Outside Enterpreneurial Capital

OUTSIDE ENTERPRENEURIAL CAPITAL*

Andy Cosh, Douglas Cumming and Alan Hughes

This article investigates factors that affect rejection rates in applications for outside finance amongdifferent types of investors (banks, venture capital funds, leasing firms, factoring firms, trade cus-tomers and suppliers, partners and working shareholders, private individuals and other sources),taking into account the non-randomness in a firm’s decision to seek outside finance. The datasupport the traditional pecking order theory. Further, the data indicate that firms seeking capital aretypically able to secure their requisite financing from at least one of the different available sources.However, external finance is often not available in the form that a firm would like.

This article engages with four interrelated empirical questions. First, what are thecharacteristics of privately held entrepreneurial firms that seek external (�outside�)finance, and what drives the request for capital from the different potential sources ofexternal finance: banks, venture capitalists, private individuals, leasing, factoring,suppliers/customers, partners/working shareholders, among other sources? Second,what are the factors that lead to rejection or acceptance of requests for external fi-nance, given this non-randomness in the types of firms that seek external finance – inthe spirit of Heckman (1976; 1979)? Third, as various forms of financing are far frombeing perfect substitutes, are there differences in the ability of firms to obtain capitalfrom the different available sources? Fourth, can firms obtain all of their desired capitalfrom the different available sources, even if it is not in the form they would like?

It is widely recognised that the decision to seek external finance and the type offinancing sought is related to information asymmetries faced by investors regarding theentrepreneurial firm’s quality; see for example, Jensen and Meckling (1976).1 Whereentrepreneurs have information that investors do not have, external equity finance(which dilutes the entrepreneurs� ownership share) is (given unused debt capacity)indicative of a low quality firm (Myers and Majluf, 1984); see also Myers (2000). Thisstream of research has derived a standard pecking order theory, in that firms prefer tofinance new projects with internal cash flows first and then, if necessary, thereafter seekexternal debt capital and lastly seek external equity capital. This rank ordering, how-ever, might be distorted by the fact that external suppliers of capital could add value totheir investee firms. For example, Garmaise (2000) shows that if investors are known topossess a greater ability to assess project quality relative to that of the entrepreneurialteam, external equity finance is indicative of a high quality firm. Also De Meza andWebb (1987, 1992) argue that banks may not be ill informed relative to new firms andshow that if we abandon the Stiglitz and Weiss (1981) assumption of mean preservingspreads of risk then their rationing results do not hold (see also De Meza and Webb,1999, 2000). In our article, we test these theoretical propositions by first taking into

* We owe special thanks to the Editor (Leonardo Felli), two anonymous referees and the seminarparticipants at York University, the 2006 Financial Management Association Annual Conference and the 2008Kauffman Foundation Conference on Entrepreneurial Finance.

1 This work also stems from seminal papers on adverse selection, including Ackerlof (1970), Stiglitz andWeiss (1981) and De Meza and Webb (1987, 1992).

The Economic Journal, 119 (October), 1494–1533. doi: 10.1111/j.1468-0297.2009.02270.x. � The Author(s). Journal compilation � Royal Economic

Society 2009. Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

[ 1494 ]

account the preliminary question of what drives the decision of a firm to seek externalfinancing. We then analyse the matching of different types of firms and investorsempirically, taking into account the non-randomness in the types of firms that soughtoutside capital.

Our analysis builds on an important but largely segmented literature in financialintermediation and entrepreneurial finance (briefly surveyed in Section 1). Academicstudies of the interaction between firms and their sources of capital focus, almostexclusively, on a single source of capital. Separate streams of literature have emerged inbank finance, lease finance, venture capital finance, private individual investor finance,supplier finance etc. In theoretical work, the need to focus on one (or in exceptionalcases, two) external capital sources is directly attributable to theoretical tractability(one cannot model everything). In empirical work, the focus on one or two capitalsources is largely attributable to data availability, since datasets are typically derivedfrom investors, particularly in the case of non-publicly traded businesses.

In building on this prior literature, the key component of our analysis is that weintroduce a very large dataset derived from private entrepreneurial firms themselvesand not from their investors. This empirical approach provides a number of advantagesfor the purpose of addressing the three central research questions outlined above.Perhaps most importantly, because we observe firms that did, and did not, seekexternal finance, we avoid Heckman (1976, 1979) sample selection problems in anal-ysing the role of external investors in facilitating the development of entrepreneurialfirms. Prior work on topic (see Section 1) is typically derived from one type of investorand thereby fails to consider the non-random selection process among those entre-preneurial firms that seek external finance, and the non-random selection processamong different types of potential investors.

A second key component of our analysis is that we have a very detailed and broad-based dataset from 2520 UK entrepreneurial firms in the period spanning 1996–7 themajority of which were formed in the 1980s and early 1990s (specific details are pro-vided in Section 3). A number of firms in the sample sought external finance from awide variety of potential outside investors and some successfully obtained externalfinance. This diversity in the data is of interest, as it allows us to carry out uniqueanalyses of the decision to seek external finance from a broad spectrum of differentinvestors. By contrast, datasets derived from investors typically provide insights into onlythose firms that applied for external finance with that investor and, further, typicallyonly those firms that were successful in obtaining finance. These limitations in allempirical prior work based on data from investors and not investees are overcome inour analysis of the new data introduced herein. We are therefore able to investigateissues that have been previously considered empirically intractable.

The data indicate the following primary key results. First, we identify factors that leadfirms to seek external finance. Controlling for a large number of other factors, we findthe most significant determinant of applications for external finance to be capitalexpenditures/profits. A firm is approximately 1% more likely to seek external financewhen capital expenditures are 20 times that of profits relative to 10 times profits. Thisindicates support for the traditional pecking order in which firms finance new projectsinternally before seeking external finance. Corporations are approximately 6% morelikely to seek external finance than partnerships or sole proprietorships. It is also quite

� The Author(s). Journal compilation � Royal Economic Society 2009

[O C T O B E R 2 0 0 9 ] 1 4 9 5O U T S I D E E N T E R P R E N E U R I A L C A P I T A L

noteworthy that, on a ranking from 1–4 with 4 being the highest growth orientation, anincrease in the ranking by 1 point tends to increase the probability of an application forexternal finance by approximately 11%. Further, firms that face 1,000 competitors areapproximately 2% more likely to apply for external finance than firms that face 100competitors.

Second, we show that systematic characteristics drive the amount of external financeactually sought, controlling for the non-randomness in the decision to seek externalfinance in the first place in the spirit of Heckman (1976, 1979). Firms seek a greateramount of external finance when they have greater capital expenditures/profits andhigher profit margins. Firms with more debt/assets also seek more external finance butthis effect is concave (that is, it increases at a decreasing rate and follows an inverted-Upattern).

In the third step of the analyses, we consider the percentage of external financeobtained. Larger firms in terms of their total assets are much more likely to obtain ahigher percentage of the total amount of finance sought. For instance, a firm with £1million in assets will obtain approximately 5% more of their desired capital than theaverage firm with £375,000 in assets. This is highly suggestive that smaller firms face acapital gap in terms of being able to obtain all of their desired capital. A firm thatrecently developed a novel innovation, by contrast, tends to obtain less of their desiredcapital. More specifically, a firm that was started within 7 years of applying for externalfinance and recently developed an innovation will obtain approximately 8% less oftheir desired capital but older firms that recently developed an innovation are notstatistically likely to receive less of their desired capital. Conversely, older firms that areless profitable and/or have not yet implemented professional directors are less likely toobtain all of their desired capital (albeit smaller firms do not face these limitations forprofits and/or professional directors).

We further explore these three main steps in the analyses outlined above by con-sidering the differences between banks and venture capital funds, as well as othersources of capital, including hire purchase or leasing firms, factoring/invoice dis-counting firms, trade customers/suppliers, partners/working shareholders, privateindividuals, often referred to as �angel investors� (Harrison and Mason, 2000), andother sources. Among the 952 of 2,520 firms (38%) in our sample that did seekexternal finance in the 1996–7 period considered, 775 approached banks, 474 ap-proached leasing firms, 151 approached factoring/invoice discounting firms, 138approached partners/working shareholders, 87 approached venture capital funds, 83approached private individuals and 53 approached trade customers/suppliers (and 67approached other sources).2 It is of interest that outright rejection rates were highestamong venture capital funds (46% rejection) and much higher than that for banks(17% outright rejection). The lowest rejection rate was among leasing firms (5%).Banks comprised the median and mean highest percentage of outside finance in termsof which type of source was approached and which type of source provided the finance.Among the 952 firms that did apply for external finance, the median number of

2 We refer to a firm �approaching� a potential source as making more than a trivial effort (for example, itinvolves more that just mailing in a business plan to the potential source for consideration), consistent withour survey evidence and related survey work undertaken by the Centre for Business Research at the Universityof Cambridge.

1496 [ O C T O B E RT H E E C O N O M I C J O U R N A L


different sources approached was 2 (and mean was 1.96). Also, firms with greatercapital expenditures/profits are more likely to approach a greater number of differenttypes of sources of finance. For the subset of the oldest firms (established before 1980in our sample), profit margins are significantly negatively related to the number ofdifferent sources of capital approached.

The new data and statistics introduced in this article provide completely newevidence on the importance of external capital for entrepreneurial firms, and thecomparative importance of different sources of capital. It is noteworthy that, amongthe 38% of firms that made non-trivial efforts to obtain external finance in oursample, the mean percentage of finance obtained (relative to the amount sought)was 80.7% and the median percentage was 100%. The data do not indicate thepresence of a substantial capital gap for the majority of firms seeking entrepre-neurial finance; rather, firms seeking capital are able to secure their requisitefinancing from at least one of the many different available sources. There are,however, differences in firms� ability to obtain finance in the form that they wouldlike. Even after controlling for selection effects, the data indicate banks are morelikely to provide the desired amount of capital to larger firms with more assets.Leasing firms, factor discounting/invoicing firms, trade customers/suppliers andpartners/working shareholders are more likely to provide the desired capital to firmswith higher profit margins. Profit margins are not statistically relevant to venturecapital funds and private individual investors; smaller firms are more likely to obtainfinance from private individuals, while young innovative firms seek external capitalfrom venture capital funds.

This article is organised as follows. Section 1 briefly reviews the related literature.Section 3 introduces the data considered in this article and presents a number of newfacts pertaining to entrepreneurial finance. The testable hypotheses within the contextof our new data and the prior literature are outlined in Section 2. Summary statisticsand multivariate analyses are provided in Section 4. Thereafter, limitations, alternativeexplanations and future research are discussed in Section 5. Concluding remarksfollow.

1. Related Literature

In this article we study the choice between internal versus external entrepreneurialfinance, with regard to a wide variety of sources of outside capital. Because we considera number of different types of investors, our work is related to a plethora of papersalong segmented streams of research in financial intermediation and entrepreneurialfinance. While it is, of course, beyond the scope of our article to review the entireliterature herein,3 we nevertheless provide a brief perspective of the contribution ofour analyses in the context of recent related prior work.

At the most generalisable level, a firm’s decision to seek external finance is related toinformation asymmetries faced by investors regarding the entrepreneurial firm’s

3 There are nevertheless recent surveys of the entrepreneurial finance and venture capital literature,including Denis (2004), Gompers and Lerner (2001), Smith and Smith (2000) and Berger and Udell (1998),as well as recent surveys of the banking literature, including Gorton and Winton (2003) and Berger andHumphrey (1997).

2009] 1497O U T S I D E E N T E R P R E N E U R I A L C A P I T A L


quality (Jensen and Meckling, 1976; Stiglitz and Weiss, 1981; De Meza and Webb, 1987,1992; Cosh and Hughes, 1994; De Meza, 2002; Carpenter and Petersen, 2002a,b). Myersand Majluf (1984) and Myers (2000) have derived a pecking order that results fromsuch information asymmetries between entrepreneurs and investors, whereby firms firstfinance new projects with internal cash flows4 and obtain external finance only wherenecessary because external finance is more costly when investors face informationasymmetries. External debt finance is preferred to external equity finance in thetraditional pecking order since equity involves a dilution of the entrepreneur’sownership share. Recent theoretical work, however, has shown this pecking order isreversed where investors have superior knowledge about the commercialisation processof an entrepreneur’s invention and/or add value to the entrepreneur’s project(Garmaise, 2000).

Pre-IPO outside entrepreneurial capital may be provided by banks, venture capi-talists, private individuals, leasing, factoring, suppliers/customers, partners/workingshareholders, among other sources. Different streams of the academic literature havebecome segmented by each of these different investor types for reasons of theoreticaltractability and data availability. Nevertheless, the fundamental questions consideredin the segmented literature significantly overlap in two primary ways that are perti-nent to our empirical analyses. First, there is a stream of literature on the ability of aninvestor to mitigate informational problems associated with small (and/or high-tech)businesses, and this has been considered in the context of banks (Berger et al., 2001;Berger and Udel, 1998), venture capital finance (Gompers and Lerner, 1999), angelfinance (Goldfarb et al., 2009; Wong, 2009; Harrison and Mason, 2000), lease finance(Porter, 1995; Myers et al., 1976, Yusopova, 2002), supplier finance (Tamari, 1970)etc. Second, there is a stream of literature on the contribution of the investor notonly in screening good from bad investees but also in assisting the development ofthe investee firm, which has been the focus of the literature in both angel finance(Wong, 2009; Harrison and Mason, 2000) and venture capital finance (Bergmannand Hege, 1998; Gompers and Lerner, 1999; Kortum and Lerner, 2000; Bascha andWalz, 2001; Mayer et al., 2005; Hege et al., 2003; Neus and Walz, 2005; Riyanto andSchwienbacher, 2006; Schwienbacher, 2008).

While some of the issues we consider have been addressed in prior work that issegmented by investor type, there is comparatively less rigorous data and evidenceabout the comparative ability of different types of investors to mitigate informationalproblems associated with outside finance and the comparative importance of differentsources of capital to different types of entrepreneurial firms. Moreover, given the factthat there is a non-randomness associated with the types of firms seeking externalcapital (in the spirit of Heckman, 1976, 1979), it is difficult to assess the importance ofdifferent types of investors to entrepreneurial firms without a comparative sample offirms that did not seek any external finance and without knowledge of alternativesources of external finance sought by the firms. We seek to overcome some of theselimitations in our analysis. We also keep in mind areas in which we face limitationsourselves, and thereby point to avenues for future research (discussed in detail inSection 5 below).

4 Relatedly, Carpenter et al. (1998) show that cash flow is a key determinant of inventory investment.



Our comparative focus is somewhat related to recent work that has made significantsteps in comparing two types of financial intermediaries. The recent literature hasprimarily been focused on comparing banks to venture capital funds.5 Theoreticalcontributions in this regard invariably conclude that venture capitalists are more skilledthan bank managers at screening potential investees, and providing greater value-added to their investee firms (Keuschnigg and Nielsen, 2004; Ueda, 2004). Thesepropositions are consistent with the large literature that focuses on venture capitalfunds themselves (Gompers and Lerner, 1999, 2001; Berger and Udell, 2002;Carpenter and Petersen, 2002a,b; Cressy, 1996, 2002; Manigart et al., 2006; Wright andLockett, 2003; Casamatta, 2003; Casamatta and Haritchabalet, 2007; Hsu, 2004; Kortumand Lerner, 2000; Kanniainen and Keuschnigg, 2003, 2004; Keuschnigg, 2004). Thiswork is also consistent with a literature comparing banks and venture capitalists;however, that empirical work does not consider other types of potential investors anddoes not consider the non-random selection process among those entrepreneurialfirms that have sought external finance nor the non-random selection process acrossdifferent types of investors. In other words, that work is not immune from the Heckman(1976, 1979) sample selection bias.6 Our dataset enables these limitations to be over-come (see also Robb and Robinson (2009) for recent work with US data) and enables abroader array of different types of investors to be considered.

Finally, it is noteworthy that our empirical analysis of outside entrepreneurial capitalis related to a very large theoretical and empirical literature on the decision of aprivately held business to go public; see Ritter and Welch (2002) for a recent review ofthat literature. Our analysis significantly differs from the IPO literature in that we focuson the much earlier decision in the life-cycle of a private business to seek externalprivate finance, long before it would be in a position to go public (and many firmsmight not want to go public). As mentioned, our data are not constrained by consid-eration of only one or two types of potential investors; rather, all types of externalsources of capital for early stage businesses are assessed. The data introduced in thisarticle are described in Section 3.

2. Testable Hypotheses

2.1. Primary Hypotheses

Our empirical analysis focuses on three primary hypotheses which are outlined in thisSection. Our first hypothesis pertains to the pecking order theory of capital structure.In particular, we are interested in knowing whether entrepreneurial firms do in factprefer to finance projects internally with their own profits prior to seeking externalfinance. Traditional pecking order theory (Myers and Majluf, 1984) is consistent withthis prediction, since entrepreneurs have information that investors do not have, and

5 In a recent theoretical contribution, Chemmanur and Chen (2002) provide an analysis on interactionbetween angel investors and venture capitalists; however, no prior paper has tested their model empirically. Ina recent empirical contribution, Cassar (2004) focuses on different sources of capital for small firms but doesnot consider selection effects in the spirit of Heckman (1976, 1979) and does not consider the interplaybetween access to outside capital and the development of the entrepreneurial firm.

6 Heckman (1976, 1979) shows that when this incidental truncation is accounted for in regression analyses,any or all of the sign, magnitude and statistical significance of regression coefficients may change.



therefore the information asymmetry faced by outside investors makes external financemore costly. By similar reasoning, external equity finance is more costly than externaldebt finance since equity involves the dilution of the entrepreneurs� ownership interestin the firm, and offers of equity finance therefore signal low quality.

Hypothesis 1a (Pecking Order and Decision to Seek External Funds): Firms withgreater profit margins are less likely to seek external finance, or at least will seek less external capitaland use internal profits to fund projects. Among firms that do seek external finance, firms withgreater profit margins will seek debt finance prior to equity finance.

Hypothesis 1b (Pecking Order and Success with Obtaining External FundsSought): Firms with greater profit margins are more likely to obtain the external financesought.

