Sie sind auf Seite 1von 31

The Journal of Financial Research • Vol. XXXIII, No.

4 • Pages 487–517 • Winter 2010

DETERMINANTS OF CAPITAL STRUCTURE IN BUSINESS


START-UPS: THE ROLE OF NONFINANCIAL STAKEHOLDER
RELATIONSHIP COSTS

Tom Franck
Katholieke Universiteit Leuven, Lessius and the National Bank of Belgium

Nancy Huyghebaert
Katholieke Universiteit Leuven

Abstract

According to the finance literature, nonfinancial stakeholders (NFS), such as


customers, suppliers, and employees, take into account their expected liquidation
costs when dealing with a firm. In this framework, firms can influence their
probability of liquidation by choosing an appropriate capital structure. Also, the
literature suggests NFS bargaining power may affect firm financing decisions.
In the current article we investigate these ideas for initial financing decisions by
business start-ups, where ex ante failure risk is high and NFS must decide whether
to make relationship-specific investments. We find that start-ups imposing larger
costs on their NFS following liquidation significantly reduce leverage. This effect
is strengthened when suppliers have greater bargaining power. We also document
a marginally negative effect of NFS liquidation costs on the proportion of bank
loans. Finally, business start-ups rely less on bank loans when customers and
suppliers are in a powerful bargaining position.

JEL Classification: C31, G21, G32, G33

I. Introduction

We empirically investigate how the relationships between a firm and its nonfinancial
stakeholders (NFS) influence capital structure. Although the finance literature so far
has paid much attention to how relationships with and among financial stakeholders

The authors thank Enrique Arzac, Hans Degryse, Antoon De Rycker, John Doukas, Evgeny Lyandres,
Phil Molyneux, Sheila O’Donohoe, Armin Schwienbacher, Anil Shivdasani, Linda Van de Gucht, Alexandra
Van den Abbeele, Steve Van Uytbergen, Frank Verboven, Josef Zechner, the editor (Jayant Kale), and
the reviewer (Gordon Phillips) at the Journal of Financial Research, and participants at the 2005 annual
conference of the European Financial Management Association (Milan), 3rd Corporate Finance Day (K.U.
Leuven), the 2005 international meeting of the Association Française de Finance (Paris), and the 2006 annual
meeting of the European Finance Association (Zurich) for valuable comments. They are also grateful to
Luc Sels, Bart Cambré, Johan Maes, and Christine Vanhoutte for providing them with the sample data.

487

c 2010 The Southern Finance Association and the Southwestern Finance Association
488 The Journal of Financial Research

affect financing decisions,1 a number of theoretical studies have argued that a firm’s
customers, suppliers, and employees may have an effect too. Titman (1984) is the
first to show that leverage determines the future liquidation decision, which in turn
affects the terms of trade between a firm and its NFS, and hence firm profitability.
More specifically, NFS take into account that they may lose their relationship-
specific investments once the firm is liquidated following a default on its debt and
thus ask compensation for that ex ante through contract terms. Alternatively, NFS
could even limit the amount of ex ante relationship-specific investments in the firm.
Firms in turn can influence this process to maximize their overall value by reducing
the probability of liquidation up front, for example, by limiting their debt ratio. This
idea is refined in subsequent models by Maksimovic and Titman (1991) and Arping
and Lóránth (2006), among others.
Besides the NFS liquidation cost channel, the bargaining power of NFS
vis-à-vis the firm may also affect firm financing decisions. Bronars and Deere
(1991), Dasgupta and Sengupta (1993), and Perotti and Spier (1993) argue that
once contracts with labor suppliers have been established, firms can lower the
amount of surplus that powerful unions can extract by increasing their debt ratio.
These authors therefore predict a positive relation between NFS bargaining power
and leverage. In contrast, Sarig (1998) shows that when firms worry about input
suppliers threatening to curtail their specialized factors of production, they will
limit leverage both before and after contracts are negotiated with NFS to prevent
potential NFS hold-up behavior. The reason is that a lower debt ratio reduces the
likelihood that a disruption of supplies will lead to the firm’s liquidation. Following
Aghion and Tirole (1994) and Allen and Phillips (2000), among others, one could
also argue that neutralizing the larger ex ante bargaining power of some NFS through
restricting the debt ratio will induce customers, suppliers, and employees to make
relationship-specific investments in the first place. Indeed, through its impact on
incentives to break off or change the terms of an agreement ex post, the bargaining
power of some NFS could influence the ex ante investment in relationship-specific
assets and knowledge by other NFS.
The empirical literature investigating the effects of NFS liquidation costs
and bargaining power on capital structure is scarce, largely because it is difficult to
operationalize these theoretical constructs using firm-level (accounting) data. Tit-
man and Wessels (1988), Opler and Titman (1994), and Welch (2004), for example,

1
Jensen (1986), for example, argues that free-cash-flow problems between managers and shareholders
can be restrained by increasing leverage. Also, agency problems between creditors and shareholders can
be reduced by raising the portion of monitored debt (e.g., Myers 1977; Diamond 1984). These theories
have been tested in numerous empirical studies using data from different countries. Overall, the literature
concludes that the collateral value of assets, profitability, growth opportunities, risk, and firm size affect the
debt ratio (e.g., Rajan and Zingales 1995; Manuel and Pilotte 1992; Wald 1999; Jiraporn and Gleason 2007)
and the proportion of bank loans (e.g., Houston and James 1996; Krishnaswami, Spindt, and Subramaniam
1999).
Determinants of Capital Structure in Business Start-Ups 489

use the ratios of selling expenses to sales and research and development (R&D) to
sales to proxy for NFS liquidation costs. A few studies have related these concepts
to the structure of input and output markets. For example, Kale and Shahrur (2007)
use concentration in customer and supplier industries to proxy for NFS bargain-
ing power. It is not surprising that this research yields mixed results regarding the
effects of NFS relationship costs on capital structure. Indeed, Titman and Wes-
sels (1988), Opler and Titman (1994), and Kale and Shahrur (2007) document a
significant negative relation between NFS liquidation costs and leverage, whereas
Welch (2004) finds no effect. Nonetheless, survey evidence by Brounen, de Jong,
and Koedijk (2006) reveals that more than 30% of chief financial officers (CFOs)
in France and the United Kingdom consider customers’ and suppliers’ concern
about bankruptcy an important determinant of their debt choice. This aspect ranks
third, after financial flexibility and the volatility of earnings and cash flows, and
thus precedes taxes, agency costs, and information asymmetries. Regarding NFS
bargaining power, Kale and Shahrur (2007) document that firms have significantly
higher debt ratios when customers and suppliers are in a strong bargaining posi-
tion, whereas Sarig (1998) shows that debt ratios are significantly lower when input
suppliers have larger bargaining power.
In this article we use unique survey data to examine the role NFS rela-
tionship costs play in the initial financing decisions made by first-time business
start-ups. These start-ups are neither the result of incorporating a previously self-
employed activity nor a new subsidiary by an existing enterprise. Therefore, these
firms need to build relationships with their customers, suppliers, and employees
from scratch. From the point of view of NFS, setting up such relationships is a risky
venture, especially when they have to make large firm-specific investments. Indeed,
start-ups typically face high ex ante failure risk, and the information asymmetries
between firm insiders and outsiders are large, given the firm’s lack of operating
history. It is well known that one out of two start-ups ceases operations within
the first five years (e.g., Berger and Udell 1998; Huyghebaert and Van de Gucht
2004). Hence, in the presence of large information asymmetries, NFS may find it
difficult to determine a start-up firm’s intrinsic quality and thus survival chances.
Next, entrepreneurs generally are owner-managers and therefore have discretionary
power regarding the firm’s strategy and operations; this also provides flexibility for
entrepreneurs to make decisions that could harm NFS once the firm is heading
for financial distress. Nonetheless, NFS may be able to benefit considerably from
the relationship-specific investments they made when start-up firms survive and
grow. From the point of view of start-ups, relationships with customers, suppliers,
and employees are crucial to ensure their profitability and survival. Start-up firms
could encourage NFS to make these relationship-specific investments by limiting
the likelihood of liquidation up front. Moreover, when liquidation risk is limited,
the terms of trade between the firm and its NFS are likely to be more favorable to
start-ups. Titman (1984) points out that choosing a capital structure with limited
debt is useful for this purpose.
490 The Journal of Financial Research

Entrepreneurs cannot ignore that NFS with bargaining power may try to
extract rents from the start-up firm when negotiating their contracting relationship.
As employees in start-up firms usually are not unionized and thus have only limited
bargaining power against their employer, we examine only the effects of customer
and supplier bargaining power in this study. Consistent with Sarig (1998), we hy-
pothesize that entrepreneurs may limit their firm’s debt ratio to reduce the negative
effects of NFS bargaining power on firm profitability and survival. This conjecture
contrasts with the arguments made for mature, listed firms, which have contracts
with NFS already in place and may not be greatly concerned about liquidation.
Entrepreneurs, however, are likely to be much more concerned about liquidation
because they usually invest substantial financial and human capital in their ven-
ture and enjoy sizable private benefits of control (e.g., Hamilton 2000). Finally,
we hypothesize that customer and supplier bargaining power may influence the
pricing of NFS liquidation costs to the start-up firm and hence its financing de-
cisions. Therefore, as an extension to the research by Titman (1984) and others,
we examine whether the relation between NFS liquidation costs resulting from
relationship-specific investments and capital structure is stronger when customers
and suppliers have greater bargaining power.
Our study extends the limited empirical research on the effects of NFS
relationship costs on capital structure in four ways. First, we examine the effects
of NFS relationship costs on the initial financing decisions of business start-ups
before NFS have made relationship-specific investments. Because start-up firms
lack an operating history, the research setting is relatively clean: the initial cap-
ital structure reflects the firm’s true financing choices at start-up and is not yet
influenced by operating performance. Second, we combine accounting data with
unique and detailed survey data to construct proxy variables for the theoretical
constructs of interest.2 The value of information that is not included in the finan-
cial statements is documented by Kale and Shahrur (2007), but only in the context
of mature, listed firms. Third, we examine whether NFS liquidation costs have a
greater effect on capital structure when customers and suppliers are in a strong
bargaining position. Finally, we investigate the impact of NFS relationship costs on
the debt ratio as well as on the debt composition. We thus explicitly recognize that
entrepreneurs may use more than one aspect of capital structure to deal with NFS re-
lationship costs. For newly established ventures in traditional industries, the external
financing sources typically consist of bank loans and trade credit (e.g., Berger and
Udell 1998; Huyghebaert and Van de Gucht 2007). Because banks enforce liquida-
tion rights more strictly upon default than suppliers, entrepreneurs may be able to

