FUNDING GAPS?

ACCESS TO THE LOAN MARKET BY HIGH-TECH START-

UPS

Massimo G. Colombo, Politecnico di Milano

Luca Grilli, Politecnico di Milano

Abstract This paper aims to shed new light on the financing of new technology-based firms (NTBFs) and the existence of

credit constraints that may negatively affect their activity. For this purpose, we analyze the different sources of

start-up financing used by NTBFs and highlight firm-, industry, and location-specific characteristics that

influence the extent of recourse to bank loans to finance the creation of high-tech start-ups. In particular, we

consider a sample composed of 386 Italian NTBFs that operate in high-tech industries, both in manufacturing

and services. First, we provide clear evidence of the existence of a financing hierarchy. In fact, NTBFs resort to

external financing only when personal financial resources are exhausted. In addition, the level of financial

leverage turns out not to be a random variable; on the contrary, it increases with an increase of the predicted

amount of firms' total initial capital, while it decreases with variables such as the number of founders, that are

indicative of a greater amount of available personal wealth to finance firms' start-up. Furthermore, we show that

credit to NTBFs is often rationed. Nonetheless, our results also indicate that the debt capital supply curve is not

vertical. In fact, the amount of bank loans is sensitive to factors that shift the demand for capital curve and so

influence the amount of total initial capital of firms.

May 2003

PRELIMINARY DRAFT. PLEASE, DO NOT QUOTE WITHOUT AUTHORS’

PERMISSION

Correspondence: Politecnico di Milano Department of Economics, Management and Industrial Engineering P.za Leonardo da Vinci, 32 20133 Milan, ITALY massimo.colombo@polimi.it

1

1. Introduction Whether capital markets are efficient or not is a crucial issue in the economic debate. In

particular, over the last two decades several authors have argued that due to adverse selection,

moral hazards, and high transactions costs, firms face financing constraints which may

negatively affect investment decisions (see Hubbard 1998 for a survey).

For a series of reasons which will be examined in next section (see Carpenter and

Petersen 2002), new technology-based firms (NTBFs) are those most likely to suffer from

capital market imperfections; this argument especially applies to debt financing. The

existence of credit constraints on the creation and growth of high-tech start-ups is very

worrisome because of the key role such firms play in assuring dynamic efficiency and

employment growth in the economic system (see Audretsch 1995). In fact, even though for

NTBFs equity financing allegedly has advantages over debt (see again Carpenter and

Petersen, 2002), empirical work shows that there is a substantial wedge between the costs of

internal and external equity financing (see for instance Lee et al., 1996); this hinders

provision of seed and start-up equity capital to most new ventures, especially in less

developed financial markets (see Berger and Udell, 1998). Therefore, poor access to bank

loans may seriously damage the competitive position and growth prospects of NTBFs.

In spite of the importance of this issue, to our knowledge no large scale empirical

evidence has been provided as to the extent of use of bank loans to finance the creation of

NTBFs. The present study aims to contribute to fill the gap in the empirical literature. For this

purpose, we consider a sample composed of 386 young Italian firms, that operate in high-tech

industries, both in manufacturing and services, were created in 1980 or later, and were

independent at start-up time. In order to detect the existence of imperfections in credit

markets, we first analyze the different sources of financing used by entrepreneurs to start the

companies. We show that most firms did not obtain any bank loan at start-up time. The

personal savings founders were able to get access to, including finance from family members

and friends, were by far the most important source of start-up financing. A very small number

of firms obtained private equity financing. However for these latter firms, the mean amount of

private equity was six time larger than the mean bank loan obtained by firms that resorted to

debt financing, and private equity accounted for a substantial percentage of total capital. In

addition, we consider firm-, industry-, and location-specific factors that may favor or hinder

use of bank debt in comparison with personal finance. For this purpose, we estimate a tobit

model of the level of financial leverage, conditioned on the amount of total start-up capital.

2

We also estimate a tobit model of the extent of use of bank loans, and compare the results

with those relating to an analogous model of the amount of internal equity (that is, equity

capital personally provided by founders). The findings of the econometric analysis reject the

hypothesis of perfect loan markets, in favor of the existence of a “financing hierarchy”

(Fazzari et al., 1988). Nevertheless, in spite of the fact that Italian banks are likely to resort to

credit rationing, the quantity of debt obtained by NTBFs appears to be sensitive to factors that

influence their investment opportunities; in other words, the supply of bank debt to NTBFs is

not completely inelastic.

The paper proceeds as follows. In Section 2 we briefly review the theoretical literature

as to the functioning of the loan market for NTBFs, and examine the findings of previous

empirical studies. In section 3 the data set is presented. In Section 4 we illustrate descriptive

statistics as to the different sources of start-up financing used by sample firms. Section 5 is

devoted to the econometric analysis. Some summarizing remarks in Section 6 conclude the

paper.

2. Access to the loan market by high-tech start-ups

In a frictionless capital market, a firm can freely borrow and lend at the market interest rate,

with this latter being determined by aggregate demand for and supply of capital. Given

interest rate, firm’s equilibrium quantity of capital depends only on the location of its demand

for capital (D) curve; in turn, this is determined by the firm’s investment opportunities.

Internal and external sources of financing are perfect substitutes. As a consequence, factors

that drive the D curve to the right also lead to an increase of firm’s (expected) quantity of

bank debt.

For instance, firms that operate in industries with greater scale economies will have a

greater desired initial scale of operations (see Mata, 1996; Mata and Machado, 1996; Gorg et

al., 2000); hence they will also ask for and obtain greater loans. On the contrary, greater

uncertainty deters investments, as there is greater risk of incurring sunk costs (Pindyck, 1991;

Dixit and Pindyck 1994; see also Cabral 1995); so uncertainty of the business environment

will have an opposite effect on the equilibrium quantity of debt. For similar reasons, firm-

specific characteristics that influence the expected profitability of invested capital will also

affect the amount of loans of a new firm. In particular, the literature on entrepreneurship (see

Evans and Jovanovic, 1989; Cressy, 1996; Xu, 1998; Åstebro and Bernhardt, 1999) has

argued that physical capital and founders' entrepreneurial ability are complements. Since new

firms founded by individuals with greater entrepreneurial talent have greater optimal initial

3

scale of operations,1 their founders will ask banks for greater loans; as the quality of their

business ideas is recognized by banks, they will obtain the requested financing.

In addition, as was originally argued by Modigliani and Miller (1958), in a frictionless

capital market there is no relation between a firm's investment decisions and its financial

structure. Following such line of reasoning, the personal wealth of entrepreneurs plays no role

in their investment decisions and has no effect on the quantity of debt obtained at start-up

time by the firms they founded. Under such circumstances, the level of leverage of firms, that

is the ratio of bank debt to total capital, will follow a random process.

Since the seminal work by Jaffee and Russell (1976) and Stiglitz and Weiss (1981),

the argument that there may be imperfections in capital markets that render external finance

expensive and constraint firms' investment decisions, has been gaining ground in the

economic literature (see Fazzari et al., 1988, and the literature mentioned in Hubbard, 1998).

