Productivity Database which also presents measures of Labour Productivity at the Total Economy level which may differ from those shown in the OECD System of Unit Labour Cost and Related Indicators database for some countries. The main source of this difference is the labour input measure used.

The OECD Productivity Database uses total hours worked as the labour input measure for all countries where this is defined as the product of series for average hours per worker or per job multiplied by total number of workers or the total number of jobs. National accounts is the default source for this data, complemented by data


http://www.oecd.org/topicstatsportal/0,2647,en_2825_30453906_1_1_1_1_1,00.html#30453948


from labour force surveys for those countries and years for which national accounts provide no information on hours worked. By contrast, all labour input data (i.e. total hours worked or total employment) used for the OECD System of Unit Labour Cost and Related Indicators database is sourced from the OECD System of Annual National Accounts database. This implies that where a country has only a short-time series of hours worked data available, the historical series will have been linked to the series on total employment. The period for which hours worked data is available in those countries where it is used as the labour input measure is shown in the country data sources.

There are other minor reasons which may also lead to discrepancies between the Labour Productivity measures presented at the Total Economy level between the two respective databases. A detailed description of these reasons and an analysis of discrepancies on a country by country basis are available on request.

Annual Labour Productivity per Person Employed


Labour productivity per person employed is defined in the OECD System of Unit Labour Cost and Related Indicators database as real output (gross value added) divided by total employed persons.

Annual Labour Productivity per Hour


Labour productivity per hour is defined in the OECD System of Unit Labour Cost and Related Indicators database as real output (gross value added) divided by total hours worked by all persons in employment.

Annual Labour Compensation per unit labour input


Labour Compensation per Unit Labour Input is defined in the OECD System of Unit Labour Cost and Related Indicators database as compensation of employees divided by total hours worked by employees (Australia, Austria, Canada, Denmark, France, Germany, Greece, Hungary, Italy, Korea, Norway, Slovak Republic, Spain and Sweden) or total employees (all other countries).

Annual Labour Compensation per Employee


Labour compensation per employee is defined in the OECD System of Unit Labour Cost and Related Indicators database as compensation of employees divided by total employees.

Annual Labour Compensation per Hour


Labour compensation per hour is defined in the OECD System of Unit Labour Cost and Related Indicators database as compensation of employees divided by total hours worked by employees.

Annual Labour Compensation per Unit Labour Input USD (PPP adjusted)


USD (PPP adjusted) Labour compensation per unit labour input is defined in the OECD System of Unit Labour Cost and Related Indicators database as compensation of employees converted from national currency to USD using





private consumption purchasing power parities, divided by total hours worked by employees (for Australia, Austria, Canada, Denmark, France, Germany, Greece, Hungary, Italy, Korea, Netherlands, Norway, Slovak Republic, Spain and Sweden) and total employees for all other countries. Note, for those countries with hours worked data, where a longer time series for total employees exists the data is linked to extend the time series.

Annual Labour Compensation per Employee USD (PPP adjusted)


USD (PPP adjusted) Labour compensation per employee is defined in the OECD System of Unit Labour Cost and Related Indicators database as compensation of employees converted from national currency to USD using private consumption purchasing power parities, divided by total employees.

Annual Labour Compensation per Hour USD (PPP adjusted)


USD (PPP adjusted) Labour compensation per hour is defined in the OECD System of Unit Labour Cost and Related Indicators database as compensation of employees converted from national currency to USD using private consumption purchasing power parities, divided by total hours worked by employees.

Annual self employment ratio


The self-employment ratio (either using 'hours worked' or 'persons') is calculated simply as: Total employment divided by employees. The resulting ratio gives the user an understanding of the proportion of the self-employed to employees in total employment. Looked across economic activities over time, the ratio can also give the user an understanding of the changing self-employed/employee composition of the country's labour force.

The ratio is based on hours worked data for Australia, Austria, Canada, Denmark, France, Germany, Greece, Hungary, Italy, Korea, Netherlands, Norway, Slovak Republic, Spain and Sweden. For all other countries data on persons is used, with the exception of Switzerland, Turkey and Iceland where data is not available to perform the calculation.


In compiling ULC's the ratio is multiplied by compensation of employees (national accounts base) to give an adjusted compensation of employees measure more suitable for use in the Unit Labour Cost (ULC) compilation process. Compensation of employees as defined in the national accounts does not include labour compensation for the self-employed which is covered in the item 'operating surplus and mixed income'. However, the output of the self-employed contributes to value added and thus introduces an inconsistency between the numerator and denominator when deriving ULC indexes. If an adjustment is not made to the labour compensation measure to account for the impact of the self-employed, this has the potential to distort the comparability of ULC indexes across countries if there are large differences in the level or, more importantly, changes over time in the number of self-employed persons across countries. Also, this impact is likely to vary across industries, as some industries are more likely to have a higher proportion of self employed (e.g. Agriculture) than others.


