当前位置: 首页 > >

The impact of firm and industry characteristics on small firms’ capital structure

发布时间:

Small Bus Econ DOI 10.1007/s11187-010-9281-8

The impact of ?rm and industry characteristics on small ?rms’ capital structure
Hans Degryse ? Peter de Goeij ? Peter Kappert

Accepted: 1 March 2010 ? The Author(s) 2010. This article is published with open access at Springerlink.com

Abstract We study the impact of ?rm and industry characteristics on small ?rms’ capital structure, employing a proprietary database containing ?nancial statements of Dutch small and medium-sized enterprises (SMEs) from 2003 to 2005. The ?rm characteristics suggest that the capital structure decision is consistent with the pecking-order theory: Dutch SMEs use pro?ts to reduce their debt level, and growing ?rms increase their debt position since they need more funds. We further document that pro?ts reduce in particular short-term debt, whereas growth increases long-term debt. We also ?nd that inter- and intraindustry effects are important in explaining small ?rms’ capital structure. Industries exhibit different average debt levels, which is in line with the trade-off theory. Furthermore, there is substantial intra-industry heterogeneity, showing that the degree of industry

competition, the degree of agency con?icts, and the heterogeneity in employed technology are also important drivers of capital structure. Keywords Capital structure ? Panel data ? Pecking-order theory ? Trade-off theory ? SME ? Industry effects JEL Classi?cations C23 ? G32 ? G30 ? L26

1 Introduction The capital structure choice is one of the most important decisions faced by ?rm management. While many studies tackle the capital structure decision, most empirical work deals with large publicly listed ?rms which often have several types of securities traded (see Frank and Goyal 2008 for a recent review). Small unlisted ?rms, however, make up for more than 90% of all existing ?rms, and are the engine of growth in most economies. In this paper we study ?rm and industry characteristics that determine the capital structure of small unlisted ?rms in The Netherlands. The capital structure decision of small ?rms comes closest to the standard textbook case which considers the choice between debt and equity. Indeed, small Dutch ?rms typically only decide from which banks to borrow and do not face other complicating issues such as the choice between private and public debt, or which type of securities to

Hans Degryse is af?liated with CentER, EBC, is a CESIfo research fellow and holds the TILEC-AFM Chair on Financial Market Regulation. Peter de Goeij is af?liated with CentER and is a senior researcher for the TILEC-AFM research network on Financial Market Regulation. H. Degryse ? P. de Goeij (&) Department of Finance, Tilburg University, Warandelaan 2, 5037AB Tilburg, The Netherlands e-mail: p.c.degoeij@tilburguniversity.nl P. Kappert Corporate Clients Nederland, Rabobank International, P.O. Box 294, 6800 AG Arnhem, The Netherlands

123

H. Degryse et al.

issue. While previous studies on industry effects focus on larger ?rms, studying industry characteristics for small ?rms is particularly important as small ?rms are more likely to be single-line businesses. We exploit a large and detailed proprietary database with ?nancial statements of Dutch SMEs from 2003 to 2005. The advantage of this proprietary dataset over publicly available datasets is that it contains detailed information on many small ?rms. Indeed, ?rms often only report partial information to public datasets, whereas they are requested to provide more details to their ?nanciers. Another unique feature of the database is its sheer size. In our analysis, we use an unbalanced panel that contains about 100,000 ?rm-year observations, covering eight different industries over 3 years. The dataset contains many very small ?rms, which distinguishes this study further from earlier SME capital structure studies that have medium-sized ?rms in their data (Michaelas et al. 1999; Sogorb-Mira 2005). SME capital structure has been investigated before for other European countries, for example the UK (Michaelas et al. 1999), Spain (Sogorb-Mira 2005), and Belgium (Heyman et al. 2008). Dutch SMEs have been considered together with a number of other European countries in a study of Hall et al. (2004). The Dutch case is particularly interesting, because compared with the USA or the UK, ?nancial markets are much less accessible for small businesses. Banks are the major ?nanciers for SMEs, and the banking sector in The Netherlands is among the most concentrated in the world (see, e.g., Cetorelli and Gambera 2001). Our dataset enables us to investigate whether the empirical results in The Netherlands are different from the results in other countries and from those of large ?rms. An additional interesting feature of our dataset is that we can test the impact of both ?rm and industry characteristics on SME capital structure (see also Michaelas et al. 1999). This allows us to investigate the importance of the pecking-order theory and trade-off theory both in general and for individual industries. Previous studies such as Balakrishnan and Fox (1993), Bradley et al. (1984), and MacKay and Phillips (2005) have found various impacts of inter- and intra-industry effects for large publicly listed ?rms. In line with Michaelas et al. (1999), we study inter-industry effects of capital structure for unlisted SMEs, but link them more closely to the importance of the pecking-order theory