2
Titman and Wessels (1988), for example, point out that the negative relation between their uniqueness
measure, which is based on the ratios of selling expenses to sales, R&D to sales, and job quit rates, and
leverage may also be caused by the relation between this variable and nondebt tax shields on the one hand
and the collateral value of assets on the other.
Determinants of Capital Structure in Business Start-Ups 491

reduce their liquidation risk by limiting the portion of debt from bank loans, ceteris
paribus.
Our results show that NFS relationship costs significantly affect capital
structure, both statistically and economically. In particular, start-ups that impose
larger costs on their NFS following liquidation significantly reduce their debt ratio.
This effect is strengthened when suppliers have greater bargaining power. We also
find a marginally negative effect of NFS liquidation costs on the proportion of bank
loans. Furthermore, our results show that when customer and supplier bargaining
power are stronger, start-ups reduce their reliance on bank loans. Overall, the last
findings are consistent with Sarig’s (1998) model and suggest that entrepreneurs,
being concerned about their firm’s liquidation and potential hold-up behavior by
customers and suppliers, reduce their vulnerability to NFS bargaining power ex
ante by adjusting the debt mix.

II. Development of Hypotheses

Ever since Modigliani and Miller (1958) developed their irrelevance propositions,
arguing that it does not matter how you slice a cake as long as its size remains
constant, researchers have analyzed a wide range of market imperfections to argue
that capital structure does have a bearing on firm value. Whereas originally this
research concentrated on taxes, financial stakeholder information, and agency costs,
more recent studies borrow insights from the industrial organization literature to
further new capital structure determinants. Most of these studies examine how the
use of debt influences strategic interactions between an enterprise and its rival
firms. A small but growing number of articles investigate how the relationship
costs between a firm and its customers, suppliers, and employees affect financing
decisions. These NFS typically have incomplete information about the firm but,
unlike shareholders and creditors, hold no direct financial stake. Yet, NFS can
influence firm value through the terms of trade they negotiate with the firm. Also,
when NFS break off or change the terms of an agreement or relationship ex post,
this will affect firm value. Firms, in turn, can reduce the negative effects of NFS
relationship costs on their value by choosing an appropriate capital structure. In this
article we examine how NFS liquidation costs resulting from relationship-specific
investments and NFS bargaining power affect leverage and debt composition.

NFS Liquidation Costs


Even when the contracts between a firm and its NFS are short term, NFS may hold
implicit long-term claims when recontracting with this same firm is less costly than
switching to another firm. As these implicit claims cannot be unbundled and sold
apart from NFS explicit contracts, the risk associated with holding these claims is
492 The Journal of Financial Research

difficult to diversify. Moreover, NFS may incur large losses if a firm is liquidated,
as their implicit claims become worthless.3 The size of the losses that NFS suffer
upon the firm’s liquidation, hereafter called NFS liquidation costs, usually rise with
the amount of relationship-specific investments made by NFS. Such investments
include the development of product-related skills by customers, the investment in
customer relations and specific supply chains by suppliers, and the acquisition
of job-specific knowledge by employees. NFS cannot recover these investments
(“sunk” investments) once the firm is liquidated.4
Therefore, before making these relationship-specific investments, NFS take
into account the firm’s likelihood of liquidation. Also, like financial stakeholders,
NFS consider the firm’s risk profile when valuing their direct and implicit claims
on the firm after contracts have been established. In other words, the value of NFS
contracts depends at least in part on the firm’s financial condition, even when an
actual default on debt is still remote. This process influences the terms of trade
between the firm and its NFS. Borenstein and Rose (1995), for example, find that
financially distressed airlines reduce their ticket prices to compensate customers
ex ante for the potential loss of frequent flyer advantages (relationship-specific
asset). From the firm’s point of view, minimizing the proportion of NFS liquidation
costs that are charged to the firm through higher input and lower output prices
thus maximizes firm value. Theoretically, this can be achieved by lowering the
probability of liquidation, reducing the size of NFS liquidation costs, and restraining
the transfer of these costs to the firm. The theoretical literature has mainly examined
how capital structure can be used to lower the probability of liquidation (e.g., Titman
1984; Maksimovic and Titman 1991). A possible explanation is that managers can
more easily adjust their financial strategy than their product market strategy. We
also follow this approach and test whether firms adapt their capital structure to the
size of NFS liquidation costs to minimize the likelihood of liquidation up front. In
addition, we examine whether customer and supplier bargaining power affect the

3
We recognize, however, that liquidation may not be a necessary condition for NFS to incur losses.
Maksimovic and Titman (1991), for example, show that firms may ex post reduce the quality of their
products to cut costs once they head for financial distress. As with liquidation costs, rational NFS will
include these expected costs of financial distress in their terms of trade.
4
These NFS liquidation costs are thus different from the bankruptcy and liquidation costs that the
firm itself incurs. Warner (1977), for example, concludes that direct costs, such as lawyers’ fees and
management time lost during bankruptcy and liquidation procedures, are moderate (from 1% of market
value seven years before bankruptcy to 5.3% immediately before bankruptcy). Altman (1984) finds that
indirect costs, which include lost business and investment opportunities and inefficient asset sales (see
also Benmelech, Garmaise, and Moskowitz 2005) are considerably larger, from 8.1% three years before
bankruptcy to 10.5% in the year of bankruptcy. Andrade and Kaplan (1998) estimate that the total costs
of financial distress are between 10% and 23% of predistress firm value for highly leveraged firms. To
calculate the expected costs of financial distress, these estimates need to be multiplied by the probability
of distress. After doing so, Almeida and Philippon (2007) conclude that the expected costs of financial
distress amount to 4.5% of predistress value for BBB-rated firms.
Determinants of Capital Structure in Business Start-Ups 493

relation between NFS liquidation costs and capital structure, as NFS bargaining
power will influence the extent to which NFS liquidation costs can be charged to
the firm. Indeed, more powerful NFS may find it easier than less powerful NFS to
transfer their liquidation costs to the firm.
In this article we study two important capital structure variables: leverage,
which is the proportion of total financing that consists of debt, and debt mix, which
is the proportion of bank loans in total debt. Prior empirical work on NFS liquida-
tion costs (Titman and Wessels 1988; Opler and Titman 1994; Kale and Shahrur
2007) concentrates on the leverage decision. This is an obvious choice, as the debt
ratio determines when a firm’s control rights are transferred from shareholders to
creditors, who may liquidate the firm following default. An increase in firm lever-
age thus increases the probability of liquidation, ceteris paribus. But the liquidation
decision for companies that default on their debt is also influenced by the composi-
tion of that debt (e.g., Gilson, John, and Lang 1990; Franks and Sussman 2005). For
business start-ups, debt largely consists of bank loans and trade credit (Berger and
Udell 1998; Huyghebaert and Van de Gucht 2007). The literature provides several
arguments why a larger proportion of bank loans relative to trade credit increases
the probability of liquidation. First, banks include restrictive covenants in their debt
contracts to reduce adverse selection and moral hazard problems after the loan is
made. However, this practice is also likely to increase the probability that funds are
cut off once firms default on their bank loans. In fact, Carey, Post, and Sharpe (1998)
find evidence consistent with the idea that banks wish to establish a reputation of
being fierce liquidators. Second, Berger and Udell (1998) report that more than
90% of bank loans to small business borrowers are collateralized. Manove, Padilla,
and Pagano (2001) show that such practices induce banks to liquidate distressed
companies prematurely following default, thereby avoiding the effort and the risk
of restructuring a distressed firm. Third, Wilner (2000) and Huyghebaert, Van de
Gucht, and Van Hulle (2007) argue that if borrowers generate a large proportion
of lender profits, creditors will be more lenient in periods of financial distress. As
banks hold only a limited implicit equity stake in their borrowers, they will liquidate
sooner than nonbank lenders, such as suppliers. Consistent with these ideas, Franks
and Sussman (2005) find that banks enforce liquidation rights more strictly than
suppliers following default by SMEs.
Based on the preceding discussion, we suggest the following hypothesis:

Hypothesis 1: Firms with larger NFS liquidation costs will use less debt
and less bank loans as a proportion of total debt, ceteris
paribus.