Annual Nominal Output


Annual nominal output is current price value added compiled according to the SNA 93.


Annual Nominal Output series are used as weights when compiling annually chain linked Real Output series for economic activities G_K (Market Services) and C_K (Business Sector). The Annual Nominal Output series for each economic activity are also used as the denominator for the Labour Income Share ratio - also known as the Real Unit Labour Cost.

Other manipulations

The real output of activity J_K is adjusted to remove the (estimated) component attributed to the services provided by a dwelling to its occupants as this activity has no associated labour input. For a detailed explanation of this issue and the methodology used to perform the adjustment, see: www.oecd.org/dataoecd/37/31/37664867.pdf. Consequently the published time series of real output for activities, J_K, G_K and C_K will differ from related national source data.



ANNEX 5: MULTI-FACTOR PRODUCTIVITY MEASURES IN OECD COUNTRIES

The multi-factor productivity estimates have been compiled by the OECD for the purpose to provide estimates as comparable and consistent across countries as possible. However, they do not necessarily constitute the best source for analysis that relate to the country only. In the recent years, a growing number of national statistical offices (NSO) experienced to develop and publish MFP calculations. This annex briefly presents official multi-factor productivity time-series

6 published and regularly updated by Australia, Canada, the Netherlands, New Zealand,

Switzerland and the United States and discusses some of the main differences from those computed by the OECD. First, the national data are often significantly more detailed and also timelier than the international data. Second, labour input measures may have been adjusted by the NSOs (Australia, Canada, United States, etc.) to reflect the composition of the labour force e.g. by age, education and experience whereas the OECD labour input data is a simple aggregate of hours worked. However, adjustment methods for labour composition may differ across countries. These and some other technical differences may lead to differences in reported MFP growth between the national and the international source. As a general rule, the national source is to be preferred over the international source for analyses that relate to country only whereas the international source is often better suited for comparisons between countries.

Australia

The Australian Bureau of Statistics (ABS) has computed and published time series of multi-factor productivity indices for several years. The headline figure, also the one that is most timely available is MFP growth for the market sector, an industry grouping comprising agriculture, forestry and fishing; mining; manufacturing; electricity, gas and water supply; construction; wholesale trade; retail trade; accommodation, cafes and restaurants; transport and storage; communication services; finance and insurance; and the cultural and recreational services industries. These are industries with marketed activities for which there are satisfactory estimates of the growth in the volume of output. In addition, MFP productivity measures are presently being developed for individual industries from the Australian and New Zealand Standard Industrial Classification (ANZSIC).

The ABS derives its estimates of MFP by forming a combined chain volume measure of labour and capital and dividing it into a chain volume measures of the gross value added of the market sector. The elements of capital input are compiled for 14 asset types for the corporate and unincorporated sectors for each of the industries included in the market sector. For each capital there is a volume indicator of the flow of capital services and a rental value to weight the service flow with the service flows of other capital inputs. An aggregate chain volume measure of capital services for the whole market sector is then combined with a measure of hours worked using estimates of capital and labour income weights. For more details see ABS (2000).

The ABS’ MFP measures differ in several aspects from the MFP measures computed by the OECD. First and importantly, national data is based on more detailed source data than the international data. Second, ABS adjusts, on an experimental basis, labour input measures to reflect the composition of the labour force e.g. by age, education and experience. There are thus two MFP series, one based on simple hours worked (akin to the OECD labour input data) and one based on hours worked adjusted for compositional changes. Both series are shown in the graph below. It is apparent that MFP based on unadjusted hours rose more quickly than MFP based on adjusted hours. This reflects the fact that unadjusted hours rose less quickly than adjusted hours which in turn means that the composition of labour input has gradually shifted towards more qualified, more experienced workers. In other

6 Note that some OECD countries contributes to the development of productivity measurement without publishing MFP time-series, as for

example the Office of National Statistics which published the ONS Productivity Handbook in 2007, available at the following address: http://www.statistics.gov.uk/about/data/guides/productivity/default.asp; or Statistics Denmark in 2005: Produktivitetsudviklingen I Danmark 1966-2003, available at: http://www.dst.dk/publikation.aspx?cid=8853&ci=true&pti=3.

http://www.statistics.gov.uk/about/data/guides/productivity/default.asp

http://www.dst.dk/publikation.aspx?cid=8853&ci=true&pti=3


words, more of output growth can be attributed to labour input, and therefore, a smaller part remains as the ‘unexplained’ MFP residual.

