URN 07/1314

Exploring gender differentials in access to business finance – an econometric

analysis of survey data

December 2006

Final Report

Report prepared by Stephen Roper, Nigel Driffield, Vania Sena, Dolores Añon Higon and Jonathan Scott (ESG Associates) for the Small Business

Service. Contact details: s.roper@aston.ac.uk

Contents

EXECUTIVE SUMMARY.......................................................................................................................3

CHAPTER 1 - AIMS AND OBJECTIVES...........................................................................................3

1.1 INTRODUCTION................................................................................................................................31.2 ORGANISATION OF THE REPORT......................................................................................................3

CHAPTER 2 - A SELECTIVE OVERVIEW OF PREVIOUS EVIDENCE.....................................3

2.1 INTRODUCTION................................................................................................................................32.2 ISSUES IN SMALL BUSINESS FINANCE.............................................................................................32.3 FINANCING START-UP – THE EVIDENCE ON GENDER DIFFERENCES.................................................32.4 KEY THEMES...................................................................................................................................3

CHAPTER 3 - ACCESS TO FINANCE AND BUSINESS START-UP.............................................3

3.1 INTRODUCTION................................................................................................................................33.2 EVIDENCE FROM GEM 2004...........................................................................................................3

3.2.1 Descriptive Analysis................................................................................................................33.2.2 Perceived financial barriers to business start-up...................................................................33.2.3 Business Start-up.....................................................................................................................33.2.4 Reasons for perceived financial barriers................................................................................33.2.5 Summary of key points.............................................................................................................3

3.3 EVIDENCE FROM THE HSE..............................................................................................................33.3.1 Descriptive Analysis................................................................................................................33.3.2 Start-up and the Financial Sourcing Decision........................................................................33.3.3 Start-up and Difficulty Obtaining Finance.............................................................................33.3.4 Key Points................................................................................................................................3

3.4 CONCLUDING REMARKS..................................................................................................................3

CHAPTER 4 - ACCESS TO FINANCE BY ESTABLISHED FIRMS..............................................3

4.1 INTRODUCTION................................................................................................................................34.2 EVIDENCE FROM THE ASBS............................................................................................................3

4.2.1 Supply Side Models.................................................................................................................34.2.2 The Demand for finance..........................................................................................................34.2.3 Concluding Points...................................................................................................................3

4.3 EVIDENCE FROM THE UKSMEF 2004 DATABASE..........................................................................34.3.1 Descriptive Analysis................................................................................................................34.3.2. Modelling the “Supply Side”: Barriers to Finance...............................................................34.3.3 Modelling the Demand for Finance across different groups..................................................34.3.4 Summary of Key Points...........................................................................................................3

4.4 CONCLUSIONS FROM THE FIRM BASED ANALYSIS...........................................................................3

CHAPTER 5 – CONCLUSIONS............................................................................................................3

5.1 INTRODUCTION................................................................................................................................35.2 GENDER EFFECTS ON ACCESSING FINANCE AND ITS IMPACT ON BUSINESS START-UP....................35.3 GENDER EFFECTS ON ACCESS TO FINANCE BY EXISTING FIRMS......................................................35.4 ISSUES FOR FUTURE RESEARCH.......................................................................................................3

ANNEX 1: MODELLING ACCESS TO FINANCE AND BUSINESS START-UP.........................3

A1.1 INTRODUCTION.............................................................................................................................3A1.2 MODELLING WITH GEM 2004......................................................................................................3A1.3 MODELLING WITH THE HSE.........................................................................................................3

ANNEX 2: MODELLING THE ACCESS TO FINANCE BY ESTABLISHED FIRMS.................3

A2.1 INTRODUCTION.............................................................................................................................3A2.2 MODELLING USING THE ASBS....................................................................................................3A2.3 MODELLING USING THE UKSMEF 2004.....................................................................................3

ANNEX 3: VARIABLE DEFINITIONS AND CONSTRUCTION....................................................3

A3.1 GEM 2004....................................................................................................................................3A3.2 HSE 2003.....................................................................................................................................3A3.3 ANNUAL SMALL BUSINESS SURVEY 2003 AND 2004...................................................................3A3.4 Variables Used in the UKSMEF 2004 Analysis..........................................................................3

2

3

Executive Summary

Background

1 The availability of finance for business start-up and expansion has attracted much attention over recent years and stimulated the development of a number of focussed policy initiatives. A particular focus of recent initiatives has been to try and support women’s enterprise given consistent evidence from the Global Enterprise Monitor (GEM) studies and other sources about lower levels of involvement in enterprise among women (Section 1.1).

2 Previous research in this area has emphasised the complexity of the issues relating to business finance and particularly the difficulty of trying to isolate and characterise any specific gender effects. In this study we use an econometric approach to analyse gender differences in access to finance in four pre-existing databases used by the Small Business Service (Section 1.1).

3 Two of these databases – the Global Entrepreneurship Monitor 2004 and Household Survey for Entrepreneurship (HSE) 2003 – are surveys of individuals and provide evidence on the role of financial constraints on business start-up. The two other surveys – the Annual Small Business Survey (ASBS) 2003 and 2004) and the UK Survey of SME Finances (UKSMEF) 2004 – provide information on access to finance by existing businesses. Each of these surveys covers the whole of the UK with the exception of the Household Survey of Entrepreneurship which covers England only (Section 1.1).

Gender effects on accessing finance and business start-up

4 Both the GEM 2004 data and the HSE provide evidence of a negative, gender-specific, finance effect which would tend to reduce start-up rates among women. The GEM dataset suggests that women are around 7.5 percentage points more likely to perceive financial barriers to business start-up than men. This in turn works to reduce start-up rates for women by 1.7-3.8 percentage points depending on the start-up indicator being used. Being a woman has an additional direct effect on each of our start-up indicators, that is not linked to financial barriers (Section 3.2)

5 Because of the structure of the data, our HSE analysis focuses on a narrower group of the population than our GEM analysis and relates specifically to those classified as either ‘Thinkers’ and ‘Doers’, i.e. those who are either engaged in or thinking about undertaking some enterprise activity. Within this group we find no evidence that women face any increased difficulty in obtaining start-up finance. We do find evidence, however, that women are less likely to seek external finance for business start-up. (Section 3.3)

6 Taking these points together suggests that women in the general population perceive stronger financial barriers to business start-up than men, and this may be discouraging them from seeking external financial support for business start-up. We find no evidence, however, that where women do seek finance for start-up they are any less likely to obtain it than men. This is suggestive of a dominant

4

demand rather than supply side effect. In either case, however, the effect is similar – that gender differences in access to finance are reducing start-up rates among women (Section 3.4).

Gender effects on the access to finance by existing firms

7 Our analysis of the ASBS for 2003 and 2004 and the UKSMEF for 2004 focuses separately on the supply and demand side of the financing relationship for existing businesses (Section 4.1)

8 In terms of the supply side our results are somewhat contradictory with evidence from the ASBS highlighting some negative gender effects but the UKSMEF suggesting that women-led businesses are less likely than men to be discouraged in their search for business finance (Section 4.2)

9 More specifically, the ASBS suggests that women-led businesses are around 2 percentage points more likely to have difficulty raising finance and also 2 percentage points more likely to find it impossible to raise the finance they are seeking. The UKSMEF 2004 data on the other hand suggests that women-led businesses encounter no difference in rejection rates compared to male-led businesses and are less likely to face discouragement when applying for external finance (Section 4.3)

10 In terms of the demand side we see a more consistent picture, however, and find evidence that broadly supports that identified earlier from the GEM and HSE analysis (Section 4.3)

11 The ASBS suggests that women-led businesses are less likely to seek external finance both on a one-off and multiple basis. The UKSMEF provides no support for this general proposition, but does suggest some more specific effects with women-led businesses 17 percentage points less likely to apply for a commercial loan or mortgage from banks or other financial institutions. These demand-side results provide one possible reason for the differences in supply side results suggested by the supply-side analysis (Section 4.4).

Acknowledgements

We are grateful to the project Steering Committee for their comments and help in progressing this study. Some of the data for this study were provided by the Global Entrepreneurship Monitor UK (GEM UK), which is part of the Global Entrepreneurship Monitor consortium. Names of the members of national teams, the global coordination team, and the financial sponsors are published in the Global Entrepreneurship Monitor 2005 Report, which can be downloaded at www.gemconsortium.org.  We thank all the researchers and their financial supporters who made this research possible. Although data used in this work are collected by GEM UK, the analysis and interpretation presented here are the sole responsibility of the authors.

5

Chapter 1 - Aims and Objectives

1.1 Introduction

1.1 The availability of finance for business start-up and expansion has attracted much attention over recent years and stimulated the development of a number of focussed policy initiatives. A particular focus of recent initiatives has been to try to support women’s enterprise given consistent evidence from the Global Entrepreneurship Monitor (GEM) studies and other sources about lower levels of involvement in enterprise among women.

1.2 Previous research in this area has emphasised the complexity of the issues relating to business finance and particularly the difficulty of trying to isolate and characterise any specific gender effects. Is it the case, for example, that lending institutions discriminate either deliberately or unwittingly against entrepreneurs who are women? Or, are women entrepreneurs simply more reluctant to seek business finance? Other factors linked to background or experience may also be important in shaping men’s and women’s access to finance.

1.3 In this study we use an econometric approach to analyse gender differences in access to finance in four pre-existing databases used by the Small Business Service. Two of these databases – the Global Entrepreneurship Monitor 2004 and Household Survey for Entrepreneurship (HSE) 2003 – are surveys of individuals and provide evidence on the role of financial constraints on business start-up. Here there are two main questions. First, how important is gender in shaping individuals access to finance for business start-up? And, secondly, how important are any such financial constraints on subsequent start-up? One important difference between GEM and the HSE is that GEM covers the whole UK while the HSE is England only. The two other surveys (both of which cover the UK) – the Annual Small Business Survey (ASBS) 2003 and 2004 and the UK Survey of SME Finances (UKSMEF) 2004 – provide information on access to finance by existing firms. Here we are interested in whether women-led firms are more likely to experience constraints in obtaining finance, distinguishing between supply and demand side influences.

1.4 Our focus here therefore is not in providing a detailed description of the survey results themselves. This has already been done effectively in the relevant survey reports. Our objective instead is to explore whether adopting an econometric approach to the survey analysis can shed additional light on the relationship between gender and access to finance, taking into account as wide a range of other factors as possible.

1.2 Organisation of the Report

1.5 The remainder of the report is organised as follows. Chapter 2, immediately following this, provides a brief overview of previous studies relating to small

business finance and, more specifically, the role of gender in shaping individuals’ and firms’ access to business finance.

1.6 The main analytical section of the report is then divided into two Chapters. Chapter 3 focuses on the two surveys of individuals and examines the impact of gender on finance for business start-up and then start-up itself. This emphasises both the general significance of gender in terms of perceptions of financial barriers to business start-up, but also the effect of such perceived barriers on start-up (GEM). Other survey data (HSE) emphasises the potential importance of demand side effects, with women generally less willing to seek external finance.

1.7 Chapter 4 then focuses on the demand for finance from existing businesses and investigates whether women-led businesses face particular barriers in accessing business finance. Demand and supply side effects are investigated separately with both surveys suggesting that women-led businesses are often reluctant to seek external finance, an effect which seems particularly strong for commercial loans or mortgages from banks or other financial institutions. Our more general results on barriers to accessing finance for existing women-led businesses are more ambiguous, however, with the ASBS suggesting some difficulties and the UKSMEF suggesting that women-led businesses may actually face less discouragement in their search for external finance. Both results, however, are likely to reflect the reluctance of some women-led businesses to seek bank finance.

1.8 Chapter 5 draws the results of the study together and suggests some possible policy implications and directions for future research and survey development.

7

Chapter 2 - A Selective Overview of Previous Evidence

2.1 Introduction

2.1 In this Chapter we provide a selective review of recent academic and policy related research on access to finance generally for smaller firms and, more specifically, gender differences. Our objective here is not to provide a comprehensive review of the relevant literatures but to highlight the key issues which have emerged from previous studies. These provide the basis for the inclusion of specific variables in the models estimated later in the report and also shape our discussion of access to finance by men and women.

2.2 Section 2.2 reviews some recent literature on small business finance in general, emphasising the range of factors which can influence individuals’ and firms’ access to finance. Section 2.3 focuses more specifically on studies of gender differences in access to finance by start-up businesses and other small firms. Section 2.4 draws out the key themes from the literature review.

2.2 Issues in Small Business Finance2.3 Financial constraints are of necessity a major issue for small firms and start-up

companies but once firms are established it is possible to over-emphasise the importance of financial constraints. The Government’s Policy Action Team (PAT 14) articulated the difficulties faced by some businesses in accessing bank finance – due primarily to their age, experience, track record or business structure. However, access to finance is often over-shadowed by other problems when businesses actually start up, with finance cited as a problem by fewer than two per cent of respondents to the NatWest Small Business Research Team’s quarterly survey. Kotey (1999) is helpful, however, in emphasising that business growth can be constrained and failure can be caused by financing constraints, and that there are both supply and demand side factors involved in shaping small firms’ access to finance.

2.4 On the supply side, Cosh and Hughes (2003) found that loans from banks provide the funding for around two thirds of UK businesses and the largest source for over 25 per cent of firms. By the end of 2004, term lending by banks had grown to nearly £35bn (16 per cent growth in 2004) and overdraft lending had grown to nearly £10bn (9 per cent growth) (British Bankers’ Association 2004). However, Kotey (1999) notes that banks are less likely to lend long-term to SMEs due to risk (which is in itself caused by SMEs “lack[ing] a track record of performance on the basis of which their credit rating could be assessed”) and cost (“administrative costs, potential interest income and to the risk of default”) and on account of collateral unavailability1.

1 More generalised literature includes articles such as those considering lending by ‘de novo’ banks in comparison with incumbent banks in 1987-1994 (Goldberg and White 1998), on relationships between SMEs and banks (Meyer 1998; Jones 2001; Strahan and Weston 1998) and with specific consideration of benefits and costs, including barriers (Bornheim and Herbeck 1998).

2.5 On the demand side, Fraser (2005) reported that some 2.9m SMEs (80 per cent) have used external finance in the last three years and that the main sources of finance for start ups are personal savings (65 per cent), bank loan (10 per cent) and friends/family loan (6 per cent). He also found that approximately 900,000 businesses (24 per cent) use term loans and that obtaining finance is reported as a major problem at start up by some 10 per cent of businesses. This generalised view of the difficulty in obtaining finance for start-up, however, reflects a number of issues relating both to start-ups’ ability to attract finance as well as their willingness to consider different types of business financing.

2.6 Research by Hamilton and Fox (1998) provides insight into the financing preferences of entrepreneurs and: “supports the view that the financing decisions of small firm owners are based on a demand-side pecking order of finance types. The resulting financial structures reflect a desire to minimise intrusion into the firms and are not entirely the consequence of persistent deficiencies in the provision of finance to small firms.” Essentially similar evidence is provided by Howorth (2001) whose evidence suggests that entrepreneurs tend to seek finance first from their own resources, and friends and families, and then from other sources such as banks. Indeed, the money from family and friends is often essential (and often regarded as quasi-equity by the banks) to unlock support from commercial institutions. Thus the issue of entrepreneurs desiring maintenance of the control of their business is also a highly relevant consideration when thinking about barriers to access to bank finance for entrepreneurs.

2.7 More generally, Winker (1999) examined the causes of finance constraints and found these to be influenced by firm age and size. Cressy and Toivanen (2001) also emphasise that, “better borrowers get larger loans and lower interest rates; collateral provision and loan size reduce the interest rate paid … the bank is shown to use qualitative as well as quantitative information in the structuring of loan contracts to small businesses.” A somewhat contrary view is emphasised by Chandler and Hanks (1998), however, who note that: “there is some feeling among scholars that competent founders will find a way of coming up with necessary resources and capital”.

2.3 Financing Start-up – the evidence on gender differences2.8 A useful starting point here is the review of the literature on women’s

entrepreneurship by Carter et al. (2001). They start by reflecting the general tenor of the literature on small business start-up, i.e.:

‘… a preoccupation with start-up permeates the female entrepreneurship literature, but is particularly noticeable within the more descriptive analyses. Within this literature there is a widespread and generally unquestioned acceptance that start-up is more difficult for women. A key debate, however, is whether the barriers encountered by women at start-up have a long-term effect on business performance or whether these constraints dissipate after start up has been successfully negotiated’. (p29)

9

2.9 The same general point is emphasised by other writers. Marlow and Watson (2004), for example,2 argue that: “female owned enterprises are more likely to be undercapitalised in a variety of forms from the outset, locate in crowded sectors and so under perform over time” Another particularly revealing quotation from other authors is that: “Not only does policy appear to concentrate on areas traditionally associated with men in self employment, but the systems of finance and advice are also firmly oriented towards them, leaving women to face a range of barriers when engaging with self employment” (Warren-Smith and Jackson 2004).3

2.10 More recent reports published by the Small Business Service (SBS) emphasise different aspects of the finance issue. The Annual Survey of Small Business (ASBS) for 2004, for example, suggests that obtaining finance was an obstacle for 15.5 per cent of all small firms but and for 16.2 per cent of women-led enterprises4. The UK Survey of SME Finances (UKSMEF) emphasises another gender related issue, noting that “female-owned businesses pay significantly higher margins on term loans than male-owned businesses (2.9 versus 1.9 percentage points over Base)” (p18).

2.11 Carter et al. (2001) stress, however, that access to finance is only one aspect of the wider set of issues which surround start-up by women. They identify a number of studies, for example, that focus on finance for start-up and 'the social systems that endowed women with a lack of business credibility.' In particular, they quote Hisrich and Brush (1986: 17), who note that there is a perception that women are “not as serious as men about business”:

“For a woman entrepreneur who lacks experience in executive management, has had limited financial responsibilities, and proposes a non-proprietary product, the task of persuading a loan officer to lend start-up capital is not an easy one. As a result, a woman must often have her husband co-sign a note, seek a co-owner, or use personal assets or savings. Many women entrepreneurs feel strongly that they have been discriminated against in this financial area”.

2.12 The empirical evidence cited in Carter et al. (2001) on the actual importance of barriers to finance for women is conflicting, however, although there is a general feeling that women may be disadvantaged in their ability to raise start up finance (Schwartz, 1976; Carter and Cannon, 1992; Johnson and Storey, 1993; Koper, 1993; Van Auken et al, 1993; Carter and Rosa, 1998). Carter and Rosa (1998), for example, based on survey of 600 firms, equally split by gender, found that there are: “quantifiable gender differences in certain areas of business financing, although intra-sectoral similarities demonstrate that gender is only one of a number of variables that affect the financing process.”

2.13 Four specific themes emerge from the literature identified by Carter et al (2001) which might provide an explanation for these difficulties:

2 Marlow, S. and Watson, E. (2003) Safety Net or Ties that Bind? Welfare Benefits, Gender and Self Employment, paper presented to the 26th Institute of Small Business Affairs National Small Firms Conference.3 Warren-Smith, I. and Jackson, C. (2004) Women Creating Wealth Through Rural Enterprise, International Journal of Entrepreneurial Behaviour & Research, vol. 10, no. 10, pp. 369-383.4 SBS (2004) Annual Small Business Survey: Executive Summary, SBS: Sheffield.

10

1. Collateral – the financial guarantees required for external financing may be beyond the scope of most women’s personal assets and credit track record (Carter et al 2001 – they refer to Hisrich and Brush, 1986; Riding and Swift, 1990). Verheul and Thurik (2001), for example, focussed on 2,000 entrepreneurial start-ups in 1994 in Holland (25 per cent of which were women) and concluded that women had less capital when starting the business but that there was no difference in the type of capital and that “the proportion of equity and debt capital (bank loans) in the businesses of women entrepreneurs is the same as in those of their male counterparts.”

2. Networks – “finance for the ongoing business may be less readily available for women-owned firms than it is for men-owned enterprises, largely due to women’s inability to penetrate informal financial networks (Olm et al, 1988; Aldrich, 1989; Greene et al, 2001).” (Carter et al 2001)

3. Discrimination – “female entrepreneurs’ relationships with bankers may suffer because of sexual stereotyping and discrimination (Hisrich and Brush, 1986; Buttner and Rosen, 1988, 1989)” (Carter et al 2001). Ennew and McKechnie (1998), for example, suggest that, “discrimination occurs amongst lenders at a more unconscious level”.

4. Financing preferences – it may be that the financial preferences of women and men owner-managers are different. However a recent study, drawing upon the results from a 400-firm telephone survey by Barclays Bank, found that "Gender appears to make little difference to the choice of finance source utilised – most settle for personal savings, but there is little difference across each source" (Irwin and Scott 2006).