Our second central hypothesis considers the decision to seek finance as a functionof entrepreneurial firm as well as investor characteristics. The costs of seeking andobtaining external capital are higher where entrepreneurial firms exhibit greaterinformational asymmetries. That is, the search costs for capital, as well as the termsoffered to the investor, are less favourable for firms for which investors have moredifficulty mitigating information problems and expected agency costs. Amongyounger and innovative firms for which these costs are more pronounced (as gen-erally viewed in the literature; see e.g., Noe and Rebello (1996), firms will only bewilling to incur these costs if they have significant growth objectives. Hence, growthorientated firms are naturally more attracted to external finance as a result of theirwillingness to incur search costs and bear the price of external capital (Kon andStorey, 2003).

Just as entrepreneurial firm characteristics matter in respect of external financingdecisions, we might also expect differences across different sources of capital. It iswidely regarded that venture capital funds have a comparative advantage in mitigatinginformation asymmetries and agency costs (relative to banks, leasing and factoringfirms) and will be more likely to finance businesses for which risks are more pro-nounced but potential returns are higher; see Berger and Udell (1998).

Hypothesis 2a (Growth): Firms for which growth is the most significant objective will be, allelse being equal, more likely to seek external finance.

Hypothesis 2b (Banks versus Venture Capital Funds): Venture capital funds are morelikely to finance riskier firms with high growth objectives than banks and other sources of externalcapital.

With the wide range of sources of external entrepreneurial capital (banks, venturecapital funds, leasing firms, factoring firms, trade customers and suppliers, partnersand working shareholders, private individuals and other sources), it is natural to expectthat many firms are likely to raise the money that they want though not in the form thatthey would like. Clearly, raising some money is better than being denied access to thecapital market but the various forms of financing are far from perfect substitutes andmay imply (undesirable) constraints of various sorts, such as obtaining value addedadvice from a more reputable source of capital. For instance, relative to other types ofsources of capital, venture capital funds typically provide more intensive financial,



administrative, marketing and strategic advice, and facilitate a network of professionaladvisors to enable the entrepreneurial firm to succeed. Our data (described below inSection 3) enable us to identify which types of firms are more likely to face certainconstraints in obtaining their desired form of capital. We may expect that firms withless favourable prospects will face greater constraints in obtaining their desired form ofcapital.

Hypothesis 3 (Desired Source of Capital): Firms seeking more capital relative to theirprofits and firms with lower profit margins will have to approach a greater number of differentforms of finance to obtain their desired capital and will have greater difficulty in obtaining theirdesired level of external finance.

2.2. Control Variables

In considering the three primary hypotheses we control for a number of potentiallyrelevant factors. First, we consider the innovativeness of the industry, as well as theinnovativeness of the firm. Innovation is associated with asset intangibility, since high-tech firms are typically more innovative and have more intangible assets (Kortum andLerner, 2000). Higher asset intangibility is associated with more pronounced infor-mation asymmetries and agency costs, as well as potential hold-up costs, therebyincreasing the costs of external finance (Gompers and Lerner, 1999). On one hand,therefore, we would expect the costs of obtaining external finance to be greater amongmore innovative firms. On the other hand, the potential benefits to external equityfinanciers are larger the more innovative the firm. This is particularly true among asmaller subset of investors that add value to their investees. This value added can comein a variety of forms, including but not limited to advice pertaining to strategy, mar-keting, financing, administrative and human resource policy, as well as facilitating anetwork of contacts for firms that includes, but is not limited to, accountants, suppliers,customers, lawyers and investment bankers. For example, one recent theoretical paper(Garmaise, 2000) predicts the traditional pecking order is exactly reversed wheninvestors do provide value added and have superior skills at assessing the value of theentrepreneurial project.7 Hence, in our empirical analyses we consider differences infinancing obtained from banks and venture capitalists.

Second, we control for the effect of industry structure on financing decisions. In hispopular work on industry structure, Porter (1998) identifies five categories of factorsthat can give rise to differences in risk-adjusted rates of return across industries. Thesefactors include supplier power, buyer power, barriers to entry, threat of substitutes andthe degree of rivalry. The economic importance of some of these factors has beenidentified in the literature on a firm’s financing decisions. In short, risk-adjusted ratesof return tend to be lower in industries for which there is more competition and astronger presence of larger dominant competitors; therefore, it is less attractive for

7 For example, as discussed in Garmaise (2000), investors might be more skilled in valuing the project thanthe entrepreneur where the investor has experience and market information pertaining to the commer-cialisation of a scientific idea, and the entrepreneur’s comparative advantage over the investor only lies in thescientific process that led to the idea.



investors to provide external capital to entrepreneurial firms in those industries andthe terms offered by investors will be less attractive.

Third, we control for the possibility that entrepreneurial firms started by females facegreater hurdles in seeking external capital (Marlow and Patton, 2005). On one level, thiscould simply be unjustified or �actual� sex discrimination. On another level, there mightbe characteristics of firms started by females that are systematically different relative tofirms started by males, such as the firm’s growth objectives, the professional qualifica-tions of the firm’s directors, the firm’s innovative activities and industry sector, amongnumerous other systematic differences. If so, then information asymmetries faced byoutside investors might systematically differ across firms started by males and females,thereby causing systematic differences in the costs and benefits of seeking externalcapital by female versus male led firms. We might label this second perspective as�apparent� (not actual) sex discrimination whereby information problems and firmcharacteristics give rise to the appearance of discrimination between males and femalesbut such discrimination is directly attributable to those characteristics and not inde-pendently related to the maleness or femaleness of the firm’s founding CEO. Either way,we do consider and control for these alternative theories in our empirical analyses.

Finally, in our empirical tests we also control for legal form (corporation versuspartnership or sole proprietorship), firm age, capital expenditures, insider ownership(including CEO and board ownership), debt levels (relative to assets) prior to seekingexternal finance in the period considered and a variety of reasons why the entre-preneurial firm was established (including reasons ranging from a desire to avoidunemployment to running a business, implementing an invention and wealth ambi-tions). Each of these variables (among others discussed immediately below) arepertinent as control variables, since they directly relate to a firm’s need/desire forexternal capital for reasons discussed in prior literature summarised above in Section 1.The specifics in our data are outlined in detail immediately below in Section 3.Empirical tests follow in Section 4.

3. Data and Summary Statistics

Our data come from a very detailed survey of 2,520 UK entrepreneurial firms in theperiod spanning 1996–7. The data were collected by the Centre for Business Researchat the University of Cambridge, as described in detail by Cosh and Hughes (1994). Themedian start-up year of the firms in the data herein was 1983, where 697 started in the1990s, 920 started in 1980s and 964 started prior to 1980.

The sampling frame for the survey was all independent businesses in manufacturingand business services with less than 500 employees in Great Britain including busi-nesses, partnerships and sole proprietors. The achieved sample was 2,520 firms basedon a size-stratified approach to avoid swamping the sample with micro businesses. Theunit response rate was 25%. A response-bias analysis in terms of age, employment,turnover, pre-tax profit and legal status revealed that older manufacturing firms weresomewhat less likely to respond. There was no response bias of any kind in servicebusiness responses, nor any bias in manufacturing firms responses in terms of age,profitability or legal status. A spatial analysis revealed that the achieved sample wasrepresentative of the regional distribution of the small business population in Great



Britain; complete details on survey design and sample selection issues are provided inBullock and Hughes (1998).8

Summary statistics of the data, as well as correlations across different variables, areprovided in Tables 1–3. Table 1 indicates that there were 37.8% (952 of 2,520) firms inthe data that did seek external finance in the 1996–7 period. The average amount ofexternal finance sought was £467,667, and the median amount sought was £100,000.The average amount obtained was almost 81% of that which was sought, and themedian percentage obtained was 100%. Overall, therefore, the data do not suggest ashortage of external capital for firms that make more than a trivial effort in applying forcapital.

Table 1 summarises a variety of explanatory variables used in the empirical analyses.In order to explain external financing decisions in the 1996–7 financial year, we uselagged (1995) balance sheet items, including lagged assets and lagged long-term debt/assets. For firms that were started in 1996–7, these items are assigned the value �0�. Wedo not use contemporaneous balance sheet values since any external financing ob-tained would directly show up in our left-hand and right-hand-side variables. Also,Table 1 reports statistics from firm income statements (including profits and profitmargin) and cash flow statements (capital expenditure). These items are measured forthe 1996–7 financial year because in applications for external finance, firms will presentthe most recent quarter’s income and cash flow statements (alongside all historicalstatements), as well as projected future income and cash flows. As such, it is appropriateto consider the most recent year’s information. There is not an endogeneity problem asapplications for external finance and obtaining external finance do not have an impacton profitability, turnover or profit margin in the same year.9 Similarly, applying for orobtaining finance will not lead to an innovation in the same year (albeit, it may lead toan innovation in subsequent years as the new capital is spent on R&D projects, asindicated by Kortum and Learner (2000) and hence our innovation variable is mea-sured in the contemporaneous period. In sum, we use lagged balance sheet items toavoid direct endogeneity in analysing external capital raised relative to debt and assets,and current cash flow and income statement items to account for the fact that firmsprovide potential sources of external capital with their most recent quarter’s

8 In some cases survey responses contained missing values. These values were interpolated with comparableavailable information such as asset sizes, profit margins, age and the like so that we make use of the fullsample of 2,520 firms in the analysis. If interpolation was not possible for a survey respondent, then medianvalues were used. We also considered other techniques to handle missing data. For example, in one speci-fication we considered a two-step probit estimate in which the first step considers whether a full set ofinformation was available and the second set then estimates the regression for alternative forms of finance.We also considered regressions with fewer explanatory variables, which in turn increases the sample sizewhere certain excluded variables have missing observations. Either way, the results are quite robust. Alter-native specifications are available on request. Also, note that an earlier version of this article (available onrequest) was distributed with a smaller sample population of 1,900 firms. The current version draws from asample of 2,520 firms. While there are some differences in the results with the larger sample, the main resultsare consistent.

9 There is one potential exception to this statement: contemporaneous capital expenditures might beendogenous to obtaining external finance if firms are cash constrained. However, the data indicate firms mostoften obtain their desired cash flows (albeit not always in the form that they would like). Hence, we use thecurrent capital expenditures to account for planned growth (i.e., not historical capital expenditures whichreflect historical growth) and to account for the fact that firms provide to potential sources of external capitaltheir most recent quarter profits and capital expenditures, as well as expected profits and capital expendi-tures.



Tab

le1

Var

iabl

eD

efin

itio

ns

and

Sum

mar

ySt

atis

tics

Var

iab

leN

ame

Defi

nit

ion

Mea

nM

edia

nSt

and

ard

Dev

iati

on

Min

imu

mM

axim

um

Ob

serv

atio

ns

Dep

ende

nt

Var

iabl

esE

xter

nal

Fin

ance

Sou

ght

Ad

um

my

vari

able

equ

alto

on

eif

the

firm

atte

mp

ted

too

bta

inex

tern

alfi

nan

ce(i

.e.

add

itio

nal

toin

tern

alca

shfl

ow

s)in

the

1996

–199

7p

erio

d.

0.37

80

0.48

50

125

20

Am

ou

nt

of

Ext

ern

alF

inan

ceSo

ugh

tT

he

(str

ictl

yp

osi

tive

)am

ou

nt

of

exte

rnal

fin

ance

sou

ght

by

the

bu

sin

ess

inth

e19

96–1

997

per

iod

.M

easu

red

inth

ou

san

ds

of

1997

po

un

ds.

467.

667

100

1631

.855

0.1

2903

0.90

795

2

Ext

ern

alF

inan

ceO

bta

ined

Th

ep

erce

nta

geo

fex

tern

alfi

nan

ceo

bta

ined

by

the

bu

sin

ess

inth

e19

96–1

997

per

iod

(as

afr

acti

on

of

the

amo

un

tso

ugh

t).

80.7

1910

033

.713

010

095

2

Ext

ern

alF

inan

ceO

bta

ined

fro

mB

anks

An

ord

ered

vari

able

equ

alto

0if

the

sou

rce

was

no

tap

pro

ach

ed,

1if

the

sou

rce

was

app

roac

hed

bu

tn

ofi

nan

ceo

ffer

ed,

2if

the

sou

rce

was

app

roac

hed

bu

to

ffer

edle

ssth

anth

efu

llam

ou

nt,

and

3if

the

sou

rce

was

app

roac

hed

and

off

ered

the

full

amo

un

t.

2.07

83

1.18

90

395

2

Ext

ern

alF

inan

ceO

bta

ined

fro

mV

entu

reC

apit

alF

un

ds

An

ord

ered

vari

able

equ

alto

0if

the

sou

rce

was

no

tap

pro

ach

ed,

1if

the

sou

rce

was

app

roac

hed

bu

tn

ofi

nan

ceo

ffer

ed,

2if

the

sou

rce

was

app

roac

hed

bu

to

ffer

edle

ssth

anth

efu

llam

ou

nt,

and

3if

the

sou

rce

was

app

roac

hed

and

off

ered

the

full

amo

un

t.

0.18

60

0.65

30

395

2

Inde

pen

den

tV

aria

bles

Firm

Fin

anci

alC

hara

cter

isti

csP

rofi

tsT

he

nat

ura

llo

go

fp

re-t

axp

rofi

ts(l

oss

es)

bef

ore

ded

uct

ion

of

inte

rest

,ta

x,an

dd

irec

tors

�,p

artn

ers�

or

pro

pri

eto

rs�e

mo

lum

ents

.M

easu

red

inth

ou

san

ds

of

1997

po

un

ds.

(sca

led

toav

oid

neg

ativ

es)

7.97

27.

953

0.20

10

9.44

925

20

Lo

g(C

apit

alE

xpen

dit

ure

s)T

he

nat

ura

llo

go

fth

efi

rm’s

tota

lca

pit

alex

pen

dit

ure

sin

1996

-199

7.M

easu

red

inth

ou

san

ds

of

1997

po

un

ds.

3.95

24.

394

2.15

40.

000

10.9

5125

20

Lo

g(C

apit

alE

xpen

dit

ure

s/P

rofi

ts)

Lo

go

fca

pit

alex

pen

dit

ure

sp

erp

rofi

ts(s

cale

dto

avo

idn

egat

ives

).�

4.02

0�

3.60

22.

129

�8.

344

6.97

425

20

Pro

fit

Mar

gin

Pro

fits

/T

urn

ove

rin

1996

–199

7.0.

909

0.11

224

.737

�2.

000

1236

.000

2520

Lo

g(T

urn

ove

r)T

he

nat

ura

llo

go

fth

efi

rm’s

tota

ltu

rno

ver

in19

96-1

997.

Mea

sure

din

tho

usa

nd

so

f19

97p

ou

nd

s.6.

691

7.42

41.

584

0.69

311

.408

2520



Tab

le1

Con

tin

ued

Var

iab

leN

ame

Defi

nit

ion

Mea

nM

edia

nSt

and

ard

Dev

iati

on

Min

imu

mM

axim

um

Ob

serv

atio

ns

Lo

ng

Ter

mD

ebt/

Ass

ets

To

tal

lon

gT

erm

Deb

t/A

sset

s(V

alu

esF

rom

the

1995

Fin

anci

alYe

arP

rio

rto

Rai

sin

gC

apit

alin

the

1996

–7P

erio

du

nd

erC

on

sid

erat

ion

)

0.89

50.

086

39.8

390.

000

2000

.000

2520

Lo

g(T

ota

lA

sset

s)L

og

of

To

tal

Ass

ets

(Val

ues

Fro

mth

e19

95F

inan

cial

Year

Pri

or

toR

aisi

ng

Cap

ital

inth

e19

96–7

Per

iod

un

der

Co

nsi

der

atio

n)

Mea

sure

din

tho

usa

nd

so

f19

97p

ou

nd

s

5.92

76.

125

1.72

60.

000

10.9

0425

20

Inn

ova

tio

nA

du

mm

yva

riab

leeq

ual

too

ne

ifth

efi

rmd

evel

op

eda

new

com

mer

cial

isab

lete

chn

olo

gy(t

ech

no

logi

call

yn

ewo

rsi

gnifi

can

tly

imp

rove

d)

inth

e19

96–7

per

iod

0.23

10

0.42

20

125

20

Firm

Age

and

Pro

file

Lo

g(A

ge)

Th

en

atu

rall

og

of

the

firm

’sag

eas

at19

97fr

om

dat

eo

fin

corp

ora

tio

n.

2.66

12.

639

0.94

20

5.62

825

20

Pro

fess

ion

alD

irec

tors

Th

ep

rop

ort

ion

of

dir

ecto

rsw

ith

anad

van

ced

deg

ree

insc

ien

ce,

engi

nee

rin

go

rso

me

oth

erp

rofe

ssio

nal

qu

alifi

cati

on

.

0.59

50.

667

0.45

90

125

20

Gen

der

Th

ege

nd

ero

fth

efi

rm’s

Ch

ief

Exe

cuti

ve/

Sen

ior

Par

tner

/P

rop

riet

or

(1¼

mal

e,2¼

fem

ale)

1.07

61

0.26

51

225

20

CE

OSh

ares

%o

fsh

ares

ow

ned

by

Ch

ief

Exe

cuti

ve51

.502

51.6

0421

.551

010

025

20B

oar

dSh

ares

%o

fsh

ares

ow

ned

by

wh

ole

Bo

ard

of

Dir

ecto

rs86

.123

86.7

7418

.105

010

025

20L

arge

stO

wn

erSh

ares

%o

fsh

ares

ow

ned

by

larg

est

sin

gle

shar

eho

lder

58.3

3358

.485

18.2

620.

100

100

2520

Gro

wth

Ob

ject

ives

Th

efi

rm’s

pla

nn

edgr

ow

tho

bje

ctiv

eso

ver

the

1997

-200

0p

erio

d(1¼

bec

om

esm

alle

r,2¼

stay

the

sam

esi

ze,

3¼gr

ow

mo

der

atel

y,4¼

gro

wsu

bst

anti

ally

)

3.01

23

0.68

71

425

20

Co

rpo

rati

on

Ad

um

my

vari

able

equ

alto

1fo

rin

corp

ora

ted

firm

s0.