and trade-off theory. Furthermore, we investigate intra-industry heterogeneity in capital structure. Our ?ndings can be summarized as follows. First, the ?rm characteristics show that the capital structure decision for Dutch SMEs is consistent with the predictions of the pecking-order theory. This is in line with previous ?ndings for, for example, the Spanish market (see Sogorb-Mira 2005). SMEs use pro?ts to reduce their debt level, since they prefer internal funds over external funds. However, if a ?rm is growing, it increases its debt position, since it needs more funds. Furthermore, we document that pro?ts affect in particular short-term debt, whereas growth affects long-term debt. This implies that, when internal funds are depleted, long-term debt is next in the pecking order. We also document that shortterm debt is more expensive and can be amortized easily. Second, we ?nd that SMEs with collateral more easily attract external ?nance. Moreover, we document that intangible assets and net debtors, which are often considered poor collateral, have a positive effect on the long-term debt level, suggesting that banks are able to employ these assets in their loangranting decisions. In addition, Dutch SMEs have a relatively large amount of long-term debt which is more risky for lenders. Third, we ?nd that long-term assets are ?nanced with long-term debt, which is consistent with the maturity-matching principle (see, e.g., Mitchel 1991). In addition, larger ?rms have relatively more longterm debt, while the impact of ?rm size on short-term debt is insigni?cant. These results are in contrast with, for example, Van der Wijst and Thurik (1993), who report that, if total debt is taken into account, most ?rm characteristics have insigni?cant effects, since the effects of long-term debt and short-term debt cancel out. Finally, we document that SME capital structure exhibits both signi?cant inter- and intra-industry variation. The inter-industry variation is in line with the trade-off theory, which suggests that industries may have different target capital structures. We further compare the role of ?rm characteristics across different industries and ?nd support for the peckingorder theory for almost all industries. The only exception is the catering and leisure sector, where more pro?table ?rms have larger debt, suggesting that the trade-off theory dominates for this sector. We

123

Firm and industry characteristics

further ?nd substantial intra-industry variation, as the ?rm ?xed effects within an industry explain a substantial fraction of the variation in capital structure. The remainder of this paper is organized as follows. In Sect. 2, we review the literature and formulate hypotheses. Section 3 presents the data as well as the econometric model. We discuss the empirical results in Sect. 4. Finally, Sect. 5 concludes.

Table 1 Capital structure theory and expected sign on leverage for explanatory variables Trade-off theory Firm characteristics Firm size Collateral Pro?tability Growth opportunities Industry characteristics Fixed effects Signi?cant ? ? ? ? ? ? ? Pecking-order theory

2 Literature review and empirical propositions 2.2 Leverage factors 2.1 Theory Modigliani and Miller (1958) argue that capital structure is irrelevant for ?rm value. In the vast stream of literature following Modigliani and Miller, the irrelevance proposition has been rejected, but a conclusive answer on what factors drive capital structure has not yet been provided. Several theories explain capital structure (for a review see, e.g., Frank and Goyal 2008). The ?rst theory is the pecking-order theory (POT) (Myers and Majluf 1984; Myers 1977, 1984), which builds upon asymmetric information between managers and investors. Firms prefer funding sources with the lowest degree of asymmetric information, since borrowing costs increase when obtaining funds from outside lenders who do not have complete borrower information. The POT implies that ?rms opt ?rst for internally generated funds (a form of inside equity), then for debt, and only as a last resort, for outside equity. This theory also states that there is no optimal debt-to-equity ratio. The second theory is the trade-off theory (TOT), which argues that a ?rm chooses the optimal capital structure by balancing the tax bene?ts of debt and the costs of ?nancial distress (see, e.g., Brennan and Schwartz 1978; DeAngelo and Masulis 1980; Bradley et al. 1984). These costs increase with the degree of leverage. Finally, the market timing theory of Baker and Wurgler (2002) states that management raises equity in hot equity markets but issues debt in cold equity markets. However, for our research, the TOT and POT are most relevant, as SMEs are typically privately held. Our empirical tests will therefore focus on these two theories. We discuss subsequently the ?rm and industry determinants of leverage as well as their relation to both capital structure theories. We summarize the predictions in Table 1 and formulate explicit hypotheses. 2.2.1 Firm characteristics Firm size is considered as an inverse proxy of bankruptcy costs. The TOT predicts a positive relationship between ?rm size and leverage, because size is assumed as a proxy for earnings volatility and larger ?rms are generally more diversi?ed and show less volatility (Fama and French 2002). Less volatile earnings reduce indirect bankruptcy costs such that ?rms can take on more debt. The POT also predicts a positive relationship between ?rm size and leverage, because more diversi?cation and less volatile earnings mitigate problems of asymmetric information. This decreases the costs of debt compared with other sources of ?nance. Several empirical studies ?nd a positive relationship for both large ?rms and SMEs (Van Dijk 1997; De Jong 1999; Fama and French 2002; Michaelas et al. 1999; Cassar and Holmes 2003; Sogorb-Mira 2005; Hall et al. 2004). Our ?rst empirical proposition (or hypothesis) based on the TOT and POT is: Proposition P1 Larger ?rms have higher leverage.

The effect of ?rm size on short-term debt has been empirically veri?ed by several authors. Michaelas et al. (1999) and Hall et al. (2004) report a negative effect, even though the effect on total leverage is positive. Sogorb-Mira (2005) ?nds similar effects for

123

H. Degryse et al.

total debt but no signi?cant effects of ?rm size on short-term debt. Ortiz-Molina and Penas (2006) ?nd that size increases the maturity of lines of credit. The high business risk and informational opacity increase if ?rms are smaller. Small ?rms then have to rely more on short-term debt. We therefore formulate the following two propositions based upon previous empirical work: Proposition P1a long-term debt. Proposition P1b short-term debt. Firm size is positively related to Firm size is negatively related to

Proposition P2a Collateral has a stronger positive effect on long-term debt than on short-term debt. Liquidity is a second dimension of a ?rm’s asset structure. Illiquid ?rms are restricted in attracting debt, as bankruptcy costs are high. The TOT then predicts a positive relationship between liquidity and leverage. We employ ‘‘net debtors’’ as a proxy for liquidity. It is particularly relevant for SMEs because small ?rms generally put less pressure on collecting payments from customers. Late payments are often ?nanced by trade credit. In the pecking order, trade credit may be on top of the preference list. Suppliers grant trade credit as they may have superior information compared with banks regarding their customers’ liquidity. This alleviates problems of asymmetric information (Berger and Udell 2006). Of course, ?rms cannot delay late payments to creditors beyond a certain point. It can therefore be expected that short-term debt increases if a ?rm suffers from late payments. Michaelas et al. (1999) report positive coef?cients of net debtors on short-term and longterm debt, although the effect on long-term debt was negligible. These results give rise to the next propositions: Proposition P3 the debt level. Net debtors is positively related to