Thirdly, capital input as computed by ABS is based on a broader scope of capital assets than used by the OECD. In particular, the national data includes agricultural land and inventories, two assets that are absent from the OECD capital computations. Productivity indexes fluctuate according to the business cycle. One way to measure the trend rate of growth of productivity is to calculate the average annual growth rate between growth cycle peaks. Growth cycle peaks for multifactor productivity are identified as local maximum positive deviations of the productivity index from its long-term trend. Growth rates of MFP between cycle peaks are shown in the figure below.

Multi-factor productivity (value-added based) in the Australian market sector

1985=100 (fiscal years)

Multi-factor productivity (value-added based) in the Australian market sector during growth cycles

Percentage changes at average annual rate between MFP growth cycle peaks (fiscal years)

90

95

100

105

110

115

120

125

130

135

Jun.1985 Jun.1988 Jun.1991 Jun.1994 Jun.1997 Jun.2000 Jun.2003 Jun.2006

MFP based on unadjusted hours worked

MFP based on hours worked adjusted for labour composition

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

Source: Australian Bureau of Statistics, Catalogue No 5204.0 - Australian System of National Accounts, 2006-07.

Sources

Australian Bureau of Statistics, available at: http://www.abs.gov.au

For further reading

Australian Bureau of Statistics (2000); Australian National Accounts: Concepts, Sources and Methods, Catalogue No 5216.0, Chapter 27.



Canada

Statistics Canada has computed and published time series of multi-factor productivity indices for a number of years. The headline figure, also the one that is most timely available is MFP growth for the business sector. In addition, MFP productivity measures are published for many 2-digit and 3-digit industries from the North American Industry Classification System (NAICS).

Multi-factor productivity (value-added based) in the Canadian business sector, 1981=100

90.0

92.0

94.0

96.0

98.0

100.0

102.0

104.0

106.0

108.0

Source: Statistics Canada, CANSIM Table 383-0021

MFP indices are computed as the ratio between value-added and combined labour and capital input, i.e., they constitute value-added based MFP measures. Statistics Canada’s MFP measures differ in several aspects from the MFP measures computed by the OECD. First and importantly, the national data is significantly more detailed and also more time than the international data. Second, labour input measures have been adjusted by Statistics Canada to reflect the composition of the labour force e.g. by age, education and experience whereas the OECD labour input data is a simple aggregate of hours worked. Thirdly, capital input as computed by Statistics Canada is based on a broader scope of capital assets than used by the OECD. In particular, the national data includes land and inventories, two assets that are absent from the OECD capital computations.

Productivity measures at the industry level are derived from a set of industry accounts. Under this approach, a variety of productivity series at the industry level are constructed using alternative measures of output along with their corresponding inputs. Industry data on outputs and inputs permit the construction of bottom-up MFP measures for major sectors as a weighted average of industry productivity growth rates. More detailed information on the methodology underlying Statistics Canada’s productivity series can be found in the series description Productivity Measures and Related Variables (CANSIM Record No 1402) and in Baldwin and Harchaoui (2005).


Multifactor productivity (value-added based) in Canada, 2-digit and selected 3-digit NAICS industries, 1995-2004 Percentage change, annual rate

-6.0 -4.0 -2.0 0.0 2.0 4.0 6.0

Oil and gas extraction

Health care and social assistance (except hospitals)

Mining and oil and gas extraction

Arts, entertainment and recreation

Leather and allied product manufacturing

Computer and electronic product manufacturing

Petroleum and coal products manufacturing

Support activities for mining and oil and gas extraction

Transportation and warehousing

Electrical equipment, appliance and component manufacturing

Clothing manufacturing

Administrative and support, waste management and remediation …

Beverage and tobacco product manufacturing

Other services (except public administration)

Finance, insurance, real estate and renting and leasing

Professional, scientific and technical services

Food manufacturing

Printing and related support activities

Machinery manufacturing

Accommodation and food services

Information and cultural industries

Fabricated metal product manufacturing

Construction

Manufacturing

Mining (except oil and gas)

Chemical manufacturing

Plastics and rubber products manufacturing

Utilities

Furniture and related product manufacturing

Wholesale trade

Transportation equipment manufacturing

Textile and textile product mills

Retail trade

Agriculture, forestry, fishing and hunting

Non-metallic mineral product manufacturing

Paper manufacturing

Primary metal manufacturing

Wood product manufacturing

Miscellaneous manufacturing

Source: Statistics Canada, CANSIM Table 383-0022

Sources

Statistics Canada, CANSIM Database, Tables 383-0021 to 383-0022.available at: http://www.statcan.ca

For further reading

Baldwin, John and Tarek Harchaoui (2005); “The Integration of the Canadian Productivity Accounts within the System of National Accounts: Current Status and Challenges Ahead”; Research Paper, Statistics Canada Economic analysis methodology paper series: National accounts; 11F0026 No 004.


http://cansim2.statcan.ca/cgi-win/CNSMCGI.EXE?Lang=E&ArrayId=383-0016..383-0019&Array_Pick=1&Detail=1&ResultTemplate=CII/CII___&RootDir=CII/&TblDetail=1


Netherlands

In 2007, Statistics Netherlands published a first set of MFP estimates at the industry branch and macro levels for the Netherlands for the period 1995-2006. The basic methodology behind these MFP estimates closely follows the methods presented in international documents such as the OECD Productivity Manual. In some specific aspects, however, the MFP measures from Statistics Netherlands differ from MFP measures computed by the OECD.