733

10.

443

01

2520

Par

tner

ship

Ad

um

my

vari

able

equ

alto

1fo

rp

artn

ersh

ips

0.14

50

0.35

20

125

20So

leP

rop

riet

ors

hip

Ad

um

my

vari

able

equ

alto

1fo

rso

lep

rop

riet

ors

hip

0.12

20

0.32

80

125

20R

easo

nFi

rmE

stab

lish

edC

om

ple

tely

New

Star

t-u

ps

Ad

um

my

vari

able

equ

alto

1fo

rfi

rms

that

wer

eC

om

ple

tely

New

Star

t-u

ps.

By

the

term

new

�sta

rt-u

ps�

we

mea

nth

isis

the

form

of

bu

sin

ess

form

atio

n,

and

do

esn

ot

nec

essa

rily

mea

nth

atth

efi

rmis

esp

ecia

lly

you

ng.

0.64

41

0.47

90

125

20



Tab

le1

Con

tin

ued

Var

iab

leN

ame

Defi

nit

ion

Mea

nM

edia

nSt

and

ard

Dev

iati

on

Min

imu

mM

axim

um

Ob

serv

atio

ns

Fo

un

ded

toav

oid

un

emp

loym

ent

of

fou

nd

er(s

)

Ad

um

my

vari

able

equ

alto

1fo

rfi

rms

fou

nd

edto

avo

idu

nem

plo

ymen

to

ffo

un

der

(s)

0.29

50

0.51

30

125

20

Fo

un

ded

for

des

ire

of

fou

nd

er(s

)to

run

ow

nb

usi

nes

s

Ad

um

my

vari

able

equ

alto

1fo

rfi

rms

fou

nd

edfo

rd

esir

eo

ffo

un

der

(s)

toru

nh

iso

rh

ero

wn

bu

sin

ess

0.73

41

0.50

00

125

20

Fo

un

ded

toim

ple

men

tan

inve

nti

on

Ad

um

my

vari

able

equ

alto

1fo

rfi

rms

fou

nd

edto

imp

lem

ent

anin

ven

tio

n0.

304

00.

515

01

2520

Fo

un

ded

du

eto

wea

lth

amb

itio

ns

of

fou

nd

ers

Ad

um

my

vari

able

equ

alto

1fo

rfi

rms

fou

nd

edd

ue

tow

ealt

ham

bit

ion

so

ffo

un

der

s0.

334

00.

554

01

2520

Com

peti

tors

/In

dust

ryL

arge

rC

om

pet

ito

rsT

he

pro

po

rtio

no

fth

efi

rm’s

pri

mar

yco

mp

etit

ors

that

are

larg

erth

anth

efi

rm.

0.68

40.

750.

536

012

2520

Lo

g(T

ota

lC

om

pet

ito

rs)

Th

en

atu

ral

log

of

the

tota

ln

um

ber

of

seri

ou

sco

mp

etit

ors

of

the

firm

.1.

934

1.79

21.

011

08.

517

2520

Ind

ust

ry(u

n)I

nn

ova

tive

nes

sT

he

per

cen

tage

of

the

firm

’ssa

les

for

wh

ich

pro

du

cts

or

serv

ices

wer

eu

nch

ange

do

ro

nly

mar

gin

ally

chan

ged

inth

ela

st3

year

s.

74.6

6980

29.5

140

100

2520

Hig

hT

ech

Man

ufa

ctu

rin

gA

du

mm

yva

riab

leeq

ual

too

ne

for

afi

rmin

ah

igh

tech

no

logy

man

ufa

ctu

rin

gin

du

stry

.0.

062

00.

241

01

2520

Co

nve

nti

on

alM

anu

fact

uri

ng

Ad

um

my

vari

able

equ

alto

on

efo

ra

firm

ina

con

ven

tio

nal

man

ufa

ctu

rin

gin

du

stry

0.40

90

0.49

20

125

20

Hig

hT

ech

Serv

ices

Ad

um

my

vari

able

equ

alto

on

efo

ra

firm

ina

hig

hte

chn

olo

gyse

rvic

esin

du

stry

.0.

039

00.

193

01

2520

Co

nve

nti

on

alSe

rvic

esA

du

mm

yva

riab

leeq

ual

too

ne

for

afi

rmin

aco

nve

nti

on

alse

rvic

esin

du

stry

.0.

299

00.

458

01

2520

Not

es.

Th

isT

able

pre

sen

tsd

efin

itio

ns

of

each

of

the

vari

able

s,al

on

gw

ith

sum

mar

yst

atis

tics

.T

he

tota

lsa

mp

lesi

zeis

2,52

0fi

rms.

Val

ues

are

in£

1997

and

fro

mth

e19

96-7

per

iod

un

less

oth

erw

ise

ind

icat

ed.

Th

ese

vari

able

sar

eu

sed

inth

esu

bse

qu

ent

tab

les

and

regr

essi

on

anal

yses

.



profitability information as well as projected cash flows and profits. In our multivariateanalyses, we do not simultaneously consider current and historical values of cash flowand income statements as these items are highly correlated over time (correlations aregreater than 0.7 for each of these variables from one year to the next).10

Table 2 reports the number of firms in our data that did seek external finance by thetype of source of finance, as well as the percentage of all their external capital obtainedfrom the source. Among the firms in our sample that did seek external finance, 775approached banks, 474 approached leasing firms, 151 approached factoring/invoicediscounting firms, 138 approached partners/working shareholders, 87 approachedventure capital funds, 83 approached private individuals, 53 approached trade cus-tomers/suppliers and 67 approached other sources. It is of interest that outrightrejection rates were highest among venture capital funds (46% rejection) and muchhigher than that for banks (17% outright rejection). The lowest rejection rate wasamong leasing firms (5%). Banks comprised the median and mean highest percentageof outside finance in terms of which type of source was approached and which type ofsource provided the finance. In fact, banks comprised the only type of source for whichthe median percentage of a firm’s total external capital was greater than 0% (for banks,the median percentage is 34%; see Table 2).

A number of the correlations in Table 2 provide insights into the different charac-teristics of the different types of investors. Banks are statistically more likely to financeprofitable businesses, while the correlations in Table 2 show private individuals arestatistically more likely to fund unprofitable businesses. Venture capital funds tend tofinance innovative firms in innovative industries; unlike banks, as would be expected,because banks provide debt finance, have very large portfolios and typically use presetcriteria in ascertaining whether the firm is a suitable risk for a loan.

The data indicate venture capital funds are more likely to finance businesses withhigher turnover and greater total assets, which is in contrast to private individuals.Consistent with Gompers and Lerner (1999), venture capital funds also finance com-panies with a greater proportion of professional directors (hence potentially bettergovernance) and higher growth objectives, which is similar to financing from privateindividuals. Venture capital funds finance firms with smaller percentages of sharesallocated to CEOs are board members and firms established to implement an inven-tion. Venture capital fund managers have small portfolios (often no more than 5investees per fund manager; see Kanniainen and Keuschingg (2003, 2004); Cumming(2006) and provide significant screening pre-investment and value-added post invest-ment in their investee firms.

Hire purchase, or leasing, firms are more likely to finance older businesses with highcapital expenditures, little innovation, a smaller proportion of professional directorsand firms started by males and not females. Leasing businesses act in ways that aresomewhat analogous to banks. The main difference is that the leasing business typicallyties the financing provided to some other product provided to the firm.

Factor/invoice discounting firms provide cash advances on the basis of an unpaidinvoice, so that a firm may meet cash shortages.11 This type of financing tends to be

10 These high correlations give rise to results that are robust when we use lagged values.11 See http://www.bizhelp24.com/business_finance/business_finance_factoring. htm.



Tab

le2

Sum

mar

ySt

atis

tics

and

Cor

rela

tion

sfo

rSp

ecifi

cSo

urc

esof

Ext

ern

alC

apit

al

Ban

ks

Ven

ture

Cap

ital

Fu

nd

s

Hir

eP

urc

has

eo

rL

easi

ng

Fir

ms

Fac

tori

ng/

Invo

ice

Dis

cou

nti

ng

Fir

ms

Tra

de

Cu

sto

mer

s/Su

pp

lier

s

Par

tner

s/W

ork

ing

Shar

eho

lder

s

Oth

erP

riva

teIn

div

idu

als

Oth

er

Nu

mb

ero

fF

irm

sth

atD

idA

pp

roac

hth

isSo

urc

efo

rE

xter

nal

Cap

ital

775

8747

415

153

138

8367

Nu

mb

ero

fF

irm

sth

atD

idN

ot

Ap

pro

ach

this

Sou

rce,

bu

tD

idSe

ekE

xter

nal

Fin

ance

Els

ewh

ere

177

865

478

801

899

814

869

885

Nu

mb

ero

fF

irm

sth

atD

idA

pp

roac

hth

isSo

urc

eb

ut

No

Fin

ance

Off

ered

133

4023

295

1220

14

Nu

mb

ero

fF

irm

sth

atA

pp

roac

hed

this

Sou

rce

bu

tL

ess

than

Fu

llA

mo

un

tO

ffer

ed13

38

2038

1019

1515

Nu

mb

erth

atD

idA

pp

roac

hth

isSo

urc

ean

dF

ull

Am

ou

nt

Off

ered

509

3943

184

3810

748

38

Mea

nP

erce

nta

geo

fT

ota

lE

xter

nal

Cap

ital

Ob

tain

edfr

om

this

Sou

rce

42.7

492.

851

23.7

925.

892

1.62

45.

313

3.91

03.

598

Med

ian

Per

cen

tage

of

To

tal

Ext

ern

alC

apit

alO

bta

ined

fro

mth

isSo

urc

e33

.580

0.00

00.

000

0.00

00.

000

0.00

00.

000

0.00

0

Stan

dar

dD

evia

tio

no

fP

erce

nta

geO

bta

ined

fro

mth

isSo

urc

e39

.451

13.9

4833

.630

18.2

619.

069

17.6

0216

.605

16.2

24

Min

imu

mP

erce

nta

geO

bta

ined

fro

mth

isSo

urc

e0.

000

0.00

00.

000

0.00

00.

000

0.00

00.

000

0.00

0M

axim

um

Per

cen

tage

Ob

tain

edfr

om

this

Sou

rce

100.

000

100.

000

100.

000

100.

000

100.

000

100.

000

100.

000

100.

000

Cor

rela

tion

sw

ith

Per

cen

tage

Obt

ain

edfr

omSo

urc

ean

dO

ther

Var

iabl

es–

––

––

––

–

Firm

Fin

anci

alC

hara

cter

isti

csL

og

(Pro

fits

)0.0

8�

0.01

0.05

�0.

030.

000.

04�

0.06

0.04

Lo

g(C

apit

alE

xpen

dit

ure

s)�

0.01

0.04

0.2

00.

010.

02�

0.05

�0.1

10.

01L

og

(Cap

ital

Exp

end

itu

res/

Pro

fits

)�

0.01

0.04

0.2

00.

010.

02�

0.05

�0.1

10.

01P

rofi

tM

argi

n�

0.03

�0.

01�

0.02

�0.

01�

0.01

�0.

01�

0.01

0.1

9L

og

(Tu

rno

ver)

0.04

0.1

20.0

80.0

8�

0.01

�0.

06�

0.05

�0.0

9L

on

gT

erm

Deb

t/A

sset

s�

0.01

0.01

�0.

010.

06�

0.01

0.0

7�

0.01

0.00

Lo

g(T

ota

lA

sset

s)0.

050.0

80.1

40.1

00.

00�

0.05

�0.1

20.

02In

no

vati

on

�0.

030.

05�

0.0

80.

000.0

70.

030.

000.0

7Fi

rmA

gean

dP

rofi

leL

og

(Age

)0.

02�

0.0

70.1

8�

0.05

�0.0

8�

0.01

�0.1

10.

02P

rofe

ssio

nal

Dir

ecto

rs�

0.02

0.0

9�

0.0

9�

0.0

80.

000.

050.0

9�

0.04



Tab

le2

Con

tin

ued

Ban

ks

Ven

ture

Cap

ital

Fu

nd

s

Hir

eP

urc

has

eo

rL

easi

ng

Fir

ms

Fac

tori

ng/

Invo

ice

Dis

cou

nti

ng

Fir

ms

Tra

de

Cu

sto

mer

s/Su

pp

lier

s

Par

tner

s/W

ork

ing

Shar

eho

lder

s

Oth

erP

riva

teIn

div

idu

als

Oth

er

Gen

der

0.05

�0.

05�

0.0

60.

010.

05�

0.01

0.01

0.02

CE

OSh

ares

�0.

02�

0.0

9�

0.01

0.00

0.00

�0.

030.

010.

02B

oar

dSh

ares

0.00

�0.1

70.

040.

030.

03�

0.01

0.01

�0.0

9L

arge

stO

wn

erSh

ares

0.01

�0.1

0�

0.02

�0.

02�

0.03

�0.

040.

000.

01G

row

thO

bje

ctiv

es�

0.02

0.1

1�

0.04

0.0

90.0

7�

0.01

0.00

�0.

01C

orp

ora

tio

n�

0.1

00.0

80.

040.1

00.

000.

04�

0.02

0.02

Par

tner

ship

0.0

6�

0.05

0.01

�0.0

80.

020.

03�

0.04

�0.

04So

leP

rop

riet

ors

hip

0.0

7�

0.06

�0.

06�

0.06

�0.

02�

0.09

0.07

0.01

Rea

son

Firm

Est

abli

shed

Co

mp

lete

lyN

ewSt

art-

up

s�

0.01

0.05

�0.

03�

0.03

0.04

0.01

0.01

0.04

Fo

un

ded

toav

oid

un

emp

loym

ent

of

fou

nd

er(s

)�

0.05

0.01

�0.

01�

0.01

�0.

030.

010.

00�

0.03

Fo

un

ded

for

des

ire

of

fou

nd

er(s

)to

run

ow

nb

usi

nes

s0.

020.

010.

06�

0.03

�0.

030.

000.

00�

0.06

Fo

un

ded

toim

ple

men

tan

inve

nti

on

�0.

020.0

6�

0.05

0.00

0.00

0.03

0.02

0.06

Fo

un

ded

du

eto

wea

lth

amb

itio

ns

of

fou

nd

ers

�0.

020.

030.

05�

0.01

0.00

0.00

�0.

01�

0.01

Com

peti

tors

/In

dust

ryL

arge

rC

om

pet

ito

rs0.

000.

01�

0.01

�0.

01�

0.02

0.00

0.00

�0.

01L

og

(To

tal

Co

mp

etit

ors

)�

0.02

�0.

060.

050.

05�

0.04

�0.

01�

0.02

�0.

01In

du

stry

(un

)In

no

vati

ven

ess

0.0

7�

0.0

70.

040.

01�

0.0

7�

0.03

�0.

05�

0.05

Hig

hT

ech

Man

ufa

ctu

rin

g�

0.03

0.03

�0.

03�

0.03

0.04

0.0

6�

0.02

�0.

01C

on

ven

tio

nal

Man

ufa

ctu

rin

g�

0.03

�0.

050.1

50.1

3�

0.03

�0.

05�

0.1

20.

03H

igh

Tec

hSe

rvic

es�

0.0

80.0

6�

0.04

�0.

050.

000.

060.

030.

00C

on

ven

tio

nal

Serv

ices

0.0

7�

0.01

�0.0

8�

0.0

7�

0.04

0.03

0.0

7�

0.05

Not

es.

Th

isT

able

pre

sen

tssu

mm

ary

stat

isti

csan

dco

rrel

atio

nco

effi

cien

tsfo

rea

chsp

ecifi

cso

urc

eo

fex

tern

alca

pit

al(b

anks

,ve

ntu

reca

pit

alfu

nd

s,h

ire

pu

rch

ase

or

leas

ing

firm

s,fa

cto

rin

g/in

voic

ed

isco

un

tin

gfi

rms,

trad

ecu

sto

mer

s/su

pp

lier

s,p

artn

ers/

wo

rkin

gsh

areh

old

ers,

oth

erp

riva

tein

div

idu

als

and

oth

er).

Ato

tal

of

952

firm

s(o

ut

of

2,52

0fi

rms

inth

ed

atas

et)

sou

ght

exte

rnal

cap

ital

fro

mat

leas

to

ne

sou

rce

inth

e19

96–7

per

iod

.T

he

nu

mb

ero

ffi

rms

that

did

and

did

no

tap

pro

ach

the

sou

rce

isin

dic

ated

,fo

llo

wed

by

the

reje

ctio

nan

dsu

cces

sra

tes

amo

ng

tho

sefi

rms

that

did

app

roac

hth

eso

urc

e.T

he

stat

isti

csfo

rth

em

ean

,m

edia

n,

stan

dar

dd

evia

tio

nan

dm

inim

um

and

max

imu

mp

erce

nta

geo

fca

pit

alo

bta

ined

fro

mth

eea

chso

urc

eis

ind

icat

ed.C

orr

elat

ion

sw

ith

vari

ou

sva

riab

les

for

the

char

acte

rist

ics

of

the

bu

sin

esse

sar

ep

rese

nte

d;

corr

elat

ion

ssi

gnifi

can

tat

the

5%le

vel

are

sho

wn

init

alic

s.



used by firms with fewer professional directors, as well as firms with high growthobjectives and lower profit levels. The type of finance provided by an invoice dis-counting firm is such that they do not have concerns with information problemsassociated with the quality of the firm that obtains the finance but rather with the firmor customer with the unpaid invoice.

Trade customers/suppliers are more likely to finance younger firms with growthobjectives and those that have recently developed an innovation. There are pro-prietary rights associated with innovative industries and smaller innovative firms facepotential hold up problems from sources of capital. The data are suggestive thatsmaller innovative firms are not concerned with these hold-up problems from thelarger suppliers in their respective industries. Suppliers can be a natural source ofcapital for some developing firms securing a supply chain for developing their ownproducts.