Following such view, high-tech start-ups are likely to suffer most from capital market

imperfections; in particular, they will have poor access to the market for bank loans (see

Carpenter and Petersen, 2002). First, business prospects are very uncertain for NTBFs; as

creditors do not share in firms' returns in good states of nature, the returns they obtain in such

states are unlikely to compensate for the high risk of failure. Second, information

asymmetries are likely to be very pronounced for NTBFs. From one side, high-tech

entrepreneurs generally are better informed than lenders as to the risks and returns of their

projects; in fact, it is quite difficult for banks to ascertain ex-ante the quality of high-tech

investments of firms that lack a track record. From the other, debt financing can lead to moral

hazard as it changes the incentive structure of entrepreneurs compared with a situation in

which they only resort to personal finance. Since for a bank it also is very difficult to monitor

ex-post the behavior of high-tech entrepreneurs, there is room for them to replace low risk-

low return projects with high risk-high return ones, to the detriment of lenders.

It has been argued in the literature that collateral may be used by banks both as a

signaling device to separate high-quality from low-quality borrowers and as an incentive

device to deter entrepreneurs' opportunism (see Berger and Udell, 1998 for a survey).2

1 Previous empirical studies have shown that the initial size of firms increases with the human capital of founders (see Mata, 1996; Åstebro and Bernhardt, 1999). In particular, Colombo and Grilli (2003a) distinguishes between the generic and specific components of founders' human capital (for greater details see section 5). They provide evidence that the latter component has a substantially larger positive effect on the start-up size of NTBFs than the former one. 2 This argument relies on the existence of a separating equilibrium; if there is only a pooling equilibrium, then recourse to collateral based lending does not allow to effectively deal with information asymmetries. For instance, Stiglitz and Weiss (1981) show that in spite of incentive aligning effects, use of collateral may also have adverse selection effects, thus lowering bank's profits. Cressy and Toivanen (1999) analyze micro-data on a

4

Nevertheless, most of high-tech investments is in intangible and/or firm-specific assets that

provide little collateral value. Unless outside collateral3 is available, high-tech start-ups will

find it difficult to resort to debt financing.

In fact, absent collaterizable assets, the supply of capital (S) curve will be horizontal

only up to an amount of capital coinciding with the quantity of available internal finance;

after this threshold, the S curve will be upward-sloping, due to constraints in the supply of

debt. The greater the imperfections in loan markets, the steeper the S curve. In a limit

situation, banks may be unwilling to provide debt financing beyond a low level of leverage;

credit rationing will cause the S curve to become vertical.

If the S curve is upward-sloping, there will be a financing hierarchy. Since the

marginal cost of debt financing exceeds the opportunity cost of internal finance, entrepreneurs

will ask for a bank loan only if personal wealth including the financial resources available

from family members and friends, is not sufficient to finance the new venture at the desired

scale. As a consequence, the equilibrium amount of debt financing of a NTBF and the value

of firm's financial leverage will decrease with greater personal wealth of firm's founders. In

addition, in the extreme situation in which credit is rationed, shifts of the D curve will have

negligible impact on the equilibrium quantity of bank loans.

Whether and to what extent capital markets are imperfect and fair access to the loan

market is inhibited to new firms – especially high-tech start-ups, are a matter of empirical

testing.

Previous studies concerned with entrepreneurship have provided evidence consistent

with the argument that new firms suffer from financial constraints. For instance, both cross-

sectoral (Meyer, 1990; Blanchflower and Oswald, 1998) and longitudinal (Evans and

Jovanovic, 1989; Evans and Leighton, 1989; Black et al., 1996) studies have shown that the

likelihood of being self-employed increases with individuals' net worth. Holtz-Eakin et al.

(1994a) have analyzed reception of an inheritance. Their results indicate that the likelihood of

establishing a new enterprise and the initial capital committed to the enterprise by the founder

significantly increase with the size of the inheritance and that such effect is more pronounced

for low net-worth individuals. In addition, if one focuses attention on entrepreneurs that

received an inheritance, the greater the amount inherited the greater the likelihood of survival

and the growth rate of the new venture (Holtz-Eakin et al., 1994b). Åstebro and Bernhardt

large sample of loans granted by a UK major bank. Their results reject the existence of a pooling equilibrium, indicating that banks in fact manage to discriminate borrowers according to the quality of their investment projects. On this issue see also Toivanen and Cressy (2000).

5

(1999) have shown that the predicted household income of US entrepreneurs positively

affects the amount of capital committed to a new venture. Lastly, the analysis of the evolution

over time of the size distribution of Portuguese firms performed by Cabral and Mata (2002)

indicates the presence of binding financial constraints that prevent new firms from attaining

their optimal initial size.

Nonetheless, the view that new firms face tight financial constraints is not

unanimously shared in the literature. In particular, it has been argued that the fact that

individuals’ assets are positively correlated with firm creation and post-entry performance

may be the effect of a spurious correlation: if assets and human capital are correlated, failure

to include into econometric models a proper specification for founders' human capital may

lead to the erroneous detection of a capital market imperfection (see Cressy 1996).4

A limited number of empirical studies have analyzed debt financing of newly created

firms. Levenson and Willard (2000) while documenting that the extent of credit rationing in

the US is fairly limited,5 show that constrained firms are smaller and younger than

unconstrained ones. Storey (1994a) analyzes use of bank loans for start-up financing by UK

firms through estimates of logit models. With a few exceptions, the independent variables

have no explanatory power; in particular, none of the personal characteristics of founders

capturing education and prior employment experience, is significantly related to recourse to

debt financing. On the contrary, use of personal savings has a negative effect, suggesting that

personal savings and bank loans are substitutes. Ǻstebro (2002) considers a sample of 893 US

firms created in 1987. The percentage of sample start-ups that obtained a loan is 34.4. He tries

to distinguish the decision of entrepreneurs to seek a bank loan at start-up time, and the

decision of a commercial bank to grant it. His findings suggest that a considerable number of

start-ups do not ask for a bank loan; however out of applicants, a large proportion does

receive a loan. Entrepreneurs with greater human capital are more likely to self-select out of

the loan market and to resort to loans from family and friends. However, the same human

capital variables that lead individuals not to seek a loan, induce banks to grant it if there is an

application.

3 Outside collateral consists of founders' personal assets that can be used to secure firm's bank loans. 4 Even in the absence of any correlation between individuals' net worth and human capital, the positive relation between net worth and the likelihood of self-employment may be explained by the lower risk aversion of richer individuals (see Cressy 2000). 5 According to their estimates only 2.14% of US firms did not get the funding for which they applied, while an additional 4.22% were discouraged from applying because of the expectation of denial. On “discouraged borrowers” see also Kon and Storey (2000).