The ?rm’s asset structure is a second factor determining capital structure. Asset tangibility is expected to be positively correlated with debt, as it provides collateral. Collateral reduces agency problems with debtholders and reduces bankruptcy costs and credit risk. Therefore, the TOT predicts a positive relationship between collateral and leverage. Collateral also mitigates information asymmetry problems such that also the POT implies a positive correspondence. De Jong (1999) con?rms the positive relationship between tangible assets and leverage, whereas Titman and Wessels (1988) report a negative, though not statistically signi?cant, relationship. The information asymmetry argument is particularly relevant for SMEs, as they are more opaque than large ?rms. Small ?rms often do not have to provide (audited) ?nancial statements or do not issue traded securities. For these reasons, collateralized lending is important for SMEs. Michaelas et al. (1999) and Sogorb-Mira (2005) ?nd a positive effect of tangible assets on leverage for SMEs. Hall et al. (2004) report a small positive relationship for Dutch SMEs. Therefore, our proposition regarding asset structure is: Proposition P2 debt ratio. Collateral has a positive effect on

Proposition P3a Net debtors has a stronger positive relationship with short-term debt than with longterm debt. Pro?tability is another determinant of capital structure. The free cash ?ow theory of Jensen (1986) states that more debt disciplines the manager if pro?ts increase. A positive relationship between debt and pro?tability is then expected. The POT predicts the opposite effect of pro?tability on leverage. Retained earnings are on top of the preference list to ?nance investments, so higher pro?ts reduce the necessity to raise debt. Studies using large-company data ?nd a negative relationship between debt and pro?tability (Titman and Wessels 1988; Van Dijk 1997; Fama and French 2002). The POT also applies to SMEs, whereas agency con?icts between managers and shareholders should be less relevant (see also Ang 1992). Studies on SMEs also ?nd a negative impact of pro?tability on debt (Van der Wijst and Thurik 1993; Michaelas et al. 1999;

Collateral may affect short-term and long-term debt differently. Previous work documents a negative relationship for short-term debt and a positive one for long-term debt (Van der Wijst and Thurik 1993; Michaelas et al. 1999; Hall et al. 2004; Sogorb-Mira 2005). Ortiz-Molina and Penas (2006) argue that collateral and maturity are substitutes in reducing agency problems. We therefore supplement proposition 2 with:

123

Firm and industry characteristics

Sogorb-Mira 2005). Therefore, our next proposition is: Proposition P4 to leverage. Pro?tability is negatively related

sales or assets. Sogorb-Mira (2005) reports a stronger positive effect on long-term debt, but a negative impact on short-term debt. Michaelas et al. (1999) ?nd a positive impact on short-term debt. Proposition P5 Growth opportunities positively relate to leverage. We also brie?y discuss expected impacts from taxation. Modigliani and Miller (1963) argue that ?rms prefer debt ?nancing because of the tax shield, so a positive relationship between the tax rate and leverage can be expected. Studies focusing on SMEs, however, ?nd a negative relationship for SMEs and argue that the tax status of a company is not informative. Sogorb-Mira (2005) show that SME managers choose other instruments to lower their tax payments, whereas Jordan et al. (1998) claim that taxes lower retained earnings. The total tax burden of a ?rm is not solely determined by the tax rate but by taxable income as well. Some authors argue that this is even more important than testing the tax rate itself (Van Dijk 1997). Interest payments reduce taxable income, but other items can do the same. These nondebt tax shields could substitute for the tax shield of debt (Titman and Wessels 1988). Hence, a negative relationship with debt ratio is expected. In the empirical section below, we also test for tax effects, but we do not formulate an explicit proposition. 2.2.2 Industry characteristics We now turn to formulating explicit propositions on industry effects. We ?rst focus on inter-industry effects. The TOT posits that ?rms target an optimal leverage ratio, and this optimal leverage may differ across industries. This can be captured by industry ?xed effects. The POT, in contrast, does not deliver a clear prediction with respect to industry ?xed effects. However, to the extent that there are unobservable factors that are correlated within an industry, then also industry ?xed effects could be signi?cant (see also Cole 2008). Finally, the TOT and the POT could also be of differential importance across industries. For example, the degree to which propositions P1– P5, particularly propositions P4 and P5, apply may be different. The empirical investigation of inter-industry effects deals with the question of the extent to which capital structure variation between ?rms is