Partners/working shareholders are more likely to be a significant source of capitalfor firms with high levels of debt/assets. Firms with excess leverage may secure addi-tional capital from sources of capital that face comparatively few informational prob-lems, such as existing partners/working shareholders.

Finally, the category of �other private individuals� in Table 2 or �angel investors�indicates similar patterns as with the venture capital funds. Angel investors financesmall, young and as-yet unprofitable businesses, albeit firms with strong managementand high quality directors.

Table 3 provides additional correlation statistics across various variables that are usedin the multivariate empirical analyses in the next Section. The matrix gives furtherinsights into the data and provides guidance in terms of considering issues of collin-earity in the regressions in Section 4.

4. Multivariate Empirical Analyses

This Section provides multivariate tests of our theoretical propositions by firstconsidering the extent of finance sought in subsection 4.1, taking into account (inthe spirit of Heckman, 1976, 1979) the first step non-randomness in the decision toseek outside capital. Subsection 4.2 considers the percentage of external capitalobtained. Thereafter, we provide tests of differences across the alternative sources ofcapital (banks, venture capital funds, leasing firms, factoring firms, trade customersand suppliers, partners and working shareholders, private individuals and othersources).

4.1. Which Firms Seek External Finance and the Amount of External Finance Sought

In this subsection we examine the factors that affect the amount of external financewhile taking into account the factors that lead firms to seek external finance in the firstplace. We report eight alternative specifications in Table 4, based on variables definedin Table 1. Models 1A, 2A, 3A and 4A are Tobit models in which the left-hand-sidevariable is the amount of external capital that was sought. The Tobit specification isused to account for the fact that the left-hand-side variable is bounded below by zero.The right-hand-side variables comprise variables for firm financial characteristics, firm



Tab

le3

Cor

rela

tion

Mat

rix

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

Dep

ende

nt

Var

iabl

es(1

)E

xter

nal

Fin

ance

Sou

ght

1.00

(2)

Am

ou

nt

of

Ext

ern

alF

inan

ceSo

ugh

t

N/

A1.

00

(3)

Ext

ern

alF

inan

ceO

bta

ined

N/

A0.

041.

00

Inde

pen

den

tV

aria

bles

Firm

Fin

anci

alC

hara

cter

isti

cs(4

)P

rofi

tM

argi

n0.

020

.23

0.02

1.00

(5)

Lo

g(C

apit

alE

xpen

dit

ure

s)0

.16

0.2

90

.23

0.01

1.00

(6)

Lo

g(C

apit

alE

xpen

dit

ure

s/P

rofi

ts)

0.1

60

.28

0.2

20.

011

.00

1.00

(7)

Lo

g(T

urn

ove

r)0

.08

0.1

70

.27�

0.1

00

.44

0.4

21.

00(8

)D

ebt/

Ass

ets

�0.

020

.17

0.06

0.00

0.02

0.01

0.0

51.

00(9

)L

og

(To

tal

Ass

ets)

0.1

10

.24

0.3

60.

000

.54

0.5

20

.71�

0.0

71.

00(1

0)In

no

vati

on

0.0

80.

05�

0.04�

0.01

0.1

50

.15

0.1

2�

0.01

0.1

61.

00Fi

rmA

gean

dP

rofi

le(1

1)L

og

(Age

)�

0.0

60

.07

0.1

5�

0.03

0.1

90

.19

0.1

50.

000

.35

0.03

1.00

(12)

Pro

fess

ion

alD

irec

tors

0.02

0.02�

0.0

90.

000.

000.

000.

010.

00�

0.03

0.04�

0.04

1.00

(13)

Gen

der

�0.

04�

0.02�

0.02�

0.01�

0.0

6�

0.0

6�

0.0

6�

0.01�

0.1

2�

0.0

6�

0.1

00.

011.

00(1

4)C

EO

Shar

es0.

00�

0.1

1�

0.1

1�

0.04�

0.1

0�

0.0

9�

0.1

40.

00�

0.1

5�

0.01�

0.0

8�

0.03

0.03

1.00

(15)

Dir

ecto

rSh

ares

�0.

01�

0.1

5�

0.0

7�

0.0

9�

0.1

2�

0.1

1�

0.1

30.

00�

0.1

5�

0.01�

0.1

1�

0.0

60.

020

.47

1.00

(16)

Gro

wth

Ob

ject

ives

0.2

10

.07

0.02

0.03

0.2

30

.23

0.1

70.

000

.20

0.2

0�

0.1

00

.07�

0.03�

0.02�

0.04

1.00

(17)

Co

rpo

rati

on

0.1

30

.10

0.03

0.01

0.2

20

.22

0.2

60.

010

.33

0.1

70

.07�

0.0

6�

0.1

00.

000.

000

.21

1.00

Rea

son

Firm

Est

abli

shed

(18)

Co

mp

lete

lyN

ewSt

art-

up

s0.

00�

0.01�

0.02�

0.03�

0.02�

0.02�

0.0

6�

0.03�

0.0

50.

020

.12

0.01

0.00

0.0

60.

040

.06�

0.0

71.

00

(19)

Fo

un

ded

toav

oid

un

emp

loym

ent

of

fou

nd

er(s

)

�0.

04�

0.02�

0.1

3�

0.01�

0.0

7�

0.0

7�

0.0

8�

0.01�

0.1

2�

0.0

5�

0.02�

0.01

0.02

0.02

0.02�

0.0

6�

0.0

7�

0.01

1.00

(20)

Fo

un

ded

for

des

ire

of

fou

nd

er(s

)to

run

ow

nb

usi

nes

s

0.00�

0.05

0.02�

0.03

0.00

0.00�

0.01

0.01

0.01�

0.04

0.1

40.

020.

020.

010.

020.

00�

0.0

80

.13

0.02

1.00



Tab

le3

Con

tin

ued

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(21)

Fo

un

ded

toim

ple

men

tan

inve

nti

on

0.02

0.01�

0.01�

0.01

0.0

80

.08

0.0

8�

0.01

0.0

90

.13

0.1

20

.06

0.01�

0.04�

0.0

80

.11

0.0

8�

0.03

0.1

60.

041.

00

(22)

Fo

un

ded

du

eto

wea

lth

amb

itio

ns

of

fou

nd

ers

0.04

0.03

0.05

0.02

0.1

00

.11

0.1

0�

0.01

0.1

20.

020

.15

0.02�

0.03�

0.0

6�

0.0

70

.08

0.04�

0.0

50

.20

0.2

30

.34

1.00

Com

peti

tors

/In

dust

ry(2

3)L

arge

rC

om

pet

ito

rs0.

01�

0.05�

0.04�

0.03�

0.0

9�

0.0

8�

0.0

50.

01�

0.1

10.

02�

0.0

70.

030.

01�

0.01

0.01

0.00

0.01

0.03�

0.01

0.01�

0.02�

0.02

1.00

(24)

Lo

g(T

ota

lC

om

pet

ito

rs)

0.04�

0.02

0.02�

0.04

0.04

0.04

0.0

60.

000

.07�

0.0

60

.05

0.0

5�

0.03

0.00�

0.01�

0.01

0.02�

0.03�

0.01

0.04�

0.0

60.

030

.12

1.00

(25)

Ind

ust

ry(u

n)I

nn

ova

tive

nes

s�

0.1

10.

020.

040.

02�

0.1

3�

0.1

3�

0.0

60.

02�

0.0

7�

0.3

50

.11�

0.1

10

.05�

0.02

0.00�

0.2

2�

0.1

1�

0.03

0.04

0.0

6�

0.1

50.

03�

0.02�

0.01

1.00

(26)

Hig

hT

ech

Serv

icesþ

Hig

hT

ech

Man

ufa

ctu

rin

g

0.0

80.

01�

0.1

4�

0.01

0.0

50

.05�

0.02�

0.01�

0.02

0.2

5�

0.0

60

.12�

0.04�

0.03�

0.0

60

.18

0.1

20

.05�

0.01�

0.02

0.1

00.

000.

02�

0.0

7�

0.2

8

Not

es.

Th

isT

able

pre

sen

tsco

rrel

atio

nco

effi

cien

tsac

ross

the

vari

able

sth

atw

ere

defi

ned

inT

able

1.C

orr

elat

ion

sth

atar

est

atis

tica

lly

sign

ifica

nt

atth

e5%

leve

lar

ein

ital

ics

fon

t.T

he

corr

elat

ion

sar

ep

rovi

ded

for

the

sub

sam

ple

of

952

firm

sth

atso

ugh

tex

tern

alca

pit

alfo

rva

riab

les

(2)

and

(3)

and

for

the

full

sam

ple

of

2,52

0fi

rms

for

all

oth

erva

riab

les.

N/

Are

fers

ton

ot

app

lica

ble

(th

ere

isn

oco

rrel

atio

nb

etw

een

vari

able

s(1

)an

d(2

)an

db

etw

een

(1)

and

(3))

.



age and profile and competitors/industry, as specified in Table 4. Model 1A encom-passes the full sample. Model 2A considers the subsample of firms started in 1990–7.Model 3A considers the subsample of firms started in 1980-9. Model 4A considers thesubsample of firms started prior to 1980. These subsamples are used to show robustnessto different time periods for two reasons. First, we only observe applications forexternal finance in 1996 and 1997, and firms started at different points will be atdifferent stages in their life-cycle and thereby may have a different likelihood ofapplying for external finance at any give point in time. Second, there may be survivor-ship bias in the typical sense that older firms observed are the ones that survived.Except where otherwise noted below, the results are quite robust (albeit in some casesless significant in the subsamples, which is likely due to the smaller sample sizes).

Models 1B, 2B, 3B and 4B are two-step Heckman corrected models whereby the firststep is a Probit model that estimates the probability that firm applied for outside capitaland the second step is the amount of external capital sought. The identification vari-ables (Puhani, 2000) in the first step (excluded in the second step) Heckman regres-sion include, first, growth objectives because it is natural that self-proclaimed growthfirms (as opposed to �life-style� firms) are more likely to apply for external capital. Thegrowth objectives variable is not included in the second step because it is a variable thatis qualitative (not easily quantifiable by the people and institutions sourcing the capi-tal) and hence more closely related to the decision to apply for finance rather than howmuch finance (entrepreneurs know that their application for capital is evaluated on thebasis of quantifiable criteria and not by their level of ambition and hence the amountsought will more likely reflect quantitative criteria).12 By contrast, the other variablessuch as capital expenditures/profits more closely capture both the decision to apply forfinance and the amount of finance sought. The second restriction is similar to the first,and refers to the wealth ambitions of the founders. The rationale for this restriction issimilar to the first restriction. The third identifying restriction was whether or not thefirm was incorporated (instead of a partnership or sole proprietorship). We wouldexpect corporations would be more likely to apply for external finance. In earlier draftswe considered other restrictions and found results consistent with those reported here;alternative specifications are available on request.

In Table 4 step (1) of Models 1B, 2B, 3B and 4B, the data indicate that firms that doseek external finance have the following characteristics: they have lower turnover(Model 4B), higher capital expenditures/profits (Models 1B, 2B, 3B and 4B) and highergrowth objectives (Models 1B, 2B, 3B and 4B). In terms of the economic significance, afirm is approximately 1% more likely to seek external finance when capital expendituresare 20 times that of profits relative to 10 times profits (see step 1 of Models 1B, 2B, 3B and4B). The relation between capital expenditures/profits and external finance applica-tions is also illustrated graphically in Figure 1. This indicates support for the traditionalpecking order in which firms finance new projects internally before seeking externalfinance (see Hypothesis 1a outlined above in subsection 2.1). Firms applying forexternal finance also tend to be smaller firms (as proxied by turnover) with a stronggrowth orientation (in support of Hypothesis 2a). The statistical significance of the

12 Regardless, note that we nevertheless considered the growth objectives variable for the second step andit was insignificant and did not materially affect the other results reported in Table 4.



effect of turnover, however, is not robust and only significant in Model 4B which showsthat a firm with £1,000,000 turnover is 1.9% more likely than a firm with £2,000,000turnover to apply for external finance. The effect of growth ambition, by contrast, isrobust in all of the specifications presented in Table 4. The economic significanceassociated with growth ambition is such that on a ranking from 1–4 with 4 being thehighest growth orientation, an increase in the ranking by 1 point tends to increase theprobability of an application for external finance by approximately 11%.

There are some differences in firm-specific and market factors that lead to applica-tions for external finance. Corporations are approximately 6% more likely to seekexternal finance that partnerships or sole proprietorships. Firms that face 1,000 com-petitors are approximately 2% more likely to apply for external finance than firms thatface 100 competitors. Also, Table 4 indicates industry differences in applications forexternal finance.

Table 4 Models 1A, 2A, 3A, 4A and step 2 of Models 1B, 2B, 3B and 4B presentregression evidence that indicates systematic characteristics drive the amount ofexternal finance actually sought. The results are presented for Tobit and Heckmanspecifications as well as different time periods for firms� founding dates. Note that wealso considered the robustness of our results to outlier observations (as identified byCook’s distances and leverage plots). The results are robust. For instance, removing theoutliers apparent in Figure 2, we do not find any material differences in the regressionresults reported in the Tables. Additional specifications are available on request.

Table 4 shows that (with the sole exception of Model 3B) greater levels of capitalexpenditures/profits give rise to firms seeking more external finance (consistent withHypothesis 1a), and this effect is the most statistically robust and economically signi-ficant driver of the amount of external finance sought. The economic significance issuch that a doubling of capital expenditure/profits gives rise to an additional £55,138in external capital sought (based on Model 1B in Table 4). The positive relationbetween capital expenditures/profits and the amount of external finance sought isalso depicted graphically in the data in Figure 2. Note that we have expressed theright-hand-side variables in logs. (A linear specification either in terms of capital

External Finance ApplicationNo External Finance Application

0

20

40

60

80

100

120

140

160

180

200

–9 –8 –7 –6 –5 –4 –3 –2 –1 0 1 2 3 4 5 6 7

Num

ber

of C

ompa

nies

Log (Capital Expenditures / Profits) [scaled]

Fig. 1. Applications for External Finance and Capital Expenditures/Profits



Tab

le4

Reg

ress

ion

An

alys

esof

Am

oun

tof

Ext

ern

alFi

nan

ceSo

ugh

t

Fu

llSa

mp

le19

90–1

997

Sub

sam

ple

1980

–198

9Su

bsa

mp

leP

re-1

980

Sub

sam

ple

Mo

del

1A(T

ob

it):

Am

ou

nt

of

Ext

ern

alF

inan

ce

Mo

del

1B

Mo

del

2A(T

ob

it):

Am

ou

nt

of

Ext

ern

alF

inan

ce

Mo

del

2B

Mo

del

3A(T

ob

it):

Am

ou

nt

of

Ext

ern

alF

inan

ce

Mo

del

3B

Mo

del

4A(T

ob

it):

Am

ou

nt

of

Ext

ern

alF

inan

ce

Mo

del

4B

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)A

mo

un

to

fE

xter

nal

Fin

ance

Step

1(P

rob

it)

¼1

ifA

pp

lied

for

Ext

ern

alF

inan

ce

Step

2(H

ecki

t)A

mo

un

to

fE

xter

nal

Fin

ance

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)A

mo

un

to

fE

xter

nal

Fin

ance

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)A

mo

un

to

fE

xter

nal

Fin

ance

Co

nst

ant

169.

455

�0.

345*

**4.

968

732.

393

�0.

712*

**13

45.7

4161

1.08

9*�

0.36

411

15.3

48**

�64

4.45

20.

098

�95

5.68

0Fi

rmFi

nan

cial

Cha

ract

eris

tics

Lo

g(T

urn

ove

r)29

.118

�0.

003

45.7

29�

2.22

10.

013

�5.

163

56.7

03*

�0.

012

90.6

24*

�11

.547

�0.

064*

*�

126.

424

Lo

g(C

apit

alE

xpen

dit

ure

s/P

rofi

ts)

100.

318*

**0.

026*

**18

3.16

5***

97.9

44**

*0.

017*

158.

355*

*26

.948

*0.

024*

*26

.092

205.

540*

**0.

039*

**37

0.26

2***

Pro

fit

Mar

gin

5.69

3***

4.38

0E-0

49.

334*

**5.

415*

**0.

001

8.96

0***

28.7

11�

0.00

648

.529

�99

.501

�0.

182*

**�

527.

760

Deb

t/A

sset

s2,

407.

133*

**�

0.00

14,

028.

959*

**1,

597.

757*

�0.

343

2,72

0.11

8*�

438.

648

0.24

4�

885.

347

21.5

400.

671*

*11

79.4

30(D

ebt/

Ass

ets)

2�

451.

036*

**�

4.20

3E�

07�

744.

353*

**�

358.

306*

0.11

4�

614.

530*

313.

773

0.00

053

2.59

63,

070.

464*

**�

0.21

839

48.7

41**

*L

og

(To

tal

Ass

ets)

35.1

540.

006

65.6

1629

.940

0.01

945

.690

37.4

870.

004

42.8

9781

.624

0.03

519

1.20

7In

no

vati

on

53.5

700.

005

104.

544

�48

.424

0.04

8�

94.0

7166

.672

0.00

776

.844

19.4

60�

0.02

42.

180

Firm

Age

and

Pro

file

Lo

g(A

ge)

5.17

8�

0.04

3***

�30

.273

17.8

930.

006

34.2

33�

139.

364

�0.

059

�16

6.64

5�

41.5

790.

003

�73

.401

Pro

fess

ion

alD

irec

tors

2.47

40.

011

15.8

43�

2.44

2�

0.01

2�

6.78

9�

49.6

940.

015

�84

.501

109.

724

0.02

721

0.01

1

CE

OSh

ares

�0.

762

2.35

5E�

0413

4.01

9�

2.74

10.

001

�4.

707

1.08

10.

000

1.61

50.

570

1.00

1E�

040.

839

Bo

ard

Shar

es�

3.57

3*1.

452E

-05

�1.

126

0.98

0�

0.00

11.

739

�4.

819*

**0.

001

�7.

527*

**�

4.78

3*�

3.70

2E-0

4�

7.34

0*G

end

er10

0.06

8�

0.03

9�

5.70

1*�

52.5

79�

0.01

3�

84.0

22�

93.6

800.