6

In the present study we focus attention on NTBFs. While several empirical studies

have documented that high-technology small firms are financially constrained (see for

instance Oakey, 1984; Egeln et al., 1997; Westhead and Storey, 1997; Giudici and Paleari,

2000. For an opposite view, see Moore, 1994), quantitative evidence on access to bank loans

in the early years of firms’ existence is more limited. Carpenter and Petersen (2002) is an

exception. They analyze a panel of over 2,400 publicly traded US firms that operate in high-

technology manufacturing industries and are observed over the period 1981-98. Most of these

firms went public during the sample period and were small at the time of the IPO. They show

that most firms obtain little debt financing prior to the IPO; in addition, the amount of debt is

negligible when compared with the new equity raised through the IPO.

Furthermore, there likely are differences across NTBFs as to the extent of use of debt

financing. Consideration of the sources of such heterogeneity may provide further insights

into the importance of credit market imperfections. Quite surprisingly, as far as we know, the

empirical literature has so far failed to provide econometric evidence based on large data sets

relating to factors that influence the quantity of bank loans used to start a NTBF.

3. The data set The sample analyzed in the present work is composed of 386 Italian start-ups. Sample firms

were established in 1980 or later, were independent at start-up time (i.e. they were not

controlled by another business organization even though other organizations may have held

minority shareholdings) and operate in the following high-tech sectors, in manufacturing and

services: computers, electronic components, telecommunication equipment, optical, medical

and electronic instruments, biotechnology & pharmaceuticals, multimedia content, software,

Internet services (e-commerce, ISP, web-based services), and telecommunication services.

The sample was extracted from the RITA database, developed at Politecnico di

Milano. The RITA database was created in 1999 and enlarged and updated in 2001. The

present release contains detailed information on more than 400 NTBFs and more than 1,000

of their founders.

The development of the database went through a series of steps. Firstly, Italian target

firms that complied with the above mentioned criteria relating to age, independence at start-up

time, and sector of operations were identified. For the construction of the target “universe” a

number of sources were used. These included lists provided by national industry associations,

on-line and off-line firm directories, and lists of participants in industry trades and

expositions. Information provided by the national financial press, specialized magazines,

7

other sectoral studies and regional Chambers of Commerce was also considered.

Unfortunately, data provided by official national statistics do not allow to obtain a reliable

description of the universe of Italian NTBFs. Altogether, around 2,000 firms were selected for

inclusion in the target universe.

Second, a questionnaire was sent to the target firms either by fax or by e-mail. The aim

of the questionnaire was to collect both quantitative information relating to the activity,

structure and performance of firms and the characteristics of their founders, and qualitative

judgements of firms' founders on specific issues. The first section of the questionnaire

contains information on characteristics of the entrepreneurs such as age, education, and prior

working experience. The second section comprises further questions concerning the

characteristics of the firms at start-up time, including the amount of initial capital and the

financial structure of the firm.

Lastly, answers to the questionnaire were checked and the questionnaires were

completed if necessary by educated personnel through phone or face-to-face interviews with

firms' owner-managers. This final step was crucial in order to obtain missing data and ensure

the accurateness of answers.

Note that there is no presumption here to have a random sample. First, as was said

earlier, the universe of Italian NTBFs is unknown; therefore there is no way to check whether

our sample is representative or not. Second, the identification process described above is

likely to have led to the oversampling of growth-oriented firms, while micro-firms are

probably underrepresented. Third, the sample was drawn in 1999; so only firms having

survived up to the survey date were included. In so far as failure rates are related to a firm’s

initial financial structure,6 there may be a sample selection bias in our data set that we cannot

correct. This notwithstanding, the sample is sufficiently large and heterogeneous to provide

adequate coverage of Italian NTBFs. In addition, the information in our dataset especially as

regards the human capital of the founding team, is much more accurate than in previous

datasets of similar size. The sample consists of 23 firms in the multimedia content sector

(5.9%), 112 software producers (28.6%), 156 Internet and telecommunication service firms

(39.9%), 19 firms in the biotechnology and pharmaceutical industry (4.8% of the sample),

while the remaining 76 firms (20.7%) operate in the following manufacturing sectors:

6 For instance Ǻstebro and Bernhardt (2001) show that failure rates of US start-ups are higher for firms that did not obtain any bank loans at start-up time. Accordingly, the data we report in section 4 possibly overestimate the extent of use of bank debt by Italian NTBFs.

8

telecommunication equipment, electronic components, computers, and optical, medical and

electronic instruments.

4. Sources of financing of Italian high-tech start-ups The aim of this section is to provide an overview of the sources of start-up financing to which

Italian NTBFs resorted. In particular, we distinguish between three sources of finance: i)

personal capital, that includes personal savings of entrepreneurs and those of family members

and friends; ii) private equity provided by business angels, venture capital firms, other

financial intermediaries, and other firms (i.e. corporate venture capital); and iii) bank debt.

As is shown in Figure 1, with a 84% share personal capital is by far the main source of

start-up finance for Italian NTBFs.7 Bank loans and private equity account for similar shares

(about 8%). The percentage of NTBFs that at start-up time resorted to debt financing is fairly

limited (22%, see Table 2), even though it is much larger than that of firms that obtained

private equity capital (3.9%). As a consequence, at 47.000 € the mean value of bank loans of

debt financed firms is almost six time smaller than the mean amount of private equity of firms

that resorted to this source of financing; it also is less than half the mean amount of personal

capital. Note also that the correlation between the amount of personal finance and that of bank

loans though positive, is very small (the Pearson correlation coefficient is equal to 0.065 and

is not statistically different from null at conventional confidence levels); the correlation

coefficient between the quantities of bank debt and private equity again is statistically

insignificant (it equals -0.007). Since most sample firms obtained neither bank loans nor

private equity, such evidence suggests that external equity may be a substitute for debt

financing for new firms that are able to gain access to it.

In Tables 2 and 3 we distinguish firms according to their sector of operations and

initial size (measured by number of employees), respectively. The initial amount of total

capital is considerably larger in manufacturing than in services; the same holds true for the

percentage of firms that obtained debt financing at start-up time, and the mean amount of

bank loans. On the contrary, the mean level of financial leverage does not substantially vary

across sectors, with the partial exception of Internet and TLC services where leverage is

above the sample average. Quite unsurprisingly, the amount of total capital and the mean

amount of bank loans increase monotonically with firms’ initial size. By contrast, there is no

consistent pattern as to the share of debt financed firms out of the total number of firms and

7 Note that such value is considerably greater than the one reported by Berger and Udell (1998) for US small firms in general and "infant" firms in particular.

9

the level of leverage of debt financed firms; in particular, the largest size class (i.e. firms with

a number of employees greater than 20) exhibits the smallest share of firms that resorted to

debt financing (less than 12%). As several firms in this size class obtained seed equity capital,

this evidence again points to a possible substitution between external equity and debt

financing.

Lastly, in Table 4 firms are classified according to the number of founders. Team

founding results in commitment of greater total initial capital. Nevertheless, the likelihood of

resorting to bank debt, the mean amount of bank loans and the mean level of leverage

conditional on having access to debt financing, are smaller for firms founded by three or more

individuals than for the remaining firms. These findings are consistent with the view that

firms founded by several entrepreneurs are more likely to self-select out of the loan market, as

they generally have greater internal financial resources.