2.4 Key Themes 2.14 The review by Carter et al. (2001) and other studies have emphasised the

potential importance for women’s start-up rates and business success of access to finance. The quantitative evidence to date, however, on the real impact of financial barriers to start-up is relatively limited and somewhat conflicting. Some key themes do emerge, however, from the literature and these underpin our empirical analysis in subsequent chapters. These are:

Access to finance is a key factor in shaping start-up rates and business development but needs to be seen in the context of other factors which may also be shaping the development of the business.

Both demand and supply-side factors may be influencing perceived financial constraints reflecting individuals’ perceptions and preferences and the types and amounts of finance being sought.

Issues of collateral and background may be important in shaping the willingness of banks or other organisations to lend to individuals or firms, and the preferences of individuals.

Particularly in terms of business start-up there is a strong process dynamic in the financing process, with availability of finance conditioning the probability of business start-up.

2.15 These themes are reflected in our methodological approach which is the subject of Chapter 3.

Chapter 3 - Access to Finance and Business Start-up

3.1 Introduction

3.1 In this chapter we consider whether there are gender differences in the perceived or actual financial constraints on individuals, and the impact of these differences on business start-up. Our analysis is based on evidence from two large-scale household surveys intended to represent the overall population of households in the UK: the 2004 Global Entrepreneurship Monitor (GEM) dataset which includes 24,006 respondents and the 2003 Household Survey of Entrepreneurship (HSE) which includes 10,002 respondents in England5. As outlined in Chapter 2, our approach is in two stages: first we consider the factors which determine perceived or actual financial barriers to business start-up, and secondly the impact of this on business start-up itself.

3.2 Comparison between the GEM and HSE results is complicated as both surveys use slightly different questions relating to both potential financial barriers to business start-up as well as the notion of business start-up itself. In the GEM 2004 study, we consider individuals’ perceived financial barriers to business start-up using the responses to a question posed to all respondents: ‘Excluding money from family and friends, would a lack of external funding prevent you from starting up a business?’. This provides a straightforward indication of the perceived lack of business finance and potential psychological or motivational barriers that this might induce to business start-up. In the population of respondents, women were significantly more likely to perceive such financial barriers to business start-up than men (Table 3.1).

3.3 In terms of individuals’ participation in business start-up, GEM 2004 provides three indicators. These are individuals’ responses to:

1. Start-up: Are you, alone or with others, currently trying to start a new business, including any type of self-employment or selling any goods or services to others?

2. Running Business: You are, alone or with others, currently the owner of a business you help manage; or you are self-employed or selling any goods or services to others

3. Expected Start-up: You are, alone or with others, expecting to start a new business, including any type of self-employment, within the next three years

3.4 As with individuals’ perceived financial barriers to business start-up, the most consistent differences of responses here relate to gender, with the proportion of women engaged in each type of business start-up activity significantly lower than that of men (Table 3.1).

5 See the Global Entrepreneurship Monitor UK (2004) report and Small Business Service (2004) for detailed descriptions of the two surveys.

Table 3.1: Perceived Shortages of Finance and Business Start-up Indicators by Gender

Women Men per cent per cent

Lack of funding preventing starting up a business 64.1 57.3

Involved in business start-up 3.1 5.9Current business owner 6.7 15.8

Expected business start-up 7.2 11.8

Notes: Figures in bold are significantly different at the 5 per cent level. See the data annex for variable definitions etc. Source: GEM 2004

3.5 The HSE adopts a different approach to both the access to finance for business start-up and the issue of business start-up itself. Here, respondents are divided into three groups according to their level of involvement in entrepreneurial activity with each group being asked different types of questions on the degree of access to external finance and whether they have encountered financial constraints. In the HSE the groups are defined as follows: Thinkers are ‘those who are thinking about becoming entrepreneurs’; Doers are ‘those who are already entrepreneurs through running their own business or by being self-employed’; and Avoiders are ‘those who are neither currently engaged in entrepreneurial activity nor thinking about doing so’. Reflecting the overall level of involvement in entrepreneurial activity from the GEM survey (Table 3.1), 76 per cent of the HSE sample was classified as Avoiders, 11 per cent as Thinkers and 13 per cent as Doers.

3.6 In terms of access to finance the questions asked for each group of HSE respondents were as follows:

For Thinkers: ‘And have you tried to obtain any finance for this new business in the past 12 months?’ and ‘did you have any difficulties in obtaining this finance from the first source you approached?’

For Doers: ‘In the past year have you tried to obtain finance for your business?’ and ‘Did you have any difficulties in obtaining this finance?’

For Avoiders: ‘And which two would you say are the biggest barriers to you starting a business or becoming self-employed?’6.

3.7 These questions reflect the access to external finance, as well as the degree of financial constraint Doers and Thinkers experience, along with the impact this may have on business start-up – reflected in individuals’ status as either a Doer or Thinker. This is not true for the Avoiders, however, where the different structure of the HSE survey questions makes comparison with Thinkers and Doers more difficult. In our analysis we therefore focus

6 For Thinkers these were survey questions 13 and 16, for Doers questions 40 and 44 and for Avoiders question 50.

14

primarily on the combined group of Thinkers and Doers and question whether difficulties in accessing finance influence the probability of becoming a Doer rather than a Thinker.

Table 3.2: Access to external funding – Thinkers and Doers by gender(Percentage of all Thinkers and Doers)

Thinkers DoersMen Women Men Women

Did try to obtain external finance ( per cent)? 6.8 4.3 15.6 13.5

Difficulties in obtaining Finance (per cent)? 2.5 1.4 3.8 2.1

Source: HSE 2003

3.8 Table 3.2 summarises the proportions of Thinkers and Doers seeking external finance and experiencing difficulties in accessing this finance (both proportions are expressed relative to the population of all Thinkers or Doers). The proportion of Thinkers trying to get access to external finance is quite small (4-6 per cent) compared to that among the Doers (13-15 per cent). Interestingly, however, a smaller overall proportion of the population of women Thinkers and Doers reported difficulties in obtaining finance than that among men although this may reflect the smaller proportion of women who actually sought finance as well as any difficulties they encountered.

3.9 The rest of this chapter is organised as follows:

Section 3.2 summarises the evidence from the GEM 2004 survey reflecting both individuals’ perceived financial barriers to start-up and the impact of these perceived financial barriers on start-up itself.

Section 3.3 then focuses on the HSE survey, and the impact of gender on access to external finance first and then on the status of individuals as either Doers or Thinkers

Section 3.4 summarises the main results and gives an indication of potential caveats and policy conclusions.

Modelling results underlying the key points made in the main text are reported in Annex 1.

3.2 Evidence from GEM 2004

3.2.1 Descriptive Analysis3.10 As indicated earlier the GEM data provides an indication of the proportion of

the UK adult population that perceive financial barriers to business start-up and the impact of this on different aspects of start-up behaviour. In this section we examine the impact of gender on both the perception of financial barriers and start-up. As a prelude to the multivariate analysis, however, it is beneficial

15

to have an understanding of the basic characteristics of the underlying GEM data. Table 3.3 summarises a number of the key characteristics of the sample of GEM respondents, which are representative of the whole working age population, by gender:

There is no significant distinction between the regional composition of the sample of GEM respondents between men and women. Four regions (South East, West Midlands, London and the Eastern region) account for around two-thirds of the national sample.

A higher proportion of men respondents have degrees (36.7 per cent), while ‘A’ levels and GCSEs are more common among women. Lower levels of qualification are equally common.

Women respondents were more likely to be in lower quartiles of the national distribution of household income.

Men respondents were more likely to be working full-time and to be either self-employed or an employer than women.

Finally, men were more likely than women to have received enterprise training and participated in work experience programmes.

The suggestion is that in the working age population men responding to the GEM survey were more likely to be highly qualified; more likely to have a stronger financial profile (i.e. are in the upper quartiles of the distribution of household incomes); and more likely to have benefited from relevant working and training experiences than women respondents. Each of these factors is likely to have a positive effect on business start-up aside from any underlying gender differences.

16

Table 3.3: Sample Characteristics: By Gender

Women Men per cent per cent

A. Home Region South West 9.3 9.5South East 25.0 25.9London 21.7 21.7Eastern 11.4 11.3Wales 1.7 1.6West Midlands 11.1 10.4East Midlands 2.3 2.4Yorkshire and the Humber 3.3 3.2North West 5.9 6.0North East 2.6 2.4Scotland 5.5 5.2Northern Ireland 0.2 0.2

B. Highest Educational LevelDegree or higher 31.6 36.7‘A’ Levels 22.4 20.4GCSE or equivalent 25.7 22.8Other vocational quals. 9.1 8.8No formal qualifications 11.2 11.3

C. National Household Income DistributionLower quartile 20.5 14.02nd quartile 25.3 24.83rd quartile 24.6 28.14th quartile 29.5 33.1

D. Age Age in years 40.5 40.3

E. Working StatusFull-time (30 or more hours) 46.4 78.6Part-time (8-29 hours) 25.3 6.2Not working (8 or less hours) 28.3 15.2

F. Employment StatusEmployee 90.9 80.7Self-employed 6.0 12.8Employer 3.1 6.5

G. Enterprise Training and Work ExperienceEnterprise training at school 11.4 13.7Enterprise training at college/university 15.8 21.4Work experience at school 34.1 35.0Work experience at college/university 12.0 14.0

Notes: Figures in bold are significantly different at the 5 per cent level. See the data annex for variable definitions etc. Source: GEM 2004

17

3.2.2 Perceived financial barriers to business start-up 3.11 Our main aim in this section is to see whether, when controlling for

individuals’ background characteristics, gender influences the perceived financial barriers to business start-up. Our approach is based on a series of probit models of the probability of perceiving financial barriers to business start-up (Table A1.1). Significant coefficients are in bold type in the table and two alternative formulations of the model are presented dropping the insignificant variable ‘Enterprise Training at school’ in the second model.

3.12 Our results here are straightforward, consistent and statistically very strong - even adjusting for a range of background characteristics. Being a woman increases the probability that an individual will perceive financial barriers to business start-up by 7.5 percentage points.

3.13 Our analysis also suggests a number of other factors which prove important in determining the probability that an individual will perceive financial barriers to business start-up. The most consistent effects were:

Respondents in the Eastern region were 3-7 pp less likely to perceive financial barriers to business start-up. No other regional differences were statistically significant.

Individuals with a degree were more likely to perceive financial barriers to business start-up than those with lower level qualifications (by 3-8 pp).

Those in higher income households were less likely to perceive financial barriers to business start-up.

Older individuals were less likely to perceive financial barriers to business start-up.

Those working part-time or not working were less likely to perceive financial barriers to business start-up than the reference group (i.e. those in full-time employment).

In general terms therefore the GEM 2004 data provide considerable support for the notion that women may perceive stronger financial barriers to business start-up than men.

3.2.3 Business Start-up3.14 The aim of this section is to investigate the potential role of perceived

financial barriers to business start-up on business start-up itself. If such perceived financial barriers to business start-up are important in influencing business start-up, then the fact that perceived financial barriers to business start-up are concentrated among women may be contributing to lower start-up rates among women. If perceived financial barriers to business start-up are not a factor in shaping business start-up then differential access to finance is likely to be less important in explaining lower start-up rates among women (e.g. Table 3.1).

18

3.15 Some significant statistical and econometric issues arise here and these are outlined in Section A1.2. Our preferred models of gender effects on business start-up activity however, are single equation Probit models (Table A1.3) . These suggest three main conclusions in relation to gender and perceived financial barriers.

Women are less likely to be involved in start-up activity, running a business and expected start-up activity than men. Start-up rates for women are reduced 1.7 pp lower than that for men, with expected start-up rates reduced by 3.5 pp (Table 3.6).

Perceived financial barriers to business start-up have a negative effect on business start-up (1.3 pp) and the probability of expected start-up (3.8 pp). The negative finance effect on start-up is around the same size as the direct gender effect.

The probability that an individual is running a business is not significantly influenced by perceived financial barriers to business start-up (Table 3.6).

3.16 Other factors also prove important in increasing business start-up rates with strong and consistently positive effects from: having a background of self employment or as an employer and experiencing enterprise training at college or university. No factors have a consistent negative effect other than being a woman.

3.17 In general therefore the GEM data suggests that gender has both direct and indirect effects on business start-up rates, with the negative effects operating through perceived financial barriers to start-up. Women are more likely to perceive financial barriers to start-up and these are likely to reduce start-up rates. In addition, there is a direct gender effect on start-up rates even allowing for education, location and personal characteristics.

3.2.4 Reasons for perceived financial barriers

3.18 GEM 2004 also provides some information on individuals’ own reasoning for why they did not obtain finance and their justification for this lack of success. This is interesting as it may inform our understanding of why perceived financial barriers are greater among potential women entrepreneurs. Sample sizes here are relatively small, however, as they relate only to a sub-group of those involved in enterprise activity in the survey. It is not possible therefore to model the effects of gender on these perceptions of success (or to identify statistically significant differences) but descriptive data is given in Table 3.4.

Table 3.4: Percentage indicating the reasons for their lack of success in obtaining finance

Women Men All per cent per cent per cent

Not investor ready 15.6 17.1 16.6Nature of the business 32.0 33.8 33.3Inadequacies in the 15.7 16.2 16.1

19

business planBusiness too small 23.9 31.7 29.1Fear of debt 21.2 24.3 23.3Unwillingness to share ownership

17.8 14.0 15.2

Cost of finance too high 34.6 27.0 29.5Weak management team 8.8 8.7 8.7

3.19 There is a broad similarity between the reasons given by women and men for their lack of success in gaining funding with the nature of the business, business size and the cost of finance predominating. Some more subtle differences are evident however, with women emphasising the cost of finance and males suggesting that business size was a more important factor in their failure to gain financial support.

3.2.5 Summary of key points 3.20 The GEM 2004 data provides a comprehensive database within which the

impact of perceived financial barriers to business start-up can be assessed. Our key finding is that being a woman impacts on start-up rates both indirectly and directly.

3.21 First, being a woman increases the probability than an individual will perceive financial barriers to business start-up by 7.5 pp with this, in turn, reducing start-up rates by 1.7-3.8 pp. Being female also has an additional direct effect on each of our start-up indicators (Table A1.3). These results derived from the models allow us to decompose the difference between men’s and women’s start-up rates into a direct ‘gender’ effect, an indirect effect due to the effect of gender on perceived financial barriers and a ‘residual’ or unexplained effect. The relative sizes of these effects then provide an indication of the importance of the overall finance effect.

3.22 Table 3.8 summarises the results for the three start-up indicators considered earlier. In the case of start-up, for example, the start-up rate for males is 2.8 percentage points higher than that for women. Of this difference, the models suggest that 1.7 percentage points can be explained by the direct gender effect with a further 1.0 percentage point being explained by the indirect gender effect due to perceived financial barriers. Here 0.1 percentage point remains of the difference in start-up rates between genders remains unexplained. For the other two business activity indicators the impact of perceived financial barriers on the difference in start-up rates suggested by the models is somewhat smaller (Table 3.5).

Table 3.5: Decomposition of differences in start-up rates

Start-upRunning Business

Expected Start-up

Start-up rates ( per cent)Women 3.1 6.7 7.2Men 5.9 15.8 11.8Difference -2.8 -9.1 -4.6

20

Contribution ( per cent)Direct gender effect -1.7 -1.0 -3.5Indirect gender effect (via finance) -1.0 -0.1 -0.3Other factors (residual) -0.1 -8.0 -0.8

Source: GEM 2004, for derivation see text

3.3 Evidence from the HSE

3.23 In this section we turn to the evidence from the Household Survey of Entrepreneurship 2003 (HSE). Here our analysis focuses on whether the probability of engaging in entrepreneurial activity is influenced by individuals’ access to finance, and whether the access to finance itself is influenced by gender. Detailed modelling results are presented in Section A1.3.

3.24 We consider two separate models reflecting different aspects of individual’s access to finance. First, reflecting the general thrust of the GEM analysis presented earlier we consider whether women are more likely to experience difficulties in obtaining finance for business start-up and whether this then influences start-up rates. Unlike the GEM analysis considered earlier, however, the HSE analysis is limited to a smaller sub-group of the population – those defined as Thinkers and Doers, and there is some difference in the question being asked7. In the GEM analysis the question relates to individuals’ general perception of financial barriers for business start-up. In the HSE the question is more specific and relates to whether individuals have actually experienced difficulty in obtaining finance for their business start-up in the last year. This contrast is important as it suggests the difference between perceived financial barriers to business start-up (GEM) and individuals’ actual experience (HSE).

3.25 A second issue considered here is whether gender actually influences individuals’ propensity to seek external finance. Women, for example, may be more reluctant to seek external financial support for their business start-up, and this may in turn influence business start-up rates. Here the question is whether this self-selection mechanism has a gender dimension and whether this then has a significant effect on start-up rates.

3.3.1 Descriptive Analysis 3.26 Before looking at the econometric results it is useful to draw some contrasts

between the characteristics of the HSE sample of Thinkers and Doers by gender (Table 3.6):

Overall, male Thinkers appear to be more highly qualified than women Thinkers. For both sexes, the most common qualifications are either a degree or GCSEs. This is also true for Doers, however, among the Doers who are men there is a large proportion that do not hold a formal qualification.

7 In some earlier analyses the Thinkers group here has been subdivided into serious and ordinary thinkers. Here these groups are treated together.

21

A large proportion of Thinkers are in the 25-44 age bracket; also this distribution does not appear to differ substantially between genders. Doers appear to be concentrated in the 25-64 age brackets and it does not appear there is any difference according to gender.

As for the employment status, a large proportion of men Thinkers appear to be currently in employment while women Thinkers are not. As for the Doers, the picture is more ambiguous with a relatively large proportion of both men and women Doers appear classified as not-employed.

Both men and women Thinkers appear to have some previous experience with self-employment. The proportion of men Doers with some previous experience in self-employment is higher than for women Doers; however, for both sexes, the proportion of Doers without any previous experience is quite high.

The proportion of Thinkers (belonging to both genders) with a positive attitude towards self-employment is large. The same applies to the Doers with the fraction of men Doers being quite substantial.

Thinkers appear to be mostly located in the Northern and Southern regions. The proportion of Thinkers located in the Midlands is small. However, for all the three areas, we can see that the fraction of men Thinkers is usually larger than that of women Thinkers. The same pattern applies to the Doers. Most Doers (men and women) are located in the North and the South of the country. Also, the proportion of men Doers is quite large.

A high proportion of Thinkers and Doers own their own house with higher home ownership proportions among male Thinkers and Doers in the sample.

22

Table 3.6: Basic Characteristics of Thinkers and Doers by Gender

Thinkers Doers  Men Women Men Women  per cent per cent per cent per centEducational Level      Degree 1.75 1.34 2.61 1.55A level 0.83 0.64 0.96 0.62GCSE 1.07 0.78 1.43 0.96Other 0.6 0.25 1.11 0.55None 0.88 0.46 2.11 0.66Age Group        16-18 years 0.26 0.16 0.09 0.0219-24 years 0.61 0.3 0.24 0.1125-34 years 1.44 1.1 1.25 0.5835-44 years 1.42 1.09 2.39 1.3945-54 years 0.89 0.57 2.21 1.2755-64 years 0.51 0.25 2.04 0.97Current Employment Status  Not Employed 1.76 2 6.61 3.74Employed 3.37 1.43 1.61 0.6Previous experience  Yes 3.26 2.48 3.72 2.21No 1.87 0.99 4.5 2.13Location        North East 0.95 0.63 1.43 0.82Yorks & Humber 0.22 0.2 0.42 0.16East Midlands 0.33 0.15 0.46 0.2East 0.15 0.1 0.32 0.2London 0.72 0.45 0.86 0.44South East 0.74 0.61 1.33 0.69South West 0.41 0.21 0.72 0.36West Midlands 0.33 0.26 0.45 0.26North West 1.28 0.86 2.23 1.21Total 5.13 3.47 8.22 4.34Home ownership    Yes 3.68 2.40 7.14 3.90No 1.45 1.07 1.08 0.44Total 5.13 3.47 8.22 4.34

3.27 These differences between the characteristics of men and women Thinkers and Doers emphasise the importance of a multivariate approach to modelling potential gender effects on access to finance and business start-up and this is the focus of the next two sections.

23

3.3.2 Start-up and the Financial Sourcing Decision3.28 Here we consider whether gender is important in individuals’ decisions

whether to seek external funding for business start-up and whether this then affects the start-up decision. Two equations are estimated: the first equation models the self-selection mechanism where we try to understand which factors affect individuals’ decisions about whether to seek external finance; the second equation models the start-up decision and is estimated only on the sample that is selected through the self-selection mechanism (see Table A1.4).

3.29 Generally speaking, the probability of seeking external finance decreases if the individual is a woman. This implies that women tend to self-select themselves out of seeking external finance. Ethnic background also makes a significant difference to the probability of seeking finance with members of the white ethnic group more likely to seek external finance for their start-up activity. These differences in the probability of seeking external finance are important for business start-up, as there is a significant link between the decision to seek external finance and the start-up decision (suggested by the significant correlation coefficient).