Pro?tability may differentially impact short-term and long-term debt. Michaelas et al. (1999) ?nd a larger effect of pro?tability on long-term debt compared with short-term debt. They argue that SMEs prefer short-term ?nancing and that long-term debt will be reduced if internal funding is available. On the other hand, short-term debt can be amortized more easily and carries higher interest rates. This suggests a stronger in?uence on short-term debt, which is validated by several SME studies (Van der Wijst and Thurik 1993; Cassar and Holmes 2003; Sogorb-Mira 2005). Therefore, proposition P4 is supplemented as follows: Proposition P4a Pro?tability has a greater negative impact on short-term debt than on long-term debt. Agency problems between managers and debtholders are particularly relevant for ?rms with growth opportunities. Myers (1977), for example, argues that managers underinvest because equity holders may not earn a pro?t on some projects with positive net present value (NPV) if interest payments are high. The TOT predicts a negative relationship between growth opportunities and leverage. Myers (1977), however, models that short-term debt could overcome the underinvestment problem and therefore is positively affected by growth opportunities. According to the POT, growth opportunities and leverage are expected to be positively related. Firms with growth opportunities are more likely to raise new funds than are ?rms without growth possibilities (De Jong 1999). Growth opportunities for larger or publicly listed ?rms are proxied by research and development (R&D) expenses, the market-to-book ratio or intangible assets. Titman and Wessels (1988), Fama and French (2002), and Graham and Harvey (2001) report a negative relationship between their proxies of growth opportunities and leverage. Another explanation for a negative link is that assets needed for future growth are poor collateral. Studies on SMEs ?nd evidence for a positive relationship of leverage with growth opportunities. Growth opportunities in these studies are proxied by intangible assets or growth in

123

H. Degryse et al.

explained by industry characteristics compared with ?rm characteristics. Balakrishnan and Fox (1993), for example, ?nd that 52% of capital structure variation is explained by ?rm effects and 11% by inter-industry differences. MacKay and Phillips (2005) report similar percentages for ?rm and inter-industry effects. Michaelas et al. (1999) use industry ?xed effects to test whether industry effects have an in?uence on SME capital structure. They ?nd significant industry ?xed effects, but the impacts are primarily on short-term debt. We therefore formulate the following two propositions: Proposition P6 Industry ?xed effects are signi?cant determinants of leverage. Proposition P7 The relevance of empirical propositions P1–P5 differs across industries. Next to heterogeneity across industries, leverage could also exhibit intra-industry heterogeneity. This may be driven, for example, by industry competition, the degree of agency con?icts, and the heterogeneity in employed technology. The degree of competition, for example, determines whether a ?rm is close to the optimal degree of leverage within an industry. In particular, in industries with low competition, ?rms face less pressure to be close to the optimal target, whereas in industries with high competition, ?rms can only survive by choosing the optimal degree of leverage in order to minimize costs (Leibenstein 1966; MacKay and Phillips 2005). Agency con?icts resulting from con?icting objectives between shareholders and managers may determine ?rms’ capital structures; for example, managers could choose too low debt ratios in order to protect their human capital (Fama 1980) or to avoid pressure from interest payments (Jensen 1986). Managers may take on too much leverage in order to signal their quality or to decrease takeover attempts (e.g., Harris and Raviv 1991 or Stulz 1988). We then expect that, in industries without agency con?icts, there should be less leverage dispersion. Finally, Maksimovic and Zechner (1991) model that industries with more technological dispersion exhibit more capital structure dispersion. We do not formulate a hypothesis on intra-industry effects, as our dataset only contains limited information on competition, technological dispersion, or agency problems within an industry.

3 Description of the data and research methodology 3.1 Dataset Our dataset has been kindly provided by Rabobank, a large Dutch ?nancial institution. The database contains ?nancial statements of the bank’s SME clients. Many clients, particularly if they have a loan, are required to provide a detailed balance sheet and income statement every year. A potential concern is that our data is self-selected, as it comes from one bank only. We believe that the dataset is highly representative for the Dutch setting for several reasons. First, Rabobank is the largest player in this SME segment with a market share of 39% (in 2008) and is active in all industries and provinces. This should reduce the potential for selection issues to impact on our sample. It is also important to mention that, when ?rms have relationships with several banks, including Rabobank, these ?rms are part of the dataset. This further increases the relevance of the dataset at hand. Second, one could argue to use a publicly available dataset based on Amadeus (i.e., REACH). We compared the descriptive statistics of REACH with those of our dataset and ?nd that our dataset contains relatively more small ?rms, as it includes information on ?rms which by law do not have to submit detailed balance sheet data. Therefore, this self-selected sample allows greater learning about the capital structure of small ?rms in comparison with using REACH. Firms are included in the dataset when they have less than 20 million annual sales over the period 2002–2005, and when they report to the bank at least two annual accounts within this period.1 We therefore have an unbalanced panel. While the bank is active in all industries, the dataset does not contain ?rms active within the agricultural sector or the energy and utilities sector. Additionally, we removed ?nancial ?rms as is common in capital structure studies, as ?nancial institutions face regulatory capital requirements and may inherently have a different capital structure. Moreover, associations (e.g., sport clubs, political organizations, labor unions) were removed, because they do not have commercial activities and
1

In the analysis, all observations for 2002 were lost because they were needed to calculate the growth variable, as discussed below.

123

Firm and industry characteristics Table 2 Descriptive statistics De?nition Dependent variables Total debt Long-term debt Short-term debt Firm characteristics Size (log) Tangible assets Net debtors ROA Intangible assets Growth (assets) Tax rate Depreciation Industries Manufacturing, construction, wholesale trade, retail trade food, retail trade nonfood, catering and leisure, transport, services Notes: The amount of taxes paid is not directly observed. The amount is derived by multiplying the return on equity (which is based on pro?ts after tax) by the amount of equity. This gives pro?ts after taxes. Deducting this from pro?ts before tax gives an implied measure of taxes paid. Bank-strategic reasons prevent us from reporting the descriptive statistics on speci?c industries. ROA, return on assets; EBITD, earnings before interest, taxes, and depreciation log of total assets tangible ?xed assets/total assets (debtors - creditors)/total assets EBITD/total assets intangible assets/total assets [tot. assets(t) - tot. assets(t - 1)]/tot. assets(t - 1) taxes paid/earnings before tax depreciation expense/total assets 6.045 0.487 0.046 0.153 0.017 0.133 0.094 0.179 1.318 0.288 0.146 0.322 0.066 0.437 0.173 0.070 0.693 0 -0.534 -14.00 -0.308 -0.599 -0.362 0 9.171 1 0.587 7.286 1 3.300 0.771 1 total debt/total assets long-term debt/total assets short-term debt/total assets 0.492 0.308 0.184 0.246 0.252 0.162 0 0 0 1.659 1.452 0.993 Mean Std. dev. Min Max