5. The econometric evidence In this section we focus attention on firm-, industry, and location-specific factors that may

influence the amount of bank loans obtained by high-technology firms at start-up time. In so

doing, we shed new light on the existence of credit constraints in the Italian financial market

and the extent to which they may bind the activity of NTBFs.

For this purpose, we follow two methodological approaches. From one side, we

consider the level of financial leverage of firms; in order to identify its determinants, we

estimate a double censored tobit model conditional on the amount of total initial capital. This

in turn depends on factors that shift the D and S curves. Note that the amount of start-up

capital and the level of leverage are simultaneously determined by firms. In order to correct

for endogeneity bias due to unobserved heterogeneity across firms, we resort to a two step

estimating procedure. It is important to remind that with perfect capital markets, debt and

equity are perfect substitutes; therefore, financial leverage at start-up time should be a random

variable. In particular, neither the predicted amount of total capital nor other firm-specific

characteristics reflecting the amount of available internal finance should be related to the level

of leverage. From the other side, we estimate a tobit model of the amount of bank loans, and

compare the results of the estimates with those of a similar model of the amount of equity

capital.8 Again if there are no financial constraints, the covariates that determine the

equilibrium amount of capital committed to a new venture should have the same effects on

8 We are aware that as far as debt and equity financing are (positively or negatively) correlated, such procedure is not efficient and we should resort to a bivariate tobit model. So the results presented in this paper are preliminary and must be considered with caution.

10

personal equity and bank debt. In order to control for private equity financing, in the

econometric exercise we do not consider firms that resorted to that particular source of start-

up financing; so the number of sample firms declines to 371. Inclusion of the remaining 15

firms does not substantially alter the findings of the estimates (results for the full sample are

available from the authors).

5.1. The explanatory variables

The explanatory variables are illustrated in Table 5. They can be subdivided in three different

groups.

The first group encompasses variables that reflect the characteristics of a firm’s team

of founders. They can be further subdivided into two sub-categories. The first one includes

variables that proxy the specific human capital of founders. Following Becker (1975), this

consists of the capabilities that individuals can directly apply to the entrepreneurial job in the

newly created firm; it is related to the industry-specific skills that founders learned in the

organization by which they were formerly employed and to the “leadership experience”

gained either through a managerial position in another firm or in prior self-employment

episodes. DManager is equal to 1 if prior to the establishment of the new firm, one or more

founders had a managerial position in a medium or large company.9 Specworkexp measures

the average value of founders’ years of professional experience in the same sector of activity

of the new firm. We also distinguish according to whether such industry specific experience

relates to the technical or to the commercial sphere; more precisely, Techworkexp refers to

years spent in research, development, engineering, and other technical positions, while

Comworkexp indicates professional experience in sale and marketing functions.

The second sub-category encompasses measures of the general knowledge acquired by

entrepreneurs through both formal education and professional experience, that is indicators of

founders’ generic human capital. The level of education of founders is measured by the mean

number of years of education (Education). As concerns graduate and post-graduate education,

we are also able to distinguish between economic and law studies (Ecoeducation) and

technical and scientific studies (Techeducation).10 As to professional experience, Genworkexp

9 In small family-owned Italian companies owner-managers generally keep control of strategic decisions, while salaried managers are assigned execution tasks (see Colombo and Delmastro 1999). So we assume entrepreneurial learning associated with such managerial positions to be fairly limited. 10 Ecoeducation measures years spent for the attainment of degrees in economics, law, management, and political sciences, while Techeducation reflects years spent for obtaining degrees in engineering, physics, biology, chemistry, medicine, pharmaceutics, and computer science. In order to properly judge the effective level of competencies of founders, we consider the minimum length of time necessary to attain a certain degree. In order to attain an Italian graduate degree in economics, law, management, political sciences and most scientific

11

refers to years of working experience in other sectors than the one of the new firm. In this

sub-category we also consider the logarithm of the number of founders (Lfounders).

As is documented by the empirical literature on the post-entry performances of start-

ups (for a survey see Storey, 1994b. See also Colombo and Grilli, 2003b), variables in the

latter sub-category are no predictors of firms’ success; hence we expect them to have no

relation with firms’ investment opportunities. However, previous studies also report that such

variables are positively related to the amount of internally available financial resources (see

for instance Ǻstebro and Bernhardt 1999). Therefore, if there is a financing hierarchy due to

capital market imperfections, they should negatively affect firms’ level of leverage, after

conditioning on the quantity of total capital initially committed to a new venture. While they

should exhibit a positive effect on total capital due to relaxation of financial constraints, the

effect on the amount of personal capital used to start the new firm should be stronger than on

bank debt; in fact this latter effect should be negative as wealthy individuals can replace high

cost bank loans with low cost personal capital.

On the contrary, variables capturing the specific component of founders’ human

capital drive to the right the D curve and if there are frictions in capital markets, also the S

curve. So they should have a stronger positive effect on initial capital than variables included

in the previous sub-category. In addition, while they should positively affect the quantity of

personal capital committed to a new firm, the effect on bank loans is uncertain as opposed

forces are at work; on the one hand greater demand for capital may lead to greater use of bank

debt, on the other greater available internal finance may replace bank debt. However, after

controlling for initial total capital, we expect specific human capital variables to negatively

influence the level of leverage.

Let us now consider the second group of variables. It includes firm-, industry-, and

location-specific factors that shift the D curve, but allegedly have no bearing on the S curve.

Mes is a proxy of the minimum efficient scale of the industry in which a new firm operates; it

is computed as the log of average employment of firms (see Gorg et al., 2000).11 As demand

for capital is likely to be positively correlated with minimum efficient scale, it is legitimate to

expect a positive impact of this variable on total initial capital. On the contrary, greater

degrees four years of studies are requested, while five years is the minimum time for a degree in engineering. Master and Ph.D. programmes require one and three additional years respectively, independently of the specific field. 11 Data source are the 1981, 1991, 1996 ISTAT Census. Due to lack of data for the Internet sector, the minimum efficient scale in this sector has been assumed to be the same as in the software sector. Alternative measures of

12

business uncertainty will lead founders to limit initial commitment of resources so as to avoid

possible sunk costs (Pindyck, 1991; Dixit and Pindyck, 1994; Cabral 1995). In order to create

a proxy for industry uncertainty we had recourse to the database on European initial public

offerings (IPO) that was jointly developed by Politecnico di Milano and Tilburg University.

Such database includes data on 482 IPOs that occurred between 1996 and 2001 in five

European new stock markets (Neuer Markt, Nuovo Mercato, Nouveau Marchè, Euro NM,

Nmax).12 Uncertainty measures the industry average of the normalized standard deviation of

the market price of newly listed firms in the 50 days following the IPO. Great variability of

post-IPO stock prices in an industry signals great ex-ante uncertainty on new firms’

performance. Hence we expect a negative impact of this variable on initial capital.13

DIncubated is a dummy variable indicating firms that are located in a technology incubator.