3.30 For the subset of individuals that decide to seek external finance the second stage of the model highlights the factors which influence the probability of becoming a Doer rather than a thinker. (Percentage effects are suggested here by the marginal effects in Table A1.4). Prior experience and educational attainment both increase the probability of becoming a doer, with those who are non-employed less likely to become a Doer by around 3.7 pp. A positive attitude towards entrepreneurship also increases the probability of becoming a Doer by around 2.9 pp. In these models, the regional variables are not significant showing that there is no locational effect at work in either the self-selection mechanism or the start-up decision. This is not surprising in the light of the descriptive analysis.

3.3.3 Start-up and Difficulty Obtaining Finance3.31 Now we consider whether the probability of becoming a Doer is affected by

the respondent’s gender and conditioned by the applicant’s probability of experiencing difficulty in obtaining finance (that in turn is affected by an additional set of variables including gender). This involves the estimation of a two-stage model as before: in the first stage we model the respondent’s probability of experiencing difficulty obtaining finance and test whether this is affected by gender, ethnic background, location, education and whether or not (s)he own a house (that can be used as collateral); in the second stage we model the probability of becoming self-employed as a function of gender, the respondent’s foregone wage income (proxied by whether the respondent has a degree8), attitudes towards entrepreneurship, location and previous experience. Detailed results are presented in Table A1.5.

3.32 Neither gender or ethnicity have any significant impact on the probability of facing difficulties obtaining external finance although their interaction (i.e. Women*White) is marginally significant. This is perhaps not surprising, however, given the fact that the proportion of respondents claiming to have

8 The assumption is that the individuals with a degree have a potential for a high income and therefore the opportunity cost of becoming self-employed is higher.

24

been financially constrained is very small. Interestingly, the two significant variables in the first stage equation are the regional variables and the dummy variable on whether or not the respondent owns a house. This last result is as expected: financial constraints are exacerbated by the lack of collateral. In the second stage, the probability of being a Doer is not affected significantly by gender, but is influenced by previous experience and by the individual’s attitude towards entrepreneurship. Marginal effects are generally not significant showing that from this sample we cannot draw conclusions regarding the whole population.

3.3.4 Key Points3.33 In this part of the report we have estimated two models of self-employment

choice and financial constraints using the Household Survey of Entrepreneurship 2003. Our key findings are:

A self-selection mechanism is at work where women decide not to go for external finance (as they may expect to encounter substantial financial constraints) and so implicitly decide not to be self-employed.

Being a woman does not, however, increase the probability that individuals will experience difficulties in obtaining start-up finance. On the contrary these are compounded by the lack of collateral (i.e. not being a home-owner) and by location.

3.4 Concluding Remarks

3.34 Data from the GEM survey and the HSE provide largely complementary perspectives on access to finance by gender and its effect on business start-up. Data from the GEM survey suggest that in the general population women are more likely to perceive finance barriers for business start-up. The HSE data suggests this leads to a decision on the part of women not to seek external finance for business start-up. In both surveys this is also linked to lower start-up rates among women.

3.35 More surprising perhaps is that the HSE data suggest that among those individuals who do seek external finance for start-up, the likelihood of obtaining finance is no different for men and women. In other words, while women in the general population are more likely to perceive greater financial barriers to start-up, and tend to be less likely to seek external finance, the HSE suggests no evidence that such gender barriers to obtaining finance actually exist. Instead, difficulties in actually obtaining finance are much more likely to be linked to a lack of collateral or individual’s location.

3.36 In short therefore, both the HSE and the GEM data suggest that being a woman does have a negative effect on both access to finance and hence business start-up. The finance effect seems more strongly influenced by the demand side than the supply side, however: women are less likely to seek external finance but those who do are no more likely than men to face financial constraints.

25

Chapter 4 - Access to Finance by Established Firms

4.1 Introduction

4.1 In this Chapter we investigate the significance of gender – i.e. women-led firms9– in shaping the ability of existing firms to access external finance using data from the Annual Small Business Survey (2003 and 2004) and the UKSMEF 200410. Our general approach here recognises the fact that when seeking to link women’s leadership to the ability to raise finance, one has to make one or two (not necessarily mutually exclusive) assumptions. Firstly, that any variation in the ability of each gender to raise finance is purely a “supply side” phenomenon. That is, that each group is equally likely to identify a particular source of finance, and see it as desirable, but that the suppliers of that finance are either deliberately or subconsciously more likely to favour a particular gender. The second less restrictive assumption is that there are also “demand side” differences across genders, that is that certain groups are more likely to identify a particular type and source of finance, and subsequently the supply side differences either do or do not apply. This type of effect was evident in Chapter 4 where the evidence from the HSE suggested that women were less likely than men to seek external finance for business start-up.

4.2 With a simple reduced form quantity equation based on whether a business has or has not raised a particular type of finance one faces the well known identification problem, of being unable to distinguish between supply side and demand side effects. However, both the ASBS and UKSMEF ask more qualitative questions from which one can make certain inferences or judgements about the relative supply and demand side effects. Of course, in all of this analysis it is necessary to control for the other factors that will impact on the likelihood of a given business to raise a particular type of finance, and these will be discussed in due course.

4.3 In the ASBS the supply side will be evaluated using the responses to a question: ‘Did you have any difficulties in obtaining this finance from the first source you approached?’ with four possible responses being identified11. The demand side will be evaluated by analysing the answers to two questions, firstly relating to whether the firm had sought to raise finance, and secondly what for, and what type. The latter would include factors such as: ‘Over the next two to three years, do you aim to grow your business?’ And, ‘How much time in a typical week would you say your business spends on paperwork relating to complying with government regulations and taxes (hours)?’ In

9 Defined here as firms in which more than 50 per cent of the directors are women.10 There were two years of the ASBS, and rather than pool the data, due to the slightly different structures of the surveys, the models were estimated separately for the two years. This is potentially revealing as there was a boost to the ethnicity sampling in 2003 but not in 2004.11 These were: Yes, was unable to obtain any finance; Yes, obtained some but not all of the finance required; Yes, obtained all the finance required but with some problem; No, had no difficulties in obtaining finance.

26

2003, of the 8693 respondents to the ASBS, 2330 had sought finance at least once; for ASBS 2004, from a sample of 7505 firms, 1627 had sought finance in the past year. Table 4.1 below analyses the pattern of responses by firms with different leadership groups. This suggests that among women-led firms the proportion of firms obtaining all of the finance they were seeking was marginally lower than that in other groups. Conversely the proportion of women-led firms unable to obtain any finance (12.2 per cent) was also relatively high compared to other sample groups. This suggests the potential for significant gender effects in firms’ ability to access finance.

Table 4.1: 2003 analysis of ability of firms to raise finance by gender and ethnicity (ASBS 2003)

Yes, was unable to

obtain any finance

Yes, obtained some but not

all of the finance

required

Yes, obtained all the finance

required but with some

problem

No, had no difficulties in

obtaining finance

Total

per cent per cent per cent per cent Number

Leadership ProfileTotally Women-led 12.2 10.2 6.1 68.5

197

Equal numbers of directors 9.4 3.5 6.2 78.0

481

Women minority among directors 9.8 5.3 8.3 73.7

338

All directors men 14.9 4.1 7.1 73.9 1283Total 11.6 4.7 7.0 74.2 2299

Note: The definitions of women-led business are taken from the database variable “smegen” .The dependent variable that is used for the analysis of whether firms reported difficulty in obtaining finance is obtained from ‘Did you have any difficulties in obtaining this finance from the first source you approached?’ Some firms did not provide a detailed response to this question and so the table total (2299) differs from the number of firms which actually sought finance in the survey (2330). Source: ASBS 2003

4.4 Our analysis of the UKSMEF database is based on existing firms attempting to obtain new finance in the last 3 years of trading, rather than at the start up phase12. As in the analysis of the ASBS, our analysis is restricted by the fact that we only observe the responses of those individuals who have been successful in starting a business. Our approach mirrors closely that adopted in the ASBS looking separately at the demand and supply sides. Essentially, the “supply side” will be evaluated using questions in relation to “discouragement” and “denial” (whether an application for finance was denied outright)13. Both variables provide a straightforward indication of barriers and difficulties in obtaining external finance. In addition, we used a constructed variable called “barrier” reflecting whether the business seeking external

12 An alternative approach would have been to base the analysis on businesses in the start-up phase. The number of start-ups in the UKSMEF sample is 149 of a total sample of 2500. This is too small a sample on which to base robust inferences. 13 A number of different measures of discouragement are available in the UKSMEF and are explored below. See the data description in Annex for details.

27

finance experienced either discouragement or denial14. The “demand side” effect will be evaluated by analysing variables which reflect whether businesses actually sought external finance as well as the type of external finance sought (in particular overdrafts, loans, asset-based finance and equity).

4.5 Table 4.2 illustrates SMEs’ external financial needs and restrictions, with a specific distinction for start-ups (business trading for less than 2 years) in terms of gender. Overall, broadly similar proportions of men and women-led businesses were seeking external funding, although some unexpected differences emerge in terms of the barriers to finance. Notably, start-up businesses being run by males seem more likely to face discouragement in applying for funds than those run by women although sample sizes here are relatively small and differences are therefore statistically insignificant.

Table 4.2: Financial Needs and Barriers to Finance (per cent of the population)

      Men Women  per cent per centSME needing Finance 46.2 43.7

Start-ups  73.0 64.7Discouraged from applying for finance1 4.1 2.0

 Start-ups 12.5 3.0Denied application for finance1 5.7 3.2

 Start-ups 3.5 1.0

Notes: 1- Applications for overdraft, term loan, asset-based finance or equity finance. Figures in bold are significantly different at the 5 per cent level. See Annex 1 for variable definitions.Source: UKSMEF 2004

4.2 Evidence from the ASBS

4.2.1 Supply Side Models 4.6 Supply side models were estimated in various forms (see Tables A2.1 and

A2.2) and perform well in general terms. In terms of gender, however, the results are conflicting. Women-led businesses were less likely to report financing difficulties based on the 2003 data, although this effect was only significant in one of the three models, but more likely to report financing difficulties based on two of the three models estimated on the 2004 data. In the 2004 models, women-led firms are approximately 2 pp more likely to have difficulty raising finance, and also 2 pp more likely to find it impossible to raise the finance that they are seeking. To set these figures in context, the corresponding figures for Owners/MDs who are over 60 however are 6.5 pp and 1 pp respectively

14 This we compare with the responses to the question: ‘On a scale of 1-10 (where 1 is no problem and 10 is critical problems) how would you rate the severity of the problems faced by your business in the following areas: Finance (by this we mean obtaining sufficient finance and the cost of finance)’.

28

4.7 The more important variables in explaining the difficulties are firm size, in that the larger firms in the sample report fewer problems in raising finance. The quintile of deprivation is included, though this variable is very sparse, as such it is included for all five categories, with zeros inserted for all missing values to prevent reducing the sample size. This limits what inferences one can make from these variables.

4.8 Both firm size and VAT registration (which to an extent may be capturing the same thing) increase the likelihood of obtaining finance, as does (to a lesser degree) firm age, while being an exporter and being in the second most deprived quintile of the deprivation distribution significantly reduces the likelihood of obtaining finance. VAT registration data were not collected for the 2003 data, so a VAT threshold dummy is used, which may not be capturing the same thing.

4.9 Other variants of these baseline models were estimated reflecting various dimensions of the supply side. These are reported in detail in Annex 2 with a summary of the key points outlined here:

(a) Amount Requested4.10 Tables A2.3 and A2.4 report results dividing the sample by the amount of

finance sought. In general, the gender effect appears stronger in the 2004 sample. This analysis is potentially important because it may be, for example, that certain parts of the market for finance work better than others, with responsibility for dealing with small requests devolved to more junior (or more local) decision makers. Based on the 2004 analysis, women-led firms are significantly less likely to obtain finance in the range of £10,000-£100,000.

4.11 Exporting firms are again less likely to obtain finance across all amounts of finance requested– perhaps due to the higher perceived risk of exporting activity. Where firms request the smaller amounts of finance, location in terms of the quintile of deprivation also appear to hamper the ability to raise finance. This may be because decisions regarding smaller amounts of finance are devolved to more local decision makers who are more likely to recognise areas as being deprived.

(b) Why finance was being sought4.12 Tables A2.5 and A2.6 shows a breakdown of why firms sought finance. It is

clear from this analysis that obtaining finance for “working capital” is problematic across all businesses, and that there are strong age and business profile effects here. Being registered for VAT for example makes it more likely to receive finance. Based on the 2004 analysis, women-led businesses are less likely than average to be given finance for working capital. Interestingly, these data are skewed to either end of the distribution, with both “impossible to obtain finance” and “no problem” far more prevalent than either of the middle groups. This suggests that many people seeking working capital are not seen by providers of finance as a good bet. The motor trade is an exception here where firms find it easier to raise working capital. These effects are not as strong in the 2003 analysis.

4.13 Across the other types of finance gender appears unimportant, while exporting firms appear far less likely to obtain any type of finance.

29

(c) By type of finance sought.4.14 Tables A2.7 and A2.8 show the breakdown by type of finance sought. Several

categories from the original questions had to be amalgamated here to obtain sufficient sample size. For the 2004 sample, the difference between the genders appears insignificant. These results, however, highlight the importance of VAT registration across all types of finance.

(d) By type of organisation4.15 Tables A2.9 and A2.10 examine the differences between sole proprietors,

partnerships and companies. Sole proprietors in the more deprived areas are less likely to obtain finance, while interestingly women-led companies are less likely than other companies to obtain finance. This difference is, however, significant only at the 15 per cent level. Again VAT registration is important for all types of firms, though firm size is not important for sole proprietors.

4.16 Another approach to the ASBS data is to model the supply side separately for male and women-led businesses and these models are reported in Tables A2.11 and A2.12, again with significant variables in bold. While approximately 12 per cent of women-led firms that applied for finance in 2003 found it impossible to raise money, the corresponding figure for firms with all male directors was just under 15 per cent (Table 4.1). More detailed econometric analysis, standardising for other characteristics, however, suggests that women are disadvantaged in finance markets, particularly where they are also ethnic minority businesses.

4.17 For the 2003 sample, ethnic minority businesses that are women-led report significantly greater problems in raising finance, while the same is true for the 2004 sample. The results for the women-led firms highlight many of the results alluded to. Ethnic minority women-led businesses are less likely to obtain finance, but the same is not true for male businesses. Partnerships of women are more likely to be successful in raising finance than women-led companies, the strongest result for type of firm through this analysis.

4.18 Location is also more important for women than men. The regional dummies in table 4.5 are collectively much more important for the women-led sample than the other sample, and women in Wales and Yorkshire and Humberside are significantly more likely to report problems. It should be stressed however that these results are based on a relatively small sample of firms reporting problems raising finance. For the 2003 sample, among the men-led businesses, firms with some women directors are more likely than the others to encounter problems in raising finance.

4.19 These results also highlight some interesting regional effects. Women-led businesses appear less likely to obtain finance if they are in the South-West of England, and also if they have a degree. They are also more likely to be able to raise finance if they are partnerships or in the least deprived areas, although this effect, while highly significant is very small.

4.20 To summarise, the results of our supply side analysis do suggest the validity of concerns about access to finance of women-led companies. What shows up more strongly is that women in ethnic minority groups are far less likely to

30

raise business finance, even when they have an existing business. This effect is multiplicative; women are far less likely to obtain finance if they are also from ethnic minorities. This also shows that among the women-led group, partnerships are less likely to face problems than companies or sole traders. It is easy to imagine that this is due to the type of sectors inhabited by partnerships, though sectoral differences are captured by dummy variables.

4.2.2 The Demand for finance 4.21 The remainder of this section turns to the differences in the demand for

finance. This utilises the information with respect to those firms that have sought finance, as to whether there are differences across gender in seeking finance, rather than in being offered it. The purpose of this is to address the problem that it may be the case that certain groups simply perceive it not to be worthwhile even applying for finance. This reflects the suggestion from the HSE that individual women considering business start-up are less likely to seek external finance than their counterparts who are men.

4.22 The demand side analysis presented in this section is similar in structure to that presented above, focussing on the types of finance sought, where from and how often, rather than the difficulties faced. Again some categories have been amalgamated from the original survey in order to build up large enough samples with sufficient within sample variation. Other variables are included in this analysis to capture further constraints, such as the number of hours per work the owner / MD claims to spend on paperwork. Detailed models are given in Annex 2.

(a) Type of finance.

4.23 One reason for carrying out this part of the analysis is that if one looks at finance overall, this included sources such as community development finance, and grants. Both of these are targeted at certain groups, so an analysis of overall finance may mask specific effects. Tables A2.13 and A2.14 illustrate the determinants of the different types of finance sought. More industry variables are included, as industry effects appear more important in the demand side than the supply side, as do regional effects. Gender effects appear to be quite weak in this analysis, though there is some evidence that women-led firms are less likely to seek HP/ factoring finance. Education, and to some extent location in terms of the deprivation measures are however more important than gender or ethnicity. In general terms therefore we find only weak support here for the contention that women-led firms are less likely to seek external finance, this varying by type of firm and by type of finance.

(b) Attempted to obtain finance more than once

4.24 Table A2.15 and A2.16 shows the results for an ordered probit, on whether firms have sought finance more than once, once or not at all. There is some evidence here that women-led firms are less likely to seek external finance even allowing for other factors.

4.2.3 Concluding Points

31

4.25 Our focus in this analysis has been on the situation of an existing firm seeking additional finance. Our results reflect our focus on the supply and demand sides of the financing relationship. Although there are some conflicts between the 2003 and 2004 results, the most consistent picture is suggested by the 2004 models which suggest that, in terms of the supply side, women-led firms are around 2 pp more likely to have difficulty raising finance and also 2 pp more likely to find it impossible to raise the finance they are seeking. In terms of the demand side we also find that women-led firms are less likely to seek external finance both on a one-off and multiple basis. This reflects the situation for individual women noted earlier from the HSE.

4.3 Evidence from the UKSMEF 2004 Database

4.26 In this section we focus on evidence from the UKSMEF 2004. Our approach follows closely that adopted in the ASBS focussing first on the supply side and then exploring women-led firms’ demand for finance.

4.3.1 Descriptive Analysis

4.27 As a prelude to the multivariate analysis of the following sections it is beneficial to have an understanding of the basic characteristics underlying the UKSMEF data, and this is the objective of this section. Table 4.7 summarises a number of key characteristics of the sample by gender.

32

Table 4.7: Sample Characteristics of the UKSMEF 2004

    Men Women per cent per cent

RegionEast 9.4 9.9 East Midlands 6.0 5.8 London 21.2 21.0 North East 3.0 2.3 Northern 2.5 4.6 North West 7.5 9.7 Scotland 5.5 4.6 South East 16.4 19.1 South West 8.5 9.4 Wales 4.9 3.7 West Midlands 7.2 4.6 Yorkshire 7.9 5.3 Business age<4 years 13.8 23.1 [4-15] yrs 33.4 32.1 >15 yrs 52.8 44.8 Legal FormSole trader 65.9 66.5 Partnership 9.7 11.8 Company 24.4 21.7 Business Size<=10 employees 93.2 95.7 >10 employees 6.8 4.3 ExportNo 89.1 92.9 Exporter 10.9 7.1 Owner CharacteristicsPrevious Experience<16 yrs of experience 35.0 56.3 >=16 yrs of experience 65.0 43.7 Education AttainmentDegree or higher 21.5 32.2 "A" levels 8.6 10.4 GCSE or equivalent 13.9 17.5 Other vocational quals. 40.6 23.7 No qualification 15.4 16.2

Notes: Figures in bold are significantly different at the 5 per cent level. See Annex 1 for variable definitions.

4.28 Comparing the group of men and women respondents:

We see no significant distinction between the regional composition of the sample between businesses run by men and women, the exception being the Northern region. Three regions (London, South East and the Eastern region) account for around a half of the national sample.

Women-led businesses were more likely to have been trading for less than 4 years than those which were male-led.

A higher proportion of the women population running an SME are relatively inexperienced (with trading experience of less than 16 years). On

33

the other hand, 65 per cent of males running businesses had previous experience of more than 16 years in business.

Finally, a higher proportion of the population of women owners of SMEs have degrees (32.2 per cent), while other vocational qualifications are more common among males. Other levels of qualification are equally common.

4.29 These factors are likely to have contrasting effects on the probability of facing barriers to finance; higher levels of education by owners of SMEs is likely to have a positive effect on accessing finance; while inexperience of both the business and the owner may have a detrimental effect.

4.3.2. Modelling the “Supply Side”: Barriers to Finance4.30 Our main aim in this section is to analyse whether, controlling for firm and

owner characteristics, gender influences the experienced barriers to finance. Our approach is based on a series of bivariate probit models of the probability of experiencing different types of barriers to accessing finance. In particular, these models were based on whether the firm had faced any barrier to finance – i.e. whether the firm faced either discouragement in seeking finance or whether it faced denial in obtaining finance. In each case we use a wide range of control factors and also include some interaction terms to capture potential contingencies between, say, gender and ethnicity.