003

�14

3.67

215

46.4

29**

*�

0.21

5**

1816

.699

***

Gro

wth

Ob

ject

ives

0.11

4***

0.15

0***

0.11

6***

0.07

7***

Co

rpo

rati

on

0.06

2**

�0.

037

0.10

7***

0.06

1C

om

pan

yfo

un

ded

du

eto

wea

lth

amb

itio

ns

of

fou

nd

ers

0.01

50.

077*

�0.

005

�0.

008

Com

peti

tors

/In

dust

ryL

arge

rC

om

pet

ito

rs�

16.7

270.

009

�18

.158

�12

8.04

70.

011

�21

8.80

0�

77.6

11�

0.00

6�

116.

707

76.4

820.

010

132.

313

Lo

g(T

ota

lC

om

pet

ito

rs)

�4.

520

0.02

2**

6.27

717

.080

0.03

9**

22.6

26�

22.6

100.

023

�45

.076

�32

.424

0.00

8�

36.5

63

Ind

ust

ry(u

n)

Inn

ova

tive

nes

s1.

958*

�0.

001*

2.58

12.

364

�0.

0002

4.03

10.

269

�0.

0004

0.85

91.

281

�0.

001

0.15

2



Tab

le4

Con

tin

ued

Fu

llSa

mp

le19

90–1

997

Sub

sam

ple

1980

–198

9Su

bsa

mp

leP

re–1

980

Sub

sam

ple

Mo

del

1A(T

ob

it):

Am

ou

nt

of

Ext

ern

alF

inan

ce

Mo

del

1B

Mo

del

2A(T

ob

it):

Am

ou

nt

of

Ext

ern

alF

inan

ce

Mo

del

2B

Mo

del

3A(T

ob

it):

Am

ou

nt

of

Ext

ern

alF

inan

ce

Mo

del

3B

Mo

del

4A(T

ob

it):

Am

ou

nt

of

Ext

ern

alF

inan

ce

Mo

del

4B

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)A

mo

un

to

fE

xter

nal

Fin

ance

Step

1(P

rob

it)

¼1

ifA

pp

lied

for

Ext

ern

alF

inan

ce

Step

2(H

ecki

t)A

mo

un

to

fE

xter

nal

Fin

ance

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)A

mo

un

to

fE

xter

nal

Fin

ance

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)A

mo

un

to

fE

xter

nal

Fin

ance

Hig

hT

ech

Man

ufa

ctu

rin

g�

136.

543

0.08

1*�

144.

619

�52

2.89

3*0.

241*

*�

920.

349*

76.6

550.

017

91.4

0024

.499

0.06

919

7.06

3

Co

nve

nti

on

alM

anu

fact

uri

ng

�18

0.50

6**

0.07

6***

�23

2.19

0�

547.

119*

**0.

108*

�93

7.04

7***

10.3

500.

017

1.76

7�

99.8

470.

124*

**79

.539

Hig

hT

ech

Serv

ices

110.

756

0.10

9*27

4.22

2�

299.

344

0.13

7�

534.

180

241.

805*

*0.

045

303.

018*

518.

451

0.21

9*11

19.7

27*

Co

nve

nti

on

alSe

rvic

es�

60.7

230.

028

�74

.776

�43

2.84

9**

�0.

051

�71

6.78

9**

�25

.254

0.01

4�

50.7

9786

.015

0.13

2**

328.

438

Hec

kman

Lam

bd

a37

2.43

0�

121.

359

�26

2.53

689

2.80

6M

odel

Dia

gnos

tics

Nu

mb

ero

fO

bse

rvat

ion

s95

22,

520

952

303

697

303

333

920

333

333

964

333

Lo

glik

elih

oo

d�

8291

.340

�1,

567.

909

�8,

280.

164�

2,71

8.60

1�

431.

343

�2,

707.

702

�2,

592.

797

�56

6.60

5�

2,58

1.19

2�

2,83

4.48

0�

576.

347

�2,

882.

751

Ad

just

edR

2

(Pse

ud

oR

2

for

Step

1P

rob

itM

od

els

and

Dec

om

pb

ased

fit

mea

sure

for

To

bit

)

0.23

30.

062

0.17

50.

250

0.09

60.

114

0.18

60.

059

0.16

30.

450

0.07

20.

508

FSt

atis

tic

(Ch

iSq

uar

edfo

rSt

ep1

Lo

git

Mo

del

san

dL

MT

est

for

To

bit

)

2787

.789

***

205.

529*

**11

.12*

**95

4.17

4***

91.6

46**

*2.

95**

*82

6.09

7***

71.1

36**

*4.

23**

*94

9.14

9***

90.0

47**

*18

.14*

**

Not

es.

Th

isT

able

pre

sen

tsT

ob

itan

dH

eckm

anco

rrec

ted

regr

essi

on

esti

mat

eso

fth

eam

ou

nt

of

exte

rnal

fin

ance

sou

ght.

Th

ed

epen

den

tva

riab

lein

Mo

del

s1A

,2A

,3A

,an

d4A

isth

e(s

tric

tly

po

siti

ve)

amo

un

to

fex

tern

alfi

nan

ceso

ugh

tb

yth

eb

usi

nes

sin

the

pas

ttw

oye

ars

(fro

m19

97),

and

ob

serv

atio

ns

wh

ere

no

exte

rnal

fin

ance

was

sou

ght

are

skip

ped

.M

od

els

1B,

2B,

3Ban

d4B

are

atw

o-s

tep

Hec

kman

corr

ecte

dm

od

els,

wh

ere

the

firs

tst

epco

nsi

der

sth

ep

rob

abil

ity

that

exte

rnal

fin

ance

was

sou

ght

and

the

seco

nd

step

acco

un

tsfo

rth

eam

ou

nt

of

exte

rnal

fin

ance

sou

ght.

Th

ein

dep

end

ent

vari

able

sar

eas

defi

ned

inT

able

1.T

he

tota

lp

op

ula

tio

no

ffi

rms

com

pri

sed

2,52

0U

Kfi

rms.

Mo

del

s1A

and

1Bco

mp

rise

the

full

sam

ple

.M

od

els

2Aan

d2B

com

pri

seth

esu

bsa

mp

leo

ffi

rms

form

edin

1990

-7.

Mo

del

s3A

and

3Bco

mp

rise

the

sub

sam

ple

of

firm

sfo

rmed

in19

80-9

.M

od

els

4Aan

d4B

com

pri

seth

esu

bsa

mp

leo

ffi

rms

star

ted

pri

or

to19

80.

Mar

gin

alef

fect

s(o

nly

)ar

ep

rese

nte

dso

that

the

eco

no

mic

sign

ifica

nce

issh

ow

nal

on

gsid

eth

est

atis

tica

lsi

gnifi

can

ce.

*,**

,**

*d

eno

tesi

gnifi

can

td

iffe

ren

cefo

rth

esa

mp

leo

fal

lo

ther

firm

sin

the

gro

up

atth

e10

%,

5%an

d1%

leve

ls,

resp

ecti

vely

.



expenditures/profits, or capital expenditures minus profits, also gives rise to a statis-tically significant and positive relation with the amount of capital sought.) The non-linear specification models the relation as a concave specification, such that a firm’sapplication for external capital increases at a decreasing rate as capital expenditures/profits increases. The intuition for this non-linear relation is that at very high levels ofcapital expenditures, firms are not as likely to apply for an equal amount of externalfinance as this would reduce the probability that such amounts would be granted by thefinancier. This issue is further addressed in the next subsection.

There is little evidence that size is correlated with the amount of external financesought. The only statistically significant evidence is a positive relation between turnoverand the amount of external finance sought but this effect is statistically significant onlyin Models 3A and 3B in Table 4. Assets are unrelated to the amount of external financesought in all specifications.

Consistent with Hypothesis 1a, the data provide evidence of a relation between profitmargins and the amount of external finance sought for the entire sample and thesubsample of firms started after 1990 (Models 1A, 1B, 2A and 2B in Table 4). Profitmargins, however, are negatively related to the extent of external finance sought forolder firms started before 1990 (Models 4A and 4B). These results are perhaps ex-plained by the fact that older firms with a longer track record exhibit comparatively lessinformation asymmetry for providers of external finance and are therefore not dis-couraged from seeking additional capital despite lower profit margins.

We further control for the capital structure of the firm by including debt/assets aswell as a squared term for debt/assets. Generally, the data indicate that for both ofthese variables there is an inverted U-shaped pattern in the relation between debt/assets and the amount of external finance sought given the positive and significantcoefficients for debt/assets and the negative and significant coefficients for the squaredvariable for debt/assets (both effects are most robust in Models 1A, 1B, 2A and 2B inTable 4). This inverted U-shaped relation may be intuitively explained as follows. Firms

0

5000

10000

15000

20000

25000

–9 –8 –7 –6 –5 –4 –3 –2 –1 0 1 2

Am

ount

of

Ext

erna

l Sou

ght (

thou

sand

s of

pou

nds)

Log (Capital Expenditures / Profits) [scaled]

Fig. 2. Amount of External Finance Sought and Captial Expenditures/Profits



with a track record with providers of external debt finance will be able to seek addi-tional capital from those sources of capital that have already carried out due diligenceto screen out bad loans. However, firms with too much debt relative to their asset basewill be discouraged from seeking significant amounts of external capital due to in-creased expected bankruptcy costs, particularly for the case of additional debt finance.In addition, external sources that had previously provided debt to the firm may havecovenants in place that restrict the amount of additional external capital that the firmcan seek.

4.2. The Percentage of External Finance Obtained

Table 5 presents evidence from OLS and Heckman-corrected regressions. Theseindicate that systematic characteristics drive the percentage of external finance ob-tained. The left-hand-side variable is transformed so that it is not bounded between 0and 100%, in a standard way of modelling fractions (Bierens, 2003). Specifically, if Y isa dependent variable that is bounded between 0 and 1 (i.e., a fraction), then a possibleway to model the distribution of Y conditional on a vector X of predetermined vari-ables, including 1 for the constant term, is to assume that

Y ¼ expðb0X þ U Þ1þ expðb0X þ U Þ ¼

1

1þ expð�b0X � U Þ

where U is an unobserved error term, then

ln½Y =ð1� Y Þ�b0X þ U

which, under standard assumptions on the error term U can be estimated by OLS.Consistent with the presentation in Table 4, Table 5 presents eight models. Models

5A, 6A, 7A and 8A are OLS specifications that do not account for Heckman selectioneffects. Models 5B, 6B, 7B and 8B are Heckman models that do account for Heckmanselection effects for the decision to seek external finance in the first place. The first stepof the Heckman corrected models in Table 5 is identical to the first step models inTable 4. As well, as in Table 4, Table 5 also reports the models for the full sample(Models 5A and 5B), the subsample for firms started in 1990–7 (Models 6A and 6B),firms started in 1980–9 (Models 7A and 7B) and firms started prior to 1980 (Models 8Aand 8B).13

The most robust result in Table 5 is that larger firms, in terms of their total assets, aremuch more likely to obtain a higher percentage of the total amount of finance sought.This effect is statistically significant in all of the different models presented in Table 5(as well as various others considered but not reported for reasons of succinctness). Interms of the economic significance, a firm with £1 million in assets will obtain

13 We considered other right-hand-side variables. For example, we considered the extent of externalfinance sought; those results (not reported for reasons of succinctness) generally showed a negative relationbetween the extent of external finance sought and the percentage of finance obtained. We do not reportspecifications with the extent of external finance sought as a right-hand-side variable in Table 5 because thatvariable is a choice variable (analysed as a dependent variable in Table 4 as discussed above). Regardless, theinclusion/exclusion of this variable does not materially impact the other results reported in Table 5 anddiscussed immediately below. Additional specifications are available on request.



Tab

le5

Reg

ress

ion

An

alys

esof

Per

cen

tage

ofE

xter

nal

Fin

ance

Obt

ain

ed

Fu

llSa

mp

le19

90–9

7Su

bsa

mp

le19

80–8

9Su

bsa

mp

leP

re–1

980

Sub

sam

ple

Mo

del

5A(O

LS)

:%

of

Ext

ern

alF

inan

ceO

bta

ined

Mo

del

5B

Mo

del

6A(O

LS)

:%

of

Ext

ern

alF

inan

ceO

bta

ined

Mo

del

6B

Mo

del

7A(O

LS)

:%

of

Ext

ern

alF

inan

ceO

bta

ined

Mo

del

7B

Mo

del

8A(O

LS)

:%

of

Ext

ern

alF

inan

ceO

bta

ined

Mo

del

8B

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)%

of

Ext

ern

alF

inan

ceO

bta

ined

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)%

of

Ext

ern

alF

inan

ceO

bta

ined

Step

1(P

rob

it)

¼1

ifA

pp

lied

for

Ext

ern

alF

inan

ce

Step

2(H

ecki

t)%

of

Ext

ern

alF

inan

ceO

bta

ined

Step

1(P

rob

it)

¼1

ifA

pp

lied

for

Ext

ern

alF

inan

ce

Step

2(H

ecki

t)%

of

Ext

ern

alF

inan

ceO

bta

ined

Co

nst

ant

1.22

6*�

0.34

5***

0.63

30.

685

�0.

712*

**0.

087

2.87

6*�

0.36

42.

296

2.15

2*0.

098

2.15

0*Fi

rmFi

nan

cial

Cha

ract

eris

tics

Lo

g(T

urn

ove

r)0.

069

�0.

003

0.06

70.

094

0.01

30.

101

0.06

9�

0.01

20.

050

0.08

7�

0.06

4**

0.08

2L

og

(Cap

ital

Exp

end

itu

res/

Pro

fits

)0.

035

0.02

6***

0.08

5*0.

009

0.01

7*0.

034

0.06

90.

024*

*0.

106

0.03

80.

039*

**0.

042

Pro

fit

Mar

gin

0.00

1***

4.38

0E-0

40.

002

0.00

10.

001

0.00

20.

092

�0.

006

0.07

61.

283*

**�

0.18

2***

1.26

3**

Deb

t/A

sset

s�

0.00

7�

0.00

10.

390

�0.

736

�0.

343

�0.

989

1.20

10.

244

1.84

81.

034

0.67

1**

1.09

4(D

ebt/

Ass

ets)

20.

055

�4.

203E

-07

0.00

40.

215

0.11

40.

293

�1.

231

�0.

0001

�1.

420

�0.

506

�0.

218

�0.

527

Lo

g(T

ota

lA

sset

s)0.

304*

**0.

006

0.32

5***

0.32

9***

0.01

90.

350*

**0.

302*

0.00

40.

336*

*0.

244*

0.03

50.

248*

Inn

ova

tio

n�

0.30

3**

0.00

5�

0.26

1*�

0.53

9**

0.04

8�

0.47

5*�

0.17

10.

007

�0.

112

�0.

116

�0.

024

�0.

117

Firm

Age

and

Pro

file

Lo

g(A

ge)

0.05

5�

0.04

3***

�0.

031

0.03

80.

006

0.01

7�

0.44

8�

0.05

9�

0.55

4�

0.15

60.

003

�0.

156

Pro

fess

ion

alD

irec

tors

�0.

227*

0.01

1�

0.20

00.

057

�0.

012

0.07

0�

0.42

2*0.

015

�0.

392*

�0.

451*

*0.

027

�0.

448*

*C

EO

Shar

es�

0.00

6**

2.35

5E-0

4�

0.00

6**

�0.

018*

**0.

001

�0.

018*

**�

0.00

8�

0.00

02�

0.00

80.

004

1.00

1E-0

40.

004

Bo

ard

Shar

es0.

003

1.45

2E-0

50.

003

0.01

0�

0.00

10.

010

0.00

70.

001

0.00

8�

0.01

0**

�3.

702E

-04

�0.

010*

*G

end

er0.

265

�0.

039

0.20

70.

547

�0.

013

0.52

8�

0.17

60.

003

�0.

161

0.39

8�

0.21

5**

0.37

8G

row

thO

bje

ctiv

es0.

114*

**0.

150*

**0.

116*

**0.

077*

**C

orp

ora

tio

n0.

062*

*�

0.03

70.

107*

**0.

061

Co

mp

any

fou

nd

edd

ue

tow

ealt

ham

bit

ion

so

ffo

un

der

s

0.01

50.

077*

�0.

005

�0.

008

Co

mp

etit

ors

/In

du

stry

Lar

ger

Co

mp

etit

ors

�0.

004

0.00

90.

015

�0.

163

0.01

1�

0.13

8�

0.18

7�

0.00

6�

0.18

20.

080

0.01

00.

081

Lo

g(T

ota

lC

om

pet

ito

rs)

�0.

022

0.02

2**

0.00

8�

0.10

00.

039*

*�

0.07

20.

099

0.02

30.

131

0.08

30.

008

0.08

4In

du

stry

(un

)In

no

vati

ven

ess

�0.

0004

�0.

001*

�0.

002

0.00

02�

0.00

02�

0.00

02�

0.00

2�

0.00

04�

0.00

4�

0.00

03�

0.00

1�

0.00

04H

igh

Tec

hM

anu

fact

uri

ng

�0.

290

0.08

1*�

0.12

6�

0.52

80.

241*

*�

0.29

8�

0.07

10.

017

�0.

012

�0.

215

0.06

9�

0.20

7C

on

ven

tio

nal

Man

ufa

ctu

rin

g�

0.03

70.

076*

**0.

089

�0.

205

0.10

8*�

0.09

1�

0.05

30.

017

�0.

016

0.13

00.