NFS Bargaining Power


NFS bargaining power is defined as the ability of NFS to appropriate a portion of
the firm’s surplus (rents). Factors that increase the bargaining power of NFS are,
494 The Journal of Financial Research

for example, the size of customers and suppliers, and employee union participation.
As employee unionization rates are low for business start-ups, we investigate only
the influence of customer and supplier bargaining power in this article.5
The theoretical literature shows that capital structure can be used to reduce
the amount of surplus NFS can extract from the firm. Yet, researchers generally
differ in their assumption as to who is more averse to liquidation: NFS or the firm
itself. As a result, they also have different conjectures about the effects of NFS
bargaining power on capital structure. Bronars and Deere (1991), Dasgupta and
Sengupta (1993), and Perotti and Spier (1993) assume the firm (its shareholders)
is not as concerned about liquidation as its NFS. Furthermore, these authors con-
centrate on situations where contracts with NFS are already in place, and financing
decisions thus can be used to rebalance the relative bargaining power of the firm
against its NFS. Bronars and Deere (1991) and Dasgupta and Sengupta (1993) argue
that firms issuing debt rather than equity are obliged to pay out a portion of their
future earnings to creditors. Hence, these obligations limit the surplus that powerful
NFS can extract without driving the firm into default. The authors thus predict that
firms will increase their leverage when dealing with strong NFS. Perotti and Spier
find a similar effect on leverage, albeit for a different reason. They show that a
highly leveraged firm can force more concessions from its NFS by threatening not
to initiate investments that are crucial for its survival.
In contrast to these early studies, Sarig (1998) examines the effects of
NFS bargaining power on financing decisions before relationships with NFS are
well established (ex ante). Also, Sarig assumes that the firm (its shareholders) is
more averse to liquidation than are its NFS. Overall, he points out that firms can
limit their vulnerability to liquidation when powerful suppliers threaten to curtail
the supply of specialized factors of production by restricting their debt ratio. This
by itself may already prevent potential NFS hold-up behavior. What’s more, when
firms can neutralize the large ex ante bargaining power of some of their NFS
through restricting their debt ratio and proportion of bank loans up front, then other
customers, suppliers, and employees may have incentives to invest in relationship-
specific assets and knowledge in the first place.
Sarig’s (1998) arguments build on the research by Aghion and Tirole (1994)
and Allen and Phillips (2000), among others. The latter authors suggest that through
its effect on incentives to break off or change the terms of an agreement ex post,
the bargaining power of some NFS could adversely affect the ex ante investment
in relationship-specific assets and knowledge. Yet, Aghion and Tirole (1994) and
Allen and Phillips (2000) examine the implications of such hold-up behavior for
corporate equity ownership, which can be used to align the incentives of the firms

5
When we use employee unionization rates to proxy for employee bargaining power, we indeed find
no effect on the initial financing decisions of business start-ups. These results are not reported but can be
obtained from the authors.
Determinants of Capital Structure in Business Start-Ups 495

that need to invest in product market relationships. For the newly established firms
in our sample, equity ownership by other corporations and the firm’s employees
are trivial in the start-up year. Nonetheless, the arguments in these papers can have
more general implications for capital structure.
Overall, we expect to find support for Sarig’s (1998) model in a sample of
business start-ups that have to decide on their initial capital structure, as (1) firms
still need to negotiate contracts with NFS and (2) entrepreneurs are likely to highly
value their venture’s survival. Indeed, ownership in business start-ups as a rule is
highly concentrated in the hands of one or a few entrepreneurs who generally hold
a largely undiversified portfolio. In our sample, 64.62% of entrepreneurs own more
than 50% of their firm’s shares. Most of the other equity is provided by family
and friends, whereas only seven firms received venture capital at the moment of
start-up. In addition, and consistent with Hamilton (2000), the entrepreneurs in our
sample highly value their private benefits of control; 56.28% indicate that their
main motivation for becoming an entrepreneur is the challenge of managing their
own business. Also, 45.58% of entrepreneurs indicate that being their own boss is
an important reason for setting up their own venture. Purely financial reasons—that
is, earning more than under wage employment—are important for only 19.53% of
the entrepreneurs in our sample.
In sum, although distinguishing between the predictions of Bronars and
Deere (1991), Dasgupta and Sengupta (1993), and Perotti and Spier (1993) on the
one hand, and Sarig (1998) on the other is a purely empirical issue, we expect to
find support for Sarig’s model when examining the initial financing decisions of
business start-ups. Therefore, when the bargaining power of customers and suppliers
is extensive, we hypothesize that entrepreneurs will reduce their debt ratio to curb the
likelihood of liquidation. Likewise, neutralizing the larger ex ante bargaining power
of some NFS through restricting the debt ratio will induce customers, suppliers,
and employees to make relationship-specific investments in the first place. From
our discussion on the relation between the debt composition and the probability
of liquidation, we also conjecture that firms can moderate the influence of NFS
bargaining power on the probability of incentive problems with NFS and firm
liquidation by limiting the proportion of bank loans in total debt. Therefore, we
suggest the following hypothesis:

Hypothesis 2: Firms that face customers and suppliers with stronger bar-
gaining power will use less debt and less bank loans as a
proportion of total debt, ceteris paribus.

Interaction between Liquidation Costs and Bargaining Power


As an extension to Hypothesis 1, and thus the theoretical models of Titman (1984),
Maksimovic and Titman (1991), among others, we argue that the effects of NFS
496 The Journal of Financial Research

liquidation costs resulting from relationship-specific investments on capital struc-


ture could be stronger when NFS bargaining power is greater. Therefore, unlike
agency costs between creditors and shareholders, which affect firm value per euro
through the price of debt and thus should be fully incorporated into financing de-
cisions (e.g., Jensen and Meckling 1976), we conjecture that firms will take into
account NFS liquidation costs especially when the liquidation costs can be easily
charged to the firm through input and product prices. As the extent to which NFS
liquidation costs can be transferred to the firm is likely to depend on—in addi-
tion to institutional characteristics—the relative bargaining power of NFS against
the firm, we hypothesize that firms may adjust their capital structure to deal with
NFS liquidation costs, especially when customers and suppliers are in a strong bar-
gaining position. Therefore, in addition to the simple term NFS liquidation costs,
as captured by Hypothesis 1, we expect its interaction term with NFS bargaining
power to influence capital structure and propose the following hypothesis:

Hypothesis 3: The effect of NFS liquidation costs on capital structure is


strengthened when customers and suppliers have greater
bargaining power. Therefore, the interaction term between
NFS liquidation costs and NFS bargaining power will be
negatively related to leverage and the proportion of bank
loans in total debt, ceteris paribus.

Equation (1) summarizes our research design:


L
CS ij = ai1 + ai2 × LC j + ai3 × BP j + ai4 × (LC j × BP j ) + bil X l j , (1)
l=1

where

CS ij = capital structure variable i (leverage: i = 1; debt mix: i = 2) of firm j,


LC j = proxy variable for NFS liquidation costs in firm j,
BPj = proxy variable for NFS bargaining power in firm j, and
X lj = set of L control variables for capital structure variable i of firm j.

III. Data

The empirical capital structure literature focuses on listed firms, largely because
data on these firms are readily available. Yet, to examine the effects of NFS rela-
tionship costs on financing decisions, a sample of newly established entrepreneurial
ventures may be more appropriate because of these firms’ high failure risk, lack
Determinants of Capital Structure in Business Start-Ups 497

of operating history, and large information asymmetries. Also, as entrepreneurs


are more averse to liquidation and firms have not yet entered into contracts with
their NFS, examining the effects of NFS relationship costs in such a sample
may yield additional insights. Whereas access to financial data on privately held
firms is not straightforward in other countries, limited liability firms in Belgium
(“corporations”)—except for financial institutions, insurance companies, foreign
exchange brokers, and hospitals—are legally required to file their annual accounts
with the National Bank as of start up. In 2006, nearly 330,000 companies submitted
their financial statements, thereby covering about 80% of gross domestic product
(only self-employed, one-person businesses are not covered). This information is
commercialized by Bureau Van Dijk Electronic Publishing via the Belfirst database.
(http://eps.bvdep.com/en/BEL-FIRST.html).
In the first stage of the sample-selection process we used the PASO START
database to select our sample of business start-ups. This database contains sur-
vey information on 638 Belgian corporations established between October 2001
and September 2002 and employing between 1 and 49 persons (http://www.econ.
kuleuven.be/projects/paso/index eng.htm). These firms represent a 23.81% response
rate from Belgian start-ups that meet the preceding criteria. The survey consists
of 91 questions, which poll entrepreneurs on their firm’s financial and ownership
structure, operations, organization, and strategic choices at start-up. Although some
questions were quantitative (e.g., the percentage of sales that are customer specific),
others required answers on a five-point Likert scale. Entrepreneurs were questioned
shortly after start-up so that there is no survivorship bias in this database. We com-
bined this unique survey information with the firms’ annual accounts collected
from Belfirst.
In the second stage we excluded all firms that were not entrepreneurial
and first-time business start-ups. This selection criterion thus removed firms that
were incorporations of a previously self-employed, one-person business or firms
that changed their type of corporation. Also, newly established subsidiaries by
existing firms, split-ups, spin-offs, and other start-ups that are affiliated with an
existing enterprise were deleted. These screening criteria reduced the sample from
638 to 223 business start-ups that had not yet built up a reputation in input and
output markets at their moment of start-up. After deleting firms with insufficient
data to perform the multivariate analyses, we obtained a final sample of 209 true
business start-ups.6 These firms are narrowly focused and report only one industry
code, the European Nomenclature des Activités Economiques dans la Communauté
Européenne (NACE) code. When considering this industry code at the five-digit
level, our sample firms are active in 90 industries. Table 1 presents our sample

6
For 14 of 223 firms, the annual accounts were not included in the Belfirst database of Bureau Van
Dijk, and hence capital structure data were missing. These firms are likely to have been liquidated before
they had to file their first financial statements. Yet, in terms of NFS relationship costs in the PASO START
database, these firms are comparable to the firms on which we do have annual accounts information.
498 The Journal of Financial Research

TABLE 1. Industry Distribution of Business Start-Ups.