(a) Facing discouragement when applying for external finance4.31 In terms of the probability of facing discouragement when applying for

external finance (Table A2.17) our results regarding gender are consistent and statistically significant: even after adjusting for a range of firm and owner characteristics, women-led businesses are about 1.3 pp less likely to perceive discouragements when applying for finance.

4.32 In addition, the control factors highlight a number of factors that proved to be important in determining the probability of SMEs facing discouragement. In particular:

Businesses with more than 15 years of trading were 2 pp less likely to experience any type of discouragement when seeking new finance.

SMEs registered for VAT are about 2 pp less likely to experience discouragement when applying for external finance.

It is intriguing that those businesses led by owners that have attained “A-levels” are less likely to experience discouragement vis-à-vis those having a degree or higher qualification.

Industry effects also proved important as determining the probability of experiencing financial discouragement.

(b) Facing refusals when applying for external finance4.33 The results presented in Table A2.18 are those modelling the probability of

facing outright denial when applying for new external finance. Here we see no evidence that women-led firms were more likely to experience outright denials of finance than other firms. Instead, there are other factors, mainly associated to the characteristics of the business that appear more relevant. These are:

34

Start up businesses (trading for less than 2 years) are 2 pp less likely to face denials when applying for external source of finance, as are exporters.

On the other hand, older businesses (trading for more than 15 years) are about 4 pp also less likely to obtain denials.

In terms of regional effects, SMEs located in London are also between 9 and 10 pp more likely to face outright denials.

4.3.3 Modelling the Demand for Finance across different groups

4.34 The remainder of this section turns to the difference in demand for finance. In particular, we focus on the type of new finance sought rather than the difficulties faced. This uses information with respect to those firms that have recently sought new finance, as to whether there are differences across gender. The main aim of this section is to address the problem that it may be the case that specific groups simply perceive it not to be worthwhile even applying for finance.

4.35 The results suggest that being an SME run by a woman has no significant effect on the probability of applying for new finance (Table A2.19). In terms of the interaction effects, we find little evidence of any reinforcing pattern of disadvantage.

4.36 The control factors, however, highlight a number of factors that prove to be more important in determining the probability to apply for new finance. Industrial and regional differences were not statistically significant. The most consistent effects were:

Businesses with more than 10 employees are 18 pp more likely to have applied for new external finance, as are businesses registered for VAT.

Business’ age is also an important factor in determining the probability to apply for new finance, with new business (those trading for less than 4 years) being more likely to apply. In particular, businesses with more than 15 years of trading are 35 pp less likely to apply for new finance. These are also the businesses that are more likely to feel discouraged when applying for finance.

Those businesses whose premises are located in towns are also less likely (vis-à-vis major conurbation) to apply for new finance.

Businesses that have sought advice are also 11 pp more likely to apply for funding.

Finally, with respect to owner characteristics, age seems to be the determinant factor, with owners aged more than 46 being less likely to apply for new finance.

4.37 In Table A2.20 we present the results according to the different types of finance which firms applied for. Particularly we look at the probability of applying for an overdraft, loan, asset-base finance and equity. The results show that, in terms of gender, women are 17 pp less likely to apply for a loan

35

or re-mortgage their home. There is no gender difference across other types of finance. Other factors which prove significant are:

Age of the business is an important factor determining the probability of applying for an overdraft and loans. Businesses trading between 4 and 15 years are 23 pp less likely to apply for overdrafts (vis-à-vis young businesses), while they are 26 pp more likely to apply for a loan. Additionally, mature businesses (with more than 15 years trading) are significantly less likely to apply for overdrafts.

The location of a business also appears important in determining the probability of applying for an overdraft or loans. Businesses located in towns and rural areas are between 26 and 35 pp more likely to apply for overdrafts (vis-à-vis those located in major conurbations). Regarding loan applications, businesses located in cities, towns and rural areas are less likely to apply for loans.

As we would expect, companies are more likely to invest in equity and less likely to apply for loans.

Finally, industry effects are significantly important regarding the demand for loans. Particularly SMEs in sectors such as construction, health, hotels, manufacturing, real estate, transport, wholesale, and other services are more likely to apply for loans (vis-à-vis businesses in the primary sector).

4.3.4 Summary of Key Points4.38 The UKSMEF 2004 data provides a comprehensive database within which the

determinants of financial barriers and types of finance can be analysed. In terms of the supply side effect, our key finding in terms of gender is that being a women-led business is likely to decrease the probability of facing discouragement when applying for external finance.

4.39 In terms of the demand side we find little evidence here that women-led firms are generally less likely to apply for finance. More specific effects are evident, however, with women-led businesses 17 pp less likely to apply for a commercial loan or mortgage from banks or other financial institutions.

4.4 Conclusions from the firm based analysis4.40 Both the UKSMEF and the ASBS have highlighted some differences in

obtaining finance between men and women-led firms. In several places the ASBS analysis highlights the additional difficulties that women-led businesses have in raising finance, and even that non-women-led businesses that have women directors are less likely to obtain finance than male businesses. These results however are rather contradicted by the UKSMEF analysis, which suggests that being a women-led business has little general effect on either the supply side or the demand side. Indeed, the UKSMEF database suggests that women-led businesses are actually less likely to face discouragement when applying for external finance.

36

4.41 One possible reason for these differences is highlighted in the demand analysis. For example, the UKSMEF analysis suggests that women-led businesses are 17 pp less likely to apply for a commercial loan or mortgage from banks or other financial institutions. These are the types of finance that are most likely to be refused, as such if women-led businesses in the UKSMEF data are less likely to seek bank financing, then they are less likely to report problems overall. Both sets of analysis do suggest that women-led businesses are less likely to seek finance, so by definition they are less likely to report problems. Equally, the women-led businesses appear (based on the 2003 ASBS) to be significantly more likely to apply for finance that is generally easier to get, such as targeted community schemes and grants, which again may highlight why women-led businesses in the UKSMEF data report fewer problems.

4.42 Overall, this highlights two issues worthy of further consideration. Firstly, whether analysis based on established firms hides many of the effects felt by new or potential firms in acquiring finance, and secondly why the gender effects should vary so much across the two samples. Our suggestion is that this is based on the demand side, that women-led businesses are less likely to seek certain types of finance, so are therefore less likely to report problems. With better data, one could do a two stage analysis of what determines why different types of firms seek different types of finance, and then secondly what the outcome of that is. With the current data, sample sizes would become too small for this analysis.

37

Chapter 5 – Conclusions

5.1 Introduction 5.1 Our key conclusions here are in two main areas relating to the impact of

gender differentials in financial constraints on (a) start-up and (b) the access to finance for existing firms. We address each of these in turn. Section 5.4 briefly outlines some potential issues which might warrant future research and data collection. Before focussing specifically on gender effects on finance, however, it is worth noting briefly the range of other factors identified in the models estimated here which also have an influence on demand for finance. At an individual level the characteristics and background of the individual prove important including factors such as education, prior experience, home ownership and in some instances ethnicity and location. At a corporate level firm size, industry, location and the ethnicity of the group of owner-managers all influence the businesses’ ability to access start-up finance. It is in the context of this complex set of factors that our results on the specific impact of gender have to be considered.

5.2 Gender effects on accessing finance and its impact on business start-up

5.2 Both the GEM 2004 data and the HSE prove interesting in terms of assessing gender effects on individuals’ access to finance and its impact on start-up. In both cases there is evidence of a negative gender finance effect which would tend to reduce start-up rates among women. These effects differ in nature, however, reflecting the structure of the datasets and the different questions asked in each of the surveys.

5.3 In the GEM dataset the focus is on whether the general perception of financial barriers to business start-up is stronger for women within the general population than men. Our key finding, here, is that being a woman increases the probability than an individual will perceive financial barriers to business start-up by 7.5 pp, controlling for a wide range of other factors which may influence the perception of financial barriers. This perception of financial barriers in turn works to reduce start-up rates by 1.7-3.8 pp depending on the start-up indicator being used. Being a woman also has an additional direct effect on each of our start-up indicators, however, not linked to financial barriers.

5.4 Our GEM analysis relates to the entire adult population and reflects the impact of financial barriers on start-up in this broad group. Our analysis of the HSE is, because of the survey structure, more focussed with a concentration on ‘Thinkers’ and ‘Doers’, i.e. those who are either engaged in or thinking about undertaking enterprise activity. Within this group we find no evidence that women do face any increased difficulty in obtaining start-up finance. We do find evidence, however, that females are less likely to seek external finance.

5.5 Taking these points together suggests an interpretation that women in the general population perceive stronger financial barriers to business start-up than

38

men, and this may be discouraging them from seeking external financial support for business start-up. We find no evidence, however, that where women do seek finance for start-up they are any less likely to obtain it than men. This is suggestive of a dominant demand rather than supply side effect. In either case, however, the final effect is similar – that gender differences in access to finance are reducing women’s start-up rates.

5.6 From a policy perspective the key points here are that women’s start-up rates are being reduced by (a) the perception among the general population of stronger financial barriers to start-up among females, and (b) their unwillingness to seek external finance for business start-up. Addressing these issues is likely to require a combination of measures designed both to redress the current perception – evident in the GEM survey - that women perceive it to be more difficult to obtain business finance and to encourage potential women entrepreneurs to be more ambitious in seeking external finance. Successfully addressing these issues alone will not, however, fully close the gap between start-up rates among men and women. We still find a further gender gap in our analysis which we cannot explain in terms of financial shortages.

5.3 Gender effects on access to finance by existing firms 5.7 Our analysis of the ASBS for 2003 and 2004 and the UKSMEF for 2004

focuses separately on the supply and demand side of the financing relationship for existing firms. In terms of the supply side our results are somewhat contradictory with evidence from the ASBS highlighting some negative gender effects, but the UKSMEF suggesting that women-led firms are less likely to be discouraged in their search for business finance. More specifically, the ASBS for 2004 suggests that women-led firms are around 2 pp more likely to have difficulty raising finance and also 2 pp more likely to find it impossible to raise the finance they are seeking. Although there is some weaker, more positive, evidence from the ASBS 2003, the main suggestion here is that even allowing for business characteristics and the characteristics of the owner-manager of the firm, women owner-managers found it more difficult to access finance than men owner-managers. The UKSMEF 2004 data on the other hand suggests that women-led businesses are less likely to face discouragement when applying for external finance and encounter no difference in rejection rates compared to male-led businesses.

5.8 In terms of the demand side we see a more consistent picture, however, and find evidence that broadly supports that identified earlier from the GEM and HSE analysis. The ASBS suggests that women-led firms are less likely to seek external finance both on a one-off and multiple basis. The UKSMEF provides no support for this general proposition, however, but does suggest some more specific effects with women-led businesses 17 pp less likely to apply for a commercial loan or mortgage from banks or other financial institutions. These demand side results provide one possible reason for the differences in supply side results suggested by the supply side analysis.

5.9 In policy terms our results again tend to provide the most consistent support for measures designed to strengthen the demand for finance from women-led businesses rather than the supply-side. In particular the UKSMEF data

39

emphasise the unwillingness of women-led firms to seek bank finance, an issue which may well be worthy of more specific research.

5.4 Issues for future research

5.10 One of the key issues highlighted in the literature review in Chapter 2 was the longer-term dynamic effects of under-capitalisation – the result of finance shortages – on the performance of women-led firms. The data analysed here are purely cross-sectional, however, and while this makes it possible to draw some inferences about the effects on start-up rates, for example, it is impossible to draw implications about any impacts on the subsequent success of those start-up companies. Similarly, it proved difficult from the existing survey data to draw any firm conclusions about the impact of finance shortages on subsequent business performance. Both require more longitudinal follow-up of individuals or firms which have participated in cross-sectional surveys and we would see this as a research priority going forwards.

5.11 In addition, our survey data provides little insight into the origin of the differences in the perceptions of males and females in terms of the availability of business finance. One possibility is that individuals’ prior experience of either business or personal dealings with banks or other financial institutions may be conditioning this view. Other factors linked more to risk-aversion, an unwillingness to take on debt or family circumstances not reflected here may also be important.

5.12 Other more specific issues were raised in each of the surveys used here in terms of their coverage of ethnic minority groups and those in disadvantaged locations. In each case, sample sizes were relatively small and inferences about either group were unlikely to be robust. The policy significance of each issue is likely to mean that developing appropriate sampling methodologies to identify ethnicity and spatial effects on financial barriers are also likely to be issues for the future.

5.13 Finally, it is worth noting that in each of these general surveys the proportion of firms and individuals reporting that they were involved in enterprise and that they had experienced difficulties in accessing finance is relatively small. In terms of this question each of the surveys therefore embodies considerable redundancy and it may be worth considering more specific data collection exercises if this is a significant enough policy issue.

40

Annex 1: Modelling Access to Finance and Business Start-up

A1.1 Introduction A1.1 In this section we present the econometric modelling on which the main

results reported in Chapter 3 are based.

A1.2 Modelling with GEM 2004A1.2 Our main aim in this section is to see whether, controlling for individuals’

background characteristics, gender influences the perceived financial barriers to business start-up. Our approach is based on a series of probit models of the probability of perceiving financial barriers to business start-up (Table A1.1). Significant coefficients are in bold type in the table and two alternative formulations of the model are presented dropping the insignificant variable ‘Enterprise Training at school’ in the second model.

Table A1.1: Probit models of Perceived Financial Barriers

Model 1 Model 2Marginals T Stat Marginals T Stat

Variables of InterestWomen 0.075 8.664 0.075 8.729

ControlsEastern Region -0.026 -2.068 -0.026 -2.053Degree or higher 0.030 2.828 0.031 2.928‘A’ Levels -0.039 -3.254 -0.037 -3.176Other vocational quals. 0.037 2.429 0.039 2.551HH Income: 2nd quartile -0.055 -4.028 -0.054 -3.960HH Income: 3rd quartile -0.041 -2.897 -0.041 -2.887HH Income: 4th quartile -0.122 -8.549 -0.121 -8.479Age in years -0.006 -16.439 -0.006 -16.647Part-time (8-29 hours) -0.042 -3.248 -0.041 -3.180Not working (8 or less hours) -0.054 -4.459 -0.054 -4.473Enterprise training at school 0.017 1.342Work experience at school 0.023 2.384 0.025 2.704Work experience at college/university -0.060 -4.897 -0.058 -4.834

Constant 0.405 18.962 0.407 19.087

No of observations 14710 14736Chi Square 634.76 639.10Estrella 0.0307 0.0304 per cent Correct 64.2 64.4

41

Notes: Marginal values suggest the increase in the probability of perceiving finance barriers when moving from dummy variable values of 0 to 1. Marginal values for the age variable are at variable means. Sample observations are weighted.Source: GEM 2004

A1.3 In terms of business start-up our aim is to investigate the potential role of perceived financial barriers to business start-up on business start-up itself. If such perceived financial barriers to business start-up are important in influencing business start-up, then the fact that perceived financial barriers to business start-up are concentrated among women may be contributing to lower start-up rates among women. If perceived financial barriers to business start-up are not a factor in shaping business start-up then differential access to finance is likely to be less important in explaining lower start-up rates among women (e.g. Table 3.1).

A1.4 Identifying the impact of perceived financial barriers to business start-up on business start-up raises some classic econometric and statistical issues. In particular, the obvious approach is to estimate a model for business start-up and include a dummy variable which takes a value of 1 if an individual perceives financial barriers to business start-up. The coefficient on this ‘treatment’ term would then suggest the significance of perceived financial barriers to business start-up in the business start-up decision. In fact, however, unless perceived financial barriers to business start-up are randomly distributed across the population – and the previous section suggests they are not – this approach will yield potentially biased estimates of the importance of perceived financial barriers. Instead, an approach is needed which corrects for so called sample selection, allowing for any connection between the factors determining perceived financial barriers to business start-up and business start-up.

A1.5 This is simply dealt with using a bivariate probit model estimating simultaneously Part A - for the perceived financial barriers to business start-up – and Part B for business start-up itself (Table A1.2). Here the main parameter of interest is the disturbance correlation coefficient. If this is significant it suggests the need to simultaneously examine the determinants of perceived financial barriers to business start-up and business start-up. If this is not significant, it suggests the validity of estimating single equation probit models for business start-up. Table A1.3 reports three bivariate probit models for the three business start-up indicators discussed earlier. In each case the disturbance correlation is insignificant suggesting the validity of examining business start-up using single equation probit models. Significant coefficients are in bold type.

42

Table A1.2: Bivariate Probit Models of Shortage of Start-up Finance and Start-up

Model 1 Model 2 Model 3Coeff T-stat Coeff T-stat Coeff T-stat

Part A: Perceived Financial Barriers ProbitConstant 1.076 16.389 1.085 16.539 1.075 16.313Women 0.233 8.858 0.234 8.878 0.234 8.877Eastern -0.119 -3.246 -0.118 -3.251 -0.113 -3.101Degree or Higher 0.064 2.471 0.065 2.495 0.059 2.255HH Income: 2nd quartile -0.110 -2.403 -0.124 -2.726 -0.114 -2.474HH Income: 3rd quartile -0.054 -1.175 -0.061 -1.328 -0.050 -1.081HH Income: 4th quartile -0.262 -5.779 -0.273 -6.046 -0.267 -5.882Age in years -0.017 -14.870 -0.017 -14.892 -0.017 -14.912Part time working -0.122 -3.497 -0.123 -3.546 -0.098 -2.810Work experience at school 0.024 0.844 0.023 0.811 0.021 0.745Work experience at college/university -0.167 -4.807 -0.167 -4.813 -0.159 -4.557B. Business Start-up Models          

Start-up Running a business Expected Start-upConstant -1.510 -3.143 -1.401 -2.651 -1.005 -2.932Lack of Finance -0.174 -0.327 -0.294 -0.567 0.128 0.307Women -0.215 -3.379 -0.099 -1.588 -0.253 -5.859London -0.051 -0.867 -0.084 -1.440 0.203 4.678Eastern -0.282 -3.472 -0.019 -0.229 -0.195 -3.025West Midlands -0.272 -3.345 -0.008 -0.099 -0.186 -3.020Yorks and Humber -0.259 -1.738 -0.347 -2.283 -0.317 -2.811North West -0.376 -3.346 -0.011 -0.117 -0.179 -2.299North East -0.296 -1.640 -0.253 -1.193 -0.245 -1.929Scotland -0.254 -2.359 -0.117 -1.052 -0.255 -3.013Age in years -0.003 -0.673 -0.006 -1.599 -0.011 -2.837Self-employed 0.773 13.720 2.752 21.629 0.357 7.001Business Owner 0.539 6.651 3.191 19.792 0.383 5.491Enterprise training at college/university 0.164 3.075 0.335 6.042 0.252 6.304Work experience at school -0.126 -2.418 -0.297 -5.382 0.088 2.423Work experience at college/university 0.387 6.233 0.004 0.065 0.184 3.932

C. Disturbance Correlation RHO(1,2) 0.012 0.038 0.219 0.714 -0.222 -0.877

Observations 11577 11580 11456Log Likelihood -9354.15 -9268.1 -10823.7

Notes: Sample observations are weighted. Source: GEM 2004

43

Table A1.3: Probit Models of Business Start-up Indicators

Model 1: Start-up

Model 2: Running Business

Model 3: Expected Start-up

Marginals T-stat Marginals T-stat Marginals T-statConstant -0.130 -23.760 -0.203 -11.735 -0.144 -11.092Women -0.017 -4.955 -0.010 -1.962 -0.035 -6.691Perceived Financial Barriers -0.013 -3.526 0.007 1.383 -0.038 -6.565South East 0.008 1.246 0.045 6.020London 0.001 0.349 -0.004 -0.590 0.076 8.282Eastern region -0.014 -3.454Wales -0.013 -1.746 -0.020 -1.291North West -0.020 -4.124North East -0.017 -2.111Scotland -0.011 -1.911 -0.028 -2.822 -0.010 -0.793Degree or higher -0.035 -4.097‘A’ Levels -0.041 -6.029 0.009 1.311GCSE or equivalent -0.026 -3.450Other vocational quals. -0.018 -2.097 0.006 0.634HH Income: 2nd quartile 0.062 4.230HH Income: 3rd quartile 0.039 3.037HH Income: 4th quartile -0.007 -2.070 0.056 4.419 -0.013 -2.435Age in years -0.001 -2.502 -0.001 -7.598Self-employed 0.106 9.887 0.802 67.580 0.071 6.026Business Owner 0.069 5.182 0.896 90.865 0.084 4.884Enterprise training at school 0.013 1.581Enterprise training at college/university 0.018 3.620 0.049 5.813 0.040 5.122Work experience at school -0.008 -2.371 -0.030 -5.868 0.014 2.364Work experience at college/university 0.037 5.729 0.026 3.158

N 11906 11601 11434chi squared 416.52 5534.89 492.19Estrella 0.034 0.481 0.031

Notes: Marginal values suggest the increase in the probability of perceiving finance shortages when moving from dummy variable values of 0 to 1. Marginal values for the age variable are at variable means. Sample observations are weighted.