The official MFP measures, the gross-output based MFP (defined as a quantity index of gross output divided by a quantity index of combined KLEMS input) and the value-added based MFP (measured as a quantity index of combined labour and capital input), are computed and then compared with calculations using alternative assumptions with regard to the volume index formula, the user cost of capital, and the labour income of self-employed (self-employed are assumed to have the same income per hour or per year as employees).

It follows that international comparisons between the national MFP results for Netherlands and OECD’s MFP data for other countries have to be made with the necessary caution. In particular, Statistics Netherlands do not adopt the behavioural and structural assumptions of the neo-classical production framework, i.e. constant scale of returns and perfect competition and choose an axiomatic approach for the calculation of aggregate quantity or volume change of inputs and outputs. As a result, MFP change cannot be interpreted as exclusively the result of technological change but may also be due to scale effects, efficiency improvements, R&D investments that lead to monopolistic behaviours on the part of producer, and other factors. More detailed information on the methodology underlying Statistics Netherlands productivity measures can be found in the method description in Balk et al. (2007).

The headline figure is MFP growth for the commercial sector, in practice defined as the whole economy except general government, defence, subsidized education, real estate activities, renting of movables, and private households with employed persons. This is an important difference on the scope with OECD MFP statistics.

Contribution of multifactor productivity to value-added volume changes in Netherlands, commercial sector, 1995=100

90

95

100

105

110

115

120

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

Source: Statistics Netherlands

Productivity statistics at the industry level are performed at two different levels of aggregation: 36 industries, 9 industries. Under this approach, a variety of productivity series are constructed using alternative measures of output along with their corresponding inputs. Industry data on outputs and inputs permit the construction of bottom-up MFP value-added based measures for major sectors as a weighted average of industry productivity growth rates.


Contribution of multifactor productivity to value-added volume changes in Netherlands, 2-digit and selected 3-digit NACE industries, 1995-2005

Average annual growth rate

-2.0 0.0 2.0 4.0 6.0 8.0 10.0

Mining and quarrying

Other service activities n.e.c.

Health and social work activities

Construction

Insurance and pension funding

Care and other service activities

Other business activities n.e.c.

Activities auxiliary to financial intermediation

Hotels and restaurants

Financial and business activities

Retail trade and repair (excl. motor vehicles/cycles)

Recreational, cultural and sporting activities

Agriculture, forestry and fishing

Supporting transport activities

Land transport

Manufacture of fabricated metal products

Publishing and printing

Commercial sector

Manufacture of rubber and plastic products

Other manufacturing

Sewage and refuse disposal services

Research and development

Manufacture of food products, beverages and tobacco

Manufacture of petroleum products

Manufacture of electrical and optical equipment

Computer and related activities

Banking

Electricity, gas and water supply

Manufacturing

Trade and repair of motor vehicles/cycles

Trade, hotels, restaurants and repair

Manufacture of machinery and equipment n.e.c.

Manufacture of textile and leather products

Manufacture of paper and paper products

Transport, storage and communication

Air transport

Manufacture of basic metals

Wholesale trade (excl. motor vehicles/cycles)

Manufacture of basic chemicals and chemical products

Manufacture of transport equipment

Water transport

Post and telecommunications

%

Source: Statistics Netherlands

Statistics Netherlands will expand the system of productivity statistics, including quality changes in labour; including capital services of R&D and ICT; and constructing complete balance sheets for non-financial assets with the extension of the coverage of assets to inventories and non-produced assets such as land and subsoil assets.

Sources

Statistics Netherlands, available at: http://www.cbs.nl

For further reading

Balk, Bert M., Bergen, Dirk van den, de Haan, Mark and Myriam van Rooijen-Horsten (2007), Productivity Measurement at Statistics Netherlands, Statistics Netherlands, Voorburg.



New Zealand

In 2006, Statistics New Zealand released, for the first time, an official time series of multi-factor productivity growth. This first dataset relates to the ‘measured sector’, consisting of industries for which estimates of inputs and outputs are independently derived in constant prices. Excluded are those industries – mainly government non-market industries whose services, such as administration, health and education, are provided free or at nominal charges – whose real value-added is measured in the national accounts largely using input methods, such as numbers of employees. Using an input series to estimate the change in outputs implicitly assumes nil productivity growth and to include these industries in the 'measured sector' would lead to a bias in the productivity series. Also excluded are a number of private sector market industries that similarly use some form of input measure to estimate real output, for example the residential and commercial property industries whose output is measured by the growth in property assets.