Industry Description Number of Firms

1. Food Food 6
2. Mines Mining and minerals 1
3. Oil Oil and petroleum products 0
4. Clothes Textiles, apparel, and footware 0
5. Durables Consumer durables 3
6. Chemicals Chemicals 4
7. Consumer goods Drugs, soap, perfumes, tobacco 3
8. Construction Construction and construction materials 28
9. Steel Steelworks, etc. 4
10. Fabricated products Fabricated products 0
11. Machines Machinery and business equipment 1
12. Cars Automobiles 7
13. Transport Transportation 17
14. Utilities Utilities 2
15. Retail Retail stores 51
16. Financial Banks, insurance companies, and other financials 13
17. Other Other 69

Note: This table presents the industry distribution of our sample of 209 business start-ups using the Fama
and French (1993) industry classification. All sample firms are established in Belgium between October
2001 and October 2002, and have between 1 and 49 employees in the start-up year. They are neither incorpo-
rations of a previously self-employed, one-person business nor firms that changed their type of corporation
or firms that are affiliated (spin-offs, subsidiaries, etc.) with an existing enterprise. The sample was
constructed from the PASO START database (http://www.econ.kuleuven.be/projects/paso/index eng.htm).

firms’ industry distribution, based on the more compact Fama and French (1993)
industry classification. Like the population, a significant portion of the firms in our
sample is active in trade and services. This contrasts with most previous studies on
newly established enterprises, which examine either manufacturing start-ups (e.g.,
Huyghebaert and Van de Gucht 2007) or high-tech ventures (e.g., Manigart and
Struyf 1997).
Table 2 reports descriptive statistics on firm size, asset structure, and fi-
nancial structure in the start-up year. The median firm employs three people in
the start-up year and its total assets equal €179,500. As median total assets is less
than median financing sources (€184,000), more than half of sample firms incur
accounting losses during the start-up year. Summary statistics on asset structure re-
veal that fixed tangible assets on average represent 35.32% of total assets, whereas
inventories and cash account for 18.41% and 14.48%, respectively. The start-ups
in our sample are highly leveraged, as outside (i.e., nonentrepreneurial) debt on
average equals 60.84% of total financing sources in the start-up year;7 the median

7
Debt is defined as the sum of long-term debt and current liabilities. In our sample, 16.20% of
entrepreneurs lent some money to their own firm, representing 17.12% of total debt in start-up accounts.
Yet, we treat these entrepreneurial loans as a source of (preferred) equity rather than debt financing and
Determinants of Capital Structure in Business Start-Ups 499

TABLE 2. Characteristics of the Start-Up Firms.

Mean Median 5th Pctl. 95th Pctl. Std. Dev.


Firm size
Number of employees 5.9882 3 1 21 11.6471
Total assets (€) 488,827 179,500 32,000 1,592,000 1,177,220
Total financing sources (€) 493,293 184,000 39,000 1,699,000 1,204,680
Asset structure
Fixed tangible assets/total assets 0.3532 0.3027 0.0200 0.8526 0.2643
Inventories/total assets 0.1841 0.1244 0.0068 0.5831 0.1846
Cash and marketable securities/total assets 0.1448 0.0966 0.0032 0.4560 0.1528
Financial structure
Leverage: outside debt/total financing sources 0.6084 0.6833 0.0914 0.9329 0.2690
Debt mix: bank debt/total debt 0.3612 0.3357 0 0.8801 0.3052
Trade credit/total debt 0.2390 0.1611 0.0049 0.6988 0.2459
Entrepreneurial loans/total debt 0.1712 0.0526 0 0.6977 0.2426
Long-term debt (>1 year)/total debt 0.3093 0.2635 0 0.8554 0.3003
Long-term bank debt (>1 year)/bank debt 0.8217 1 0 1 0.3271

Note: This table provides descriptive statistics for the sample of 209 business start-ups that is constructed
from the PASO START database. (http://www.econ.kuleuven.be/projects/paso/index eng.htm). All sample
firms are established in Belgium between October 2001 and October 2002, and have between 1 and 49
employees in the start-up year. They are neither incorporations of a previously self-employed, one-person
business nor firms that changed their type of corporation or firms that are affiliated (spin-offs, subsidiaries,
etc.) with an existing enterprise. The descriptive characteristics are based on the financial statements of
the first accounting year. Total financing sources is the sum of outside (i.e., nonentrepreneurial) debt and
entrepreneurial loans and equity. The other variables are self-contained.

debt ratio equals 68.33%. We use total financing sources rather than total assets as
the scaling variable because it abstracts from the earnings generated and retained
during the first accounting year. In our sample, the average firm raises 36.12% of
funds from banks and 23.90% from suppliers (trade credit). We find that 72.25%
of firms rely on bank loans whereas 97.26% use trade financing in the start-up
year (not reported). Other creditors include the workforce and tax authorities. Only
30.93% of total debt outstanding matures after one year, which reflects the impor-
tance of current liabilities (including trade credit) for business start-ups. For bank
debt, this percentage is 82.17%, however.
We extended the operating and financial information on the business start-
ups in our sample with annual accounts—again using Belfirst—of all incumbent
firms in the corresponding five-digit NACE industries from 1996 to 2001 (i.e.,
during the six years preceding the sample year). This resulted in data for 35,528

thus do not include them in the calculation of the debt ratio. The reason is that entrepreneurs are unlikely to
voluntarily file for bankruptcy when the debt-service payments on these loans can no longer be met. Indeed,
unlike the United States, Belgium has a creditor-oriented bankruptcy law (see also La Porta et al. 1998).
Therefore, debtors have no incentive to seek protection under it. Furthermore, most bankruptcy procedures
in Belgium entail the firm’s liquidation rather than a reorganization of the distressed firm.
500 The Journal of Financial Research

additional firms. As explained in the next section, data on recently established


start-up firms are used to calculate historical, exogenous industry-level variables
to measure some of the control variables.

IV. Variable Measurements and Methodology

To examine the influence of NFS relationship costs on capital structure, we need to


construct a set of proxy variables that are closely related to the concepts of interest.
One methodological challenge is resolving the endogeneity problem between cap-
ital structure and product market strategy. Arping and Lóránth (2006) discuss the
example of Apple, which made its software more compatible with that of Microsoft
when it became financially distressed. In this way, customer liquidation costs were
reduced. Therefore, for mature firms, the size of NFS liquidation costs and, more
generally, the firm’s operating strategy could be affected by its financial history. In
their robustness checks, Kale and Shahrur (2007) take this potential endogeneity
problem into account by lagging explanatory variables and estimating a simul-
taneous equations model for leverage and customer and supplier R&D intensity,
respectively. Our study takes a different approach by examining business start-ups
that do not have an operating or financial history and that need to decide on their
capital structure before entering the product market.
As all firms in our sample are private enterprises, we use book value mea-
sures to define the capital structure variables, that is, the dependent variables. Start-
ing from first-year balance sheets, we recalculate the initial financial structure as
closely as possible to the moment of start-up by disregarding the changes in equity
due to retained earnings (mostly losses). Also, entrepreneurial loans are considered
equity rather than debt financing and thus are not included in our definition of the
debt ratio (see footnote 7). Leverage is then calculated as the ratio of outside debt
to total initial financing sources. Debt mix is the ratio of bank loans to total debt
outstanding.
The variables NFS liquidation costs and NFS bargaining power are diffi-
cult to measure. Furthermore, proxy variables used in prior empirical research (e.g.,
R&D expenses) are not always available for European firms because of some minor
differences in financial reporting. Nonetheless, Kale and Shahrur (2007) point out
that information that is generally not included in the annual accounts may also ade-
quately capture NFS relationship costs. In this study, we therefore combine financial
statement information with a set of new proxy variables that are calculated from
the PASO START survey database. (http://www.econ.kuleuven.be/projects/paso/
index eng.htm). As this survey contains unique and detailed information on the
start-up’s contracts with NFS and input and output market strategies, it is well suited
to test our hypotheses. Table 3 provides descriptive statistics on the survey questions
that are relevant for our study on NFS liquidation costs and bargaining power. For the
Determinants of Capital Structure in Business Start-Ups 501

TABLE 3. Characteristics of the Proxy Variables.

NFS Statement/Question Likert Scale: Disagree on Left; Agree on Right N


NFS Liquidation Costs
C/S/E Our products/services are unique in 209
comparison to those of competitors

C/S/E It is difficult for our competitors to 209


copy our products

C/S/E Our firm has no problems 209


differentiating itself from its main
competitors

C/S/E Our firm strongly emphasizes 209


new/advanced processes and
technologies
C/S/E Our firm uses new or advanced 209
technologies

C/S/E Percentage of firm goods that are 209


customer specific

NFS Bargaining Power


C Management frequently discusses 209
changes in consumersí needs with
employees
C At least once a year we invite 209
customers in order to find out which
products/services they need in the
future
C We pay little attention to changes in 209
the preferences of our customers

C We have a tendency to neglect 196


important changes in our customers’
needs

C At least once a year we check 209


whether customers are satisfied with
the quality of our products/services

S The firm has only one supplier for its 40.67% of firms have only one supplier for their 203
main inputs (0 = no, 1 = yes) main inputs