A1.3 Modelling with the HSE

A1.6 In this section we turn to the evidence from the Household Survey of Entrepreneurship 2003 (HSE). Here our analysis focuses on whether the probability of engaging in entrepreneurial activity is influenced by individuals’ access to finance, and whether the access to finance itself is influenced by gender.

A1.7 First, we consider whether gender is important in individuals’ decisions whether to seek external funding for business start-up and whether this then affects the start-up decision. The model will be estimated by using the Heckman two-stage procedure where two equations are estimated: the first equation models the self-selection mechanism where we try to understand

44

which factors affect individuals’ decisions about whether to seek external finance; the second equation models the start-up decision and is estimated only on the sample that is selected through the self-selection mechanism. In the self-selection equation, we will model the probability of seeking external funds as a function of gender, education and location. In the main equation, we will model the probability of start-up as a function (among the others) of previous experience (proxying entrepreneurial ability), employment status (measuring indirectly the foregone wage income) and attitudes towards self-employment (as a measure of the disutility cost attached to running an entrepreneurial project). Table A1.4 presents the results for the Heckman model, with marginal effects given for the start-up equation and significant coefficients highlighted in bold type.

Table A1.4: Heckman Two Stage Models of the Probability of Seeking External Finance and Start-up

Coefficient T-ratioStage 1: Probability of Seeking External Finance

Region 0.002 0.12Degree level qualification 0.096 0.99Women -0.387 -2.54White 0.233 3.27Sex*White 0.529 3.17Constant -1.460 -10.29

Stage 2: Probability of Start-upPrevious experience 0.261 2.94Degree 0.199 1.73Attitude towards entrepreneurship (negative) -0.174 -2.55Employment Status -0.295 -2.04Region -0.006 -0.24Constant -1.632 -9.41

Correlation coefficient 3.803 2.20

Marginal Effects for Stage 2 (Start-up) Previous experience 0.041 2.72Degree 0.033 1.62Attitude towards entrepreneurship (negative) -0.029 -2.31Employment Status -0.047 -2.02Region -0.001 -0.24

Notes: N=2106. The observations are weighted by WEIGHT_1. Variable definitions are given in Annex 1. Source: HSE

A1.8 Now we consider whether the probability of becoming a Doer is affected by the respondent’s gender and conditioned by the applicant’s probability of experiencing difficulty in obtaining finance (that in turn is affected by an additional set of variables including gender). Econometrically, this involves

45

the estimation of a two-stage Heckman model: in the first stage we model the respondent’s probability of experiencing difficulty obtaining finance and test whether this is affected by gender, ethnic background, location, education and whether or not (s)he own a house (that can be used as collateral); in the second stage we model the probability of becoming self-employed as a function of gender, the respondent’s foregone wage income (proxied by whether the respondent has a degree15), attitudes towards entrepreneurship, location and previous experience.

Table A1.5: Heckman two-stage model: Model 1

Coefficient T-ratio

Stage 1: Difficulties Obtaining External Finance

Region 0.09 2.41Degree 0.17 0.72Not home-owner 0.02 4.17Women 1.13 1.54White -0.91 -1.61Women*White -1.29 -1.69

Stage 2: Probability of Start-upPrevious experience 1.93 1.99Degree 0.49 0.71Attitude towards entrepreneurship -7.79 -5.29Region -0.18 -1.72Women 0.25 0.50

Correlation coefficient -1.118 -1.52

Marginal Effects for Stage 2 (Start-up)Previous experience 0.011 1.32Degree 0.003 0.66Attitude towards entrepreneurship -0.073 -1.53Region -0.001 -1.07Women 0.002 0.39

15 The assumption is that the individuals with a degree have a potential for a high income and therefore the opportunity cost of becoming self-employed is higher.

46

Annex 2: Modelling the Access to Finance by Established Firms

A2.1 Introduction

A2.1 In this annex we report the econometric analysis underlying the discussion in Chapter 4. This relates to the access to finance by established firms.

A2.2 Modelling Using the ASBSA2.2 Separate models are estimated here for the supply and demand side (see the

discussion in Section 4.2). The baseline supply side results were derived by employing a general-to-specific approach, commencing with all of the factors assumed to impact on the ability of a firm to raise finance. This involves running a regression including all variables theoretically assumed to contribute to the explanation of the dependent variable, and dropping all those that prove insignificant. This includes all the industry and regional effects, firm size, exporting and growth. The variables capturing ownership characteristics, including age, education, as well as gender were then added16. This was then tested across three versions of the finance question, firstly for the four categories, based on a multivariate probit, (ranging from “no difficulty in obtaining finance”, to “obtained some but not all of the finance”, “obtained all of the finance required” and finally “no difficulty in raising finance” and then for two other versions of these answers based on a bivariate probit, divided into “any problems” and “no problem” and “impossible” compared with all other categories. These were based on whether the firm had eventually been able to raise all of the finance or not, and secondly whether it had been simply impossible to raise any finance. See Question 84 of the 2003 survey.

A2.3 Tables A2.1 and A2.2 present the baseline models for 2003 and 3004 based on the alternative measures of the “supply side” model with significant variablesc in bold17. A positive coefficient here is associated with an increased difficulty in raising finance. The models generally perform well. The percentage of correct predictions is high, and not due simply to under- or over-predicting. Industry and regional effects are perhaps surprisingly unimportant in explaining the ability of a firm to raise finance.

A2.4 In addition to the baseline models in A2.1 and A2.2, we also estimate supply-side models for various sub-groups of the ASBS 2003 and 2004. These are reported in Tables A2.3-A2.10 and discussed in Section 4.2

16 Models were also estimated including a deprivation index, though these data are relatively sparse, with 6478 observations for 2003, but only 3309 observations from the sample of 7505. As a result, many of the industries have to be dropped, and there are no observations for Scotland, and the Welsh data is reported in a different format. Results from these models with respect to gender were consistent with those presented below.17 Throughout this report, the marginal effects are reported for all of the bivariate regressions. For the multivariate regressions, the usual coefficients are reported.

47

Table A2.1: Baseline Results from the ASBS - 2003

Multivariate probit 1-4

Bivariate “any problem”

Bivariate “impossible”

Parameter Estimate t-statistic Marginal effect

t-statistic Marginal effect

t-statistic

Constant -1.061 -14.641 -0.212 -16.882 -0.149 -16.574

Women-led -0.111 -2.116 -0.001 -0.176 -0.008 -1.176

Control VariablesEmployment (no) 0.003 7.566 0.000 0.846 0.000 -0.740Growth Aspiration 0.377 10.167 0.021 3.413 0.006 1.341VAT registered 0.220 3.519 0.029 2.661 0.015 1.987Firm Age=1 0.144 1.562 0.079 6.064 0.038 4.169Firm Age=2 -0.095 -1.124 0.047 3.918 0.027 3.237Firm Age=3 0.033 0.471 0.043 4.133 0.023 3.150Firm Age=4 0.134 1.487 0.051 3.728 0.012 1.086Firm Age=5 0.036 0.430 0.044 3.493 0.017 1.795Firm Age=6-10 -0.015 -0.302 0.028 3.366 0.018 3.061Agriculture 0.275 3.022 -0.058 -2.370 -0.031 -1.710Construction 0.014 0.260 -0.030 -2.954 -0.018 -2.244Retail -0.202 -3.911 -0.032 -3.485 -0.010 -1.534Wholesale -0.005 -0.081 -0.027 -2.368 -0.020 -2.206Hotels & Rest’nts -0.135 -2.269 -0.024 -2.337 -0.011 -1.569Post & communication

-0.071 -0.374 0.035 1.397 0.009 0.458

Exporting Firm -0.007 -0.186 0.017 2.886 0.019 4.363Sole Proprietor -0.222 -4.741 -0.022 -2.895 -0.003 -0.517Partnership 0.039 0.918 -0.025 -3.215 -0.007 -1.242EMB owned/directors

0.086 2.431 0.011 2.175 0.008 2.283

Some women directors

0.091 1.922 0.018 2.236 0.004 0.733

Regional DummiesNorth East 0.042 0.524 0.001 0.056 0.008 0.828North West 0.067 1.347 -0.010 -1.183 0.000 -0.069Yorks and Humber 0.054 1.055 -0.008 -0.875 0.002 0.295East Midlands -0.018 -0.250 -0.043 -3.021 -0.009 -0.940West Midlands -0.024 -0.387 -0.016 -1.486 -0.008 -1.071Eastern Region -0.031 -0.467 -0.027 -2.215 -0.011 -1.199South West 0.141 2.199 -0.029 -2.354 -0.008 -0.920Wales 0.090 1.772 0.017 2.070 0.017 2.911 per cent correct 64 per

cent94 per cent

97 per cent

Reported “problems”

0 17291 108 2 162 3 272

542 272

48

Table A2.2: Baseline Results from the ASBS 2004

Multivariate probit, 1-4

Bivariate “any problems”

Bivariate “impossible”

Parameter Estimate t-statistic Marginal effect

t-statistic Marginal effect

t-statistic

Constant -1.105 -7.577 -0.231 -10.100 -0.212 -10.787

Women-led 0.136 1.253 0.020 2.543 0.022 2.097

Control VariablesLog(Employment)18 -0.112 -3.700 -0.028 -3.424 -0.012 -2.085Growth Aspiration 0.183 1.847 0.054 1.939 0.023 1.191VAT registered -0.255 2.701 -0.067 2.514 -0.034 2.691Firm age = 4-10 0.423 2.600 0.095 1.373 0.072 0.414Firm age = 11-20 0.160 1.019 0.063 2.959 0.015 2.022Firm age > 20 -0.394 2.954 -0.109 -1.505 -0.052 -0.804Wholesale -0.236 -1.508 -0.058 -1.292 -0.021 -0.642Motor Retail -0.272 -1.165 -0.069 -0.527 -0.027 -0.144Financial services -0.084 -0.401 -0.056 -1.414 -0.017 -1.158Education -0.539 -1.549 -0.149 -1.832 -0.093 -1.248Post-graduate qualificat. 0.085 0.753 0.032 0.783 -0.008 0.331Degree 0.071 0.767 0.021 1.242 0.005 -0.532A-level 0.099 0.807 0.036 -1.216 -0.015 -0.362GCSE -0.138 -1.088 -0.039 4.373 -0.008 3.590Exporting Firm 0.360 4.440 0.097 1.095 0.058 2.149Partnership 0.065 0.598 -0.003 -1.219 0.025 -1.120Partnership -0.174 -1.620 -0.053 1.894 -0.034 0.422OWNER > 60 0.205 1.598 0.065 1.604 0.011 1.454EMB directors/owners 0.235 1.854 0.040 1.081 0.027 1.459Have sought business advice

0.232 2.898 0.054 -0.741 0.032 -1.018

Deprivation Quintile 1 -0.129 -0.822 -0.032 0.006 -0.035 0.249Deprivation Quintile 2 -0.008 -0.058 0.000 -1.235 0.005 -0.586Deprivation Quintile 3 -0.179 -1.189 -0.055 2.140 -0.020 1.099Deprivation Quintile 4 0.267 1.930 0.073 0.328 0.025 0.597Deprivation Quintile 5 0.055 0.421 0.007 2.259 0.010 1.909Regional DummiesNorth West -0.285 -1.828 -0.080 2.053 -0.042 1.229South West 0.266 1.975 0.073 2.428 0.031 0.885West Midlands 0.320 2.140 0.090 1.299 0.021 0.517Scotland 0.107 0.980 0.036 1.200 0.010 -0.248 per cent correct 63 85 91Reported problems 0 1258

1 1072 743 141

352 141

NB in deprivation quintiles 5 is worst

18 Where employment of the firm is given as zero, this is replaced by 1 before logging. This seems reasonable as a sole proprietor is employed by the business.

49

Table A2.3: Testing for differences across amount requested 2003

Amount Borrowed

Less than £10,000 £10,000-£100,000 £100,000+

Parameter Estimate t-statistic Estimate t-statistic Estimate t-statisticConstant -1.410 -3.669 -0.151 -0.292 -1.100 -1.908Employment -0.008 -0.853 -0.010 -2.180 -0.001 -0.873Growth Aspiration

0.072 0.265 0.157 0.837 -0.254 -1.722

VAT registered

0.173 0.622 -0.784 -1.710 -0.219 -0.395

FIRM AGE=1 0.628 1.239 -0.040 -0.102 0.136 0.304FIRM AGE=2 0.398 0.854 0.330 0.938 0.002 0.004FIRM AGE=3 0.475 1.076 -0.254 -0.765 0.653 2.618FIRM AGE=4 -0.255 -0.431 -0.385 -0.678 0.377 1.085FIRM AGE=5 -0.185 -0.404 -1.170 -1.513 0.775 2.182FIRM AGE=6-10

0.154 0.521 -0.335 -0.932 0.446 2.015

Agriculture -5.394 -0.001 -0.317 -0.742 -5.136 -0.001Construction -0.970 -1.949 -0.213 -0.702 -0.006 -0.026Retail 0.046 0.145 -0.684 -1.688 -0.005 -0.018Wholesale -0.149 -0.219 -0.601 -1.727 -0.386 -1.406Hotels & Rest’nts

-0.051 -0.109 -0.003 -0.010 -0.082 -0.294

Post & communication

-4.798 0.000 -5.428 0.000 0.396 0.760

North East 0.016 0.023 0.136 0.421 -0.220 -0.572North West 0.287 0.812 -0.144 -0.535 -0.516 -2.087Yorks and Humber

0.243 0.701 -0.425 -1.402 -0.003 -0.016

East Midlands 0.067 0.150 -0.614 -1.059 -0.149 -0.483West Midlands -0.103 -0.273 -0.113 -0.292 -0.391 -1.205Eastern region -0.124 -0.201 -0.462 -1.169 -0.338 -1.090South West -0.538 -0.856 0.003 0.009 -1.064 -2.557Wales -0.363 -1.048 0.100 0.388 0.194 1.030Exporting firm 0.619 2.156 0.512 2.928 0.353 2.459Sole Proprietor 0.355 1.161 -0.348 -1.076 0.474 2.116Partnership 0.084 0.224 -0.183 -0.781 -0.270 -1.260African -0.394 0.000 0.114 0.000Pakistani -0.604 0.000 -0.064 -0.101 0.804 1.918Bangladeshi 0.080 0.000 0.689 1.028 -0.723 0.000Other Asian 0.916 0.000 0.212 0.383Caribbean -0.691 0.000 -0.435 -0.001 0.236 0.378White and Black Caribbean

0.565 2.009 0.162 2.630 0.830 1.696

Women- led 0.048 0.154 0.171 0.556 -0.053 -0.207Some directors women

0.140 0.278 -0.026 -0.101 0.001 0.005

Number of obs 262 (43 positive) 483(59 positive) 885 (87 positive) per cent correct

86 88 90

50

Table A2.4: Testing for differences across amount requested 2004

Amount borrowed19

Less than £10,000 £10,000-£100,000 £100,000 +

Parameter Estimate t-statistic Estimate t-statistic Estimate t-statisticConstant -1.098 -7.532 -1.208 -5.248 -0.860 -2.948Employment -0.109 -3.606 -0.084 -1.609 -0.182 -3.251Growth Aspiration

0.176 1.778 0.243 1.512 0.060 0.312

VAT registered -0.258 2.777 -0.233 2.278 -0.237 2.172FIRM AGE 4-10 0.421 2.585 0.703 3.245 0.633 1.345FIRM AGE 11-20 0.186 1.193 0.155 0.568 0.297 1.046AGE OF FIRM > 20

0.404 3.034 0.382 1.729 0.626 2.730

Wholesale -0.227 -1.453 -0.068 -0.318 -0.171 -0.598Motor Retail -0.277 -1.185 -0.415 -0.979 0.185 0.584Financial services -0.066 -0.317 -0.116 -0.358 -0.317 -0.618Education -0.540 -1.550North West -0.280 -1.799 -0.413 -1.678 -0.304 -0.979South West 0.260 1.935 0.202 0.858 0.383 1.644West Midlands 0.323 2.157 0.181 0.769 0.042 0.127Scotland 0.100 0.920 0.124 0.723 -0.131 -0.613Post grad Qualification

0.093 0.824 0.165 0.900 0.052 0.268

Degree 0.079 0.858 0.048 0.317 0.147 0.885A-level 0.091 0.737 0.048 0.233 0.160 0.623GCSE -0.141 -1.112 -0.067 -0.361 -0.574 -1.869Exporting Firm 0.356 4.382 0.388 2.866 0.502 3.605Sole Proprietor 0.073 0.666 0.004 0.024 0.130 0.580Partnership -0.166 -1.545 -0.091 -0.561 -0.388 -1.489OWNER > 60 0.205 1.592 0.011 0.051 0.208 0.962Women - led 0.146 1.346 0.368 2.037 0.204 0.921Firm sought advice

0.226 2.827 0.279 2.105 0.399 2.733

Caribbean 0.585 1.248 0.431 0.708 0.856 0.882White and Black African

0.106 0.438 0.242 0.589 -0.207 -0.477

Indian 0.210 0.311Bangladeshi 0.605 0.964Deprivation Quintile 1

-0.145 -0.919 -0.569 -1.718 0.041 0.152

Deprivation Quintile 2

-0.010 -0.072 0.221 1.011 -0.228 -0.909

Deprivation Quintile 3

-0.184 -1.219 0.048 0.195 -0.333 -1.163

Deprivation Quintile 4

0.270 1.960 0.546 2.439 -0.144 -0.561

Deprivation Quintile 5

0.059 0.447 0.103 0.509 0.120 0.418

per cent correct 64 90 91

19 There are more categories in the data, but these were dictated by having enough variation in the dependent variable within each group.