Incubated firms are forced to operate at a relatively small scale, given the limited physical

space at disposal in Italian incubators; 14 with everything else being equal, they will require

less financial resources. DMother company denotes firms that at start-up time benefited from

the support of a sponsoring firm, which provided the new venture with access to such

resources as complementary technological know-how, physical equipment, skilled personnel,

or distribution channels. Such favorable situation is likely to result in an increase of start-up

capital. Lastly, Infrastructure reflects the infrastructure endowment in the region in which the

new firm is located. Location in an area with efficient infrastructure may make founders more

confident on the future prospects of the firm and convince them to demand more capital in

order to start operations at greater size. While the above mentioned variables are expected to

have a significant effect on the total amount of both initial debt and personal capital (and

hence on the quantity of total start-up capital), once we condition on initial capital they should

not have any further effect on financial leverage; so they are not introduced as independent

covariates in this latter equation.

In the last group we consider other variables that may be responsible for shifts of the S

curve in addition to the effects attributable to founders’ characteristics. Rreal is the real

interest rate in the year of firms’ foundation (source: Banca d’Italia 2001). Defaults is the

Mes such as the one proposed by Caves et al. (1975) and Lyons (1980) could not be used because of lack of data. 12 Data on IPOs have been collected primarily through IPOs brochures and companies web sites, while data on market prices have been obtained from Datastream database and the web sites of the above cited new markets. For further details see Giudici and Roosenboom (2002). 13 Actually, it may be argued that uncertainty in the business environment amplifies information asymmetries between banks and high-tech entrepreneurs, thus making the S curve steeper. In order to control for such effect, we have introduced Uncertainty also in the leverage equation.

13

ratio between the number of firms’ defaults and the number of existing business activities in

the geographic area where NTBFs are located (source: Istituto Tagliacarne, “Databank

Starter”); it aims to control for heterogeneity in local capital market conditions, Greater cost

of capital negatively affects the equilibrium quantity of investments; therefore, the predicted

impact of such variables on total initial capital, amount of bank loans and personal capital is

negative.15

In Table 6 we illustrate descriptive statistics and the correlation matrix of independent

variables. Correlation across variables are generally low, suggesting absence of any relevant

problem of multicollinearity.

5.2. Results of the estimates

The results of the econometric estimates are illustrated in Tables 7 and 8. In the former table

we report the OLS estimates of the equation relating to firms’ amount of total initial capital

and the double censored tobit estimates of the level of leverage. In the tobit model we

consider among the covariates the predicted quantity of total initial capital; we report

marginal effects evaluated at the sample mean. In the latter table we present the estimates of

the tobit models of the amount of bank loans and personal capital used for start-up financing;

again we report the marginal effects of the covariates.

Let us consider first the amount of total initial capital (see Table 7). Our findings

indicate that with the only exception of Uncertainty, which has the predicted negative

coefficient but is insignificant in both models I and II, factors that allegedly drive to the right

(left) the D curve have a substantial, statistically significant, positive (negative) effect on the

quantity of start-up capital of Italian NTBFs. In fact, initial capital is greater in high-tech

industries characterized by greater minimum efficient scale, in regional areas equipped with

developed infrastructure (actually, the coefficient of this latter variable in model II is only

close to statistical significance), and for firms that at start-up time benefited from support

provided by a “mother” company. On the contrary, firms located in a technology incubator

start operations with less capital: the coefficient of DIncubator always is negative as

expected, even though it is only significant at 90%.

As to factors that shift upward the S curve, the real interest rate turns out to negatively

affect the amount of capital initially invested by NTBFs; its coefficient is negative and

14 See Colombo and Delmastro (2002) for a description of technology incubators in Italy. 15 For reasons similar to the ones mentioned in footnote 12, we have introduced Defaults also in the leverage equation. In addition, if there are imperfections in capital markets, the wedge between the costs of internal finance and bank debt may be sensitive to the level of real interest rates. So we also consider Rreal among the covariates in the leverage equation.

14

significant at conventional confidence levels in both models. On the contrary, failure rates in

the regional areas where NTBFs are located prove to have negligible effects.

Lastly, let us consider the characteristics of the team of founders. As to working

experience, the quantity of total initial capital increases with founders’ professional

experience in the same sector of the new firm, while working experience in other sectors

plays a negligible role. In particular, what really matters is industry-specific technical

experience. In fact, while Genworkexp in model I and Comworkexp in model II exhibit

positive but insignificant coefficients, the coefficients of Specworkexp in model I and

Tecworkexp in model II are positive and statistically significant at 99%. As to the remaining

human capital variables, founders’ managerial experience always is insignificant. The same

applies to years of education. Nonetheless, if we focus on university education, firms started

by individuals with technical or scientific degrees start operations with less capital, while the

opposite holds true for firms established by individuals with degrees in economics,

management, law or political sciences.

Let us now turn attention to the leverage equation. The results clearly support the view

that personal capital and bank loans are no perfect substitutes. First of all, the null hypothesis

that all covariates have no explanatory power is rejected at 99% in both models I and II by a

LR test (χ2=66.9 and 66.1, d.f=9 and 11, respectively); in other words, in contrast with the

contention that there are no imperfections in capital markets, financial leverage is no random

variable. On the contrary, its level increases with the predicted value of the amount of initial

capital of firms, with the coefficient of such variable being significant at 95% in both models.

In addition, the number of founders always exhibits a negative coefficient, significant at 99%.

Genworkexp also has a negative coefficient, even though it is significant (at 90%) only in

model I. The other human capital variables have negligible effects on the level of leverage,

even though their coefficients generally are negative. Altogether these findings suggest that

firms that have decided to start operations with an amount of invested capital that exceeds

available personal finance, are forced to resort to bank debt, while other firms are much less

likely to do so. In fact, the larger the predicted amount of initial capital, the larger the amount

of bank loans in comparison with that of personal finance. As it may reasonably be assumed

that the personal wealth founders may have access to increases with their number and

possibly with their human capital, this may explain the negative effects of such variables on

leverage. Lastly, the level of real interest rate at start-up time and the relative number of

defaults in the regional area where NTBFs are located turn out to have no impact on leverage.

15

The evidence illustrated above is confirmed by the results of the estimates of the

personal capital and bank loans tobit models reported in Table 8. As to personal capital

equations, the coefficients of all covariates have the same sign (and with few exceptions

similar significance levels) as those in the corresponding total capital equations illustrated in

Table 7. In other words, factors that explain the amount of total initial capital of Italian

NTBFs, also explain the amount of personal finance committed to such firms. On the

contrary, in the bank debt equations only Mes and DIncubated have coefficients of the same

sign (positive and negative, respectively) as in the corresponding total capital equations,

significant at conventional confidence levels. Note in particular that DIncubated is not

significant in the personal capital equation; while the amount of personal capital invested in a

high-tech start-up proves not to depend on whether the firm is located in a technology

incubator or not, with such location the need for debt financing is considerably reduced. This

probably is a consequence of the small available physical space. Lastly let us consider in

greater detail the effects of founders’ characteristics. Lfounders, which has a positive,

statistically significant effect on the amount of personal capital (as was the case for total

initial capital), exhibits a negative coefficient, statistically significant (at 90%) in both models

I and II. As to the remaining variables, those that turned out to have a significant positive

effect on total initial capital and personal finance (i.e. Specworkexp and Techworkexp), have a

positive though insignificant coefficient; those that had no significant impact on total initial

capital and personal finance (in particular, Education, Genworkexp and Comworkexp) have a

negative though insignificant coefficient. Such evidence again points to the existence of a

substitution relation between personal capital and bank loans.