Note: This table contains descriptive statistics on nonfinancial stakeholder (NFS) liquidation costs and
NFS bargaining power variables for the sample of 209 business start-ups. All variables are collected from
the PASO START database (http://www.econ.kuleuven.be/projects/paso/index eng.htm) and represent
features of a firm’s relationship with customers (C), suppliers (S), and employees (E). Block diagrams
represent the frequency of answers based on five-item Likert scales. The latter scales vary from completely
disagree (at the left) to completely agree (at the right). The last column in the table (N) reports the number
of respondents that answered that particular survey question.
502 The Journal of Financial Research

questions where the response rate was less than 100% (as reported in the last column
of Table 3), we imputed the sample mean to avoid losing data from the questions
that had been answered by these respondents.
Each of the questions capturing NFS liquidation costs is related to the con-
cept of NFS relationship-specific investments and has implications for all types of
NFS (customers, suppliers, and employees): if products and services are unique
and production processes and technologies are new or advanced, all NFS have to
make firm-specific investments. As the bargaining power of customers and suppli-
ers is not necessarily closely related, we calculate separate measures for each. This
approach is consistent with theoretical models on NFS liquidation costs, which do
not distinguish between customers, suppliers, and employees (e.g., Titman 1984).
In contrast, studies linking NFS bargaining power to capital structure do focus on a
specific NFS type, for example, employees (Bronars and Deere 1991) and suppli-
ers (Sarig 1998). To effectively summarize the relevant information in the survey
questions, we perform factor analyses on the questions measuring NFS liquidation
costs and customer bargaining power. This technique is based on the idea that the
correlations within a set of selected variables are due to some common underlying
(and unobservable) force(s). Hence, factor analysis extracts the force(s) that best
explain the correlations among the variables (see also Titman and Wessels 1988).
The results of the factor analyses for NFS liquidation costs and customer bargain-
ing power are reported in Table 4. We retain the factor with the highest eigenvalue
from each analysis and investigate whether the sign of the factor loadings, which
measure the importance of a particular survey question for the underlying factor,
is consistent with the concept of interest.
Table 4 shows that firms scoring high on the NFS liquidation cost variable
have unique products and services that are difficult to copy. These firms have no
problem differentiating themselves from their competitors and strongly emphasize
as well as use new or advanced processes and technologies. Overall, these results
support our conjecture that NFS have to make substantial relationship-specific
investments and thus face high liquidation costs when the proxy variable for NFS
liquidation costs is high. Next, firms with a high value for customer bargaining
power involve customers in their strategic and product market decisions. These
firms regularly consult their customers and incorporate information on customer
needs, tastes, and preferences in their operating decisions. Furthermore, these firms
pay close attention to changes in customer preferences and are responsive to changes
in customer needs.
For supplier bargaining power, we use a dummy variable that equals 1
when the firm buys its main inputs from only one supplier (“single sourcing”), and
0 otherwise. This bargaining power variable is related to the measure used by Kale
and Shahrur (2007), who calculate the concentration ratio in supplier industries. As
expected, our measures of NFS liquidation costs and bargaining power of customers
and suppliers are not highly correlated (below 0.3).
Determinants of Capital Structure in Business Start-Ups 503

TABLE 4. Factor Analysis for NFS Liquidation Costs and NFS Bargaining Power.

Factor: NFS Liquidation Cost Factor Loading Kaiser’s Measure

Our products/services are unique in comparison to those of 0.7447 0.7213


competitors
It is difficult for our competitors to copy our products 0.6224 0.7254
Our firm has no problems differentiating itself from its main 0.5548 0.7326
competitors
Our firm strongly emphasizes new/advanced processes and 0.7213 0.5931
technologies
Our firm uses new/advanced processes and technologies 0.7244 0.5948
Percentage of firm goods/services that are customer specific 0.1982 0.7138
Kaiser’s measure of sampling adequacy = 0.6548
Number of factors with eigenvalue > 1: 2
Eigenvalue factor “NFS Liquidation Costs” = 2.3340; Eigenvalue factor 2 = 1.2478
Factor: Customer Bargaining Power Factor Loading Kaiser’s Measure

Management frequently discusses changes in consumers’ needs 0.7484 0.6772


with employees
At least once a year we invite customers in order to find out 0.6958 0.6707
which products/services they need in the future
We pay little attention to changes in the preferences of our −0.4822 0.6164
customers
We have a tendency to neglect important changes in our −0.2350 0.5283
customers’ needs
At least once a year we check whether customers are satisfied 0.7275 0.6541
with the quality of our products/services
Kaiser’s measure of sampling adequacy = 0.6493
Number of factors with eigenvalue > 1: 2
Eigenvalue factor “Customer Bargaining Power” = 1.8611; Eigenvalue factor 2 = 1.1791

Note: This table reports the results of the factor analyses for nonfinancial stakeholder (NFS) liquidation
costs and customer bargaining power. These analyses are performed on the sets of variables that are
included in Table 3. The factor loading of a variable describes the relation between that variable and the
underlying (unobservable) theoretical construct (i.e., factor). Kaiser’s measure indicates the sampling
adequacy for each factor.

Finally, based on the existing literature we select control variables for the
leverage and debt mix equations. Prior studies find that collateral value of assets,
profitability, growth opportunities, risk, and firm size are all significant determi-
nants of capital structure in listed firms (see footnote 1). These variables are also
likely important for business start-ups, as agency problems with creditors cannot be
ignored given the high failure risk and large information asymmetries in start-ups
(see also Huyghebaert and Van de Gucht 2007). Because start-ups have no history,
historical firm-level data are not available to calculate these control variables. Using
data on asset structure from the start-up year could result in serious endogeneity
problems, however. Therefore, except for firm size, we measure the control vari-
ables at the corresponding five-digit NACE industry level. We obtained information
for business start-ups that started one year before our sample, from October 2000 to
504 The Journal of Financial Research

TABLE 5. Description of the Control Variables Used in the Estimations of the Financial Structure
Determinants.

Variable Name Description Mean Median Std. Dev. Leverage Debt Mix

Collateral value Industry mean of fixed tangible 0.3784 0.3720 0.1331 + +


assets to total assets for all
historical start-up firms
(between Oct. 2000 and Oct.
2001) in the corresponding
industry
Profitability Industry mean of EBITDA to total 0.0646 0.0585 0.0779 − +/−
assets for all historical start-up
firms (between Oct. 2000 and
Oct. 2001) in the corresponding
industry
Growth Industry mean of sales growth for 0.3167 0.1691 0.7405 − +/−
opportunities all historical start-up firms
(between Oct. 2000 and Oct.
2001) in the corresponding
industry
Risk Percentage of start-up firms with a 0.2032 0.2043 0.0857 − −
negative cash flow during the
first start-up year for all start-up
firms during 1996–2001 in the
corresponding industry
Firm size Logarithm of total financing 5.2076 5.0173 1.2936 + +
sources in the start-up year
(firm-level variable)

Note: This table presents the control variables that are used in the multivariate regressions of leverage
and debt mix. Except for firm size, these control variables are measured at the corresponding five-digit
Nomenclature des Activités Economiques dans la Communauté Européenne (NACE) industry level. In
addition, we report summary statistics (mean, median, and standard deviation) for these control variables
and the expected sign of their relation with leverage and debt mix. The data to calculate these control
variables were collected from the Belfirst database (http://eps.bvdep.com/en/BEL-FIRST.html), which
includes the annual accounts of all corporations in Belgium.

October 2001. Collateral value is measured as fixed tangible assets to total assets.
Profitability is calculated as the ratio of earnings before interest, taxes, depreciation,
and amortization (EBITDA) to total assets. Growth opportunities are proxied by
the one-year lagged sales growth rate. Risk is captured by the percentage of start-up
firms with negative cash flows in the corresponding industry; we calculate this
variable over 1996–2001 to improve its reliability. Finally, Firm size is measured
by the log of total financing sources in the start-up year. Table 5 summarizes the
control variables and presents descriptive statistics and their expected relation to
the capital structure variables.
Recent studies (e.g., MacKay 2003; Huyghebaert and Van de Gucht 2007)
incorporate the potential interdependencies among various capital structure com-
ponents in their estimations using a simultaneous equations model. The conditions
Determinants of Capital Structure in Business Start-Ups 505

of such a model are demanding, however. MacKay (2003) and Huyghebaert and
Van de Gucht (2007) conclude that even though various aspects of capital structure
are jointly determined, using ordinary least squares (OLS) to estimate the effect
of exogenous variables does not largely affect the results. Hence, we use a cross-
sectional OLS regression model to examine the role of NFS relationship costs and
test the robustness of our findings after accounting for the potential interactions
between leverage and debt composition.

V. Empirical Results

In this section, we look at how NFS liquidation costs and NFS bargaining power
affect a start-up’s initial capital structure. Then, we test whether these variables also
have a joint effect on financing decisions and discuss our findings for the control
variables. Finally, we report the results from various additional analyses and robust-
ness checks. Overall, given the size of our sample, we use a 10% cutoff to determine
the statistical significance of the variables in our models. Multicollinearity is never
a problem, as variance inflation factors are always below 5.

NFS Liquidation Costs


In line with earlier findings by Titman and Wessels (1988), Opler and Titman
(1994), and Kale and Shahrur (2007) for listed firms, Table 6, column 1 reveals
that business start-ups have significantly lower leverage when NFS liquidation
costs are large (p-value of .0180). This result is both statistically and economically
significant. A firm with NFS liquidation costs 1 standard deviation higher than an
otherwise identical firm has 3.97% less debt outstanding. This relation is robust after
including industry-level dummy variables based on the Fama and French industry
classification in column 2 (p-value of .0087).8 Furthermore, this result remains
robust as each survey question, one by one, is excluded from the factor analysis for
NFS liquidation costs (not reported). Overall, these findings confirm Hypothesis 1
that newly established entrepreneurial ventures choose a capital structure that lowers
their probability of default and thus liquidation when NFS face high liquidation
costs, ceteris paribus.
We find a similar, albeit somewhat weaker, negative effect of NFS liquida-
tion costs on the debt mix in Table 6 (see column 1 of the debt mix equation), again
supporting Hypothesis 1. A firm with NFS liquidation costs 1 standard deviation
higher than an otherwise identical firm has 3.66% fewer bank loans as a proportion
of total debt. Again, this relation continues to hold after excluding each survey

8
For this purpose, we include an industry dummy variable for each industry—as reported in Table
1—that contains at least five business start-ups (see also Lally 2004).
506

TABLE 6. Determinants of Leverage and Debt Mix for Start-Up Firms: Base Model.