51

Table A2.5: Breakdown by why finance was requested 2003Working capital Capital

investmentR&D and IP Other

Parameter Est t-stat Est t-stat Est t-stat Est t-stat

Constant -0.185 -13.192 -0.237 -13.340 -0.052 -8.312 -0.086 -10.582Employment 0.001 0.898 0.008 3.425 0.001 1.613 0.005 4.043Growth Aspiration 0.013 5.480 0.075 9.183 0.011 3.526 0.016 4.328VAT registered 0.030 0.069 0.103 0.352 0.141 1.613 0.038 0.347Manufacturing 0.011 0.745 0.061 2.481 -0.007 -0.509 -0.007 -0.492Construction 0.008 0.824 0.047 2.691 0.001 0.402 -0.016 -2.099Retail -0.005 -0.381 0.024 1.046 -0.009 -1.363 -0.017 -1.971Wholesale 0.009 0.487 -0.007 -0.308 -0.008 -1.052 0.002 0.227Motor retail 0.003 0.194 0.013 0.554 -0.007 -0.922 -0.006 -0.751Hotels -0.002 -0.384 0.010 0.027 -0.011 -1.130 -0.042 -2.092Financial services -0.009 -0.581 0.016 0.745 -0.002 -0.247 -0.002 -0.327Real Estate -0.006 -0.037 0.094 3.612 -0.004 -0.319 -0.028 -1.250Health & social work

-0.016 -0.433 -0.008 -0.227 0.005 0.481 -0.008 -0.416

manufacturing -0.013 -1.095 -0.039 -1.662 -0.014 -1.453 -0.011 -1.303construction -0.008 -0.540 0.028 1.464 -0.001 -0.214 -0.008 -1.189Retail -0.009 -0.177 0.032 0.925 0.009 1.343 0.006 0.558Wholesale -0.010 -0.398 0.048 2.383 -0.001 -0.022 -0.002 -0.136Eastern Region -0.015 -1.225 -0.013 -1.142 -0.008 -1.057 0.003 -0.148London 0.000 -0.575 -0.063 -3.840 -0.002 -0.718 -0.003 -0.660North East 0.006 0.089 -0.019 -0.935 0.008 1.056 -0.006 -0.506North West 0.015 0.867 -0.015 -0.706 -0.004 -0.628 0.003 0.419South East 0.005 -0.149 -0.025 -1.826 0.003 0.456 -0.001 -0.695South West 0.001 -0.126 -0.003 -0.419 -0.003 -0.477 0.008 0.677West Midlands -0.001 -0.311 -0.026 -1.333 0.005 0.676 -0.004 -0.441Yorks and Humber 0.004 0.121 -0.007 -0.366 -0.001 -0.214 0.005 0.534East Wales -0.002 0.040 -0.016 -0.670 0.003 0.631 0.007 0.865West Wales -0.003 0.060 -0.003 0.052 0.005 0.970 0.008 1.041Scotland -0.004 -0.196 -0.005 -0.258 -0.001 -0.146 -0.008 -0.923Highlands 0.027 1.419 0.009 0.551 -0.001 -0.001 0.001 0.216Post Graduate 0.038 4.272 0.026 2.834 0.008 2.591 0.008 1.957Degree 0.032 4.433 0.005 1.160 0.005 1.784 0.006 1.587A levels 0.025 2.512 0.002 0.431 -0.004 -0.616 0.004 0.720GCSE 0.023 2.591 0.021 2.186 0.004 1.273 0.000 0.347Exporting Firm 0.019 2.311 -0.034 -3.324 0.002 0.548 0.004 0.728EMB owner/direct. 0.010 1.022 0.003 0.483 0.003 0.732 0.002 0.616Women- led -0.003 -0.525 -0.027 -2.322 -0.003 -0.740 0.009 1.839Deprivation Quintile 1

0.025 -1.202 0.018 -0.773 0.004 -0.306 0.004 -1.626

Deprivation Quintile 2

-0.013 -0.241 -0.008 -0.180 -0.001 0.806 -0.014 -0.304

Deprivation Quintile 3

-0.003 -0.903 -0.002 -1.866 0.003 -0.409 -0.002 -1.279

Deprivation Quintile 4

-0.007 -1.436 -0.025 -1.079 -0.001 -0.595 -0.008 -0.899

Deprivation Quintile 5

-0.016 -0.236 -0.013 1.353 -0.002 -1.279 -0.005 1.710

Firm sought advice -0.005 3.995 0.020 2.654 -0.008 1.673 0.010 1.464Positives 520 819 73 154

52

Table A2.6: Breakdown by why finance was requested 2004

Working capital/ cash flow

Capital Investment R&D and IP Other

Parameter Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statisticConstant -1.387 -4.106 -0.973 -4.920 -1.437 -0.005 -1.392 -2.682Employment -0.038 -0.702 -0.150 -3.457 -0.107 -0.596 -0.034 -0.291Growth Aspiration

0.235 1.350 0.137 1.002 -0.050 -0.151

VAT registered -0.253 2.835 -0.255 2.235 -0.248 2.207 -0.232 2.135FIRM AGE 4-10 0.735 2.858 0.326 1.358FIRM AGE 11-20 0.177 0.735 0.443 1.867 0.187 0.194AGE OF FIRM > 20

0.338 1.228 0.365 1.970 0.901 1.245 0.768 1.951

Wholesale -0.801 -2.626 -0.012 -0.056 -0.213 -0.432Motor Retail -1.050 -1.850 -0.015 -0.053Financial services 0.036 0.107 -0.130 -0.382 0.141 0.267Education 0.310 0.253 -0.330 -0.439North West 0.031 0.117 -0.198 -0.917South West 0.315 1.245 0.298 1.643 0.983 1.436 -0.010 -0.017West Midlands 0.554 2.180 0.330 1.513 -0.160 -0.237Scotland 0.290 1.470 0.193 1.298 0.494 0.628 -1.005 -1.617Post grad qualification

-0.035 -0.170 0.020 0.125 0.468 1.151

Degree -0.209 -1.202 0.097 0.758 0.442 1.519A-level 0.257 1.200 -0.018 -0.106 0.068 0.163GCSE -0.112 -0.490 -0.172 -1.010Exporting firm 0.306 2.171 0.448 3.695 0.974 2.053 0.618 2.111Sole Proprietor 0.252 1.276 -0.083 -0.538 0.144 0.231 0.606 1.405Partnership -0.407 -1.892 -0.075 -0.534 -11.645 -0.004 0.293 0.703OWNER > 60 0.331 1.556 0.239 1.270 -1.039 -1.332 -0.405 -0.736EMB director/owner.

0.536 2.608 0.053 0.263 0.531 0.749 -0.158 -0.353

Women - led 0.345 1.910 0.044 0.271 0.518 0.642 0.317 0.869Firm sought advice 0.259 1.752 0.176 1.568 0.808 1.363 0.153 0.586Deprivation Quintile 1

-0.133 -0.468 -0.070 -0.327 -0.995 -1.170 -0.547 -0.859

Deprivation Quintile 2

-0.006 -0.025 -0.039 -0.196 -0.524 -0.685 0.038 0.084

Deprivation Quintile 3

-0.349 -1.274 -0.174 -0.811 1.134 1.486 -0.702 -0.892

Deprivation Quintile 4

0.337 1.381 0.338 1.746 0.124 0.242

Deprivation Quintile 5

0.217 0.893 0.114 0.614 -0.491 -0.486 -0.459 -1.055

Observations(1400 responses)

615 732 87 177

53

Table A2.7: By type of finance sought – 2003Bank HP / Factor Other

Parameter Estimate t-statistic Estimate t-statistic Estimate t-statisticConstant 1.447 7.413 1.648 3.075 -1.061 -14.641

Women - led 0.170 1.159 -0.398 -1.107 -0.111 -2.116

Control VariablesEmployment 0.002 2.271 0.002 1.050 0.003 7.566Growth Aspiration

0.155 1.757 0.003 0.016 0.377 10.167

VAT registered

-0.309 -1.801 0.255 0.516 0.220 3.519

FIRM AGE=1 -0.702 -3.673 -0.046 -0.088 0.144 1.562FIRM AGE=2 -0.727 -4.351 -0.982 -2.045 -0.095 -1.124FIRM AGE=3 -0.295 -1.972 -0.586 -1.688 0.033 0.471FIRM AGE=4 -0.328 -1.683 -0.223 -0.466 0.134 1.487FIRM AGE=5 -0.191 -0.919 -0.608 -1.357 0.036 0.430FIRM AGE=6-10

-0.238 -1.859 0.232 0.687 -0.015 -0.302

Agriculture 0.129 0.572 5.381 0.001 0.275 3.022Construction 0.223 1.575 -0.081 -0.314 0.014 0.260Retail -0.017 -0.131 5.726 0.001 -0.202 -3.911Wholesale 0.453 2.818 0.091 0.226 -0.005 -0.081Hotels & Rest’nts

-0.113 -0.807 0.133 0.288 -0.135 -2.269

Post & communication

-0.345 -0.849 -0.915 -1.093 -0.071 -0.374

North East -0.055 -0.280 5.205 0.001 0.042 0.524North West 0.099 0.806 0.262 0.814 0.067 1.347Yorks and H. 0.152 1.169 -0.091 -0.339 0.054 1.055East Midlands 0.234 1.168 0.021 0.055 -0.018 -0.250West Midlands -0.005 -0.034 -0.667 -1.871 -0.024 -0.387Eastern region 0.061 0.384 -0.307 -0.813 -0.031 -0.467South West 0.490 2.799 -0.336 -1.061 0.141 2.199Wales 0.187 1.328 -0.190 -0.718 0.090 1.772Exporter -0.266 -3.049 -0.068 -0.330 -0.007 -0.186Sole Proprietor -0.072 -0.620 0.446 1.252 -0.222 -4.741Partnership 0.410 3.446 0.178 0.710 0.039 0.918EMB owner/director.

-0.191 -2.504 -0.298 -1.487 -0.086 -2.431

Some women directors

0.155 1.279 -0.155 -0.618 0.091 1.922

Total respondents =2330

0 1621 582 903 1076

0 181 11 2 213 411

0 331 1 1082 1623 1729

54

Table A2.8: By type of finance sought - 2004Equity Bank HP/factor Overdraft Other

Parameter Esti t-stati Esti t-stati Esti t-stati Esti t-stati Esti t-statiConstant 1.989 1.069 -1.235 -5.705 -2.137 -4.930 -0.871 -2.633 0.599 1.406

Women - led -0.356 -0.001 -0.111 -0.634 0.362 1.032 -0.080 -0.305 -0.007 -0.030

Control Variables Employment -0.693 -1.690 -0.095 -2.071 -0.002 -0.028 -0.065 -0.925 -0.220 -2.864Growth Aspiration

0.185 0.160 0.201 1.294 0.421 1.457 0.205 0.907 -0.211 -0.875

VAT registered -0.266 2.586 -0.241 2.318 -0.245 2.256 -0.251 2.433 -0.229 2.280FIRM AGE 4-10

-2.928 0.000 0.718 3.278 1.354 2.251 0.603 1.880 -0.720 -1.217

FIRM AGE 11-20

-7.602 -0.001 0.119 0.476 0.858 1.957 -0.024 -0.075 0.046 0.094

AGE OF FIRM > 20

-0.808 -0.635 0.506 2.430 0.611 1.700 0.828 2.915 0.125 0.396

Wholesale -8.481 -0.001 0.119 0.551 -0.182 -0.358 -0.520 -1.644 -1.066 -2.253Motor Retail -2.642 -1.698 0.006 0.018 0.060 0.103 -0.251 -0.624Financial services

-6.610 -0.001 0.332 1.073 0.531 0.840 -0.118 -0.294 -0.075 -0.124

Education -0.262 -0.377 -6.288 -0.001 -0.554 -1.191North West -0.033 -0.038 -0.572 -2.031 -0.071 -0.184 0.067 0.198 -0.529 -1.389South West -0.191 -0.175 0.367 1.918 0.328 0.992 -0.107 -0.340 0.203 0.532West Midlands -8.692 -0.001 0.324 1.446 0.035 0.073 0.182 0.419 -0.001 -0.004Scotland -1.610 -1.378 0.026 0.161 -0.232 -0.617 0.300 1.278 0.356 1.367Post grad -0.868 -0.752 0.014 0.085 -0.395 -1.181 -0.155 -0.586 0.511 1.721Degree -2.033 -1.302 -0.103 -0.739 -0.079 -0.301 -0.323 -1.493 -0.026 -0.119A-level -0.915 -0.448 0.161 0.890 -0.783 -1.768 0.137 0.543 -0.490 -1.488GCSE -0.035 -0.020 -0.238 -1.280 -0.443 -1.202 -0.447 -1.531 -0.294 -0.800Exporter 0.465 0.465 0.444 3.709 0.513 2.151 0.383 2.137 0.335 1.585Sole Proprietor -6.934 -0.001 0.239 1.531 0.238 0.712 -0.028 -0.115 -0.270 -0.975Partnerships -2.692 -1.660 -0.149 -0.926 -0.150 -0.464 -0.178 -0.696 -0.269 -1.052OWNER > 60 0.969 0.634 0.250 1.305 0.757 2.136 0.292 0.989 -0.516 -1.426EMB owner/director

-0.783 0.000 0.181 0.956 0.605 1.409 -0.120 -0.342 0.264 0.918

Firm sought advice

1.960 2.061 0.197 1.651 0.668 2.876 0.148 0.806 -0.074 -0.339

Deprivation Quintile 1

0.033 0.142 -0.071 -0.147 -0.148 -0.430 -0.264 -0.545

Deprivation Quintile 2

-0.031 -0.154 0.063 0.161 -0.451 -1.373 0.133 0.340

Deprivation Quintile 3

-0.096 -0.426 0.247 0.671 -0.249 -0.720 -0.595 -1.366

Deprivation Quintile 4

0.127 0.605 0.828 2.128 0.482 1.611 0.125 0.335

Deprivation Quintile 5

0.001 0.006 0.082 0.216 -0.101 -0.326 0.365 1.156

Freq: 1234

40377

657 52 27 66

256111314

231 30 16 28

129182245

55

Table A2.9: Differences by type of organisation – 2003SP Partnership Company

Parameter Estimate t-statistic Estimate t-statistic Estimate t-statisticConstant -1.318 -12.908 -1.248 -7.408 -1.168 -8.379

Women - led -0.150 -1.496 -0.120 -0.877 -0.105 -1.459

Control VariablesEMPS 0.010 2.743 0.008 4.898 0.003 6.297Growth Aspiration

0.556 5.133 0.248 2.644 0.365 8.209

VAT registered

0.046 0.542 0.277 1.773 0.422 3.137

FIRM AGE=1 0.042 0.233 0.510 2.668 0.036 0.265FIRM AGE=2 -0.022 -0.109 -0.130 -0.496 -0.108 -1.079FIRM AGE=3 -0.361 -1.315 0.133 0.605 0.031 0.399FIRM AGE=4 0.127 0.657 0.245 1.022 0.104 0.909FIRM AGE=5 0.225 1.439 -0.068 -0.316 -0.036 -0.311FIRM AGE=6-10

0.031 0.347 0.139 1.376 -0.130 -1.580

Agriculture 0.130 0.692 0.549 3.540 0.270 1.790Construction 0.080 0.630 0.187 1.420 -0.051 -0.791Retail 0.024 0.207 -0.109 -0.943 -0.307 -4.514Wholesale 0.281 0.940 -0.040 -0.185 -0.040 -0.585Hotels & Rest’nts

0.041 0.369 -0.134 -1.278 -0.248 -2.290

Post & communication

-0.573 -1.283 -0.202 -0.453 0.273 1.026

North East 0.186 0.777 -0.334 -1.438 0.068 0.735North West 0.043 0.352 0.131 0.995 0.040 0.652Yorks and H 0.079 0.540 0.178 1.307 0.018 0.291East Midlands -0.291 -1.566 0.233 1.592 -0.070 -0.716West Midlands -0.040 -0.315 0.112 0.854 -0.080 -0.890Eastern region 0.144 0.928 -0.045 -0.322 -0.062 -0.711South West 0.267 1.774 0.345 2.448 0.050 0.602WALES 0.225 1.887 0.011 0.094 0.081 1.249Exporter 0.118 0.952 0.041 0.399 -0.047 -1.114EMB owner/d. -0.051 -0.564 -0.172 -1.809 -0.071 -1.637Some women directors

-0.257 -0.562 0.054 0.496 0.079 1.469

Total respondents 2330

0 2351 20 2 133 54

0 3461 182 113 38

0 1140 1 1292 723 180

per cent correct

88 90 78

56

Table A2.10: Differences across type of organisation – 2004SP partnership company

Parameter Estimate t-statistic Estimate t-statistic Estimate t-statisticConstant -1.019 -4.152 -1.090 -3.038 -1.117 -6.044

Women - led 0.120 0.570 -0.049 -0.114 0.229 1.630

Control Variables Employment -0.054 -0.646 -0.221 -2.608 -0.114 -3.077Growth Aspiration

0.139 0.687 0.370 1.283 0.132 1.011

VAT registered -0.258 2.687 -0.255 2.329 -0.243 2.183FIRM AGE 4-10

0.625 1.613 0.508 1.012 0.402 2.026

FIRM AGE 11-20

0.355 0.989 0.483 1.194 0.066 0.330

AGE OF FIRM > 20

0.402 1.018 1.228 3.099 0.286 1.792

Wholesale 0.057 0.155 0.763 1.913 -0.504 -2.389Motor Retail -0.675 -1.149 -0.039 -0.141Financial services

-0.299 -1.234

Education -0.540 -1.514North West -0.028 -0.076 -0.820 -1.248 -0.291 -1.529South West -0.002 -0.005 -6.157 -0.001 0.492 2.968West Midlands -0.068 -0.148 1.254 2.820 0.295 1.689Scotland 0.161 0.590 0.541 2.022 0.024 0.169Post graduate -0.215 -0.645 0.003 0.010 0.180 1.345Degree 0.068 0.279 -0.265 -0.985 0.145 1.289A-level 0.036 0.117 0.228 0.714 0.159 1.034GCSE 0.152 0.600 -0.461 -1.263 -0.221 -1.260Exporter 0.746 2.959 0.013 0.044 0.355 3.849OWNER > 60 0.113 0.322 0.218 0.566 0.236 1.560EMB owner 0.238 0.733 0.391 1.104 0.266 1.712Sought advice -0.038 -0.188 -0.002 -0.008 0.289 2.969Deprivation Quintile 1

-0.532 -0.860 0.698 1.495 -0.245 -1.341

Deprivation Quintile 2

0.376 1.198 0.349 0.879 -0.180 -1.051

Deprivation Quintile 3

0.806 2.168 -5.910 -0.001 -0.349 -1.914

Deprivation Quintile 4

0.691 2.026 0.470 1.235 0.139 0.789

Deprivation Quintile 5

-0.096 -0.260 0.344 0.720 0.094 0.611

Choice Frequency Frequency Frequency0 208 229 8651 17 17 732 9 11 543 35 15 91

57

Table A2.11: Differences By Gender 2003 (Multivariate Probit Models)Women-led Male-led

Parameter Estimate t-statistic Estimate t-statisticConstant -1.451 -6.806 -1.023 -13.228Employment (no) 0.002 1.391 0.003 7.431Growth Aspiration

0.422 3.396 0.379 9.711

VAT registered 0.323 1.858 0.199 2.943Firm age =1 0.230 0.975 0.120 1.193Firm age =2 -0.168 -0.633 -0.084 -0.939Firm age =3 0.164 0.779 0.019 0.252Firm age =4 0.490 1.772 0.098 1.024Firm age =5 0.017 0.071 0.045 0.500Firm age =6-10 0.174 1.286 -0.050 -0.905Agriculture 0.203 0.529 0.268 2.860Construction 0.248 0.805 -0.002 -0.030Retail 0.107 0.749 -0.250 -4.475Wholesale 0.278 1.006 -0.022 -0.340Hotels & Rest’nts -0.125 -0.721 -0.140 -2.202Post & communication

-0.195 -0.332 -0.066 -0.330

Exporting Firm -0.174 -1.167 0.003 0.071Sole Proprietorship

-0.319 -2.430 -0.210 -4.168

Partnership -0.068 -0.443 0.049 1.109EMB directors/owners

0.234 1.947 -0.112 -2.998

Some directors women

0.087 1.834

Regional DummiesNorth East 0.355 1.415 0.007 0.089North West 0.181 1.084 0.058 1.107Yorks and Humber

0.340 2.027 0.025 0.474

East Midlands -0.153 -0.562 -0.005 -0.072West Midlands -0.170 -0.846 -0.002 -0.028Eastern Region 0.277 1.210 -0.057 -0.830South West 0.246 1.179 0.130 1.919Wales 0.327 1.948 0.067 1.245

Choice: 0 135 9481 12 912 20 523 24 165 per cent correct 67 84

58

Table A2.12: Differences By Gender 2004 (Multivariate Probit Models)

Women-led Male-ledParameter Estimate t-statistic Estimate t-statisticConstant -0.780 -1.694 -1.126 -7.261Employment -0.255 -2.351 -0.104 -3.247Growth Aspiration

0.079 0.262 0.198 1.862

VAT registered -0.254 2.609 -0.237 2.156Firm age = 4-10 0.308 0.671 0.393 2.202Firm age = 11-20

-0.504 -0.750 0.207 1.278

Firm age > 20 0.055 0.117 0.463 3.291Wholesale -0.050 -0.111 -0.242 -1.417Motor Retail 0.952 0.899 -0.341 -1.385Financial services

-0.029 -0.138

Education -0.500 -0.688 -0.693 -1.652Post Graduate 0.463 1.237 0.042 0.345Degree 0.584 1.918 -0.002 -0.018A LEVEL 0.228 0.606 0.084 0.633GCSE -0.020 -0.052 -0.156 -1.142Exporting Firm -0.257 -0.938 0.436 5.017Sole Proprietor -0.283 -0.976 0.111 0.915Partnership -1.127 -2.536 -0.102 -0.909OWNER > 60 0.293 0.832 0.179 1.284EMB owner/directors

0.620 2.034 0.163 1.137

Firm sought advice

0.490 1.930 0.210 2.443

Deprivation Quintile 1

-0.001 -6.653 -0.076 -0.473

Deprivation Quintile 2

-0.108 -0.267 0.056 0.386

Deprivation Quintile 3

0.210 0.489 -0.195 -1.179

Deprivation Quintile 4

0.817 2.127 0.227 1.503

Deprivation Quintile 5

-0.012 -0.028 0.083 0.585

Regional DummiesNorth West 0.083 0.223 -0.322 -1.837South West 0.679 1.969 0.209 1.395West Midlands -0.068 -0.130 0.364 2.268Scotland -0.196 -0.511 0.132 1.143Choice0 144 11581 12 952 10 643 24 117

59

A2.5 The remainder of this section turns to the differences in the demand for finance. This utilises the information with respect to those firms that have sought finance, as to whether there are differences across gender in seeking finance, rather than in being offered it. The purpose of this is to address the problem that it may be the case that certain groups simply perceive it not to be worthwhile even applying for finance. This reflects the suggestion from the HSE that individual women considering business start-up are less likely to seek external finance than their counterparts who are men.

A2.6 The demand side analysis presented in this section is similar in structure to that presented above, focussing on the types of finance sought, where from and how often, rather than the difficulties faced. The procedure is similar to that outlined above, one first determines the best performing baseline model, and then adds the variables of interest. The estimation here initially are logit models, so the marginal effects rather than the simple coefficients are reported. Again some categories have been amalgamated from the original survey in order to build up large enough samples with sufficient within sample variation. Other variables are included in this analysis to capture further constraints, such as the number of hours per work the owner / MD claims to spend on paperwork.