6. Concluding remarks The aim of this paper was to shed new light on the financing of NTBFs and the existence of

credit constraints that may negatively affect their activity. In fact, NTBFs are likely to be

those that suffer most from imperfections in the loan market; this especially applies to

countries like Italy, where financial institutions (and commercial banks in particular)

generally did not develop specialized competencies and procedures to deal with the financial

needs of this category of firms. We focus on start-up financing. As most high-tech start-ups

do not resort to external equity financing, the presence of credit constraints is very worrisome,

due to the key role allegedly played by NTBFs in assuring innovation and growth in the

economic system. In spite of the existence of a rich stream of theoretical literature on

imperfections in financial markets and of a growing number of empirical studies documenting

16

their impact on firms' investment decisions, there is a lack of evidence based on large datasets

relating specifically to NTBFs.

In this work, we are interested in the different sources of start-up financing used by

NTBFs. In particular, we aim to highlight firm-, industry, and location-specific characteristics

that influence the extent of recourse to bank loans to finance the creation of high-tech start-

ups. For this purpose, we analyze a sample composed of 386 Italian firms that operate in high-

tech industries, both in manufacturing and services, are less than 22 years old, and were

independent at start-up time. First, we consider the relative importance of different sources of

start-up capital, distinguishing between i) personal finance, that is capital personally provided

by entrepreneurs, their family, and friends, ii) external equity financing (i.e. equity capital

from business angels, venture capitalists, other financial intermediaries and other firms), and

iii) bank debt. Then, we estimate a series of econometric models to assess factors that have a

bearing on the extent of use of bank loans at start-up time and the level of financial leverage

of firms.

The main findings of the empirical analysis can be summarized as follows. First, we

provide clear evidence of the existence of a financing hierarchy. In fact, our results are

consistent with the view that NTBFs resort to external financing only when personal financial

resources are exhausted. In particular, personal capital and bank loans are no perfect

substitutes. In fact, a relatively small number of firms (22%) obtained access to debt financing

at start-up time, and an even smaller number (less than 4%) resorted to private equity. In

addition, the level of financial leverage turns out not to be a random variable. On the contrary,

it increases with an increase of the predicted amount of firms' total initial capital, while it

decreases with variables such as the number of founders, that are indicative of a greater

amount of available personal wealth to finance firms' start-up.

Furthermore, the findings illustrated in the paper suggest that credit to NTBFs is often

rationed. Firms that managed to get access to bank debt obtained on average an amount of

bank loans (47,000 €) that was less than half the mean amount of personal capital and six

times smaller than the mean amount of private equity obtained by firms that did use that

source of financing.

Nonetheless, our results also indicate that the debt capital supply curve is not vertical.

In fact, the amount of bank loans is sensitive to factors that shift the demand for capital curve

and so influence the amount of total initial capital of firms. For instance, the amount of bank

debt is substantially greater for firms that operate in industries with considerable scale

17

economies, while it is smaller for firms that being located in a technology incubator, face

physical constraints which hinder achievement of greater scale of operations.

We think that this paper offers an interesting contribution to the empirical literature

concerned with financing of NTBFs. Nonetheless, we are aware that this only is a preliminary

step. In particular, there are two avenues for future research that seem to us especially

promising. On the one hand, we did not consider here firms' investment decisions. Therefore,

we are unable to determine whether and to what extent credit constraints actually bind

NTBFs' investments. If they do, one should expect more financially constrained firms to

exhibit inferior post-entry performances, after controlling for firms' investment opportunities.

In order to gain further insights into such issue, it would be helpful to investigate the relation

between use of different sources of financing at start-up time and in the early years after firms'

foundation, and firms' growth and survival rates. On the other hand, it is often claimed in the

capital market imperfection literature that it is difficult to discriminate ex-ante between

"constrained" and "unconstrained" firms (see for instance Hubbard 1998, p. 200). While all

NTBFs are potentially constrained in that they face high information costs, the implications of

such constraints for the financing behavior of firms are likely to differ according to the extent

of firms' investment opportunities and the amount of personal finance founders may be able to

collect. In our dataset we only have indirect proxies of the personal wealth on which

entrepreneurs can tap to finance the creation of a firm. A more direct assessment of whether

personal finance would be sufficient to face the financial needs of a new firm would allow a

more precise analysis of the effects of credit constraints on the financing behavior of firms.

18

References (incomplete)

Ǻstebro, T., Bernhardt, I., 1999, “The Winner’s Curse of Human Capital,” Center for Economic Studies, Working Paper CES 99-5, U.S. Department of Commerce.

Audretsch, D. B., 1995, Innovation and industry evolution, MIT Press, Cambridge, Mass.

Banca d’Italia, 2001, “Relazione generale sulla situazione economica del Paese e bollettino economico” (in Italian).

Becker, G. S., Human Capital, 1975, New York: National Bureau of Economic Research.

Black, J., de Meza, D., Jeffreys, D., 1996, “House pricing, the supply of collateral and the enterprise economy”, Economic Journal, vol. 106, pp. 60-75.

Blanchflower, D., Oswald, A., 1998, “What Makes an Entrepreneur”, Journal of Labor Economics, vol. 16, pp. 26-60.

Cabral L., 1995, “Sunk costs, firm size and firm growth”, Journal of Industrial Economics, vol. 43, pp. 161-172.

Cabral, L., Mata, J., 2002, “On the evolution of firm size distribution: facts and theory”, forthcoming in American Economic Review.

Carpenter, R. E, Petersen, B. C., 2002, “Capital market imperfections, high-tech investment, and new equity financing”, Economic Journal, vol. 112, pp. F54-F72.

Caves, R. E., 1998, “Industrial organization and new findings on the turn-over and mobility of firms”, Journal of Economic Literature, vol. 36, pp. 1947-1982.

Caves, R. E., Khalilzadeh-Shirazi, J., Porter, M.E., 1975, “Scale economies in statistical analyses of market power”, Review of Economics and Statistics, vol. 57, pp. 133-140.

Colombo, M. G., Delmastro M., 1999, “Some stylized facts on organization and its evolution, Journal of Economic Behaviour & Organization, vol. 40, pp. 252-274.

Colombo, M. G., Delmastro, M., 2002, “How effective are technology incubators? Evidence from Italy", Research Policy, vol. 31, pp. 1103-1122.

Cressy, R., 1996, “Are business start-ups debt-rationed?”, Economic Journal, vol. 106, pp. 1253-1270.

Cressy, R., 2000, “Credit rationing or entrepreneurial risk aversion? An alternative explanation for the Evans and Jovanovic finding”, Economics Letters, vol. 66, pp. 235-240.