often rely on governmental funding. Finally, we removed all entries with data errors, or which take values which are unreasonable on economic grounds, and drop observations with extreme values such as very large ?rms. Our ?nal dataset contained 99,031 ?rm-year observations. The number of observations in 2005 decreased substantially (by more than 30%) compared with 2003 and 2004. This stems from the collection efforts by the bank. From 2005 onwards they started to rely more on REACH, making the 2005 dataset more comparable to that dataset. We investigate the robustness of this in Sect. 4.3 below. We employ different proxies for capital structure. The most commonly used measure is total debt ratio, i.e., the relative amount of debt (leverage), de?ned as total debt over total assets. We also consider the short-term and long-term debt ratio separately. Definitions and descriptive statistics are presented in Table 2. Debt is measured by its book value. Market values are not known for SMEs, such that most SME managers have to base their ?nancing decisions on book values. For short-term debt, we include bank loans and other short-term debt.2 Following other
2

In a previous version of the paper we employed only shortterm bank debt. The empirical results were qualitatively similar.

studies, we excluded trade credit as it does not carry an explicit interest rate and is under the in?uence of completely different determinants (for example, de Jong et al. (2008) focus for that reason only on longterm debt in their cross-country analysis of the determinants of capital structure, and do entirely drop short-term debt). Table 2 shows that Dutch SMEs have more longterms loans than short-term loans (63% of total debt is long-term debt). These numbers are in contrast to those of Hall et al. (2004), who report an average long-term debt level of 2% for Dutch SMEs. This difference can be explained as follows. The long-term debt de?nition in our dataset is based upon loans given with a long maturity but not necessarily a long duration. That is, some short-term loans may be classi?ed as long-term debt as the debt is given within the framework of a line of credit but with a revisable loan rate. We further checked with previous work that focused on the capital structure of Dutch ?rms. For example, Chen et al. (1999) report a longterm to total credit ratio of 77%. Our numbers are in the same ballpark, showing that the Hall et al. (2004) results on short-term versus long-term debt should be seen as an outlier. The descriptive statistics in Table 2 also differ from those of other countries.

123

H. Degryse et al.

Sogorb-Mira (2005) reports for Spanish SMEs that 15% of total debt is long-term debt, and Michaelas et al. (1999) ?nd for UK SMEs that the ratio is 29%. Table 2 also provides descriptive statistics on our determinants of capital structure: ?rm size, tangible ?xed assets, net debtors, pro?tability, intangible assets, asset growth, effective corporate tax rate, depreciation, and industry characteristics. Firm size is measured as the log of total assets. A measure for asset structure is tangible assets. Tangible assets are all ?xed assets except intangible ?xed assets and inventories (Titman and Wessels 1988; Sogorb-Mira 2005). As opposed to real estate and equipment, inventories are short-term assets and therefore expected to be poor collateral. Net debtors is measured by the difference between debtors and creditors, scaled by total assets (Michaelas et al. 1999). Table 2 shows that the ?rms in our sample have much more tangible than intangible assets. In addition, on average, net debtors is small. To measure the effect of pro?tability, we use return on assets (ROA), which is de?ned as earnings before interest, taxes, and depreciation (EBITD) scaled by total assets. The pro?t numbers of nonincorporated business are corrected for an owner’s wage.3 Depreciation is not deducted in all empirical studies, but if the aim is to test how managers change their debt position with pro?ts, managers will very likely take into account the cash position. Moreover, depreciation is already used as a measure for nondebt tax shield. The proxy for growth opportunities is intangible assets scaled by total assets (Michaelas et al. 1999; Sogorb-Mira 2005). Intangible assets refer to assets that are expected to pay off in the future, such as brand names, goodwill, or research and development expenses. Current growth is measured by the relative yearly change in total assets, implying that the ?rst year of our analysis becomes 2003. We have data on eight industries. Bank-strategic reasons however prevent us from reporting descriptive statistics on those industries. The effective corporate tax is measured as the amount of company taxes divided by the pro?t before tax. This variable is not scaled by total assets, since the amount of taxes depends on pro?ts. Nondebt tax shields lower taxable income and can therefore
3

substitute for the tax bene?ts of debt. Titman and Wessels (1988) introduced depreciation as a proxy for nondebt tax shields, but did not ?nd signi?cant effects. A problem with depreciation as a proxy for nondebt tax shields is that it can also be an indicator for ?xed assets. Van Dijk (1997) reported a high correlation (i.e., 0.495) between depreciation and ?xed assets. Since he ?nds a signi?cant negative relationship between depreciation and leverage, he argues that it is unlikely that a ?rm’s collateral value (for which depreciation can be a proxy as well) has a positive in?uence on leverage. Nevertheless, depreciation was used in many other empirical studies (e.g., Fama and French 2002; Sogorb-Mira 2005). Table 3 shows the correlations between all variables of interest. As expected, there is a large correlation between long-term debt and total debt. In addition, the highest correlation is between long-term debt and tangible assets, suggesting that long-term debt goes together with physical collateral. Tangible assets and intangible assets exhibit a slightly negative correlation (-0.185), suggesting that they are (weak) substitutes. The other correlations are quite low, showing that multicollinearity is not a concern. 3.2 Econometric model We employ panel data analysis, as our dataset includes observations over several years. Some ?rms appear twice, while others appear for all 4 years, which makes the dataset unbalanced. We index all variables with an i for the individual (i = 1,…, N) and a t for the time period (t = 1,…, T). Depending upon our model below, the individuals i may be ?rms or industries. The general static panel data regression model can then be written as yit ? b0 ? x0it b ? eit ; i ? 1; . . .; N and t ? 1; . . .; T; ?3:1? where xit is a K-dimensional vector of explanatory variables, which does not contain an intercept term. This model imposes that the intercept b0 and the slope coef?cients in b are identical for all individuals (i.e., ?rms or industries) and time periods. A frequently employed panel data model assumes that eit = ai ? uit, where ai denotes the unobservable individual-speci?c effect (i.e., industry or ?rm) that is time invariant, and uit is the random error. In our