124*

**0.

141



Tab

le5

Con

tin

ued

Fu

llSa

mp

le19

90–1

997

Sub

sam

ple

1980

–198

9Su

bsa

mp

leP

re–1

980

Sub

sam

ple

Mo

del

5A(O

LS)

:%

of

Ext

ern

alF

inan

ceO

bta

ined

Mo

del

5B

Mo

del

6A(O

LS)

:%

of

Ext

ern

alF

inan

ceO

bta

ined

Mo

del

6B

Mo

del

7A(O

LS)

:%

of

Ext

ern

alF

inan

ceO

bta

ined

Mo

del

7B

Mo

del

8A(O

LS)

:%

of

Ext

ern

alF

inan

ceO

bta

ined

Mo

del

8B

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)%

of

Ext

ern

alF

inan

ceO

bta

ined

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rE

xter

nal

Fin

ance

Step

2(H

ecki

t)%

of

Ext

ern

alF

inan

ceO

bta

ined

Step

1(P

rob

it)

¼1

ifA

pp

lied

for

Ext

ern

alF

inan

ce

Step

2(H

ecki

t)%

of

Ext

ern

alF

inan

ceO

bta

ined

Step

1(P

rob

it)

¼1

ifA

pp

lied

for

Ext

ern

alF

inan

ce

Step

2(H

ecki

t)%

of

Ext

ern

alF

inan

ceO

bta

ined

Hig

hT

ech

Serv

ices

�0.

802*

**0.

109*

�0.

586*

�0.

603

0.13

7�

0.43

5�

1.07

7**

0.04

5�

0.93

1*�

0.74

90.

219*

�0.

729

Co

nve

nti

on

alSe

rvic

es�

0.03

30.

028

0.01

7�

0.05

4�

0.05

1�

0.08

4�

0.09

20.

014

�0.

055

0.00

60.

132*

*0.

017

Hec

kman

Lam

bd

a0.

830*

0.59

20.

713

0.04

6M

odel

Dia

gnos

tics

Nu

mb

ero

fO

bse

rvat

ion

s95

22,

520

952

303

697

303

333

920

333

333

964

333

Lo

glik

elih

oo

d�

1,87

1.68

6�

1,56

7.90

9�

1,85

9.06

8�

617.

614

�43

1.34

3�

606.

296

�65

8.71

8�

566.

605

�64

7.25

5�

595.

574

�57

6.34

7�

584.

727

Ad

just

edR

2(P

seu

do

R2

for

Step

1P

rob

itM

od

els)

0.13

30.

062

0.13

50.

093

0.09

60.

093

0.09

10.

059

0.09

10.

150

0.07

20.

147

FSt

atis

tic

(Ch

iSq

uar

edfo

rSt

ep1

Lo

git

Mo

del

s)8.

65**

*20

5.52

9***

8.44

***

2.63

***

91.6

46**

*2.

54**

*2.

74**

*71

.136

***

2.66

***

4.08

***

90.0

47**

*3.

86**

*

Not

es.T

his

Tab

lep

rese

nts

OL

San

dH

eckm

anco

rrec

ted

regr

essi

on

esti

mat

eso

fth

ep

erce

nta

geo

fex

tern

alfi

nan

ceo

bta

ined

.Th

ed

epen

den

tva

riab

lein

Mo

del

s5A

,6A

,7A

,an

d8A

and

step

(2)

of

Mo

del

s5B

,6B

,7B

,an

d8B

isln

[Y/

(1�

Y)]

,wh

ere

Yis

the

per

cen

tage

of

exte

rnal

fin

ance

sou

ght

by

the

bu

sin

ess

inth

ep

ast

two

year

s(f

rom

1997

).T

his

tran

sfo

rmat

ion

of

the

dep

end

ent

vari

able

enab

les

OL

Sto

be

use

dw

ith

ou

tb

ias

fro

mb

ein

gb

ou

nd

edb

elo

wb

yze

roo

rb

ou

nd

edab

ove

by

100%

.Ob

serv

atio

ns

wh

ere

no

exte

rnal

fin

ance

was

sou

ght

are

skip

ped

.Mo

del

s5B

,6B

,7B

,an

d8B

are

atw

o-s

tep

Hec

kman

corr

ecte

dm

od

el,w

her

eth

efi

rst

step

con

sid

ers

the

pro

bab

ilit

yth

atex

tern

alfi

nan

cew

asso

ugh

t,an

dth

ese

con

dst

epac

cou

nts

for

ln[Y

/(1�

Y)]

wh

ere

Yis

the

per

cen

tage

of

exte

rnal

fin

ance

ob

tain

ed.

Th

ein

dep

end

ent

vari

able

sar

eas

defi

ned

inT

able

1.T

he

tota

lp

op

ula

tio

no

ffi

rms

com

pri

sed

2,52

0U

Kfi

rms.

Th

eva

lues

pre

sen

ted

for

step

(1)

of

Mo

del

s5B

,6B

,7B

,an

d8B

are

no

tth

est

and

ard

logi

tco

effi

cien

ts;

rath

er,

they

are

the

mar

gin

alef

fect

sso

that

the

eco

no

mic

sign

ifica

nce

issh

ow

nal

on

gsid

eth

est

atis

tica

lsi

gnifi

can

ce.

*,**

,**

*Si

gnifi

can

td

iffe

ren

cefo

rth

esa

mp

leo

fal

lo

ther

firm

sin

the

gro

up

atth

e10

%,

5%an

d1%

leve

ls,

resp

ecti

vely

.M

od

els

5Aan

d5B

com

pri

seth

efu

llsa

mp

le.

Mo

del

s6A

and

6Bco

mp

rise

the

sub

sam

ple

of

firm

sfo

rmed

in19

90-7

.M

od

els

7Aan

d7B

com

pri

seth

esu

bsa

mp

leo

ffi

rms

form

edin

1980

–9.M

od

els

8Aan

d8B

com

pri

seth

esu

bsa

mp

leo

ffi

rms

star

ted

pri

or

to19

80.M

argi

nal

effe

cts

(on

ly)

are

pre

sen

ted

soth

atth

eec

on

om

icsi

gnifi

can

ceis

sho

wn

alo

ngs

ide

the

stat

isti

cal

sign

ifica

nce

.*,

**,

***

den

ote

sign

ifica

nt

dif

fere

nce

for

the

sam

ple

of

all

oth

erfi

rms

inth

egr

ou

pat

the

10%

,5%

and

1%le

vels

,re

spec

tive

ly.



approximately 5% more of their desired capital than the average firm in the samplewhich has assets of £375,000 (see Table 1). Overall, this shows generally that small firmstend to face a problem in obtaining all of their desired external capital. This evidence ishighly suggestive of a capital gap. It is possible that size is correlated with unobservedfeatures of the quality of the application for external finance (that is, unobservableover-and-above the variables considered in the regressions presented in Table 5). Thecorrelation between assets and the residuals in the regressions in Table 5 is 0.00 andthe residuals have statistical properties that are not materially different from thenormal distribution; therefore, there is no reason to believe that size is pickingup unobserved features of the quality of the loan applications. We may infer from thedata that smaller applicants in terms of size are disadvantaged in terms of obtaining allof their desired capital.

Table 5 further indicates that firms that recently developed a novel innovation (atechnologically new or significantly improved product) tend to obtain less of their desiredcapital. This effect is statistically significant in the full sample (Models 5A and 5B and thesubsample of firms started after 1990 (Models 6A and 6B). The economic significance issuch that a firm that was started within 7 years of applying for external finance andrecently developed an innovation will obtain approximately 8% less of their desiredcapital. Older firms started prior to 1990 that recently developed an innovation, bycontrast, do not have the same difficulties in obtaining their desired capital.

Table 5 shows older firms started prior to 1990 do nevertheless face problems inobtaining their desired finance when they have few professional directors and/or lowerprofit margins. First, older firms started prior to 1990 that do not have professionaldirectors tend to obtain approximately 1% less of their desired finance, and this effectis statistically significant at the 10% level in Models 7A and 7B (for the subsample offirms started in the 1980s), and significant at the 5% level in Models 8A and 8B (for thesubsample of firms started prior to 1980). Second, in regards to profit margins for firmsstarted prior to 1980, higher profit margins are positively and significantly related tothe percentage of finance obtained, consistent with Hypothesis 1b. If profit marginsincrease from 0 to the mean level of 0.909 (Table 1), then a firm is likely to increase thepercentage of external capital obtained by approximately 4%.

In sum, the evidence in Table 5 shows that asset size matters most for obtaining thedesired level of external finance. Small firms that have recently developed a novelinnovation tend to obtain less of their desired capital. This is expected as smallerinnovative firms exhibit more pronounced information asymmetries when applying forexternal capital. Larger firms that do not have professional directors and/or are lessprofitable tend to obtain less of their desired finance. This is also expected as largerfirms that have not put in place professional governance mechanisms and/or are lessfinancially successful are less attractive candidates for external finance. The next Sec-tion below considers whether there are differences in firms� ability to obtain theirdesired finance across the different sources of capital.

4.3. Differences Across Alternative Sources of Capital

Before formally examining specific sources of capital, we first consider ordered logitevidence on the number of different sources of capital sought in Table 6. This evidence



Table 6

Regression Analyses of The Number of Sources of Capital Sought

Model 9(Ordered Logit):no. of DifferentTypes of Sources

Approached(Full Sample)


Approached(Post-1990

Subsample)


Approached(1980-9

Subsample)


Approached(Pre-1980

Subsample)

Constant �0.840*** �1.994*** �0.058 0.142Firm FinancialCharacteristics

Log (Turnover) �0.018 0.023 �0.038 �0.147**Log (CapitalExpenditures/Profits)

0.072*** 0.049** 0.064*** 0.109***

Profit Margin �0.0001 0.0001 �0.030 �0.510***Debt/Assets 0.436** �0.385 0.336 1.765***(Debt/Assets)2 �0.011 0.115 �0.001 �0.663Log (Total Assets) 0.031 0.058* 0.023 0.091Innovation 0.054 0.182 0.075 �0.089Firm Age and ProfileLog (Age) �0.130*** 0.008 �0.393** �0.036Professional Directors 0.017 �0.013 0.011 0.080CEO Shares 1.272E-03 0.005** �1.033E-04 �0.0001Board Shares 3.141E-04 �1.058E-04 9.616E-04 0.0002Gender �0.068 0.040 �0.070 �0.541**Growth Objectives 0.266*** 0.363*** 0.287*** 0.216***Corporation 0.160** �0.026 0.285*** 0.124Company foundeddue to wealthambitions of founders

0.017 0.142 0.016 �0.051

Competitors/IndustryLarger Competitors 0.021 0.019 �0.019 0.019Log (Total Competitors) 0.043* 0.060 0.029 0.057Industry (un)Innovativeness �0.002** �0.001 �0.001 �0.003*High Tech Manufacturing 0.311*** 0.578*** 0.108 0.186ConventionalManufacturing

0.219*** 0.307** 0.038 0.296***

High Tech Services 0.344*** 0.266 0.081 0.664**Conventional Services 0.072 �0.135 �0.025 0.300**Ordered LogitCut-off Parameters

Mu(1) 0.427*** 0.465*** 0.401*** 0.430***Mu(2) 1.120*** 1.106*** 1.123*** 1.165***Mu(3) 1.762*** 1.788*** 1.821*** 1.755***Mu(4) 2.383*** 2.372*** 2.376*** 2.518***Model DiagnosticsNumber of Observations 2,520 697 920 964Loglikelihood �2,738.542 �821.418 �966.516 �971.817Pseudo R2 0.045 0.069 0.041 0.049Chi-squared Statistic 258.514*** 120.946*** 82.964*** 100.288***

Notes. This Table presents ordered logit estimates of the number of different sources of finance approached.The dependent variable in Models 9–12 is the number of different types of sources approached, where Model9 is for the full sample, Model 10 is for the subsample of firms started after 1990, Model 11 is for thesubsample of firms started between 1980 and 1989, and Model 12 is for the subsample of firm started prior to1980. The independent variables are as defined in Table 1. The total population of firms comprised 2,520 UKfirms. *, **, *** denote significant difference for the sample of all other firms in the group at the 10%, 5%and 1% levels, respectively.



in Table 6 provides a context in which to examine differences across different sourcesof capital. The evidence in Table 6 is strongly consistent with Hypothesis 2a. Firms withsignificant growth objectives are more likely to approach a greater number of sourcesof capital; this effect is statistically significant at the 1% level in each of the Models 9–12. A firm with a higher ranking by 1 on a scale of 1–4 is approximately 1% more likelyto apply to an additional type of source of capital. This is economically significant inview of the fact that among the 952 of 2,520 firms that did apply for external finance,the median number of different sources approached was 2 (and mean was 1.96)(Tables 1 and 2).

Also, much of the evidence in Table 6 is strongly consistent with Hypothesis 3. First,firms are more likely to approach a greater number of different sources of capital whentheir capital expenditures/profits are greater, as shown in all specifications in Table 6.This is consistent with Hypothesis 3. Second, there is evidence from Model 12 inTable 6 for firms started before 1980 with lower profit margins must approach a greaternumber of different sources of external capital. It is natural that new firms may not yethave achieved profitability and hence current profit margins are less important forexternal finance for young firms relative to their more established counterparts. Third,firms with higher levels of debt/assets approach a greater number of different sourcesof capital for the full sample (Model 9) and the subsample of firms started before 1980(Model 12). As with profit margins, we would expect debt/asset levels to play a pro-nounced role in external capital for more established firms.

Table 7(a) and (b) explores further the issues addressed above by considering thedifferences between banks, venture capital funds, hire purchase/leasing firms, factor-ing/invoice discounting firms, trade customers/suppliers, partners/working share-holders, private individuals and other sources of capital.14 As discussed above insubsection 2.1 (Hypothesis 2b), among the approximately 38% of 2,520 firms in oursample that did seek external finance in the 1996–7 period considered, 775 ap-proached banks and only 87 approached venture capital funds (the other sourcesapproached were identified in Section 2 and Table 2 above). Outright rejection rateswere highest among venture capital funds (46% rejection), and much higher than thatfor banks (17% outright rejection). Table 7 addresses in a multivariate context whetherthe factors leading to financing by these distinct sources were materially differentacross each of the 8 different sources of capital.

Table 7(a) presents OLS models for the percentage of external capital obtainedfrom each particular source of finance. Also, to show robustness, Table 7(b) presentsHeckman-corrected estimates of the percentage of external capital obtained from eachof the 8 different sources of capital. The first step of Part (b) of each Model for theHeckman regressions is a probit model that determines whether the firm soughtexternal capital from the particular source in question. The percentage term left-hand-side variables make use of the transformation as described in Table 5 and theaccompanying text above. Likewise, the right-hand-side variable specifications areconsistent with Table 5.

14 Models 13–20 are presented for the full sample of firms. The results from the subsamples by differentyears in which the firms were founded did not give rise to materially different inferences that could be drawnfrom the data and are hence excluded for reasons of succinctness (otherwise there were be an extra 24regression models). These regression results are nevertheless available on request.



Tab

le7

Reg

ress

ion

An

alys

esof

Per

cen

tage

ofE

xter

nal

Fin

ance

Obt

ain

edfr

omSp

ecifi

cSo

urc

esof

Cap

ital

(a)

Ban

ksV

entu

reC

apit

alH

ire

Pu

rch

ase

or

Lea

sin

gF

irm

sF

acto

rin

g/In

voic

eD

isco

un

tin

gF

irm

s

Mo

del

13A

(OL

S):

%o

fB

ank

Fin

ance

Ob

tain

ed

Mo

del

13B

Mo

del

14A

(OL

S):

%o

fV

entu

reC

apit

alF

inan

ceO

bta

ined

Mo

del

14B

Mo

del

15A

(OL

S):

%o

fL

easi

ng

Fin

ance

Ob

tain

ed

Mo

del

15B

Mo

del

16A

(OL

S):

%o

fF

acto

rin

gF

inan

ceO

bta

ined

Mo

del

16B

Step

1(P

rob

it)

¼1

ifA

pp

lied

for

Ban

kF

inan

ce

Step

2(H

ecki

t)%

of

Ban

kF

inan

ceO

bta

ined

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rV

entu

reC

apit

alF

inan

ce

Step

2(H

ecki

t)%

of

Ven

ture

Cap

ital

Fin

ance

Ob

tain

ed

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rL

easi

ng

Fin

ance

Step

2(H

ecki

t)%

of

Lea

sin

gF

inan

ceO

bta

ined

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rF

acto

rin

gF

inan

ce

Step

2(H

ecki

t)%

of

Fac

tori

ng

Ob

tain

ed

Co

nst

ant

�2.

684*

*�

0.44

1***

�4.

336*

**1.

768

�0.

043*

**2.

136

�1.

253

�0.

398*

**�

4.72

3**

�5.

300*

*�

0.17

6***

�1.

082

Firm

Fin

anci

alC

hara

cter

isti

csL

og

(Tu

rno

ver)

0.05

2�

0.00

30.

050

0.20

6�

0.00

040.

197

�0.

026

�0.

007

�0.

072

0.32

7*�

0.00

10.

384*

Lo

g(C

apit

alE

xpen

dit

ure

s/P

rofi

ts)

�0.

051

0.02

0***

0.03

9�

0.21

80.

0004

�0.

232

0.18

8**

0.03

0***

0.46

0***

0.02

4�

0.00

01�

0.00

2

Pro

fit

Mar

gin

0.18

6�

0.00

80.

160

�0.

175

�0.

007*

�0.

110

0.17

5�

0.00

90.

119

0.37

1***

�0.

014

0.66

8*D

ebt/

Ass

ets

�0.

357

0.09

10.

187

�2.

895

�0.

002

�2.

923

�0.

211

�0.

0004

0.11

7�

1.94

90.

028*

�3.

123

(Deb

t/A

sset

s)2

0.07

3�

4.63

2E-0

50.

023

0.48

0�

1.39

8E-0

50.

588

Lo

g(T

ota

lA

sset

s)0.

302*

*0.

008

0.34

5***

0.29

10.

002

0.28

00.

103

0.01

20.

250

0.29

40.

006*

*0.

070

Inn

ova

tio

n�

0.18

10.

004

�0.

061

�1.

552*

*0.

005*

�1.

584*

*�

0.32

2�

0.03

4**

�0.

542*

�0.

069

�0.

001

�0.

103

Firm

Age

and

Pro

file

Lo

g(A

ge)

0.05

9�

0.04

3***

�0.

133

�0.

001

�0.

004*

**0.

049

0.42

2***

�0.

002

0.31

1*�

0.13

5�

0.01

4***

0.25

7P

rofe

ssio

nal

Dir

ecto

rs�

0.07

00.