60

Table A2.13: Demand effects: What sort of finance have firms sought 2003?

equity O/D bank HP/ factor loan other

Param Est t-stat Est t-stat Est t-stat Est t-stat Est t-stat Est t-stat

Constant -0.037 -0.004 -0.099 -11.983 -0.169 -13.774 -0.120 -8.400 -0.033 -8.117 -0.128 -11.304

Women - led -0.026 -0.004 -0.002 -0.681 -0.009 -1.262 -0.046 -3.437 0.002 0.582 0.029 2.610

Control VariablesEmployment 0.000 0.175 0.000 1.586 0.000 5.563 0.000 5.915 0.000 1.880 0.000 3.998

Growth Aspiration

0.002 2.146 0.014 3.637 0.052 6.953 0.016 4.406 0.003 1.898 0.029 6.055

VAT registered

0.027 0.003 0.004 1.175 0.008 1.736 0.049 3.135 0.002 0.837 0.022 2.249

FIRM AGE=1 -0.023 -0.003 -0.009 -0.343 0.059 3.004 -0.016 -0.495 0.013 3.802 0.039 2.662

FIRM AGE=2 0.002 1.355 0.024 2.354 0.003 0.204 -0.013 -0.818 0.006 1.400 -0.010 -.515

FIRM AGE=3 0.003 2.117 0.011 1.304 0.027 1.869 -0.009 -0.434 0.005 1.438 0.016 1.659

FIRM AGE=4 0.005 3.854 0.032 3.159 0.018 1.034 -0.007 -0.236 0.000 0.111 -0.002 .0995

FIRM AGE=5 0.003 1.728 0.012 1.030 0.011 0.614 -0.023 -1.173 -0.001 -0.139 0.017 1.237

FIRM AGE=6-10

0.001 0.343 -0.002 -0.262 0.000 0.067 -0.005 -0.496 0.001 0.304 0.008 .8034

Agriculture -0.025 -0.003 0.034 3.287 0.057 2.565 0.011 0.950 0.002 0.514 -0.043 -1.587

Construction -0.026 -0.005 0.019 2.310 0.006 -0.064 -0.004 -0.939 -0.004 -1.066 -0.024 -2.562

Retail -0.027 -0.004 0.013 0.874 0.010 -0.875 -0.058 -4.922 -0.009 -2.123 -0.037 -4.354

Wholesale 0.000 -0.190 0.020 2.297 0.000 -0.317 -0.026 -2.603 -0.004 -0.978 -0.014 -1.501

Hotels & Rest’nts

-0.026 -0.004 -0.027 -2.559 0.042 1.575 -0.056 -3.450 -0.008 -1.930 -0.015 -2.016

Post & communication

0.005 1.738 -0.699 -0.001 0.129 0.441 0.018 -0.488 -0.114 -0.001 0.074 .7485

North East 0.002 0.891 -0.009 -0.640 -0.020 -0.721 0.004 0.420 0.008 1.856 0.027 2.148

North West 0.000 -0.305 -0.009 -0.961 0.012 1.236 0.012 1.906 0.007 2.409 0.003 .6015

Yorks and Humber

-0.001 -0.360 0.000 0.106 0.012 1.096 0.013 1.867 0.002 0.597 -0.006 -.3882

East Midlands -0.027 -0.003 0.002 0.001 -0.012 -0.767 0.015 1.209 0.007 1.759 -0.009 -.7398

West Midlands -0.027 -0.003 -0.025 -2.279 0.013 0.244 -0.012 -1.153 0.010 2.838 -0.004 -.6423

Eastern region -0.026 -0.003 0.008 0.586 0.010 0.179 -0.006 -0.707 0.004 0.881 -0.022 -1.640

South West 0.000 -0.210 0.007 0.867 0.037 2.343 0.001 0.244 0.002 0.354 -0.022 -1.270

Wales -0.025 -0.006 -0.008 -0.857 -0.010 -0.264 0.018 2.824 0.006 2.010 0.055 7.393

Exporter 0.002 2.029 0.011 2.185 0.005 0.682 -0.013 -2.072 0.000 0.172 0.007 1.229

Sole Proprietor 0.002 0.968 -0.003 -1.113 -0.019 -2.585 -0.019 -2.861 -0.002 -1.051 -0.039 -4.628

Partnership 0.001 1.017 0.005 0.746 0.013 0.884 -0.015 -2.165 0.004 1.508 -0.018 -2.255

EMB owner/d. 0.000 0.301 0.001 -0.063 0.002 -0.291 -0.020 -2.385 -0.001 -0.342 -0.004 -.9122

Some directors women

0.001 0.529 0.010 1.810 0.002 0.541 -0.004 -0.308 0.001 0.552 0.018 2.535

Obs (5552=0) 12 277 884 301 49 430

per cent correct 83 68 72 71 86 78

61

Table A2.14: Demand effects: What sort of finance have firms sought 2004?equity O/D Bank HP/ factor loan other

Para Est t-stat Est t-stat Est t-stat Est t-stat Est t-stat Est t-stat

Constant -0.034 -7.927 -0.086 -11.55 -0.211 -15.82 -0.103 -12.31 -0.030 -7.391 -0.106 -12.71

Women - led -0.003 -0.890 0.000 -0.151 -0.011 -1.283 -0.020 -2.467 -0.002 -0.776 0.005 0.733

Control Variables

Employment 0.002 2.702 -0.002 -0.705 0.006 2.711 0.008 4.737 0.000 -0.407 0.000 0.133

Growth Aspiration

0.004 1.992 0.012 3.352 0.074 8.939 0.013 3.611 0.001 0.962 0.023 5.408

VAT registered

0.034 0.925 0.054 2.204 0.002 1.106 0.088 1.015 0.048 1.228 0.089 -2.669

Manufacturing -0.001 -0.400 0.002 0.313 0.021 1.646 0.006 0.966 0.004 1.211 -0.012 -1.695

Construction 0.000 0.014 0.012 1.390 0.005 0.026 -0.011 -1.274 -0.003 -0.577 -0.031 -3.411

Retail -0.002 -0.383 0.010 0.889 0.014 0.361 -0.020 -1.735 0.005 1.038 -0.053 -3.362

Wholesale 0.000 -0.141 0.000 -0.148 -0.006 -0.543 -0.018 -1.626 0.006 1.717 -0.016 -1.951

Motor retail 0.008 1.104 0.030 1.200 0.051 -1.019 0.004 -1.660 0.002 -0.308 -0.554 -0.002

Hotels -0.007 -1.238 -0.010 -1.131 0.024 1.370 -0.010 -1.253 -0.001 -0.383 -0.018 -2.738

Financial services

0.004 0.656 0.008 0.349 -0.049 -2.477 -0.029 -2.211 -0.001 -0.394 -0.039 -2.957

Real Estate 0.004 1.424 0.004 0.514 0.007 0.331 -0.005 -0.722 0.000 -0.123 -0.028 -4.028

Health & social work

-0.003 -0.594 -0.017 -1.424 0.000 -0.066 -0.005 -0.512 0.000 0.060 0.010 1.513

Eastern region 0.003 0.626 -0.021 -2.009 0.008 -0.477 -0.016 -1.454 0.004 0.664 -0.020 -1.634

London 0.001 -0.218 -0.012 -1.694 -0.033 -2.742 -0.024 -2.566 0.000 -0.290 -0.003 -0.647

North East 0.001 -0.167 0.007 0.724 -0.029 -1.751 -0.023 -1.599 0.007 1.181 0.015 1.390

North West 0.000 -0.230 -0.005 -0.522 -0.010 -0.681 0.000 0.137 0.005 1.012 0.011 1.021

South East 0.000 -0.407 -0.007 -0.971 -0.008 -1.488 0.001 -0.018 0.001 0.048 -0.009 -1.239

South West 0.004 0.630 -0.007 -0.616 -0.007 -0.891 0.005 0.626 0.004 0.776 0.009 0.570

West Midlands 0.001 0.051 -0.022 -1.841 -0.022 -1.503 -0.018 -1.314 0.007 1.368 0.021 1.950

Yorks and Humber

0.004 0.646 -0.008 -0.762 -0.022 -1.312 0.004 0.427 -0.003 -0.509 0.017 1.534

East Wales 0.001 0.276 -0.005 -0.603 -0.023 -1.030 -0.004 -0.595 0.004 0.853 0.010 1.030

West Wales -0.001 -0.002 -0.012 -1.173 -0.018 -0.613 0.000 -0.146 0.004 0.983 0.026 2.984

Scotland -0.001 -0.188 0.000 -0.145 -0.008 -0.307 -0.006 -0.893 -0.004 -0.628 0.000 0.005

Post Graduate 0.003 1.063 0.008 1.558 0.037 3.714 0.011 2.096 0.003 1.392 0.012 2.129

Degree 0.004 2.066 0.003 0.874 0.012 1.699 -0.004 -0.299 0.002 0.872 0.021 4.555

A-level 0.002 0.769 0.009 1.456 0.000 0.211 -0.001 0.004 0.002 0.611 0.009 1.359

GCSE -0.001 -0.168 0.003 0.593 0.023 2.143 0.003 0.733 0.002 0.829 0.003 0.552

Exporter 0.002 1.142 0.007 1.275 -0.005 -0.776 -0.017 -2.864 0.004 2.054 -0.001 -0.300

EMB owner/d. 0.000 -0.005 0.005 0.653 0.018 1.514 -0.010 -0.789 -0.003 -0.632 0.008 1.472

Deprivation Quintile 1

0.002 -0.924 0.003 -0.222 0.019 -2.012 0.003 0.098 0.000 -0.102 0.023 -0.579

Deprivation Quintile 2

-0.003 -0.032 0.000 0.478 -0.030 -0.850 0.003 0.636 0.000 0.178 -0.005 -0.665

Deprivation Quintile 3

0.000 -0.583 0.004 -0.054 -0.012 2.895 0.006 0.296 0.001 -1.096 -0.006 -1.361

Deprivation Quintile 4

-0.001 -0.003 0.002 -0.064 -0.039 -1.638 0.005 -0.170 -0.004 -0.756 -0.010 -0.558

Deprivation Quintile 5

-0.094 -0.962 0.004 0.470 -0.011 0.343 0.004 1.704 -0.002 -0.762 0.000 1.120

Sought advice -0.004 1.455 0.003 0.929 0.002 3.029 0.014 0.979 -0.003 0.376 0.007 4.995

Positives 38 224 746 283 44 214

62

Table A2.15: Multiple attempts: “how many times have you sought finance” 2003?

Parameter Estimate t-statisticConstant -0.084 -0.613

Women - led 0.066 0.678

Control VariablesEmployment 0.000 0.066Growth Aspiration 0.039 0.673VAT registered -0.071 -0.580AGE=1 0.149 1.014AGE=2 0.262 1.937AGE=3 0.211 1.953AGE=4 0.258 1.761AGE=5 0.365 2.592AGE=610 0.136 1.502Agriculture 0.215 1.500Construction 0.117 1.333Retail 0.405 4.322Wholesale 0.225 2.165Hotels & Rest’nts 0.254 2.356Post & communication 0.139 0.436North East -0.222 -1.599North West -0.074 -0.890Yorks and Humber -0.336 -3.732East Midlands -0.095 -0.739West Midlands -0.088 -0.799Eastern region -0.079 -0.678South West -0.086 -0.803Wales -0.227 -2.730Exporter 0.114 1.886Sole Proprietor -0.129 -1.497Partnership 0.077 1.053Some directors women 0.072 0.961ChoicesZero 1079once 668More than once 321

63

Table A2.16: Ordered Probit “have you tried to obtain finance” no, yes once, more than once 2004?

Once More than onceParameter Estimate t-statistic Estimate t-statisticConstant -1.262 -17.301 -1.261 -17.292

Women-led -0.097 -1.875 -0.095 -1.853

Control VariablesEmployment 0.002 3.867 0.002 3.875Growth Aspiration 0.058 1.416 0.056 1.414VAT registered 0.449 11.287 0.448 11.269Eastern region -0.176 -1.965 -0.172 -1.916London -0.213 -2.763 -0.201 -2.614North East -0.001 -0.006 0.003 0.030North West -0.002 -0.021 0.004 0.051South East -0.143 -1.813 -0.137 -1.742South West -0.040 -0.463 -0.037 -0.423West Midlands -0.107 -1.144 -0.107 -1.148Yorks and Humber -0.017 -0.177 -0.012 -0.129East Wales -0.099 -1.106 -0.096 -1.070West Wales 0.055 0.634 0.056 0.640Scotland -0.068 -0.871 -0.066 -0.841Highlands -0.001 -0.008 0.000 -0.001Post graduate 0.314 5.568 0.317 5.620Degree 0.131 3.015 0.134 3.093A-level 0.072 1.213 0.072 1.225GCSE 0.125 2.232 0.124 2.222Exporting firm 0.008 0.185 0.006 0.158EMB owner/director 0.187 2.771Mixed: White and Black African

0.506 1.586

Pakistani 0.293 2.498Black African 0.045 0.147Chinese -0.008 -0.026Sought advice 0.200 4.813 0.200 4.818PAPERHRS 0.000 0.364 0.000 0.377

Positives (total sample 7227)

1111 516

64

A2.3 Modelling Using the UKSMEF 2004A2.7 In this section we focus on evidence from the UKSMEF 2004. Our approach

follows closely that adopted in the ASBS focussing first on the supply side and then exploring women-led firms’ demand for finance.

A2.8 In terms of the supply-side, our main aim in this section is to analyse whether, controlling for firm and owner’s characteristics, gender influences the experienced barriers to finance. Our approach is based on a series of bivariate probit models of the probability of experiencing different types of barriers to accessing finance. In particular, these models were based on whether the firm had faced any barrier to finance – i.e. whether the firm faced either discouragement in seeking finance or whether it faced denial in obtaining finance. In each case we use a wide range of control factors and also include some interaction terms to capture potential contingencies between, say, gender and ethnicity.

A2.9 The coefficients reported ( ) here are the marginal effects, (i.e. the ceteris paribus change in the probability of an establishment facing discouragement with respect to a change in each determining variable). Note a stepwise regression procedure was adopted with variables retained in the model that had associated parameter estimates significant at the 30 per cent or better level, while restricting the main variables of interest to be included in the final specification. Two models are considered: the probability of facing discouragement (Table A2.17) and the probability of facing refusals when applying for finance (Table A2.18). Both are discussed in Section 4.3.2. The demand side is dealt with in Tables A2.19 and A2.20 and discussed in Section 4.3.3.

65

Table A2.17: Modelling the Probability of Facing Discouragement when Applying for Finance

Model 1 Model 2t-stat T-stat

Women (d) -0.012*** (-3.556) -0.013*** (-3.373)

Ethnic Minority (d) 0.020 (1.413) 0.024 (1.142)Women*Ethnic (d) 0.009 (0.330)Ethnic* Deprived (d) -0.005 (-0.610)Start-up (d) 0.036 (1.312) 0.036 (1.308)Firm CharacteristicsDeprived Area(d) -0.006 (-1.026) -0.004 (-0.696)Age business (+15) (d) -0.022** (-2.886) -0.022** (-2.916)City (d) 0.017 (1.163) 0.015 (1.135)Town (d) 0.009 (1.229) 0.009 (1.158)Advice (d) 0.004 (0.738) 0.004 (0.724)Member Org. (d) 0.005 (0.766) 0.005 (0.745)VAT Registered (d) -0.016* (-2.018) -0.016* (-2.054)Company (d) 0.008 (1.102) 0.008 (1.150)Owner CharacteristicsA-levels (d) -0.010** (-3.073) -0.010** (-3.004)Vocational qualif. (d) -0.007 (-1.519) -0.008 (-1.609)Age+46 (d) 0.004 (0.787) 0.005 (0.840)

Health (d) -0.010** (-3.059) -0.010** (-3.077)Hotels (d) -0.008* (-2.274) -0.008* (-2.262)Real Estate (d) -0.019** (-3.083) -0.018** (-3.083)Wholesale (d) -0.019*** (-4.144) -0.019*** (-4.100)East Midlands (d) 0.112 (1.347) 0.114 (1.348)London (d) 0.080* (2.058) 0.080* (2.049)North East (d) -0.011** (-3.141) -0.011** (-3.165)Northern Ireland (d) 0.182 (1.414) 0.182 (1.360)North West (d) 0.069 (1.215) 0.071 (1.218)Scotland (d) 0.020 (0.775) 0.020 (0.771)South East (d) 0.061 (1.360) 0.061 (1.363)South West (d) 0.059 (1.057) 0.060 (1.065)Wales (d) 0.042 (0.849) 0.041 (0.839)West Midlands(d) 0.064 (1.192) 0.065 (1.197)Yorks (d) -0.008 (-1.336) -0.008 (-1.299)

Log-L -272.313 -271.927N 2039 2039chi2 94.05555 95.10849r2_p .2450003 .2460727Notes: weighted regression using the ‘probit’ procedure in STATA 9.1 was used with UKSMEF data. For variable definitions, see appendix. (d) marginals for discrete change of dummy variable from 0 to 1 * p<0.05, ** p<0.01, *** p<0.001

66

Table A2.18: Modelling the Probability of Experiencing Denials when Applying for Finance

Model 1 Model 2

t-stat t-stat

Women (d) -0.011 (-1.021) -0.003 (-0.226)

Ethnic Minority (d) 0.006 (0.321) 0.043 (1.034)Ethnic*Deprived (d) -0.022 (-1.632)Start-up business (d) -0.020** (-2.686) -0.021** (-2.841)Firm CharacteristicsDeprived Area (d) -0.002 (-0.095) 0.004 (0.196)Size (+10 employees) (d) 0.051* (2.489) 0.051* (2.402)Export (d) -0.023*** (-3.399) -0.024*** (-3.363)Age business (+15) (d) -0.046** (-2.651) -0.044** (-2.598)Growing_b (d) 0.023 (1.687) 0.026 (1.812)VAT registered (d) -0.011 (-0.840) -0.012 (-0.924)City (d) -0.017 (-1.551) -0.021 (-1.674)Town (d) -0.020 (-1.845) -0.028 (-1.786)Rural (d) -0.012 (-0.724)Owner CharacteristicsNo qualification (d) 0.010 (0.700) 0.010 (0.739)Experience +16 (d) 0.033** (2.587) 0.034** (2.666)

Construction (d) 0.030 (0.660) 0.030 (0.651)Health & Social Work (d) 0.048 (0.643) 0.045 (0.630)Hotels & Restaurants(d) 0.089 (0.966) 0.080 (0.897)Manufacturing (d) 0.073 (0.934) 0.064 (0.873)Other Community Actv.(d) 0.178 (1.404) 0.164 (1.333)Real Estate (d) 0.039 (1.044) 0.038 (0.995)Transport (d) 0.117 (1.101) 0.105 (1.040)Wholesale & Retail (d) 0.036 (0.753) 0.035 (0.736)East Midlands (d) 0.047 (1.391) 0.049 (1.404)London (d) 0.113** (2.588) 0.096** (2.656)North West (d) -0.013 (-1.107) -0.014 (-1.181)South East (d) 0.028 (0.971) 0.030 (1.003)South West (d) 0.048 (1.165) 0.051 (1.219)Wales (d) 0.080 (1.328) 0.078 (1.291)

Log-L -352.828 -348.799 N 2039 2009 chi2 116.4745 116.9478 r2_p .1918502 .2003651 Notes: weighted regression using the ‘probit’ procedure in STATA 9.1 was used with UKSMEF data. For variable definitions, see appendix. (d) marginals for discrete change of dummy variable from 0 to 1 * p<0.05, ** p<0.01, *** p<0.001

67

Table A2.19: Modelling the Probability of Applying for New External Finance

Model 1 Model 2

t-stat t-stat

Women (d) -0.004 (-0.080) -0.005 (-0.096)

Ethnic Minority (d) 0.021 (0.290) -0.002 (-0.020)Woman*Ethnicity (d) 0.009 (0.047)Ethnic*deprived (d) 0.044 (0.304)Firm CharacteristicsDeprived Area (d) -0.009 (-0.170) -0.015 (-0.272)Size [+10 employees] (d) 0.181*** (4.792) 0.182*** (4.824)Age-business [4-15] (d) -0.181** (-2.740) -0.181** (-2.739)Age-business [+15] (d) -0.351*** (-5.205) -0.352*** (-5.205)Towns (d) -0.112** (-2.623) -0.112** (-2.612)Advice (d) 0.112* (2.520) 0.112* (2.515)Growing-bus (d) 0.081 (1.910) 0.079 (1.885)VAT-registered (d) 0.210*** (4.380) 0.210*** (4.388)Partnership (d) 0.065 (1.177) 0.065 (1.180)Owner CharacteristicsAge+46 (d) -0.128* (-2.422) -0.128* (-2.427)Experience [+15] (d) 0.075 (1.356) 0.075 (1.358)A-levels (d) -0.061 (-0.757) -0.062 (-0.772)GCSE (d) -0.070 (-0.925) -0.070 (-0.937)Vocational qualif. (d) -0.059 (-0.963) -0.059 (-0.956)No qualification (d) 0.062 (0.828) 0.062 (0.824)