Dixit, A. K., Pindyck, R. S., 1994, Investment under uncertainty, Princeton University Press, Princeton, New Jersey.

Evans, D., Jovanovic, B., 1989, “An estimated model of entrepreneurial choice under liquidity constraints”, Journal of Political Economy, vol. 97, pp. 808- 827.

Evans, D., Leighton, L., 1989, “Some empirical aspects of entrepreneurship”, American Economic Review, vol. 79, pp. 519- 535.

19

Giudici, G., Paleari, S., 2000, “The provision of finance to innovation: a survey conducted among Italian technology-based small firms”, Small Business Economics, vol. 14, pp. 37-53.

Giudici, G., Roosenboom, P., 2002, “Pricing initial public offerings on European ‘new’ stock markets”, XIII Tor Vergata Financial Conference, Rome Italy.

Gorg, H., Strobl, E., Ruane, F., 2000, “Determinants of firm start-up size: an application of quantile regression for Ireland”, Small Business Economics, vol. 14, pp. 211-222.

Holtz-Eakin, D., Joulfaian, D., Rosen, H. S., 1994a, “Entrepreneurial decisions and liquidity constraints”, Rand Journal of Economics, vol. 25, pp. 334-347.

Holtz-Eakin, D., Joulfaian, D., Rosen, H. S., 1994b, “Sticking it out: entrepreneurial survival and liquidity constraints”, Journal of Political Economy, vol. 102, pp. 53-75.

Levenson, A.R., Willard, K.L., 2000, “Do firms get the financing they want ? Measuring credit rationing experienced by small businesses in the U.S.”, Small Business Economics, vol. 14, pp. 83-94.

Lyons, B., 1980, “A new measure of minimum efficient plant size”, Economica, vol. 47, pp. 19-34.

Mata, J., 1996, “Market, entrepreneurs and the size of new firms”, Economic Letters, vol. 52, pp. 89-94.

Mata, J., Machado, J. A. F., 1996, “Firm start-up size: A conditional quantile approach”, European Economic Review, vol. 40, pp. 1305-1323.

Meyer, B.D., 1990, “Why are there so few black entrepreneurs?”, NBER Working paper n. 3537, Cambridge, Mass.

Pindyck, R. S., 1991, “Irreversibility, uncertainty, and investment”, Journal of Economic Literature, vol. 29, pp. 1110-1148.

Storey, D. J., 1994b, Understanding the small business sector, Thomson Learning, London.

Westhead, P., Storey, D. J., 1997, “Financial constraints on the growth of high technology small firms in the United Kingdom”, Applied Financial Economics, vol. 7, pp. 197-201. Xu, B., 1998, “A reestimation of the Evans-Jovanovic entrepreneurial choice model”,

Economics Letters, vol. 58, pp. 91-95.

20

Figure 1. Financing of Italian NTBFs at start-up time.

Composition of start-up capital of Italian NTBFs (thousands of Euro)

4,117 (8%)3,980 (8%)

40,451 (84%)

personal capitalprivate equitybank debt

Source: RITA database Table 1. Sources of start-up financing of Italian NTBFs

Financing sources Firms Amount of capital (‘000 €)

N. % Meana S.D.a

Personal capital 378 97.9 107.0 265.3

Private equity 15 3.9 274.5 660.4

Bank debt 85 22.0 46.8 92.0

Totalb 386 100 125.8 303.1 Legend a For every source of financing, mean and standard deviation values refer only to firms that obtained that particular type of financing. b The sum of the number firms that resorted to the different sources of financing is greater than the total number of firms as some firms used more than a single financing source. Mean and standard deviation values refer to total amount of capital (i.e. the sum of personal capital, private equity and bank debt).

Source: RITA database

1

Table 2. Sources of start-up financing of Italian NTBFs, by industry

Sector Firms Firms obtaining debt financing Amount of total initial capital

(‘000 €) Firms obtaining debt

financing Amount of bank loans

(‘000 €) Leverage (%)

N. Mean S.D. N. % Mean S.D. Mean S.D

Internet and TLC services 156 85.6 235.0 28 17.9 27.1 19.4 54.9 24.5

Multimedia content 23

54.4 35.6 7 30.4 17.8 14.7 44.7 29.1

Software houses 112 66.8 93.9 20 17.9 37.5 83.7 47.5 30.4

ICT manufacturing 76 309.7 536.9 23 30.3 89.0 151.3 47.9 26.5

Biotechnology/Pharmaceutics 19 153.7 215.3 7 36.8 42.8 20.9 44.4 22.2

Total 386 125.8 303.1 85 22.0 46.8 92.0 49.6 26.4Source: RITA database Table 3. Sources of start-up financing of Italian NTBFs, by firm size

Size class (n. of employees) Firms Firms obtaining debt financing Amount of total initial capital

(‘000 €) Firms obtaining debt

financing Amount of bank loans (‘000 €) Leverage (%)

N. Mean S.D. N. % Mean S.D. Mean S.D

1-2 107 30.9 55.1 22 20.6 11.0 5.6 52.4 29.0

3-5 162

84.9 152.6 38 23.5 30.8 28.4 51.8 27.7

6-10 73 154.4 294.1 16 21.9 86.4 166.5 42.0 18.2

11-20 27 329.4 634.3 7 25.9 106.3 99.1 44.7 17.8

>20 17 665.9 672.5 2 11.8 221.4 238.2 54.5 64.3

Total 386 125.8 303.1 85 22.0 46.8 92.0 49.6 26.4Source: RITA database

2

Table4. Sources of start-up financing of Italian NTBFs, by size of the founding team

N. of founders Firms Firms obtaining debt financing Amount of total initial capital

(‘000 €) Firms obtaining debt

financing Amount of bank loans (‘000 €) Leverage (%)

N. Mean S.D. N. % Mean S.D. Mean S.D

1-2 202 116.1 252.8 54 26.7 49.1 103.7 50.2 25.2

≥ 3 184

136.4 350.5 31 16.8 42.8 68.3 48.5 28.8

Total 386 125.8 303.1 85 22.0 46.8 92.0 49.6 26.4Source: RITA database

3

Table 5 - The explanatory variables of the econometric models Variable Description

DManager One for firms with one ore more founders with a prior management position in a large or medium company (i.e. number of employees greater than 100)

Specworkexp Average number of years of working experience gained by founders in the same sector of the start-up before firm’s foundation

Techworkexp Average number of years of technical working experience gained by founders in the same sector of the start-up before firm’s foundation

Comworkexp Average number of years of commercial working experience gained by founders in the same sector of the start-up before firm’s foundation

Education Average number of years of founders’ education

Ecoeducation Average number of years of founders’ economic, law and/or managerial education at graduate and post-graduate level

Techeducation Average number of years of founders’ scientific and/or technical education at graduate and post-graduate level

Genworkexp Average number of years of working experience gained by founders in other sectors than the one of the start-up before firm’s foundation

Lfounders Logarithm of the number of founders

Mes Minimum efficient scale in the sector of the start-up in the year in which the firm was created (or in the nearest year for which data were available) measured by the log of the average number of employeesa

Uncertainty Industry average of the normalised standard deviation of the market price of newly listed firms in the 50 days following the IPO

DIncubated One for firms located in a technology incubator

DMother company One for firms that at start-uptime, received some kind of aid by a “mother” company

Infrastructure Value of the index measuring regional infrastructures in 1992 (mean value among Italian regions=100)

Rreal Real interest rate in the year of firm’s foundation Defaults Ratio between the number of firms’ defaults and the number of existing business

activities in the geographic area where the start-up is located

Legend a Data are available for 1981, 1991, 1996.