Net pro?t (before tax) of nonincorporated ?rms has been adjusted with a proxy for the average Dutch income for a small business director, which is 40,000.

123

Firm and industry characteristics Table 3 Correlations among variables employed in regressions Total debt Long-term debt Short-term debt Size (log) Tangible assets Net debtors ROA Intangible assets Growth (assets) Tax rate Depreciation 0.787 0.293 0.101 0.389 -0.046 -0.064 0.050 0.016 -0.111 0.049 -0.359 0.114 0.604 -0.157 -0.060 0.025 0.020 -0.182 0.011 -0.023 -0.346 0.174 -0.004 0.038 -0.007 0.114 0.058 0.034 -0.033 0.033 0.042 0.035 0.427 -0.347 -0.182 -0.063 -0.185 -0.047 -0.225 0.154 0.177 -0.004 0.017 0.129 0.047 0.031 0.026 0.023 0.155 0.018 0.090 0.091 0.017 -0.178 -0.123 Long-term debt Short-term debt Size (log) Tangible assets Net debtors ROA Intangible assets Growth (assets) Tax rate

empirical analysis we assume a ?xed-effects model for the unobservable individual effects for two reasons. First, the ?xed-effects model introduces an individual-speci?c intercept term (i.e., ?rm speci?c or industry speci?c) that could capture speci?c entrepreneurial skills or industry-speci?c factors. Berger and Udell (2006), for example, argue that the management capabilities of the entrepreneur are a crucial factor in SME ?nancing. We follow the approach of several SME capital structure studies which also use a ?xed-effects panel data model (Van der Wijst and Thurik 1993; Michaelas et al. 1999; Sogorb-Mira 2005). Second, the nature of the unobserved effects has been statistically veri?ed with a (not reported) Hausman test. This test rejects the null hypothesis that the explanatory variables and the individual effects (i.e., ?rm or industry) are uncorrelated. A ?xed-effects model can cope with correlation between explanatory variables and individual effects (i.e., ?rm or industry) and therefore it is statistically preferred (see also Verbeek 2008, pp. 367–369).

4.1 Firm characteristics The results of panel data regressions for total debt, long-term debt, and short-term debt are reported in Table 4. All regressions include seven industry ?xed effects to which we turn in Sect. 4.2. In all models, most of the individual variables are statistically signi?cant. The estimates presented in Table 4 con?rm proposition P1, as larger ?rms exhibit higher leverage. A one standard deviation change in log size implies a 3.03 percentage point increase in the ratio of total debt to total assets. Proposition P3a is con?rmed as well: the coef?cient for size in the long-term debt regression is positive, statistically signi?cant, and economically relevant. Proposition P3b is rejected, as ?rm size appears with a signi?cant positive coef?cient in the short-term debt regression. Its economic relevance, however, is very small. These results show that larger ?rms rely more on long-term ?nance and less on short-term ?nance. The results on total debt and long-term debt are in line with previous studies on SMEs (see, for example, Van der Wijst and Thurik 1993 and SogorbMira 2005). Larger ?rms are more aware of better ?nancing methods, since they employ more ?nancial and administrative staff and may have a stronger bargaining position towards lenders. Strong support is found for proposition P2 concerning the positive relationship between total debt and collateral. A one standard deviation increase in tangible assets implies a 10.08 percentage point increase in the ratio of total debt to assets. The

4 Empirical results Section 4.1 discusses the results on the ?rm characteristics as drivers of capital structure, using the entire sample pooling all industries. The results of the industry characteristics are discussed in Sect. 4.2. Section 4.3 investigates the issue of limited liability and summarizes the results of several robustness checks.

123

H. Degryse et al. Table 4 Industry ?xed-effects panel regressions with ?rm characteristics Total debt Estimate Firm characteristics Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation Industry ?xed effects Manufacturing Construction Wholesale trade Retail trade food Retail trade nonfood Catering and leisure Transport Services R
2

Long-term debt Std. error Estimate Std. error

Short-term debt Estimate Std. error

0.023* 0.350* 0.486* 0.161* -0.040* 0.024* -0.109* 0.070 Omitted -0.034* 0.014* 0.036* 0.076* 0.009 0.017* -0.009* 0.202