008

�0.

014

0.16

70.

004*

0.12

2�

0.55

6**

�0.

008

�0.

581*

*�

0.50

8�

0.01

1*�

0.36

2C

EO

Shar

es�

3.76

0E-0

31.

830E

-04

�0.

003

�0.

018

0.00

0�

0.01

8�

0.00

42.

466E

-05

�0.

005

�0.

017*

*0.

0002

**�

0.02

2**

Bo

ard

Shar

es�

2.78

4E-0

33.

753E

-04

�0.

001

�0.

036*

�0.

0002

***�

0.03

4*�

0.00

30.

001*

**0.

008

8.98

9E-0

37.

996E

-05

0.00

6G

end

er0.

787*

�0.

008

0.76

0*�

2.19

7**

�0.

011

�2.

155

�0.

014

�0.

051

�0.

405

0.24

50.

003

0.14

0G

row

thO

bje

ctiv

es0.

107*

**0.

008*

**0.

062*

**0.

017*

**C

orp

ora

tio

n0.

030

0.00

9**

0.01

70.

017*

*C

om

pan

yfo

un

ded

du

eto

wea

lth

amb

itio

ns

of

fou

nd

ers

�0.

001

0.00

10.

010

0.00

2

Com

peti

tors

/In

dust

ryL

arge

rC

om

pet

ito

rs�

0.03

90.

016

0.03

90.

339

�0.

001

0.34

80.

115

0.00

30.

169

0.01

8�

0.00

50.

142

Lo

g(T

ota

lC

om

pet

ito

rs)�

0.10

20.

017*

�0.

049

�0.

332

�0.

002*

�0.

295

0.20

6*0.

010

0.26

5**

�0.

045

0.00

7***

�0.

213

Ind

ust

ryIn

no

vati

ven

ess

0.00

4�

0.00

040.

0020

�0.

014

1.91

4E-0

5�

0.01

40.

001

�0.

0004

�0.

002

0.00

50.

000

0.00

8H

igh

Tec

hM

anu

fact

uri

ng

�0.

996*

*0.

074

�0.

698

0.31

20.

002

0.21

70.

039

0.09

4**

0.82

8�

0.48

90.

024

�0.

923

Co

nve

nti

on

alM

anu

fact

uri

ng

�0.

367

0.05

8**

�0.

162

0.62

2�

0.00

30.

609

0.25

60.

080*

**0.

892*

0.50

40.

023*

**�

0.00

4



Tab

le7

Con

tin

ued

(a)

Ban

ksV

entu

reC

apit

alH

ire

Pu

rch

ase

or

Lea

sin

gF

irm

sF

acto

rin

g/In

voic

eD

isco

un

tin

gF

irm

s

Mo

del

13A

(OL

S):

%o

fB

ank

Fin

ance

Ob

tain

ed

Mo

del

13B

Mo

del

14A

(OL

S):

%o

fV

entu

reC

apit

alF

inan

ceO

bta

ined

Mo

del

14B

Mo

del

15A

(OL

S):

%o

fL

easi

ng

Fin

ance

Ob

tain

ed

Mo

del

15B

Mo

del

16A

(OL

S):

%o

fF

acto

rin

gF

inan

ceO

bta

ined

Mo

del

16B

Step

1(P

rob

it)

¼1

ifA

pp

lied

for

Ban

kF

inan

ce

Step

2(H

ecki

t)%

of

Ban

kF

inan

ceO

bta

ined

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rV

entu

reC

apit

alF

inan

ce

Step

2(H

ecki

t)%

of

Ven

ture

Cap

ital

Fin

ance

Ob

tain

ed

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rL

easi

ng

Fin

ance

Step

2(H

ecki

t)%

of

Lea

sin

gF

inan

ceO

bta

ined

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rF

acto

rin

gF

inan

ce

Step

2(H

ecki

t)%

of

Fac

tori

ng

Ob

tain

ed

Hig

hT

ech

Serv

ices

�1.

641*

**0.

091*

�1.

209*

*0.

359

0.01

6*0.

173

�0.

323

0.08

9*0.

416

�0.

820

0.00

8�

1.12

2C

on

ven

tio

nal

Serv

ices

0.09

90.

009

0.13

90.

784

0.00

20.

722

0.41

10.

009

0.51

70.

112

�0.

002

0.22

2H

eckm

anL

amb

da

1.65

5*�

0.26

92.

436*

*�

1.71

6M

odel

Dia

gnos

tics

Nu

mb

ero

fO

bse

rvat

ion

s77

52,

520

775

872,

520

8747

42,

520

474

151

2,52

015

1

Lo

glik

elih

oo

d�

2000

.767

�14

80.3

04�

1988

.089

�20

5.82

0�

321.

398

�19

4.45

2�

1117

.485

�11

23.5

55�

1104

.289

�32

8.10

1�

514.

613

�31

5.68

4A

dju

sted

R2

(Pse

ud

oR

2

for

Step

1P

rob

itM

od

els)

0.03

60.

054

0.04

00.

085

0.15

80.

072

0.06

20.

085

0.07

10.

081

0.10

80.

088

FSt

atis

tic

(Ch

iSq

uar

edfo

rSt

ep1

Lo

git

Mo

del

s)

2.55

***

170.

456*

**2.

63**

*1.

4512

0.51

7***

1.36

2.75

***

209.

822*

**2.

93**

*1.

71**

124.

573*

**1.

73**

*



Tab

le7

Con

tin

ued

(b)

Tra

de

Cu

sto

mer

s/Su

pp

lier

sP

artn

ers/

Wo

rkin

gSh

areh

old

ers

Oth

erP

riva

teIn

div

idu

als

Oth

er

Mo

del

17A

(OL

S):

%o

fT

rad

eF

inan

ceO

bta

ined

Mo

del

17B

Mo

del

18A

(OL

S):

%o

fP

artn

erF

inan

ceO

bta

ined

Mo

del

18B

Mo

del

19A

(OL

S):

%o

fP

riva

teIn

div

idu

alF

inan

ceO

bta

ined

Mo

del

19B

Mo

del

20A

(OL

S):

%o

fO

ther

Fin

ance

Ob

tain

ed

Mo

del

20B

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rT

rad

eF

inan

ce

Step

2(H

ecki

t)%

of

Tra

de

Fin

ance

Ob

tain

ed

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rP

artn

erF

inan

ce

Step

2(H

ecki

t)%

of

Par

tner

Fin

ance

Ob

tain

ed

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rP

riva

teF

inan

ce

Step

2(H

ecki

t)%

of

Pri

vate

Fin

ance

Ob

tain

ed

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rO

ther

Fin

ance

Step

2(H

ecki

t)%

of

Oth

erF

inan

ceO

bta

ined

Co

nst

ant

�3.

246

�0.

073*

**�

1.64

6�

4.35

2**

�0.

086*

*�

3.26

08.

789*

*�

0.06

2*5.

120

�1.

377

�0.

042

�22

.407

Firm

Fin

anci

alC

hara

cter

isti

csL

og

(Tu

rno

ver)

1.10

0***

0.00

01.

109*

**0.

331*

�0.

008*

**0.

499*

�0.

173

�0.

001

�0.

220

Lo

g(C

apit

alE

xpen

dit

ure

s/P

rofi

ts)

0.15

50.

001

0.13

5�

0.28

4*0.

004*

*�

0.39

2**

�0.

056

�0.

001

�0.

036

�0.

134

0.00

20.

401

Pro

fit

Mar

gin

2.84

4***

�0.

0004

2.86

3***

0.72

4***

�0.

014*

*1.

026*

�0.

603

�0.

007

�0.

970

0.00

3**

0.00

020.

027

Deb

t/A

sset

s2.

236

�0.

003

2.25

92.

186*

**�

3.21

6E-0

52.

120*

*�

0.43

8�

2.38

4E-0

5�

0.42

3�

3.06

4�

0.00

01�

2.05

8L

og

(To

tal

Ass

ets)

�0.

418

0.00

2�

0.44

9�

0.25

20.

004

�0.

343

�0.

405*

*0.

0001

�0.

397

0.24

8�

0.00

2�

0.14

2In

no

vati

on

�0.

127

0.00

6�

0.22

5�

0.42

70.

011

�0.

627

�0.

708

0.01

0�

0.20

7�

0.07

60.

022*

*4.

898

Firm

Age

and

Pro

file

Lo

g(A

ge)

0.60

3**

�0.

007*

**0.

730

0.36

1�

0.00

60.

545*

0.22

0�

0.00

9***

�0.

290

1.31

4***

�0.

001

0.52

5P

rofe

ssio

nal

Dir

ecto

rs�

0.70

10.

0002

�0.

695

�0.

277

0.00

9�

0.42

5�

0.12

90.

012*

*0.

523

�0.

955

�0.

009

�2.

649

CE

OSh

ares

�0.

019*

*4.

726E

-05�

0.02

0*�

0.00

5�

0.00

01�

0.00

2�

0.01

86.

557E

-05�

1.42

6E-0

28.

150E

-04

0.00

03**

0.07

0B

oar

dSh

ares

�7.

760E

-03

0.00

03*

�0.

013

0.01

2�

4.93

7E-0

50.

014

�0.

016

2.56

0E-0

5�

1.49

0E-0

2�

0.04

3**

�0.

0003

*�

0.11

3G

end

er0.

197

0.00

60.

081

�0.

957*

�0.

001

�0.

904

�1.

004

�0.

001

�1.

299

�0.

199

0.00

040.

531

Gro

wth

Ob

ject

ives

0.00

30.

013*

*0.

005

0.00

5C

orp

ora

tio

n�

0.00

30.

018*

*0.

011*

*�

0.00

1C

om

pan

yfo

un

ded

du

eto

wea

lth

amb

itio

ns

of

fou

nd

ers

�0.

002

0.00

4�

0.00

2�

0.00

3

Com

peti

tors

/In

dust

ryL

arge

rC

om

pet

ito

rs�

0.56

3�

0.00

1�

0.57

4�

0.45

6�

0.00

05�

0.48

4�

1.72

4**

�0.

004

�1.

808*

0.16

6�

0.00

1�

0.09

2L

og

(To

tal

Co

mp

etit

ors

)�

0.69

8***

0.00

1�

0.70

9***

0.27

20.

002

0.25

90.

314

�0.

001

0.23

00.

126

�0.

001

�0.

103

Ind

ust

ryIn

no

vati

ven

ess

�0.

001

�0.

0001

**0.

001

0.00

1�

1.56

0E-0

40.

005

�0.

011

�0.

0001

�0.

016

�0.

001

�0.

0002

�0.

047

Hig

hT

ech

Man

ufa

ctu

rin

g�

2.41

5**

0.00

4�

2.43

4**

0.24

90.

044*

�0.

281

�3.

785*

**�

0.00

2�

3.80

6***

2.49

8�

0.01

5**

�1.

935

Co

nve

nti

on

alM

anu

fact

uri

ng

�2.

060*

**0.

005

�2.

120*

**0.

411

0.00

40.

447

�2.

768*

**�

0.01

3**

�3.

497*

**0.

021

0.00

21.

084

Hig

hT

ech

Serv

ices

�0.

376

0.00

3�

0.37

71.

705*

0.03

01.

327

�3.

277*

**0.

002

�3.

163*

*�

1.35

40.

007

0.54

0C

on

ven

tio

nal

Serv

ices

�1.

966*

**0.

001

�1.

961*

**0.

015

0.01

6�

0.15

2�

1.62

9**

0.00

4�

1.59

8*�

0.54

2�

0.00

6�

1.29

2H

eckm

anL

amb

da

�0.

540

�1.

505

2.56

013

.955



Tab

le7

Con

tin

ued

(b)

Tra

de

Cu

sto

mer

s/Su

pp

lier

sP

artn

ers/

Wo

rkin

gSh

areh

old

ers

Oth

erP

riva

teIn

div

idu

als

Oth

er

Mo

del

17A

(OL

S):

%o

fT

rad

eF

inan

ceO

bta

ined

Mo

del

17B

Mo

del

18A

(OL

S):

%o

fP

artn

erF

inan

ceO

bta

ined

Mo

del

18B

Mo

del

19A

(OL

S):

%o

fP

riva

teIn

div

idu

alF

inan

ceO

bta

ined

Mo

del

19B

Mo

del

20A

(OL

S):

%o

fO

ther

Fin

ance

Ob

tain

ed

Mo

del

20B

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rT

rad

eF

inan

ce

Step

2(H

ecki

t)%

of

Tra

de

Fin

ance

Ob

tain

ed

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rP

artn

erF

inan

ce

Step

2(H

ecki

t)%

of

Par

tner

Fin

ance

Ob

tain

ed

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rP

riva

teF

inan

ce

Step

2(H

ecki

t)%

of

Pri

vate

Fin

ance

Ob

tain

ed

Step

1(P

rob

it)¼

1if

Ap

pli

edfo

rO

ther

Fin

ance

Step

2(H

ecki

t)%

of

Oth

erF

inan

ceO

bta

ined

Mo

del

Dia

gno

stic

sN

um

ber

of

Ob

serv

atio

ns

532,

520

5313

82,

520

138

882,

520

8867

2,52

067

Lo

glik

elih

oo

d�

92.1

36�

238.

258

�79

.634

�31

8.31

9�

513.

335

�30

7.14

6�

206.

114�

349.

761

�19

4.40

9�

165.

650

�30

4.41

0�

152.

205

Ad

just

edR

2

(Pse

ud

oR

2fo

rSt

ep1

Pro

bit

Mo

del

s)

0.26

20.

087

0.24

10.

044

0.07

50.

042

0.02

60.

084

0.02

00.

157

0.06

90.

197

FSt

atis

tic

(Ch

iSq

uar

edfo

rSt

ep1

Lo

git

Mo

del

s)

2.05

**45

.366

**1.

88*

1.37

82.8

53**

*1.

331.

1363

.791

***

1.09

1.78

*45

.074

***

1.97

**

Not

es.

Th

isT

able

pre

sen

tsO

LS

and

Hec

kman

corr

ecte

dre

gres

sio

nes

tim

ates

of

the

per

cen

tage

of

exte

rnal

fin

ance

ob

tain

ed.

Th

ed

epen

den

tva

riab

lein

step

(1)

of

each

of

Par

tB

of

Mo

del

s13�

20is

ad

um

my

vari

able

equ

alto

1fo

rfi

rms

that

app

lied

toth

ere

spec

tive

sou

rce

of

cap

ital

ind

icat

edan

d0

oth

erw

ise.

Th

ed

epen

den

tva

riab

lein

Par

tA

and

step

(2)

of

Par

tB

of

Mo

del

s13

-20

isln

[Y/

(1�

Y)]

,w

her

eY

isth

ep

erce

nta

geo

fex

tern

alfi

nan

ceo

bta

ined

fro

mth

esp

ecifi

cso

urc

eo

fca

pit

alin

dic

ated

inth

ep

ast

two

year

s(f

rom

1997

).T

his

tran

sfo

rmat

ion

of

the

dep

end

ent

vari

able

enab

les

OL

Sto

be

use

dw

ith

ou

tb

ias

fro

mb

ein

gb

ou

nd

edb

elo

wb

yze

roo

rb

ou

nd

edab

ove

by

100%

.T

he

ind

epen

den

tva

riab

les

are

asd

efin

edin

Tab

le1.

Var

iab

les

skip

ped

wh

ere

coll

inea

rity

cau

sed

esti

mat

ion

pro

ble

ms.

Th

eto

tal

po

pu

lati

on

of

firm

sco

mp

rise

d2,

520

UK

firm

s.*,

**,

***

Sign

ifica

nt

dif

fere

nce

for

the

sam

ple

of

all

oth

erfi

rms

inth

egr

ou

pat

the

10%

,5%

and

1%le

vels

,re

spec

tive

ly.

*,**

,**

*d

eno

tesi

gnifi

can

td

iffe

ren

cefo

rth

esa

mp

leo

fal

lo

ther

firm

sin

the

gro

up

atth

e10

%,

5%an

d1%

leve

ls,

resp

ecti

vely

.



The Step 1 regressions in Table 7(a) show firms with significant growth objectives aremore likely to apply for external finance from banks, venture capital funds, leasingfirms, factoring/invoice discounting firms and partners/working shareholders. This isconsistent with Hypothesis 2a that growth objectives are a significant determinant ofthe decision to seek external capital. In Step 1 regressions in Table 7(b) there are alsonotable differences in terms of which firms apply for different types of finance. Firmswith greater capital expenditures/profits are only more likely to apply for bank finance,lease finance and partner/working shareholder finance and not finance from othertypes of sources of capital. These results are consistent with Hypotheses 1a and 2b. Bankand lease finance are sources of debt capital and we conjectured that entrepreneursgenerally seek debt prior to equity in order to avoid dilution of their ownership shares.Partner/working shareholder finance is external capital but capital from preexistingsources that had already financed the firm. Partners/working shareholders face fewerinformation asymmetries relative to other sources of capital in carrying out due dili-gence as partners/working shareholders have a relationship with the firm. The cost ofexternal capital is smaller when information problems are less pronounced. As well,note that less profitable firms are more likely to apply for venture capital and partner/working shareholder finance. This is expected since venture capital funds and part-ners/working shareholders are better able to carry out due diligence and mitigateadverse selection problems than other sources of capital.

Table 7(a) and Step 2 of the regressions in Table 7(b) consider the percentage offinance obtained. Consistent with Hypothesis 1b, profit margins are a statisticallysignificant variable for successfully obtaining the desired amount of external financefrom factoring/invoice discounting firms, trade customers/suppliers and partners/working shareholders but not the other sources of capital.

Larger firms in terms of asset size are more likely to obtain bank finance but asset sizedoes not impact on the percentage of finance obtained from other sources. In terms ofthe economic significance, a firm with £1 million in assets will obtain approximately 8%more of their desired capital than the average firm in the sample which has assets of£375,000 (see Table 1). Problems associated with capital gaps for smaller firms aretherefore pronounced with banks but this positive relation between assets and per-centage of desired capital obtained is not statistically significant with other sources ofcapital. By contrast, in the case of private individuals as a source of capital there is evenevidence that smaller firms obtain a higher proportion of their desired capital (Model19A) but this effect is not statistically significant after controlling for selection effects.The models are more robust in showing that private individuals provide a greaterpercentage of their desired capital to smaller firms facing larger competitors.