Leverage Debt Mix


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

Intercept 0.3917 0.3134 0.4061 0.1873 0.4039 0.3882 0.0948 −0.0174 0.1044 −0.0155 0.0887 −0.0565
(.0010) (.0067) (.0006) (.1810) (.0012) (.0142) (.500) (.9173) (.4522) (.9249) (.5412) (.7625)
NFS liquidation −0.0411 −0.0461 −0.0306 −0.0478 −0.0366 −0.0483 −0.0353 −0.0305 −0.0259 −0.0195 −0.0320 −0.0288
costs (.0180) (.0087) (.0837) (.0062) (.0426) (.0077) (.0924) (.1505) (.1324) (.2462) (.0812) (.0848)
Customer −0.0317 −0.0210 −0.0500 −0.0589
bargaining (.0739) (.2598) (.0186) (.0078)
power
Supplier −0.0626 −0.0575 −0.1330 −0.1296
bargaining (.2701) (.3110) (.0481) (.0558)
power
Collateral value 0.0304 0.1049 −0.0015 0.1807 0.0207 0.1248 0.6005 0.5135 0.5520 0.4809 0.5839 0.5229
(.8349) (.4884) (.9917) (.2244) (.8916) (.4273) (.0007) (.0018) (.0017) (.0066) (.0013) (.0057)
Profitability −0.0043 −0.0046 −0.0037 −0.0016 −0.0040 −0.0045 −0.0079 −0.0038 −0.0074 −0.0037 −0.0071 −0.0054
(.1115) (.1583) (.1750) (.5652) (.1465) (.1734) (.0159) (.2456) (.0219) (.3511) (.0307) (.1680)
Growth −0.0180 −0.0098 −0.0167 0.0123 −0.0169 −0.0064 −0.0347 −1.2259 −0.0313 −0.0364 −0.0283 −0.0257
opportunities (.4631) (.7162) (.4927) (.4064) (.4961) (.8140) (.2378) (.2363) (.2824) (.2535) (.3367) (.4279)
Risk −0.4823 −0.2156 −0.4650 −0.0362 −0.5135 −0.1961 −0.7677 −0.7677 −0.7379 −0.5850 −0.7427 −0.5516
The Journal of Financial Research

(.0270) (.4082) (.0314) (.8850) (.0227) (.4782) (.0034) (.0034) (.0044) (.0010) (.0054) (.0249)
Firm size 0.0674 0.0643 0.0677 0.0693 0.0677 0.0672 0.0458 0.0458 0.0488 0.0570 0.0472 0.0541
(<.0001) (<.0001) (<.0001) (<.0001) (<.0001) (<.0001) (.0043) (.0043) (.0022) (.0005) (.0036) (.0015)

(Continued)
TABLE 6. Continued.

Leverage Debt Mix


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

Food −0.2390 −0.1438 −0.2449 −0.0520 −0.0425 −0.0573


0.0462 (.2224) (.0423) 0.7169 (.7598) (.6880)
Construction 0.0198 0.1085 0.0282 0.0535 0.0123 0.0664
0.8182 (.1694) (.7464) 0.6056 (.8946) (.5227)
Cars −0.0163 0.0934 −0.0200 0.1820 0.2163 0.1804
0.8868 (.3987) (.8617) 0.1864 (.0986) (.1888)
Transport −0.0950 −0.0013 −0.0915 0.0021 0.0700 0.0124
0.3061 (.9980) (.3279) 0.9848 (.4959) (.9114)
Retail −0.1441 −0.0364 −0.1501 0.1060 0.0877 0.1109
0.0714 (.6328) (.0653) 0.2687 (.3316) (.2511)
Financial −0.0544 0.0692 −0.0473 0.2713 0.2377 0.2513
0.6133 (.5100) (.6776) 0.0370 (.0565) (.0644)
Other −0.1103 −0.0330 −0.1094 0.0757 0.0713 0.0736
0.1460 (.6266) (.1521) 0.4055 (.3738) (.4174)
Adjusted R2 14.45% 14.45% 15.27% 17.51% 14.56% 17.92% 18.73% 19.66% 20.62% 22.17% 19.04% 20.84%

Note: This table presents the ordinary least squares (OLS) regression estimates of the determinants of (1) leverage and (2) debt mix in the start-up year for 209
first-time business start-ups. Leverage is the ratio of outside debt to total financing sources. Debt mix is the ratio of bank debt to total debt. The measurement of
the test variables is presented in Table 4, and the control variables are defined in Table 5. The p-values are reported in parentheses.
Determinants of Capital Structure in Business Start-Ups
507
508 The Journal of Financial Research

variable from the NFS liquidation cost factor analysis (not reported). After the
Fama and French (1993) industry dummy variables are included (see column 2 of
the debt mix equation), the p-value of NFS liquidation costs rises to .1505. How-
ever, in other models including these industry-level dummy variables, we find that
NFS liquidation costs remains statistically significant at the 10% level, confirming
a weak negative relation between this variable and the debt mix.
Overall, our findings indicate that for first-time business start-ups, limiting
the debt ratio does not suffice to curb the negative effect of NFS liquidation costs
on firm value. Indeed, and consistent with our hypotheses, these firms also adjust
their debt composition when NFS are likely to face high liquidation costs. More
specifically, when NFS liquidation costs are potentially large, business start-ups
rely less on bank loans, consistent with the idea that banks in general are more
aggressive liquidators than suppliers. In unreported analyses, we have examined
whether business start-ups also adjust the maturity structure of their bank loans to
deal with NFS relationship costs, but found no such relation.

NFS Bargaining Power


In this section we test Hypothesis 2, that is, that start-up firms reduce their leverage
and proportion of bank loans in total debt to deal with powerful NFS. The discussion
is presented first for customers and then suppliers.
The results for customers are presented in columns 3 and 4 of each capi-
tal structure equation in Table 6. Customer bargaining power is significantly and
negatively related to the debt ratio (p-value of .0739). This finding, however, is not
robust to the inclusion of industry dummy variables (p-value of .2598). Nonethe-
less, customer bargaining power significantly and negatively affects the proportion
of bank loans in total debt (p-value of .0186). A 1 standard deviation increase in
customer bargaining power lowers the proportion of bank loans by 4.95%. This re-
lation remains robust after including industry dummy variables (p-value of .0078).
Overall, these results support Hypothesis 2 but reveal that firms prefer to adjust
their debt composition rather than overall debt ratio to reduce the likelihood of
liquidation when customers are more powerful.
The results for supplier bargaining power are reported in columns 5 and
6 of each equation in Table 6. They reveal that firms do not adjust their leverage
to deal with more powerful suppliers but rather change the composition of their
debt. Indeed, we find that supplier bargaining power is significantly and negatively
related to the debt mix (p-value of .0481), which continues to hold after controlling
for industry fixed effects (p-value of .0558). A start-up firm that buys its inputs from
a single supplier has 13.30% fewer bank loans outstanding relative to total debt,
ceteris paribus. Again, these results are consistent with Hypothesis 2, suggesting
Determinants of Capital Structure in Business Start-Ups 509

firms adjust their debt mix to reduce their probability of liquidation when suppliers
can threaten firm survival.9
Overall, our findings appear to contrast with those of Kale and Shahrur
(2007). Indeed, these authors document a positive relation between customer and
supplier concentration and leverage. Their results are thus consistent with the mod-
els of Bronars and Deere (1991), Dasgupta and Sengupta (1993), and Perotti and
Spier (1993), whereas ours support Sarig (1998). However, Kale and Shahrur exam-
ine listed firms, where contracts with NFS are already in place. Established firms
may have incentives to change their capital structure ex post to reduce the amount
of surplus NFS can extract. In contrast, newly established ventures still need to
persuade NFS to invest in a long-term relationship. Furthermore, for entrepreneurs
in business start-ups, liquidation is likely to be a serious threat and firms are likely
to worry about fierce bank liquidation policies. Hence, to restrain the negative ef-
fects of customer and supplier bargaining power on firm survival, firms appear to
limit their proportion of debt that consists of bank loans. In sum, our findings do
not contradict those of Kale and Shahrur but rather illustrate the unique context of
entrepreneurial start-ups.

Interaction between Liquidation Costs and Bargaining Power


As an additional test of the NFS liquidation cost theory, we hypothesize that the
effects of NFS liquidation costs on capital structure are strengthened when NFS
are in a strong enough bargaining position to influence firm profitability (Hypoth-
esis 3). In Table 7, we include interaction terms between NFS liquidation costs
and customer and supplier bargaining power, respectively, to investigate this idea.
We find some limited support for Hypothesis 3, as the interaction term between
NFS liquidation costs and supplier bargaining power is negative and significant in
the leverage equation (p-values of .0594 and .0592 after including industry fixed
effects). Yet, the interaction terms between NFS liquidation costs and customer bar-
gaining power are never significantly related to capital structure. An explanation
could be that the latter NFS can fully charge their expected liquidation costs when
entering into a contracting relationship with the start-up firm, independent of their
bargaining power.

Control Variables
The discussion of the control variables is based on the parameter estimates in
Table 6 (Table 7 offers similar conclusions). First, we find that business start-ups

9
An alternative explanation for this negative relation between supplier bargaining power and debt mix
could be that suppliers with strong bargaining power grant less trade credit. We find no support for this
conjecture, as the correlation coefficient between supplier bargaining power and the number of days of
supplier credit is positive and insignificant.
510

TABLE 7. Determinants of Leverage and Debt Mix for Start-Up Firms: Models with Interaction Terms.