Construction (d) -0.116 (-1.738) -0.117 (-1.746)Health (d) -0.152 (-1.745) -0.152 (-1.741)Hotels (d) -0.089 (-1.303) -0.089 (-1.294)Other (d) -0.105 (-1.328) -0.105 (-1.331)Real Estate (d) -0.085 (-1.326) -0.086 (-1.351)Wholesale (d) -0.094 (-1.457) -0.095 (-1.463)East Midlands (d) 0.063 (0.902) 0.063 (0.900)London (d) 0.120 (1.714) 0.121 (1.720)Northern Ireland (d) 0.114 (1.139) 0.112 (1.107)South East (d) 0.085 (1.168) 0.085 (1.163)South West (d) 0.113 (1.538) 0.113 (1.533)Wales (d) 0.134 (1.767) 0.135 (1.783)Yorks (d) 0.088 (1.148) 0.089 (1.152)Log-L -1205.614 -1205.460 N 2039 2039 chi2 151.2654 152.3215 r2_p .1404645 .1405744

Notes: weighted regression using the ‘probit’ procedure in STATA 9.1 was used with UKSMEF data. For variable definitions, see appendix. (d) marginals for discrete change of dummy variable from 0 to 1 * p<0.05, ** p<0.01, *** p<0.001

68

Table A2.20: Demand by Type of Finance sought

DependentVariable

Overdraft Loan Asset-based finance

Equity

t-stat t-stat t-stat t-stat

Woman (d) -0.06 -0.93 -0.17* -2.48 -0.00 -1.75 -0.00 -0.78

Ethnic Minor. (d) 0.08 0.85 -0.14 -1.60 0.01 0.78 0.01 0.71Start-up (d) 0.17 1.48 -0.00 -1.71Firm CharacteristicsDeprived Area (d) -0.03 -0.45 0.00 0.03 -0.00 -1.77 -0.00 -0.70Size [+15] (d) 0.06 1.21 0.04 0.82 0.02 1.61 -0.00 -0.95Age_b [4-15] (d) -0.23*** -3.75 0.26** 2.85 -0.00 -1.42 Age_b [+15] (d) -0.27*** -4.37 0.14 1.46 -0.01 -1.76 0.01 1.15Cities (d) 0.21 1.71 -0.29*** -3.46 0.02 1.23 -0.00 -1.44Town (d) 0.26* 2.42 -0.23* -2.23 0.02 1.23 -0.00 -0.83Rural (d) 0.35** 3.03 -0.27* -2.55 0.01 0.78 0.01 0.66Partnerships 0.00 0.76 0.08 1.72Companies 0.08 1.21 -0.16* -2.45 0.01 1.35 0.15*** 3.80Advice (d) 0.01* 2.49 -0.00 -0.84Member Org. (d) -0.01 -1.60Export (d) -0.07 -0.90 -0.08 -0.98 0.01 0.97Growing-b (d) -0.05 -0.93 0.10 1.93 0.01* 2.54 0.00 0.91VAT-reg.(d) -0.26*** -3.39 0.09 1.18 0.01 1.51Owner CharacteristicsAge46 (d) -0.06 -0.98 0.00 1.73 0.01 1.87Experience [+15] -0.00 -1.22 -0.01 -1.59A-levels (d) -0.11 -1.45 -0.11 -1.40 -0.00 -0.77 0.01 0.98GCSE’s (d) -0.06 -0.89 Vocational (d) -0.04 -0.71 -0.00 -1.39 Others (d) 0.02 1.17

Construction 0.13 1.32 0.24* 2.27 -0.01* -2.55 Health 0.34** 2.79 -0.00 -1.08 0.06 0.96Hotels 0.16 1.33 0.51*** 8.50 Manufacturing 0.17 1.73 0.28* 2.41 0.03 0.82Other 0.28* 2.40 0.37*** 3.44 -0.00 -1.09 -0.01 -1.77Real Estate 0.10 1.16 0.38*** 3.68 -0.01 -1.32 0.02* 1.97Transport 0.25* 2.02 0.01 1.03 -0.00 -1.19Wholesale 0.10 1.11 0.42*** 4.74 -0.00 -1.42 East Midlands (d) 0.10 0.87 0.01 0.63 -0.00 -1.39London (d) 0.16 1.35 -0.15 -1.53 0.01 1.04 0.02 1.23North East (d) 0.22 1.76 -0.18 -1.88 North. Ireland (d) 0.17 1.12 -0.12 -1.13 0.02 0.93 North West (d) -0.15 -1.78 -0.00 -1.18 -0.00 -1.23Scotland (d) 0.10 0.76 -0.12 -1.36 -0.00* -2.09 -0.00 -0.96South East (d) 0.14 1.31 -0.16 -1.82 0.00 0.58 South West (d) 0.11 0.96 -0.15 -1.85 0.02 1.15Wales (d) 0.13 1.00 0.01 0.89 -0.00 -0.84West Midlands(d) 0.13 1.02 0.01 0.68 0.02 1.04Yorks (d) 0.14 1.04 -0.18* -2.09 0.02 0.77

Log-L -640.76 -701.53 -93.43 -171.92 N 1191 1191 1191 1191 chi2 73.5027 79.3934 245.447 117.543 r2_p .149192 .134350 .298973 .423439Notes: weighted regression using the ‘probit’ procedure in STATA 9.1 was used with UKSMEF data. For variable definitions, see appendix. (d) marginals for discrete change of dummy variable from 0 to 1 * p<0.05, ** p<0.01, *** p<0.001

69

Annex 3: Variable Definitions and Construction

A3.1 GEM 2004A3.1 Data was taken from the GEM 2004 data file with sample observations being

weighted to give results representative of the UK working age population (variable: weight_L). This option was chosen as being indicative of the population most likely themselves to be involved in business start-up. The gender indicator is derived directly from the database. The derivation of other variables used in the analysis are summarised in Table A3.1.

Table A3.1: Variable Descriptions from the GEM 2004 Database

Variable name Description Variable Type

GEM Question No.

Dependent Variables SLFUND Excluding money from family and

friends, would a lack of external funding prevent you from starting up a business?

Dummy variable: 1=yes.

q1o2

SSTUP You are, alone or with others, currently trying to start a new business, including any type of self-employment or selling any goods or services to others

Dummy variable: 1=yes.

Q1a

SOM You are, alone or with others, currently the owner of a business you help manage; or you are self-employed or selling any goods or services to others

Dummy variable: 1=yes.

Q1c

SESTUP You are, alone or with others, expecting to start a new business, including any type of self-employment, within the next three years

Dummy variable: 1=yes.

Q1e

Factors of InterestGENDER Gender Indicator (Male=1) Dummy

variable: 1= male

SETH Ethnicity (White British = 0) Dummy variable: 1= ethnic minority

SIMD_A Index of multiple deprivation, combined variable for England and Wales and Scotland. Index value for SOA in which postcode of respondent is located. 1 is most deprived

IMD value

70

SDEP15 Postcode in most deprived 15 per cent of SOAs

Dummy variable: 1= deprived ward

Control FactorsSREGDUM1.12 Regional dummies: 1 South West; 2

South East; 3 London; 4 Eastern; 5 Wales; 6 West Midlands; 7 East Midlands; 8 Yorkshire and Humberside; 9 North West; 10 North East; 11 Scotland; 12 Northern Ireland

Dummy variable: 1= home region

Region

SEDUM1…5 Educational Attainment: 1 degree or higher; 2 A levels; 3 GCSEs; 4 other vocational qualification; 5 no qualifications

Dummy variable: 1= home region

UKREDUC

SINCDUM1..4 Household income quartiles: 1 lowest quartile; 2 second quartile; 3 3rd quartile; 4 upper quartile.

Dummy variable: 1= income quartile

UKHHINC

SAGE Age in Years Years AGESWSTDUM1..3 UK Working Status: 1 full-time (30

hours per week or more); 2 part-time (8-29 hours per week); 3 not working (8 hours or less per week).

Dummy variable: 1= working status

UKRWSTAT

SESTDUM1..3 Employment Status: 1 employee; 2 self employed; 3 business owner.

Dummy variable: 1=employment status

Q7b

SETDUM1 Have you ever taken part in business or enterprise training at school?

Dummy variable: 1=yes.

q6e01

SETDUM2 (443)Have you ever taken part in business or enterprise training atcollege or university?

Dummy variable: 1=yes.

q6e02

SETDUM3 Have you ever taken part in work experience in a small or mediumsized business whilst at school or college?

Dummy variable: 1=yes.

q6e03

SETDUM4 Have you ever taken part in a Government or public sector trainingcourse in business or enterprise skills ?

Dummy variable: 1=yes.

q6e04

WeightsWeight_L Weights to adjust to the 18 to 64

year old working population in the 2004 Census

Weight Weight_L

71

A3.2 HSE 2003 A3.2 The data for the empirical analysis have been drawn from the Household

Survey on Entrepreneurship 2003 data file with sample observations being weighted to give results representative of the UK working age population (variable: weight_1).

Table A3.2: Variable Descriptions for the HSE 2003 Database

Variable name CommentsDependent VariablesIndividual is a Doer

A dummy variables taking the value of 1 if the respondent is a Doer and 0 otherwise. This variable has been constructed by combining the dummy variables indicating Thinkers and Doers in the dataset.

Individual has sought finance

A dummy variable taking the value of 1 if the respondent is trying to get access to external finance and 0 otherwise. This variable has been constructed by combining the answers to Q13 for the Thinkers and Q40 for the Doers.

Difficulties obtaining Finance

A dummy variable taking the value of 1 if the respondent has experienced financial constraints and 0 otherwise. This variable has been constructed by combining the answers to Q16 (for the Thinkers) and Q44 (for the Doers).

Independent Variables

Education:A series of dummies indicate whether a particular level of academic educational qualification is the highest achieved by the individual concerned. The five dummies are, respectively, DEGREE, A LEVEL, GCSE, OTHER (taking the value of one if the respondent has other qualifications) and NONE (set to one if the respondent has no qualification - this is the omitted category).

Region : This enters as a continuous variable with a broad South-North orientation with the highest value indicating the North West.

Gender: This is a dummy variables set to one for women.Age: This is a set of dummy variables taking the value of 1 for each

respondent who is in a specific age bracket and 0 otherwise. Employment status:

This is a dummy variable taking the value of 1 if the respondent is NOT in either full- or part-time employment.

15 per cent most deprived wards:

This is a dummy taking the value of 1 if the respondent is located in one of the most deprived areas and 0 otherwise.

Attitude towards entrepreneurship:

This is a dummy variable taking the value of 0 if the respondent has a positive attitude towards entrepreneurship and 1 otherwise.

Past Experience: Variable takes the value of 1 if the respondent has previous entrepreneurial experience and 0 otherwise.

White: This is a dummy variable taking the value of 1 if the respondent is White and 0 otherwise.

Landlord: This is a dummy variable taking the value of 1 if the respondent does not own a house and 0 otherwise.

72

A3.3 Annual Small Business Survey 2003 and 2004A3.3 Data for the ASBS was analysed for both 2003 (including the ethnic minority

boost) and 2004. The 2003 data sample includes 8693 respondents of whom 2330 had sought finance at least once. The 2004 ASBS includes 7505 firms of which 1627 had sought finance at least once. ASBS analysis is undertaken unweighted.

Table A3.3: List of variables included in the ASBS models

Firm sizeLog(EMPS)20 Log of total employmentGROW Are you trying to grow the business? 1/0 1=yesVAT Is the business VAT registered?Firm age In years, omitted category is the oldest, >10 yearsRegionsEM East MidlandsNE North East NW North West England dummySCOT Scotland dummy (not H&I, included as HI)SW South West England dummyWALES WalesWM West Midlands England dummyYH Yorkshire and HumbersideEducation of owner/ MD (not in 2003 survey)PG Highest qualification of Owner / MD is a post graduate qualificationDEGREE Highest qualification of Owner / MD is a first degreeALEVEL Highest qualification of Owner / MD is A levelsGCSE Highest qualification of Owner / MD is GCSEEXPORT Does the firm sell overseas?Type of firm (default is company)SP Sole ProprietorPARTNERSHIP PartnershipOwner > 60 Owner / MD >60 (not in 2003 survey)Ethnic Groups White and Black Caribbean

White and Black AfricanMixed: White and Asianother mixed IndianPakistaniBangladeshiOther Asian Caribbean African other Black or Black BritishChinese

EMB Is the firm owned / directed by ethnic minoritiesFEMALE LED Is the firm women- led?SDW Some directors are women (data very sparse in 2004 survey)ADVICE Have you sought business advice?Quintiles of deprivationQUINT1 Least deprivedQUINT2QUINT3QUINT4QUINT5 Most deprivedDEPWALES “Deprived within the Welsh classification

20 Where employment of the firm is given as zero, this is replaced by 1 before logging. This seems reasonable as a sole proprietor is employed by the business.

73

A3.4 Variables Used in the UKSMEF 2004 AnalysisA3.4 Data was taken form the UKSMEF 2004 data file with sample observations

being weighted to give results representative of the UK population (variable: pweight). The gender indicator is derived directly from the database. Ethnicity is derived from a more detailed breakdown in the UKSMEF. All respondents other than White British are included in the “ethnic” group.

A3.5 We control also for the location of the business in a deprived area. The deprivation variable was constructed using the data on deprivation scores for England, Scotland and Wales. In particular, the variable “Deprived Area” represents those firms in the 4th quartile of the distribution of deprivation scores. The derivation of other variables used in the analysis are summarised in Table A1.4.

Table A3.4: Variable Descriptions used in the UKSMEF analysis

Variable Definitions UKSMEF

Dependent VariableSought-f Whether the business sought new external finance in the last 3 years

(1=yes) –see UKSMEF general notessought_f

Need-fin Whether the business sought finance or was discouraged to do so Need_fin

Discouraged Whether the business was discouraged from appling for external finance, excluding friends and family (1=yes)

discoura

Denial Whether the application for external finance the business made was totally denied

denied

Problems to finance

Business that rate the cost or the availability of finance as a problem to the business

n3_4

Barrier Whether the business was discouraged to apply for external finance or its application was totally denied

Variables of Interest

Gender Dummy coded 1 if business’ principal owner is a woman Sex

Ethnicity Dummy coded 1 if business owner belongs to an ethnic minority(White British=0)

s8_eth

Control VariablesFirm CharacteristicDeprived Area Dummy coded 1 if business is located in the most deprived area (in the

last quartile according to the multiple index of deprivation, combined variable for England, Wales and Scotland.)

ideng_scidsct_scidwls_sc

Age Business Business age broken down into age-bands (less than 4 years=1, [4,15]years =2, more than 15=3]

s11, s11ran

Size Business size broken down into size-bands (less than 10 employees=1, 10 or more than 10 employees=2)

s4quot_s

Legal status Legal status of the business (sole trader=1, partnership=2, company=3,) s7_statuExport Whether the business sold goods and services outside the UK (coded 1

if export share is > 0 ) or not s6_a

Location Location of business (major conurbation=1, city=2, town=3, rural=4) b3Member Org. Whether the business is member of any business organization (1=yes) b7Growing-bus. Growing business according to recent turnover growth (coded 1 if last

year turnover was inferior)Advice Dummy coded 1 if business has seek external advice n1VAT-register. Dummy coded 1 if the business is VAT registered (1=yes) S9Industry Dummy variable =1 if business belongs to a particular industry industryRegion Dummy variable =1 if business located in particular region region

74

Owner characteristicsAge Principal owner age broken down into age-bands age_rangQualifications Highest degree attained (Degree or higher=1, A-level=2, GCSE’s or

equivalent=3, Vocational qualify.=4, No qualification=5)a3, a14, a26

Experience Years of experience of principal owner broken down into bands (less than 1 year=1, [1-9]=2, [10-15]=3, >15=4)

a4, a15, a27

75

References

Bornheim, S.P. and Herbeck, T.H. (1998) A Research Note on the Theory of SME – Bank Relationships, Small Business Economics, 10 (4): 327-331

British Bankers’ Association, press release, 18 April 2004

Carter, S. and T. Cannon. (1992), Women as Entrepreneurs, London: Academic Press

Carter, S., Anderson, S. and Shaw, E. (2001) Women’s Business Ownership: A Review of the Academic, Popular and Internet Literature, Report to the Small Business Service, Department of Marketing, University of Strathclyde: Glasgow

Carter, S. and Rosa, P. (1998) The Financing of Male- and Female-owned Businesses, Entrepreneurship & Regional Development, 10 (3): 225-241

Centre for Women’s Business Research (2004) Women-owned Business in the US: Trends in the U.S. and 50 States, Centre for Women’s Business Research: Washington, DC

Chandler, G.N. and Hanks, S.H. (1998) An Examination of the Substitutability of Founders’ Human and Financial Capital in Emerging Business Ventures, Journal of Business Venturing, 13 (5): 353-369

Cosh, A. and Hughes, A. eds (2003) Enterprise Challenged: Policy and Performance in the British SME Sector 1999-2002, ESRC Centre for Business Research: Cambridge

Cressy, R. and Toivanen, O. (2001) Is there Adverse Selection in the Credit Market?, Venture Capital 3, 3 215-238

Curran J. and Blackburn R. (1993) Enterprise and the High St Bank, Small Business Research Centre, Kingston Business School, Kingston University.

Ennew, C. and McKechnie, S. (1998) The Nature of the Banking Relationship: A Comparison of the Experiences of Male and Female Small Business Owners, International Small Business Journal, 16 (3): 39-55

Fraser, S. (2005) Finance for Small and Medium Sized Enterprises: A Report on the 2004 UK Survey of SME Finances, Warwick Business School: Coventry

Goldberg, L.G. and White, L.J. (1998) De Novo Banks and Lending to Small Businesses: An Empirical Analysis, Journal of Banking & Finance, 22: 851-867

Green, P.G., Brush, C.G., Hart, M.M. and Saparito, P. (2001) Patterns of Venture Capital Funding: Is Gender a Factor?, Venture Capital, 3 (1): 63-83

Hamilton, R.T. and Fox, M.A. (1998) The Financing Preferences of Small Firm Owners, International Journal of Entrepreneurial Behaviour and Research, 4 (3): 239-248.

76

Hisrich, Robert D. and Brush, Candida, (1986) ‘The Woman Entrepreneur: Starting, Financing, and Managing a Successful New Business’ (Lexington, MA: Lexington Books, Inc.

Howorth, C.A. (2001) Small Firms’ Demand for Finance, International Small Business Journal, 19 (4): 78-86

Irwin, D. and Scott, J.M. (2006) The Barriers Faced by SMEs in Raising Finance from Banks, paper to be presented at the ISBE conference, November, Cardiff.

Johnson, S and Storey, D (1993) ‘Women Entrepreneurs’, London: Routledge.

Jones, E. (2001) The Industrial Finance Department: An Australian Experiment in Small Business Finance, Australian Economic History Review, 41 (2): 176-197

Koper, G (1993), "Women entrepreneurs and the granting of business credit", in Allen, S, Truman, C. (Eds),Women in Business: Perspectives on Women Entrepreneurs, Routledge, London.

Kotey, B. (1999) Debt Financing and Factors Internal to the Business, International Small Business Journal, 17 (3): 11-29

Meyer, L.H. (1998) The Present and Future Roles of Banks in Small Business Finance, Journal of Banking & Finance, 22 (7): 1109-1116

Olm, K, Carsrud, A.L., Alvey, L (1988), "The role of networks in new venture funding of female entrepreneurs: a continuing analysis", in Kirchoff, B.A, Long, W.A, McMullan, E., Vesper, K.H, Wetzel, W.E (Eds),Frontiers of Entrepreneurship Research, Babson College, Wellesley, MA, .

Riding, A.L, Swift, C.S (1990), "Women business owners and terms of credit: some empirical findings of the Canadian experience", Journal of Business Venturing, Vol. 5 No.5, pp.327-40.

Small Business Service (2004) ‘SBS Household Survey of Entrepreneurship 2003’, April.

Strahan, P.E. and Weston, J.P. (1998) Small Business Lending and the Changing Structure of the Banking Industry, Journal of Banking & Finance, 22 (7): 821-845

Van Auken, H.E, Gaskill, L.R, Kao, S (1993), "Acquisition of capital by women entrepreneurs: patterns of initial and refinancing capitalisation", Journal of Small Business and Entrepreneurship, Vol. 10 No.4, pp.44-55.

Verheul, I. and Thurik, R. (2001) Start-Up Capital: “Does Gender Matter?”, Small Business Economics, 16 (4): 329-345

Warren-Smith, I. and Jackson, C. (2004) Women Creating Wealth Through Rural Enterprise, International Journal of Entrepreneurial Behaviour & Research, vol. 10, no. 10, pp. 369-383Winker, P. (1999) Causes and Effects of Financing Constraints at the Firm Level, Small Business Economics, 12 (2): 169-181

77

78