4

Table 6. Descriptive statistics and correlation matrix of the explanatory variablesa

Mean S.D. DManager Specwork. Techwork. Comwork. Edu. Ecoedu. Techedu. Genwork. Lfounders Mes Uncertainty Dincubated Dmother company

DManager 0.094 0.293 1.00

Specworkexp 4.444 6.963 0.12 1.00

Techworkexp 2.884 5.926 0.09 0.80 1.00

Comworkexp 1.273 3.840 0.06 0.50 -0.06 1.00

Education 14.782 2.565 0.15 -0.15 -0.13 -0.06 1.00

Ecoeducation 0.341 0.898 0.02 -0.15 -0.12 -0.09 0.28 1.00

Techeducation 1.456 1.920 0.15 0.01 0.03 0.00 0.75 -0.13 1.00

Genworkexp 7.903 8.472 0.14 -0.43 -0.34 -0.21 -0.06 0.08 -0.09 1.00

Lfounders 0.887 0.533 0.09 -0.13 -0.14 -0.04 0.08 0.00 0.05 -0.09 1.00

Mes 0.972 0.371 0.09 0.19 0.15 0.08 0.04 -0.09 0.10 0.00 -0.11 1.00

Uncertainty 0.035 0.003 -0.09 -0.19 -0.18 -0.06 -0.07 0.12 -0.19 0.06 0.07 -0.45 1.00

Dincubated 0.124 0.330 -0.01 -0.04 0.00 -0.06 0.15 -0.02 0.14 0.09 0.09 0.13 -0.10 1.00

Dmother company 0.108 0.311 0.01 0.13 0.13 0.03 0.13 -0.02 0.14 0.01 -0.05 0.06 -0.08 0.03 1.00

Infrastructure 116.096 27.604 0.09 0.00 0.02 -0.02 0.03 0.13 -0.02 0.09 -0.10 0.11 -0.06 0.11 0.02

Legend: aFor the sake of sinthesys we omit control variables relating to the S curve.

5

Table 7 – Determinants of total start-up capital and financial leverage Model I Model II

OLS Tobit (Marginal effects) OLS Tobit (Marginal effects)

Variables Start-up capital (log) Financial leverage Start-up capital (log) Financial leverage

Costant 3.849 (0.737)*** -38.493 (25.503) 3.995 (0.679)*** -43.961 (25.062)*

DManager 0.069 (0.158)

0.297 (4.195) 0.105 (0.154) -0.890 (4.212)

Specworkexp 0.022 (0.008)*** -0.126 (0.201) - - - -

Techworkexp - - - - 0.029 (0.008)*** -0.063 (0.219)

Comworkexp - - - - 0.007 (0.012) -0.005 (0.301)

Education -0.018 (0.018) -0.321 (0.446) - - - -

Ecoeducation - - - - 0.124 (0.049)** -0.251 (1.326)

Techeducation - - - - -0.057 (0.024)** 0.464 (0.614)

Genworkexp 0.004 (0.006) -0.281 (0.160)* 0.002 (0.006) -0.231 (0.156)

Lfounders 0.417 (0.085)*** -7.225 (2.389)*** 0.425 (0.084)*** -7.156 (2.401)***

Mes 1.049 (0.132)*** - - 1.079 (0.130)*** - -

Uncertainty -17.766 (15.008) 375.289 (415.863) -24.339 (14.971) 415.332 (430.942)

Dincubated -0.259 (0.137)* - - -0.244 (0.135)* - -

Dmother company 0.563 (0.143)*** - - 0.584 (0.140)*** - -

Infrastructure 0.004 (0.002)** - - 0.003 (0.002) - -Rreal -0.042 (0.023)* 0.115 (0.587) -0.046 (0.023)** 0.082 (0.590)

Defaults -0.003 (0.547) -11.480 (14.279) -0.121 (0.543) -11.028 (14.469)

Start-up capital (predicted value) - - 6.305 (2.755)** 5.881 (2.709)**

N° of observations 371 371 371 371

R2 0.31 0.33

Log-likelihood -563.80 -564.15

Legend. Standard errors in parentheses. * stands for a significance level greater than 90%, ** stands for a significance level greater than 95% and *** stands for a significance level greater than 99%.

6

Table 8 – Determinants of the amount of banl loans and equity capital at start-up time (tobit models, marginal effects) Model I Model II Variables

Personal capital (log) Bank loans (log) Personal capital (log) Bank loans (log)

Costant 3.857 (1.026)*** -1.075 (1.394) 4.334 (0.946)*** -1.306 (1.317)

DManager 0.027 (0.220)

-0.012 (0.304) 0.108 (0.215) -0.078 (0.302)

Specworkexp 0.026 (0.011)** 0.009 (0.013) - - - -

Techworkexp - - - - 0.031 (0.011)*** 0.019 (0.014)

Comworkexp - - - - 0.006 (0.016) -0.002 (0.022)

Education -0.007 (0.025) -0.019 (0.033) - - - -

Ecoeducation - - - - 0.157 (0.068)** 0.074 (0.092)

Techeducation - - - - -0.075 (0.033)** 0.022 (0.044)

Genworkexp 0.016 (0.008)* -0.012 (0.012) 0.011 (0.008) -0.009 (0.011)

Lfounders 0.492 (0.118)*** -0.307 (0.163)* 0.499 (0.117)*** -0.297 (0.163)*

Mes 0.886 (0.185)*** 0.662 (0.243)*** 0.926 (0.181)*** 0.663 (0.241)***

Uncertainty -23.547 (20.917) 15.547 (29.352) -33.276 (20.864) 15.437 (29.479)

Dincubated -0.062 (0.191) -0.653 (0.311)** -0.019 (0.188) -0.709 (0.311)**

Dmother company 0.376 (0.199)* 0.067 (0.267) 0.430 (0.195)** 0.003 (0.267)

Infrastructure 0.002 (0.002) 0.000 (0.003) 0.001 (0.002) 0.000 (0.003)

Rreal -0.050 (0.032) -0.009 (0.042) -0.055 (0.032)* -0.013 (0.041)

Defaults -0.035 (0.762) -1.050 (1.051) -0.148 (0.757) -1.204 (1.060)

N° of observations 371 371 371 371

Log-likelihood -581.46 -367.94 -575.57 -367.23 Legend. Standard errors in parentheses. * stands for a significance level greater than 90%, ** stands for a significance level greater than 95% and *** stands for a significance level greater than 99%.

7