Leverage Debt Mix


(1) (2) (3) (4) (1) (2) (3) (4)

Intercept 0.4072 0.4241 0.4123 0.3847 0.10343 −0.0151 0.0884 −0.0567


(.0006) (.0057) (.0009) (.0143) (.4577) (.9329) (.5436) (.7625)
NFS liquidation costs −0.0321 −0.0430 −0.0241 −0.0358 −0.0245 −0.0182 −0.0323 −0.0278
(.0826) (.0221) (.2044) (.0605) (.1044) (.4098) (.1654) (.2308)
Customer bargaining power −0.0320 −0.0220 −0.0497 −0.0581
(.0723) (.1057) (.0198) (.0092)
Supplier bargaining power −0.0434 −0.0381 −0.1335 −0.1282
(.4478) (.5055) (.0514) (.0632)
NFS liquidation costs × 0.0044 0.0015 −0.0040 −0.0046
Customer bargaining power (.7739) (.9215) (.8236) (.7996)
NFS liquidation costs × −0.1061 −0.1060 0.0030 −0.0075
Supplier bargaining power (.0594) (.0592) (.9644) (.9111)
Collateral value −0.0037 0.0714 0.0490 0.1556 0.5541 0.4289 0.5831 0.5250
(.9797) (.6436) (.7470) (.3207) (.0017) (.0197) (.0015) (.0058)
Profitability −0.0037 −0.0046 −0.0039 −0.0041 −0.0074 −0.0064 −0.0071 −0.0054
(.1711) (.1540) (.1627) (.2077) (.0234) (.0949) (.0312) (.1722)
Growth opportunities −0.0172 −0.0108 −0.0189 −0.0068 −0.0308 −0.0360 −0.0282 −0.0258
The Journal of Financial Research

(.4831) (.6880) (.4461) (.8013) (.2910) (.2608) (.3393) (.4285)


Risk −0.4640 −0.2186 −0.5040 −0.1781 −0.7389 −0.5828 −0.7430 −0.5504
(.0322) (.4027) (.0244) (.5168) (.0044) (.0605) (.0056) (.0097)

(Continued)
TABLE 7. Continued.

Leverage Debt Mix


(1) (2) (3) (4) (1) (2) (3) (4)

Firm size 0.0675 0.0650 0.0634 0.0636 0.0489 0.0548 0.0473 0.0539
(<.0001) (<.0001) (<.0001) (<.0001) (.0022) (.0010) (.0041) (.0018)
Food −0.2238 −0.2367 −0.0108 −0.0567
(.0637) (.0482) (.9396) (.6922)
Construction 0.0126 0.0376 0.0390 0.0671
(.8844) (.6648) (.7044) (.5204)
Cars −0.0022 0.0129 0.2274 0.1809
(.9852) (.9104) (.0995) (.1890)
Transport −0.0826 −0.0997 0.0420 0.0118
(.3821) (.2837) (.7072) (.9159)
Retail −0.1406 −0.1434 0.1182 0.1114
(.0799) (.0762) (.2125) (.2508)
Financial −0.0471 −0.0399 0.2935 0.2518
(.6635) (.7236) (.0228) (.0647)
Other −0.0979 −0.1095 0.1122 0.0736
(.2034) (.1490) (.2184) (.4185)
Adjusted R2 14.87% 16.83% 15.72% 17.90% 20.23% 22.41% 18.61% 19.08%

Note: This table presents the ordinary least squares (OLS) regression estimates of the determinants of (1) leverage and (2) debt mix in the start-up year for 209
first-time business start-ups. Leverage is the ratio of outside debt to total financing sources. Debt mix is the ratio of bank debt to total debt. The measurement of
the test variables is presented in Table 4 whereas the control variables are defined in Table 5. The p-values are reported in parentheses.
Determinants of Capital Structure in Business Start-Ups
511
512 The Journal of Financial Research

in industries where assets have a higher collateral value raise a larger proportion
of bank debt. From the supply side, banks may be wary of lending to first-time
business start-ups. Indeed, adverse selection and moral hazard problems are likely
to be an important consideration for banks, as these firms exhibit high failure rates
and information asymmetries. However, when assets have a high collateral value,
banks can reduce their exposure to these problems by securing their loans. The
collateral value of assets for business start-ups is not significantly related to the
debt ratio, which is at odds with the positive and significant relation found for
mature, listed firms (e.g., Rajan and Zingales 1995; Wald 1999). Yet, business
start-ups that cannot obtain bank loans typically rely on noncollateralized trade
credit (see also Huyghebaert and Van de Gucht 2007).
Firms operating in more profitable industries have a lower proportion of
bank loans. However, profitability does not significantly affect overall leverage.
Growth opportunities, when measured at the industry level, have no significant
effect on initial financing decisions of business start-ups. Risk is significantly and
negatively related to leverage. We also find that banks limit their lending to business
start-ups in risky industries. These findings are robust to alternative definitions of
risk, such as the industry failure rate and the variance of cash flows within the
corresponding industry during 1996–2001.10 Overall, our results emphasize the
importance of failure risk as a determinant of initial capital structure for business
start-ups. Indeed, given these firms’ high default risk, potential agency problems
with creditors cannot be ignored (see also Huyghebaert and Van de Gucht 2007).
When we use the entire population of industry incumbents during 1996–2001 to
measure the control variables, our conclusions continue to hold (not reported). Al-
ternatively, when we calculate the control variables using firm-specific data from the
start-up year whenever possible—in particular, collateral value and profitability—
we find that NFS relationship variables continue to have a robust influence on
start-up capital structure, whereas risk loses some of its influence. If anything,
these results point out that firm-level variables (collateral value and profitability)
can better capture a firm’s risk profile than an industry-level proxy variable.
Consistent with research for mature, listed firms (e.g., Rajan and Zingales
1995; Wald 1999), larger business start-ups have significantly more debt outstand-
ing. In the bankruptcy costs literature, this is explained by the negative relation
between firm size and the probability of bankruptcy and by the notion that direct
bankruptcy costs represent a larger fraction of firm value for smaller firms. Yet,
start-ups with a greater need for external funds may have no alternative other than
to raise debt, given that equity sources are not readily available. Finally, we find that
larger business start-ups have a larger proportion of bank loans, ceteris paribus.

10
Peterkort and Nielsen (2005) consider the book-to-market ratio as an alternative measure of risk, but
this measure cannot be calculated for private enterprises.
Determinants of Capital Structure in Business Start-Ups 513

Additional Analyses and Robustness Checks


In this section, we discuss the findings from various additional analyses and robust-
ness checks. These results are not reported but are available from the authors upon
request. First, we examine the robustness of our conclusions when using the individ-
ual survey questions to measure the NFS liquidation costs and customer bargaining
power variables. For this purpose, we calculate the correlation coefficients of these
survey variables with leverage and debt mix. In addition, we estimate the multi-
variate models after replacing either NFS liquidation costs or customer bargaining
power with the corresponding survey question.
The correlation coefficients and OLS parameter estimates for the NFS
liquidation cost survey variables largely confirm our earlier conclusions. When
products or services are more unique, difficult to copy, or differentiated from those
of competitors, or when entrepreneurs emphasize or use new or advanced processes
and technologies, start-up firms raise less debt. The results for the debt mix equation
reveal the more limited effect of NFS liquidation costs on the proportion of bank
loans in total debt outstanding. Business start-ups rely less on bank loans only when
their products are differentiated from those of competitors or when firms emphasize
or use new or advanced processes and technologies.
The individual survey questions for customer bargaining power show that
firms have lower leverage and a smaller proportion of bank loans when customers
are more involved in future product decisions. The ratio of bank loans to total debt is
also significantly lower when the firm explicitly investigates whether its customers
are satisfied with the quality of products and services, and when the management
frequently discusses changes in consumers’ needs with employees. Overall, results
are strongest for the debt mix equation, which is consistent with our earlier findings
in Tables 6 and 7.
Next, we analyze whether the results differ when using customer identity
(individuals vs. firms and governments) as a proxy for NFS liquidation costs.
Professional customers are likely to have higher relationship-specific investments
and thus liquidation costs, as the products and services they buy are used as an
intermediate input in their own production process. In our sample, 32.84% of start-
ups realize more than 25% of their sales from dealing with other corporations
or the government. We find that the percentage of sales to professional clients is
significantly and negatively related to leverage and the proportion of bank loans in
total debt. Therefore, these results support our earlier findings and conclusions.
As we find that NFS relationship costs significantly affect both leverage
and the debt mix, and as these capital structure components are often jointly de-
termined, we estimate a model that incorporates the potential interactions between
leverage and debt mix. For this purpose, we first instrumented start-up leverage and
debt mix on their corresponding industry-level variables. This approach resulted in
multicollinearity problems, rendering the capital structure components and control
514 The Journal of Financial Research

variables insignificant. Hence, we reestimated the simultaneous equations model


using as instruments the residuals of a first-step auxiliary regression that removes
the effects of the industry-level control variables on the corresponding industry-
level capital structure components. Although these results must be interpreted with
caution, they make clear that our main conclusions are robust.
Split-sample regressions—using firm size, time of start-up within the sam-
pling period, and degree of competition within the industry—do not reveal any
significant differences either.
Finally, we consider the possibility that NFS relationship costs also affect
equity ownership by NFS in the start-up year. In our sample, 7.5% of firms have
a corporate owner, holding on average 25.53% of start-up equity. Unfortunately,
we are unable to ascertain whether this corporate owner holds a product market
relationship with the start-up. Likewise, 6% of sample firms have employee equity
ownership, representing 20.50% on average in these firms. We therefore estimate
various logit models where the dependent variable equals 1 if at least part of the
start-up’s stock is held by a corporate owner (employees), and 0 otherwise. The
explanatory variables in these models are those in Tables 6 and 7. The results reveal
that NFS liquidation costs, NFS bargaining power, and their interactions are not
significantly related to the likelihood of start-up equity being held in part by other
corporations and employees, respectively. So, we find no evidence that business
start-ups facing large NFS relationship costs attempt to align the incentives of NFS
with their own through involving these NFS in their ownership structure.