Banks provide high-tech services firms with 28% less of their desired capital. Thisresult is consistent with Hypothesis 2b in that banks provide less capital to smaller riskyventures. Further, the data indicate that banks provide male-founded firms with 18%less of their desired capital relative to female-founded firms. There is much evidence(Jianakoplos and Bernasek, 1998) that females are more risk averse than males and,hence, this evidence is consistent with the prediction in Hypothesis 2b that banksprovide less capital to riskier ventures.

It is noteworthy from the Part B Step 1 regressions in Table 7 that firms thatrecently developed a new innovation are more likely to apply for venture capital finance,



consistent with Hypothesis 2b. Venture capital funds attract innovative companies thatrequire value-added assistance and exhibit high information asymmetries and growthprospects (Berger and Udell, 2002; Carpenter and Petersen, 2002a,b; Cressy, 1996, 2002).Also, the data show venture capital is more sought after by younger firms with professionaldirectors and more intensive growth objectives (Step 1 of Model 14B). This is intuitive asprofessional directors provide the governance required to mitigate information asym-metries and agency costs associated with financing early stage high-tech companies(Gompers and Lerner, 1999). However, venture capital funds provide firms with 4% lessof their desired capital when they have recently developed a new innovation. Venturecapital funds invariably stage their investments, and stage more frequently for high-techinnovative firms, in order to mitigate information asymmetries and agency costs.

When the board members hold a greater percentage of the equity shares in the firm,venture capital funds provide a smaller percentage of the desired capital. A changefrom the mean level of 86% equity held by the board to 50%, for example, increases thepercentage of desired finance provided by venture capitalists by approximately 8%.This result is likely attributable to the fact that venture capitalists demand significantequity shares in their firms and a reduction in equity held by board members impliesthere is greater scope for outside equity to be sold to the venture capitalists. A greaterequity share held by the CEO, however, does not reduce the percentage of venturecapital obtained as it is relatively more important to maintain a high CEO ownershipshare to mitigate moral hazard problems.

In sum, the evidence in Table 7 shows there are systematic differences in firms� ability toobtain their desired capital in the form that they would like. Even after controlling forselection effects, the data indicate banks are more likely to provide the desired amount ofcapital to larger firms with more assets. Leasing firms, factor discounting/invoicing firms,trade customers/suppliers and partners/working shareholders are more likely to providethe desired capital to firms with higher profit margins. The data further show smallerfirms are more likely to obtain finance from private individuals, and innovative firms seekexternal capital from venture capital funds. These results are generally consistent with thehypotheses set out in subsection 2.1 above.

5. Additional Robustness Checks, Extensions and Future Research

This article introduced a very expansive and detailed new dataset which significantlyextends the literature on the financing decisions of entrepreneurial firms. The data areunique in that they are derived from the entrepreneurial firms themselves and not thefinancial institutions that provided the financing. This enables a broad perspective onwhich firms seek external finance, from which types of institutions do they seek capital,how much is sought, and how much is obtained. The data enable selection effects inthe spirit of Heckman (1976, 1979) to be considered in a unique way in the literatureon entrepreneurial capital sourcing decisions.

The introduction of new data invariably gives rise to questions about sample selectionand robustness. However in this case we have a very large sample spanning the full sizerange of small and medium-sized entrepreneurial firms which is free from responsebias in terms of key finance-related factors such as size age or profitability. Completedetails on survey design and sample selection issues are provided in Bullock and



Hughes (1998). In our econometric tests we explicitly considered sample selectionissues with regard to different types of firms applying for external finance among otherthings. The robustness of our results was explicitly considered. Additional detailsregarding the data and additional econometric tests are available on request.

It is noteworthy that our data are derived from a �snapshot� of a cross-section ofentrepreneurial firms in period leading up to 1997. The CBR data used here arehowever part of a unique panel data set which tracks the same firms through time. Infuture work we will to exploit this aspect of the data to see what happened to thebusinesses in subsequent years after they obtained external finance. How was themoney spent? Did some firms fail in subsequent years due to the fact that they did notobtain all of their requisite capital? Did obtaining external finance enable the busi-nesses to become more profitable in the years subsequent to obtaining externalfinance?

Our data indicated an absence of a capital gap in that most firms applying forexternal finance were able to obtain their requisite capital. As such there does notappear to be a generic finance gap in the UK entrepreneurial business sector, at least interms of the quantity of external finance available in the market. Nevertheless, theex post analysis of the impact of external finance on subsequent entrepreneurial firmperformance (which would be post-1997 in our data, for example), would shed light onthe quality of external finance – in the spirit of Storey and Wynarczyk (1996). Whilethat type of ex post question poses some interesting issues worthy of future data col-lection efforts, we must leave it for future research.

6. Summary and Conclusion

This article investigated the internal versus external financing decisions among aunique dataset comprising 2,520 early stage privately held UK firms in 1996–7. We firstidentified factors that lead firms to seek external finance. The most robust evidenceindicated that firms with higher capital expenditures/profits and firms with strongergrowth objectives were much more likely to seek external finance. We then demon-strated the existence of systematic differences in the amount of external financeactually sought across businesses. We found a strong and robust relation between afirm’s capital expenditures/profits and the amount of external finance actually sought.Overall, we view the evidence as providing strong support for the traditional peckingorder theory which predicts that firms prefer to finance new projects internally prior toseeking external capital.

The data also indicated systematic differences in the percentage of external financeobtained relative to that sought in ways that are consistent with information asymme-tries faced by providers of external capital. Smaller firms are more likely to obtainfinance from private individuals. By contrast, even after controlling for selection effects,one of the most important factors in regards to obtaining the desired level of bankfinance is a firm’s assets. Small firms without significant assets have difficulty obtainingbank finance. Young, high growth innovative firms without significant assets or profitsseek capital from venture capital funds but venture capital is not widely available andrejection rates are quite high. High profits are not important for obtaining venturecapital but are important for obtaining the desired capital from leasing firms, factor



discounting/invoicing firms, trade customers/suppliers and partners/working share-holders.

We noted that, among the 38% of firms that made non-trivial efforts to obtainexternal finance in our sample, the mean percentage of finance obtained (relative tothe amount sought) was 84.5% and the median percentage was 100%. Hence, only fewfirms face a problem in obtaining all of their desired external capital. This evidence issomewhat counter to the conventional wisdom that entrepreneurial firms face acomparative dearth of capital. However, firms are not always able to obtain the type ofcapital that they want. This suggests capital gaps in terms of the quantity of capitalavailable are not overly pronounced in the UK entrepreneurial finance sector. Ques-tions pertaining to the quality of capital available, among other things, could prove tobe a fruitful avenue for future empirical studies.

University of CambridgeYork UniversityUniversity of Cambridge

Submitted: 16 October 2007Accepted: 3 August 2008

ReferencesAckerlof, G.A. (1970). �The market for �lemons�: Quality uncertainty and the market mechanism�, Quarterly

Journal of Economics, vol. 84, pp. 48–500.Bascha, A. and Walz, U. (2001). �Convertible securities and optimal exit decisions in venture capital finance�,

Journal of Corporate Finance, vol. 7, pp. 285–306.Berger, A. and Humphrey, D.B. (1997). �Efficiency of financial institutions: international survey and direc-

tions for future research�, European Journal of Operational Research, vol. 98, pp. 175–212.Berger, A. and Udell, G. (1998). �The economics of small business finance: the roles of private equity and

debt markets in the financial growth cycle�, Journal of Banking and Finance, vol. 22, pp. 613–73.Berger, A. and Udell, G.F. (2002). �Small business credit availability and relationship lending: the importance

of bank organizational structure�, Economic Journal, vol. 112, pp. F32–53.Berger, A., Klapper, L.F. and Udell, G.F. (2001). �The ability of banks to lend to informationally opaque small

businesses�, Journal of Banking and Finance, vol. 25, pp. 2127–67.Bergmann, D. and Hege, U. (1998). �Venture capital financing, moral hazard, and learning�, Journal of Banking

and Finance, vol. 22, pp. 703–35.Bierens, H.J. (2003). �Modeling fractions�, Research note posted online at: http://econ.la.psu.edu/�hbierens/

EasyRegTours/FRACTIONS.PDF <accessed 1 June 2004>Bullock, A. and Hughes, A. (1998). �Survey design, response bias and sample characteristics in the CBR SME

Survey�, in (A.D. Cosh and A.Hughes, eds) Enterprise Britain :Growth Innovation and Public Policy I the Smalland Medium Sized Enterprise Sector 1994–1997, pp. 125–32, Cambridge ESRC Centre for Business Research,University of Cambridge

Carpenter, R.E. and Petersen, B.C. (2002a). �Capital market imperfections, high-tech investment, and newequity investment�, Economic Journal, vol. 112 (2), pp. F54–72.

Carpenter, R. and Petersen, B.C. (2002b). �Is the growth of small firms constrained by internal finance?�,Review of Economics and Statistics, vol. 84, pp. 298–309.

Carpenter, R.E., Fazzari S.M. and Petersen B.C. (1998). �Financing constraints and inventory investment: acomparative study with high-frequency panel data�, Review of Economics and Statistics, vol. 80, pp. 513–9.

Casamatta, C. (2003). �Financing and advising: optimal financial contracts with venture capitalists�, Journal ofFinance, vol. 58, pp. 2059–86.

Casamatta, C. and Haritchabalet, C. (2007). �Learning and syndication in venture capital investments�, Journalof Financial Intermediation, vol. 16, pp. 368–98.

Cassar, G. (2004). �The financing of business start-ups�, Journal of Business Venturing, vol. 19, pp. 261–83.Chemmanur, T. and Chen, Z. (2002). �Angels, venture capitalists, and entrepreneurs: a dynamic model of

private equity financing�, Working Paper, Boston College.



Cosh, A. and Hughes A. (1994). �Size, financial structure and profitability: UK companies in the 1980s�, in(A. Hughes and D.J. Storey, eds.), Finance and the Small Firm, pp.18–63, London: Routledge.

Cressy, R. (1996). �Are business startups debt-rationed?�, Economic Journal, vol. 106, pp. 1253–70.Cressy, R. (2002). �Funding gaps: a symposium�, Economic Journal, vol. 112, pp. F1–16.Cumming, D. (2006). �The determinants of venture capital portfolio size: empirical evidence�, Journal of

Business, vol. 79, pp. 1083–126.Denis, D. (2004). �Entrepreneurial finance: an overview of the issues and evidence�, Journal of Corporate Finance,

vol. 10, pp. 301–26.De Meza, D. (2002). �Overlending?, Economic Journal, vol. 112, pp. F17–31.De Meza, D. and Webb, D.C. (1987). �Too much investment: a problem of asymmetric information�, Quarterly

Journal of Economics, vol. 102, pp. 281–92.De Meza, D. and Webb, D.C. (1992). �Efficient credit rationing�, European Economic Review, vol. 36, pp. 1277–

90.De Meza, D. and Webb, D.C. (1999). �Wealth, enterprise and credit policy�, Economic Journal, vol. 109,

pp. 153–63.De Meza, D. and Webb, D.C. (2000). �Does credit rationing imply insufficient lending?�, Journal of Public

Economics, vol. 78, pp. 215–34.Garmaise, M., (2000). �Informed investors and the financing of entrepreneurial projects�, Working Paper,

University of Chicago Graduate School of Business.Goldfarb, B.D., Hoberg, G., Kirsch, D. and Triantis, A.J. (2009). �Does angel participation matter? An Analysis

of early venture financing�, Robert H. Smith School Research Paper no. 06-072.Gompers, P.A. and Lerner, J. (1999). The Venture Capital Cycle, Cambridge: MIT Press.Gompers, P.A. and Lerner, J. (2001). �The venture capital revolution�. Journal of Economic Perspectives, vol. 15,

pp. 145–68.Gorton, G. and Winton, A. (2003). �Financial intermediation�, in (G. Constantinides, M. Harris and R. Stulz),

Handbooks of the Economics and Finance, vol. 1A., pp. 431–552, Boston: Elsevier.Harrison R.T. and Mason, C. (2000). �Venture capital market complementarities: the links between business

angels and venture capital funds in the United Kingdom�, Venture Capital: An International Journalof Entrepreneurial Finance, vol. 2, pp. 223–42.

Heckman, J. (1976). �The common structure of statistical models of truncation, sample selection, and limiteddependent variables and a simple estimator for such models�, Annals of Economic and Social Measurement,vol. 5, pp. 475–92.

Heckman, J. (1979). �Sample selection bias as a specification error�, Econometrica, vol. 47, pp. 153–61.Hege, U., Palomino F. and Schwienbacher, A. (2003). �Determinants of venture capital performance: Europe

and the United States�, Working Paper, HEC School of Management.Hsu, D. (2004). �What do entrepreneurs pay for venture capital affiliation?�, Journal of Finance, vol. 59,

pp. 1805–44.Jensen, M.C. and Meckling, W. (1976). �Theory of the firm: managerial behaviour, agency costs, and capital

structure�, Journal of Financial Economics, vol. 3, pp. 305–60.Jianakoplos, N.A. and Bernasek, A. (1998). �Are women more risk averse?�, Economic Enquiry, vol. 36, pp. 620–

30.Kanniainen, V. and Keuschnigg, C. (2003). �The optimal portfolio of start-up firms in venture capital finance�,

Journal of Corporate Finance, vol. 9, pp. 521–34.Kanniainen, V. and Keuschnigg, C. (2004). �Start-up investment with scarce venture capital support�, Journal of

Banking and Finance, vol. 28, pp. 1935–59.Keuschnigg, C. (2004). �Taxation of a venture capitalist with a portfolio of firms�, Oxford Economic Papers,

vol. 56, pp. 285–306.Keuschnigg, C. and Nielsen, S.B. (2004). �Start-ups, venture capitalists, and the capital gains tax�, Journal

of Public Economics, vol. 88, pp. 1011–42.Kon, Y. and Storey, D.J. (2003). �A theory of discouraged borrowers�, Small Business Economics, vol. 21, pp. 37–

49.Kortum, S. and Lerner, J. (2000). �Assessing the contribution of venture capital to innovation�, RAND Journal

of Economics, vol. 31, pp. 647–92.Mayer, C., Schoors, K. and Yafeh, Y. (2005). �Sources of funds and investment strategies of VC funds:

evidence from Germany, Israel, Japan and the UK�, Journal of Corporate Finance, vol. 1, pp. 586–608.Manigart, S., Lockett, A., Meuleman, M., Wright, M., Landstrom, H., Bruining, H., Desbrieres, P. and

Hommel, U. (2006). �Venture capitalists� decision to syndicate�, Entrepreneurship Theory and Practice, vol.30, pp. 131–53.

Marlow, S. and Patton, D. (2005). �All credit to men? Entrepreneurship, finance and gender�, EntrepreneurshipTheory & Practice, vol. 29, pp. 717–35.

Myers, S.C. (2000). �Outside equity�, Journal of Finance, vol. 55, pp. 1005–37.Myers, S.C. and Majluf, N. (1984). �Corporate financing and investment decisions when firms have infor-

mation that investors do not have�, Journal of Financial Economics, vol. 13, pp. 187–222.



Myers, S.C., Dill, D.A. and Bautista, A.J. (1976). �Valuation of financial lease contracts�, Journal of Finance,vol. 31, pp. 799–819.

Neus, W. and Walz, U. (2005). �Exit timing of venture capitalists in the course of an initial public offering�,Journal of Financial Intermediation, vol. 14, pp. 253–77.

Noe, T.H. and Rebello, M.J. (1996). �Asymmetric information, managerial opportunism, financing andpayout policies�, Journal of Finance, vol. 51, pp. 637–60.

Porter, M. (1998). Competitive Strategy: Techniques for Analyzing Industries and Competitors, New York: Free Press.Porter, R.H. (1995). �The role of information in U.S. offshore oil and gas lease auctions�, Econometrica, vol. 63,

pp. 1–27.Puhani, P.A. (2000). �The Heckman correction for sample selection and its critique�, Journal of Economic

Surveys, vol. 14, pp. 53–68.Ritter, J.R. and Welch, I. (2002). �A review of IPO activity, pricing and allocations�, Journal of Finance, vol. 57,

pp. 1795–828.Riyanto, Y.E. and Schwienbacher, A. (2006). �The strategic use of corporate venture financing for securing

demand�, Journal of Banking and Finance, vol. 30, pp. 2809–33.Robb, A. and Robinson, D. (2009). �The capital structure decisions of new firms�, Working paper, University of

California Santa Crus and Duke University.Schwienbacher, A. (2008). �Innovation and venture capital exits�, Economic Journal, vol. 118(533), pp.

1888–916.Smith, R.L. and Smith, J.K. (2000). Entrepreneurial Finance, New York: Wiley.Stiglitz, J. and Weiss, A. (1981). �Creditrationing in markets with imperfect information�, American Economic

Review, vol. 73, pp. 393–409.Storey, D. and Wynarczyk, P. (1996). �The survival and non-survival of small firms in the UK�, Review of

Industrial Organisation, vol. 11, pp. 211–29.Tamari, M. (1970). �The nature of trade credit�, Oxford Economic Papers, vol. 22, pp. 406–19.Ueda, M. (2004). �Bank versus venture capital: project evaluation, screening and expropriation�, Journal of

Finance, vol. 59, pp. 601–21.Wong, A. (2009). �Angel finance: the ��other�� venture capital�, in (D.J. Cumming, ed.), Comparison to Venure

Capital, (chapter 5), Chichester: Wiley (forthcoming).Wright, M. and Lockett, A. (2003). �The structure and management of alliances: syndication in the venture

capital industry�, Journal of Management Studies, vol. 40, pp. 2073–104.Yusopova, A. (2002). �Lease finance development in Russian transition economy�, Working paper, Novosibirsk

State University.



Documents

Outside Enterpreneurial Capital