Labour input is measured as the total number of hours paid, the number of ordinary and overtime hours for which an employee is paid. It excludes unpaid overtime but may include some hours that are not actually worked, such as paid leave and statutory holidays. While Statistics New Zealand states a conceptual preference of a measure of hours worked over hours paid, it has greater confidence in the quality of its hours paid data which are also available for longer time series. These data considerations led to the choice of hours paid over hours worked. A full description of data sources and methodology can be found in Statistics New Zealand (2006).

Statistics New Zealand follows standard methodology and derives its estimates of MFP by subtracting a volume growth measure of labour and capital inputs from a volume growth measure of output, constant price gross value added.

The elements of capital input are compiled at a level of 24 types of assets, and 22 industries. This is more detailed and broader in scope than the OECD’s own estimates of capital services (which exclude, for example, residential buildings). In the national series for New Zealand, for each type of asset in every industry there is a volume indicator of the flow of capital services and a rental value to weight the service flow with the service flows of other capital inputs. An aggregate chain volume measure of capital services for the whole economy is then combined with a measure of hours worked using estimates of capital and labour income weights.

The basic methodology behind the New Zealand MFP estimates closely follows the methods presented in international documents such as the OECD Productivity Manual. In some specific aspects, however, the national measures differ from MFP measures computed by the OECD (scope of capital assets, sector coverage). It follows that the OECD’s estimates of New Zealand’s MFP growth cannot be directly compared with the national estimates published by Statistics New Zealand.

Sources

Statistics New Zealand (2006); Productivity Statistics: Sources and Methods; available at http://www2.stats.govt.nz For further reading



Multi-factor productivity (value-added based) in New Zealand 1988=100

Growth accounts for New Zealand, 1990-95, 1995-2000, 2000-2005 and 1995-2005 In percentage points

90

95

100

105

110

115

120

125

130

135

140

MFP total economy

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

1990-1995 1995-2000 2000-2005 1995-2005

Contribution of labour input Contribution of capital input

Source: Statistics New Zealand

Switzerland

In 2006, the Swiss Federal Statistical Office published a first set of MFP estimates for Switzerland. This first dataset relates to the total economy and comprises thus the private and the public sector. This differs from other national MFP measures (e.g. Canada, Australia or the United States) but is similar to the OECD MFP statistics.

Labour input measures also correspond to the simple unadjusted series of hours worked that are used by the OECD. This reflects a constraint on data availability that does not presently permit the Swiss Federal Statistical Office to adjust hours worked for compositional change although the desirability of such an adjustment is clearly recognised by the Statistical Office.

The Swiss Federal Statistical Office follows standard methodology and derives its estimates of MFP by deducting, from the volume measure of GDP a weighted volume growth measure of labour and capital inputs. The elements of capital input are compiled at a level of 16 types of assets. For each type of asset there is a volume indicator of the flow of capital services and a rental value to weight the service flow with the service flows of other capital inputs. An aggregate chain volume measure of capital services for the whole economy is then combined with a measure of hours worked using estimates of capital and labour income weights.

The basic methodology behind the Swiss MFP estimates closely follows the methods presented in international documents such as the OECD Productivity Manual. In some specific aspects, however, the Swiss MFP measures differ from MFP measures computed by the OECD. The national data is based on more detailed source data than the six-way asset classification used by the OECD. There is also important difference in the scope of capital measures. In particular, the estimates by the Swiss Federal Statistical Office include residential assets which are excluded from the OECD data. It follows that international comparisons between the national MFP results for Switzerland and OECD’s MFP data for other countries have to be made with the necessary caution.

Sources

Swiss Federal Statistical Office http://www.bfs.admin.ch For further reading

Swiss Federal Statistical Office (2006), Analyse des contributions à la croissance des facteurs de production, de la productivité multifactorielle et du rôle joué par l'intensité capitalistique de 1991 à 2004, Neuchâtel. Swiss Federal Statistical Office (2006), « Stock de capital non financier », Rapport méthodologique, Neuchâtel. Swiss Federal Statistical Office (2006), « Productivité multifactorielle », Rapport méthodologique, Neuchâtel. OECD (2001), Measuring Productivity – OECD Manual, OECD, Paris.

http://www.bfs.admin.ch/


Multi-factor productivity (value-added based) in Switzerland 1991=100

Growth accounts for Switzerland, 1995-2000, 2000-2005 and 1995-2004

In percentage points

94

96

98

100

102

104

106

108

110

MFP total economy

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

1995-2000 2000-2005 1995-2005

Contribution of labour input Contribution of capital input

Source: Swiss Federal Statistical Office

United States

The United States Bureau of Labor Statistics has computed and published time series of multi-factor productivity indices for a number of years. The headline figure, also the one most time available are MFP growth for the business and the non-farm business sector. In addition, MFP productivity measures are published for 18 3-digit SIC manufacturing industries and 86 4-digit SIC manufacturing industries, railroad transportation, and air transportation and the utility and gas industry.