VI. Conclusions

Business start-ups lack history and reputation in input and output markets, and face
a high ex ante failure risk. Establishing relationships with these firms is therefore
a high-risk venture for NFS, especially when they have to make large relationship-
specific investments that will be lost if the firm is liquidated. Titman (1984) argues
that firms with large NFS liquidation costs may limit their debt ratio to reduce
the likelihood of future liquidation. Other theoretical studies show that firms may
adjust their capital structure to the size of NFS bargaining power, either to reduce
the surplus to be bargained (e.g., Bronars and Deere 1991; Dasgupta and Sengupta
1993; Perotti and Spier 1993) or to increase their survival chances and/or prevent
potential NFS hold-up behavior (e.g., Sarig 1998). We provide compelling empirical
evidence that NFS relationship costs are indeed a significant determinant of the
initial financing decisions of business start-ups.
First, we find that the size of NFS liquidation costs is significantly and
negatively related to leverage and, to a smaller extent, to the proportion of bank
loans in total debt. Second, supplier bargaining power affects the relation between
NFS liquidation costs and capital structure, as start-ups rely on even less debt
Determinants of Capital Structure in Business Start-Ups 515

when NFS liquidation costs are high and suppliers are powerful. In contrast, the
interaction terms between NFS liquidation costs and customer bargaining power
are never significant. Third, customer and supplier bargaining power negatively
affect how much financing a start-up obtains from banks. Overall, these last results
support Sarig’s (1998) conjecture that entrepreneurs, being concerned about strict
bank liquidation policies, limit their vulnerability to NFS strategic actions ex ante
by adjusting their capital structure. Indeed, to reduce the effects of customer and
supplier bargaining power on the likelihood of NFS hold-up behavior and firm
liquidation, firms limit their proportion of debt that consists of bank loans. Overall,
our findings do not contradict those of earlier studies on mature, listed firms but
rather illustrate the unique context of entrepreneurial business start-ups.
Our empirical findings suggest some avenues for future research. Survival
analysis could be used to determine whether start-ups that pay more attention to
NFS liquidation costs and NFS bargaining power have a higher probability of being
successful. It might also be instrumental to determine whether there are ways other
than adjusting capital structure to decrease the effects of NFS relationship costs on
firm value and survival. Arping and Lóránth (2006) show that firms can mitigate
NFS liquidation concerns by reducing the uniqueness of their products. This notion
could be important in the context of business start-ups, especially in traditional
industries, as debt usually is the only available source of outside financing. For
these firms, the opportunities to adjust capital structure to mitigate NFS relationship
costs thus are not inexhaustible.

References

Aghion, P., and J. Tirole, 1994, The management of innovation, Quarterly Journal of Economics 109,
1185–1209.
Allen, J. W., and G. M. Phillips, 2000, Corporate equity ownership, strategic alliances, and product market
relationships, Journal of Finance 55, 2791–2815.
Almeida, H., and T. Philippon, 2007, The risk-adjusted costs of financial distress, Journal of Finance 62,
2557–86.
Altman, E., 1984, A further empirical investigation of the bankruptcy question, Journal of Finance 39,
1067–89.
Andrade, G., and S. Kaplan, 1998, How costly is financial (not economic) distress? Evidence from highly
leveraged transactions that become distressed, Journal of Finance 53, 1443–93.
Arping, S., and G. Lóránth, 2006, Corporate leverage and product differentiation strategy, Journal of
Business 79, 3175–3207.
Benmelech, E., M. J. Garmaise, and T. J. Moskowitz, 2005, Do liquidation values affect financial con-
tracts? Evidence from commercial loan contracts and zoning regulations, Quarterly Journal of
Economics 120, 1121–54.
Berger, A. N., and G. F. Udell, 1998, The economics of small business finance: The roles of private equity
and debt markets in the financial growth cycle, Journal of Banking & Finance 22, 613–73.
Borenstein, S., and N. L. Rose, 1995, Bankruptcy and pricing behavior in US airline markets, American
Economic Review 85, 397–402.
Bronars, S. G., and D. R. Deere, 1991, The threat of unionization, the use of debt, and the preservation of
shareholder wealth, Quarterly Journal of Economics 106, 231–54.
516 The Journal of Financial Research

Brounen, D., A. de Jong, and K. Koedijk, 2006, Capital structure policies in Europe: Survey evidence,
Journal of Banking & Finance 30, 1409–42.
Carey, M., M. Post, and S. Sharpe, 1998, Does corporate lending by banks and finance companies differ?
Evidence on specialization in private debt contracting, Journal of Finance 53, 845–78.
Dasgupta, S., and K. Sengupta, 1993, Sunk investment, bargaining and choice of capital structure, Interna-
tional Economic Review 34, 203–20.
Diamond, D. W., 1984, Financial intermediation and delegated monitoring, Review of Economic Studies 51,
393–414.
Fama, E. F., and K. R. French, 1993, Common risk factors in the returns on stocks and bonds, Journal of
Financial Economics 33, 3–56.
Franks, J., and O. Sussman, 2005, Financial distress and bank restructuring of small to medium size UK
companies, Review of Finance 9, 65–96.
Gilson, S. C., K. John, and L. H. P. Lang, 1990, Troubled debt restructuring: An empirical study of private
reorganization of firms in default, Journal of Financial Economics 27, 315–53.
Hamilton, B., 2000, Does entrepreneurship pay? An empirical analysis of the returns to self-employment,
Journal of Political Economy 108, 604–31.
Houston, J., and C. James, 1996, Bank information monopolies and the mix of private and public debt
claims, Journal of Finance 51, 1891–1908.
Huyghebaert, N., and L. M. Van de Gucht, 2004, Incumbent strategic behavior in financial markets and the
exit of entrepreneurial start-ups, Strategic Management Journal 25, 669–88.
Huyghebaert, N., and L. M. Van de Gucht, 2007, The determinants of financial structure: New insights
from business start-ups, European Financial Management 13, 101–33.
Huyghebaert, N., L. M. Van de Gucht, and C. Van Hulle, 2007, The choice between bank debt and trade
credit in business start-ups: Financing costs versus liquidation policy, Small Business Economics
29, 435–55.
Jensen, M. C., 1986, Agency costs of free cash flows, corporate finance and takeovers, American Economic
Review 76, 323–29.
Jensen, M. C., and W. H. Meckling, 1976, Theory of the firm: Managerial behavior, agency costs, and
ownership structure, Journal of Financial Economics 3, 305–60.
Jiraporn, P., and K. C. Gleason, 2007, Capital structure, shareholder rights, and corporate governance,
Journal of Financial Research 30, 21–33.
Kale, J. R., and H. Shahrur, 2007, Corporate capital structure and the characteristics of suppliers and
customers, Journal of Financial Economics 83, 321–65.
Krishnaswami, S., P. Spindt, and V. Subramaniam, 1999, Information asymmetry, monitoring, and the
placement structure of corporate debt, Journal of Financial Economics 51, 407–34.
Lally, M., 2004, The Fama–French model, leverage, and the Modigliani–Miller propositions, Journal of
Financial Research 27, 341–49.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny, 1998, Law and finance, Journal of Political
Economy 106, 1113–55.
MacKay, P., 2003, Real flexibility and financial structure: An empirical analysis, Review of Financial
Studies 16, 1131–65.
Maksimovic, V., and S. Titman, 1991, Financial policy and reputation for product quality, Review of Financial
Studies 4, 175–200.
Manigart, S., and C. Struyf, 1997, Financing high technology start-ups in Belgium: An explorative study,
Small Business Economics 9, 125–35.
Manove, M., J. Padilla, and M. Pagano, 2001, Collateral versus project screening: A model of lazy banks,
RAND Journal of Economics 32, 726–44.
Manuel, T., and E. Pilotte, 1992, Production technology, nondebt tax shields, and financial leverage, Journal
of Financial Research 15, 167–80.
Modigliani, F., and M. Miller, 1958, The cost of capital, corporation finance, and the theory of investment,
American Economic Review 48, 261–97.
Myers, S. C., 1977, Determinants of corporate borrowing, Journal of Financial Economics 5, 147–75.
Opler, T. C., and S. Titman, 1994, Financial distress and corporate performance, Journal of Finance 49,
1015–40.
Determinants of Capital Structure in Business Start-Ups 517

Perotti, E. C., and K. E. Spier, 1993, Capital structure as a bargaining tool: The role of leverage in contract
renegotiation, American Economic Review 83, 1131–41.
Peterkort, R. F., and J. F. Nielsen, 2005, Is the book-to-market ratio a measure of risk? Journal of Financial
Research 28, 487–502.
Rajan, R. G., and L. Zingales, 1995, What do we know about capital structure? Some evidence from
international data, Journal of Finance 50, 1421–60.
Sarig, O., 1998, The effect of leverage on bargaining with a corporation, Financial Review 33, 1–16.
Titman, S., 1984, The effect of capital structure on a firm’s liquidation decision, Journal of Financial
Economics 13, 137–51.
Titman, S., and R. Wessels, 1988, The determinants of capital structure choice, Journal of Finance 43,
1–19.
Wald, J. K., 1999, How firm characteristics affect capital structure: An international comparison, Journal
of Financial Research 22, 161–87.
Warner, J., 1977, Bankruptcy costs: Some evidence, Journal of Finance 32, 106–31.
Welch, I., 2004, Capital structure and stock returns, Journal of Political Economy 112, 106–31.
Wilner, B., 2000, The exploitation of relationships in financial distress: The case of trade credit, Journal of
Finance 55, 153–78.