0.0029 0.0278 0.0449 0.0353 0.0155 0.0050 0.0281 0.0517

0.019* 0.546* 0.631* 0.021 -0.006 0.022* -0.141* -0.287* Omitted

0.0024 0.0148 0.0456 0.0114 0.0047 0.0029 0.0255 0.0239

0.004* -0.195* -0.145* 0.140* -0.034* 0.002 0.032* 0.357* Omitted

0.0015 0.0170 0.0224 0.0346 0.0113 0.0055 0.0097 0.0371

0.0026 0.0043 0.0056 0.0086 0.0059 0.0063 0.0007

-0.024* 0.002 0.019* 0.056* 0.008* 0.021* -0.022* 0.422

0.0014 0.0022 0.0024 0.0031 0.0036 0.0036 0.0003

-0.009* 0.012* 0.017* 0.020* 0.001 -0.003 0.012* 0.156

0.0013 0.0030 0.0045 0.0069 0.0033 0.0042 0.0006

Notes: This table provided the estimation results for Eq. 3.1 using the complete sample. * Signi?cant at the 5% level. Variable de?nitions are presented in Table 2

interpretation of a change is such that it is induced by the numerator and that total assets as a scaling variable remains unaffected. Collateral is very important for SMEs, since it helps to overcome informational problems. The positive effect on total debt stems entirely from long-term debt, as short-term debt is negatively affected by collateral, partly con?rming proposition P2a. Since collateral is a way to mitigate risk of SMEs, these ?rms can fully use their collateral to attract long-term debt. For the ?rm, the costs of long-term debt are lower because banks charge (relatively) higher interest rates on short-term loans. These ?ndings are in accordance with the maturity-matching principle that long-term assets are ?nanced with long-term ?nancing and short-term assets are ?nanced with short-term funds. There is also strong support for propositions P3 and P3a. Net debtors, ?nanced with both long-term and short-term debt, positively affect the total debt level. Firms with low net debtors have lower debt ratios (ceteris paribus). A one standard deviation

decrease in net debtors lowers the debt ratio by about 2.4 percentage points. The empirical results show that the effect is only statistically signi?cant and positive for short-term debt. This also provides evidence for the maturity-matching principle. Pro?tability is negatively related to the total debt ratio, and this supports proposition P4. A one standard deviation increase in ROA lowers the total debt ratio by 1.3 percentage points. Debt levels are lower if a ?rm generates pro?ts. This suggests that SME managers prefer internal ?nancing ?rst, as predicted by the POT. The most likely reason is that they want to stay in control and avoid debt as much as possible (Vos et al. 2007). This result shows that the agency problem of free cash ?ow is nonexistent in SMEs, because they do not have public equity and typically ownership is concentrated. We investigate this further in Sect. 4.3. The negative relationship between pro?tability and debt is only signi?cant for shortterm debt, providing support for proposition P4. This ?nding is consistent with previous studies by Van der

123

Firm and industry characteristics

Wijst and Thurik (1993), Cassar and Holmes (2003), and more recently Sogorb-Mira (2005) for Spanish data. Short-term debt can be amortized easily. Support for proposition P5 on growth opportunities is provided, as ?rms with more intangible assets have a greater total debt ratio: a one standard deviation increase implies a 3.2 percentage point increase in total debt ratio. The agency theory of Myers (1977) is therefore not supported by the results for growth opportunities. Support for the POT, however, is provided by the results of growth opportunities and asset growth. Firms with a lot of intangible assets have less short-term debt and are very well able to ?nance their future growth with long-term debt. It is, however, important to note that many ?rms in the database have no intangible assets on their balance sheet (Table 2). Also comparing economic relevance, tangible assets seem more important. The results for asset growth do not change the conclusion drawn for proposition P5. The coef?cients on asset growth are low, but a positive effect of asset growth on long-term debt is found. Therefore, our empirical results support proposition P5, which is in line with Michaelas et al. (1999). In the period under investigation (2003–2005) the average total assets per ?rm increased. The growth in total assets is mainly due to an increase in ?xed assets, which implies that ?rms invested more and could attract external ?nancing for this. However, in the same period, interest rates have declined, making it likely that ?rms used that opportunity to opt for long-term loans. Unfortunately, the effect of loan rates cannot be studied more in depth due to lack of detailed data. The results in Table 4 indicate that the tax rate has a signi?cant negative effect on total and long-term debt, but a slightly positive effect on short-term debt. In particular the results imply that a one standard deviation increase in tax decreases the long-term debt ratio of SMEs by 2.4 percentage points (ceteris paribus), while the short-term debt ratio increases by 0.45 percentage points (ceteris paribus). This ?nding is in line with Michaelas et al. (1999), who also report negative but small effects of taxes. A possible explanation is that high taxes stem from high pro?ts, which in turn decreases the need for debt (Jordan et al. 1998). The second measure of the tax effect, i.e., depreciation, is not signi?cant for total debt. It

shows a signi?cant positive coef?cient for short-term debt and a negative coef?cient for long-term debt. 4.2 Investigating industry effects We ?rst focus on inter-industry differences and test proposition P6. The bottom panel of Table 4 reveals that all industry dummies are signi?cant. This shows that all industries have a different capital structure compared with manufacturing, which is our base case. This holds for total debt, long-term debt, as well as short-term debt, providing support for proposition P6. These results show that differences in ?rm characteristics cannot explain all differences between industries for SMEs. In other words, this is evidence that some other characteristics of an industry are important determinants of the SME debt ratio. The industries with the strongest ?xed effects are retail trade nonfood and food, with 7.6 and 3.6 percentage points greater total debt ratio than the base case. These industries have a leverage ratio that is above average, while important ?rm characteristics such as pro?tability and collateral are below average. The retail food industry is known for its low equity ratio, since it is an extremely competitive industry. This is probably the reason why higher debt ratios are observed. The construction sector exhibits the lowest total debt ratio. Interestingly, the results also show that there is a ?xed industry effect that differs in sign between long-term and short-term debt for the catering and leisure, and transport sectors. In order to test proposition P7, we estimate a model for each industry separately. We now include both ?rm ?xed effects and ?rm characteristics. These industry sample regressions compute a coef?cient for each ?rm characteristic per industry (results are presented in Table 6 in the Appendix). Before turning to the heterogeneity in coef?cients on our ?rm characteristics, we mention that almost all the conclusions regarding the hypotheses are the same for all industries individually, suggesting that the POT is most relevant for all industries studied. To test proposition P7, we then investigate whether there is signi?cant cross-sectional variation in the estimated coef?cient for each ?rm characteristic. The standard deviation of the cross-section of the individual estimates for the eight different industries is used as a measure for this variation. This helps us to investigate whether the relevance of