Availability of the U.S. BLS productivity measures for major sectors and sub-sectors of the economy

Productivity measure Input(s) Index available

Labour productivity1

Business Labour Quarterly

Non-farm business Labour Quarterly

Non-financial corporations

Manufacturing, total

Durable

Non-Durable

Multi-factor productivity1

Private business Labour, capital Annually

Private non-farm business Labour, capital Annually

KLEMS Multi-factor productivity

Manufacturing and 20 2-digit SIC manufacturing industries services

Labour, capital, energy, materials, services

Annually

1 Includes government enterprises; multi-factor productivity measures exclude such enterprises


Business sector indices of MFP are computed as the ratio between value-added and combined labour and capital input, i.e., as value-added based MFP measures.

Major sectors and manufacturing industries of the economy in the United States

Multi-factor productivity (value-added based) in the business sector

1987=100

KLEMS multi-factor productivity in manufacturing, 1987=100

95

100

105

110

115

120

125

1987 1989 1991 1993 1995 1997 1999 2001 2003 2005

Private business sector Private non-farm business sector

95

100

105

110

115

120

125

130

135

1987 1989 1991 1993 1995 1997 1999 2001 2003 2005

Manufacturing industries

Source: United States Bureau of Labor Statistics.

Industry-level multifactor productivity measures are constructed by dividing an index of output by an index of combined inputs. Combined inputs are a weighted average of employee hours, capital services (land, structures, equipment and inventories), and intermediate purchases (materials, energy, and purchased services).

KLEMS multi-factor productivity in the US, 3-digit SIC manufacturing industries, 1987-2005 Percentage change, annual rate

-2.0 0.0 2.0 4.0 6.0 8.0 10.0

Elec.Equip. & Appl.

Machinery

Food, Beverages & Tobacco

Transportation Eq.

Petroleum & Coal

Wood Products

Chemical Products

Non-durable goods

Paper Products

Furniture & Rel.

Fabricated Metal

Printing & Support

Nonmetallic Mineral

Plastics & Rubber

Primary Metal

Apparel & Leather

Misc. Manufacturing

Textile Mills

Durable goods

Computer & Elec.

%

Source: United States Bureau of Labor Statistics.


The indexes of output and of combined inputs are Törnqvist indexes, developed for each industry by computing a weighted average of the growth rates of the various outputs or inputs between two periods, with weights based on relative cost shares. The weight for each item equals its average value share in the two periods. Thus, industry-level MFP measures follow a KLEMS-type approach which is different from the value-added based approach for the total business sector. For a more complete discussion of the Törnqvist methodology see "Industry Productivity Measures," Chapter 11 of the BLS Handbook of Methods.

BLS MFP measures differ in several aspects from the MFP measures computed by the OECD. First and importantly, the national data is significantly more detailed and also more time than the international data. Second, labour input measures have been adjusted by BLS to reflect the composition of the labour force e.g. by age, education and experience whereas the OECD labour input data is a simple aggregate of hours worked.

Sources

United States Bureau of Labor Statistics, Multifactor Productivity statistics, available at: http://www.bls.gov/mfp/home.htm.

For further reading

United States Bureau of Labor Statistics (1997), Handbook of Methods; BLS Bulletin 2490.


http://www.bls.gov/opub/hom/homch11_a.htm


REFERENCES

Ahmad, Nadim (2003), “Measuring Investment in Software”, STI Working Paper 2003/6, OECD, Paris

Ahmad, Nadim (2003), “Towards More Harmonised Estimates of Investment in Software”, OECD Economic Studies, No. 37, 2003/2, pp. 185-200.

Ahmad, Nadim, Francois Lequiller, Pascal Marianna, Dirk Pilat, Paul Schreyer and Anita Wölfl (2003), “Comparing Labour Productivity Growth in the OECD Area: The Role of Measurement”, STI Working Paper 2003/14 also published as OECD Statistics Working Papers 2003/5, OECD, Paris.

Australian Bureau of Statistics (2000), Australian National Accounts: Concepts, Sources and Methods, Catalogue N°5216.0, Chapter 27.

Balk, Bert, Bergen, Dirk van der, de Haan, Mark and Myriam van Rooijen-Horsten, (2007), Productivity Measurement at Statistics Netherlands, Statistics Netherlands, Voorburg.

