Upload
independent
View
0
Download
0
Embed Size (px)
Citation preview
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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
1498 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1499O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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,
1500 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1501O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
1502 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1503O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
1504 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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
2009] 1505O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
.
1506 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1507O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
1508 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1509O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
1510 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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
2009] 1511O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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))
.
1512 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1513O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
1514 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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
2009] 1515O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
.
1516 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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
2009] 1517O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
1518 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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
2009] 1519O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
1520 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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
2009] 1521O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
Table 6
Regression Analyses of The Number of Sources of Capital Sought
Model 9(Ordered Logit):no. of DifferentTypes of Sources
Approached(Full Sample)
Model 10(Ordered Logit):no. of DifferentTypes of Sources
Approached(Post-1990
Subsample)
Model 11(Ordered Logit):no. of DifferentTypes of Sources
Approached(1980-9
Subsample)
Model 12(Ordered Logit):no. of DifferentTypes of Sources
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.
1522 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1523O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
1524 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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**
*
2009] 1525O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
1526 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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
.
2009] 1527O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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,
1528 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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
2009] 1529O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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
1530 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1531O 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
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
1532 [ O C T O B E RT H E E C O N O M I C J O U R N A L
� The Author(s). Journal compilation � Royal Economic Society 2009
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.
2009] 1533O 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
� The Author(s). Journal compilation � Royal Economic Society 2009