Baldwin, John and Tarek Harchaoui (2005), ”The Integration of the Canadian Productivity Accounts within the System of National Accounts: Current Status and Challenges Ahead”, Research Paper, Statistics Canada Economic Analysis Methodology Paper Series: National Accounts, 11F0026 N°004.

Bassanini, Andrea and Stefano Scarpetta (2001), ”Does Human Capital Matter for Growth in OECD Countries”, Economics Department Working Papers No. 282, OECD, Paris.

Bassanini, Andrea and Stefano Scarpetta (2001), “The Driving Forces of Economic Growth: Panel Data Evidence for the OECD Countries”, OECD Economic Studies, No. 33, 2001/II, pp. 9-56, OECD, Paris.


Colecchia, A. (2005), “A World of Heterogeneous Workers. What Implications for Human Capital and Productivity Comparisons in the G7 Countries”, STI Working Paper, OECD, Paris, forthcoming.

Colecchia, Alessandra and Paul Schreyer (2001), “The Impact of Information Communications Technology on Output Growth”, STI Working Paper 2001/7, OECD, Paris.

Colecchia, Alessandra and Paul Schreyer (2002), “The Contribution of Information and Communications Technologies to Economic Growth in Nine Countries”, OECD Economic Studies No. 34, pp. 153-171, OECD, Paris.


Criscuolo, Chiara (2005), “The Contribution of Foreign Affiliates to Productivity Growth: Evidence from OECD Countries”, STI Working Paper 2005/8, OECD, Paris.

De Serres, Alain (2003), ”Structural Policies and Growth – A Non-Technical Overview”, Economics Department Working Papers No. 355, OECD, Paris.

Evans, John, Douglas Lippoldt and Pascal Marianna (2002), ”Trends in Working Hours in OECD Countries”, Labour Market and Social Policy Occasional Paper 45, OECD, Paris.

Guellec, Dominique and Bruno van Pottelsberghe de la Potterie (2001), “R&D and Productivity Growth: Panel Data Analysis of 16 OECD Countries”, OECD Economic Studies, No. 33, 2001/II, pp. 103-126, OECD, Paris.

Hall, Robert E. (1990), “Invariance Properties of Solow’s Productivity Residual”, in Peter Diamond (ed.), Growth/Productivity/Unemployment: Essays to Celebrate Bob Solow’s Birthday, MIT Press, pp 71-112.


Kohli, Ulrich (2004), “Real GDP, Real Domestic Income and Terms of Trade Changes”, Journal of International Economics, 62.

Lequiller, Francois, Nadim Ahmad, Seppo Varjonen, William Cave, Kil-Hyo Ahn (2003), ”Report of the OECD TaskForce on Software Measurement in the National Accounts”, OECD Statistics Working Papers 2003/1, OECD, Paris.

Maddison, Angus (2001), The World Economy: A Millennial Perspective, Development Centre Studies, OECD, Paris.

Nicoletti, Giuseppe and Stefano Scarpetta (2003), “Regulation, Productivity and Growth: OECD Evidence”, OECD Economics Department Working Papers, No. 347, OECD, Paris.

OECD (2007), OECD Science, technology and Industry Scoreboard 2007, OECD, Paris.


OECD (2005), Enhancing Services Sector Performance, OECD, Paris.

OECD (2005), Measuring Globalisation: Economic Globalisation Indicators. OECD, Paris.

OECD (2004), The Economic Impacts of ICT – Measurement, Evidence and Implications, OECD, Paris.

OECD (2004), “Recent Labour Market Developments and Prospects – Special Focus on: Clocking in (and out): Several Facets of Working Time”, OECD Employment Outlook, Chapter 1, OECD, Paris.

OECD (2003), The Sources of Economic Growth in OECD Countries, OECD, Paris.

OECD (2003), ICT and Economic Growth – Evidence from OECD Countries, Industries and Firms, OECD, Paris.

OECD (2003), OECD Science, technology and Industry Scoreboard 2003, section D, Annex 1., OECD, Paris.

OECD (2001), Measuring Productivity - OECD Manual, OECD, Paris.

OECD (2001), OECD Science, technology and Industry Scoreboard 2001, D.4., OECD, Paris.

OECD (1996), Industry Productivity: International Comparison and Measurement Issues, OECD, Paris

OECD (forthcoming), “Measuring Capital – OECD Manual: Measurement of Capital Stock, Consumption of Capital and Capital Services”, revised edition, OECD, Paris.

Office for National Statistics (2007), The ONS Productivity Handbook: A Statistical Overview and Guide, edited by Dawn Camus, first published by Palgrave MacMillan, Basingstoke and New York.

Oliveira Martins, Joaquim, Scarpetta, Stephano and Dirk Pilat (1996), “Mark-up Pricing, Market Structure and the Business Cycle”, OECD Economic Studies No. 27, 1996/22, OECD, Paris.

Oliveira Martins, Joaquim, Scarpetta, Stephano and Dirk Pilat (1996), “Mark-up Ratios in Manufacturing Industries: Estimates for 14 OECD Countries”, OECD Economics Department Working Papers No. 162, OECD, Paris.

Oliveira Martins, Joaquim and Stephano Scarpetta (1999), “The Levels of and Cyclical Behaviour of Mark-up Across Countries and Market Structure”, OECD Economics Department Working Papers No. 213, OECD, Paris.

Pilat, Dirk (2004), “The ICT Productivity Paradox: Insights from Micro Data”, OECD Economic Studies, 2004/1, pp. 37-65, OECD, Paris.

Pilat, Dirk, Cimper, Agnès, Olsen, K., Webb, Colin (2006), “The Changing Nature of Manufacturing in OECD Economies”, STI Working Paper 2006/9, OECD, Paris.

Pilat, Dirk, and Frank Lee (2001), “Productivity growth in ICT-producing and ICT-using industries: A source of growth differentials in the OECD?”, STI Working Paper 2001/4, OECD, Paris.

Pilat, Dirk, Frank Lee and Bart van Ark (2002), “Production and Use of ICT: A Sectoral Perspective on Productivity Growth in the OECD Area”, OECD Economic Studies No. 35, pp. 47-78, OECD, Paris.


Pilat, Dirk and Paul Schreyer (2004), “The OECD Productivity Database – An Overview”, International Productivity Monitor, Number 8, Spring, pp. 59-65.

Scarpetta, Stefano, Andrea Bassanini, Dirk Pilat and Paul Schreyer (2000), “Economic Growth in the OECD Area: Recent Trends at the Aggregate and Sectoral Level”, Economics Department Working Papers No. 248, OECD, Paris.

Schreyer, Paul (2003), “Capital Stocks, Capital Services and Multi-Factor Productivity Measures”, OECD Economic Studies, No. 37, 2003/2, pp. 164-184.

Schreyer, Paul (2001), “Computer Price Indices and International Growth and Productivity Comparisons”, Economics of Innovation and New technology Vol. 10.

Schreyer, Paul (2000), “The Contribution of ICTs to Output Growth: A Study of the G7 Countries”, STI Working Paper 2000/2, OECD, Paris.

Schreyer, Paul and Francette Koechlin (2002), “Purchasing power parities – measurement and uses”, Statistics Brief, March 2002, No. 3, OECD, Paris.

Schreyer, Paul and Dirk Pilat (2001), “Measuring Productivity”, OECD Economic Studies No. 33, pp. 127-169, OECD, Paris.

Schreyer, Paul, Pierre-Emmanuel Bignon and Julien Dupont (2003), “OECD Capital Services Estimates: Methodology and A First Set of Results”, OECD Statistics Working Papers 2003/6, OECD, Paris.

Skyttesvall, Tomas and Hans-Olov Hagen (2007), Growth in the Nordic Business Sector, Nordic Council of Ministers, Copenhagen.

Statistics Denmark (2005), Produktivitetsudviklingen I Denmark 1966-2003, Statistics Denmark, Copenhagen.

Swiss Federal Statistical Office (2006), Analyse des contributions à la croissance des facteurs de production, de la productivité multi-factorielle et du rôle joué par l’intensité capitalistique de 1991 à 2004, OFS, Neuchâtel.

Swiss Federal Statistical Office (2006), « Stock de capital non financier », Rapport méthodologique, OFS, Neuchâtel.

Swiss Federal Statistical Office (2006), « Productivité multi-factorielle », Rapport méthodologique, OFS, Neuchâtel.

Triplett, Jack (2004), “Handbook on Hedonic Indexes and Quality Adjustment in Price Indexes: Special Application to Information Technology Products”, STI Working Papers 2004/9, OECD, Paris.

United States Bureau of Labor Statistics (1997), Handbook of Methods, BLS Bulletin 2490.

Wölfl, Anita (2005), “The Service Economy in OECD Countries”, STI Working Papers 2005/3, OECD, Paris.

Wölfl, Anita (2003), “Productivity Growth in Service Industries: An Assessment of Recent Patterns and the Role of Measurement”, STI Working Papers 2003/7, OECD, Paris.

OECD COMPENDIUM OF PRODUCTIVITY INDICATORS · 2021. 4. 25. · 8 OECD Compendium of Productivity...

Documents

Transcript of OECD COMPENDIUM OF PRODUCTIVITY INDICATORS · 2021. 4. 25. · 8 OECD Compendium of Productivity...