123

H. Degryse et al. Table 5 Industry effects and leverage: variation of parameter estimates across industries Total debt St. dev. of estimates Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation 0.012* 0.084* 0.117* 0.088* 0.062* 0.014* 0.091* 0.132 Long-term debt St. dev. of estimates 0.008* 0.055* 0.068* 0.069* 0.034* 0.010* 0.091* 0.065* Short-term debt St. dev. of estimates 0.006* 0.047* 0.099* 0.096* 0.039* 0.008* 0.032 0.098*

Notes: This table presents the standard deviation of the estimates for eight industries as reported in the Appendix. The Wald tests indicate whether the individual estimates are the same across industries. * Signi?cant at the 5% level. De?nitions of all variables are presented in Table 2

propositions P1–P5 differs across industries. In other words, it allows us to investigate which capital structure theories are most relevant for which industries. Table 5 presents the estimates for the variation measure as well as the results for the individual Wald tests. These tests investigate whether all the coef?cients for a ?rm characteristic are equal across industries. If the Wald test hypothesis is rejected, the relationship is different for at least one industry. The Wald test indicates that, for most ?rm characteristics, the relationship with the debt level varies signi?cantly across industries. This variation is most pronounced and signi?cant for net debtors, tangible assets, intangible assets, tax rate, and pro?tability. We are most interested in the results for pro?tability and intangible assets, as these are related to propositions P4 and P5, respectively. The reason is that the TOT and POT have opposite predictions. Table 6, in the Appendix, shows that the coef?cient on pro?tability is only positive for the catering and leisure sector. This suggests that the pecking-order theory dominates for the other seven sectors, whereas the TOT dominates for the catering and leisure sector. The effect of pro?tability on leverage is particularly negative in the wholesale trade, retail trade food and nonfood, and transport sectors, suggesting that the POT dominates more for these sectors. Also note that the retail trade food sector is the only industry in which the effect on longterm debt is larger than on short-term debt: pro?ts reduce long-term debt more than short-term debt. The coef?cient on intangible assets is positive for all

sectors, suggesting that proposition P5 applies for all sectors. In other words, the POT dominates the TOT for all sectors. The coef?cient on intangible assets is largest for the transport sector. Finally, we investigate the role of intra-industry characteristics. We evaluate this by considering the impact of the ?rm ?xed effects on the R2 (R2 ?rm ?xed effects versus R2 pooled) for the regressions studying the different industries separately. A reading of the results in the Appendix shows two important results. First, ?rm ?xed effects are important in all industries, suggesting that within-industry heterogeneity is important. As the F-tests clearly show, the null hypothesis that all ?rm ?xed effects are equal to zero is rejected at all conventional signi?cance levels for all industries. These F-tests take into account the difference in degrees of freedom between the ?rm ?xed-effects model and the pooled regression model. Second, the values of the F-tests are highest for the retail trade nonfood and transport industries, which implies the largest increase in R2 (corrected for the difference in degrees of freedom) after adding the ?rm ?xed effects. This indicates that, within these industries, ?rm ?xed effects which pick up the individual variability in leverage ratios are very important. This shows that industry competition, the degree of agency con?icts, and the heterogeneity in employed technology are important drivers of capital structure. Unfortunately, as our dataset only contains limited information on competition, technological dispersion, or agency problems within an industry, we cannot further explore this issue and leave this for future research.

123

Firm and industry characteristics

4.3 Limited liability and robustness checks The dataset provides us with information on whether a ?rm has limited-liability protection or not. Based on Cole (2008), we expect that ?rms with limitedliability protection have higher leverage than otherwise similar ?rms. Such protection may also partly capture a decrease in ownership concentration compared with sole proprietorships, for example. To conserve space, we brie?y discuss our ?ndings without reporting them in tables. When adding a dummy for limited liability to the speci?cations reported in Table 3, we ?nd that it is not signi?cant for total debt, positive for short-term debt, and negative for long-term debt. This suggests that limited liability only increases the degree of short-term leverage. This is in contrast with the results in Cole (2008), where he ?nds that limitedliability ?rms exhibit higher total leverage (he does not investigate short- and long-term debt separately). We further investigated whether the ?rm and industry effects differ between limited-liability ?rms or not, by interacting all explanatory variables with the limited-liability dummy. We ?nd that the coef?cients on the ?rm characteristics are different for limitedliability ?rms. In particular, total debt of limitedliability ?rms is less exposed towards ?rm size, and tangible and intangible assets. In addition, both short term and long term debt of limited liability ?rms are less exposed towards tangible and intangible assets. This suggests that limited-liability ?rms have different means of raising capital for (future) investments or running their business, which is in line what we expect for those types of ?rms. We now brie?y summarize the results of four robustness checks. First, as indicated above, 2005 contains about 30% fewer observations than the other years. We investigated whether the results for 2005 were different from those in other years by running a model where we included interaction terms of all ?rm characteristics with a 2005 dummy. We found that almost all the interaction terms were insigni?cant. Using a Wald test, the null hypothesis that the coef?cients of all 2005 interaction terms were jointly zero could not be rejected. This indicates that our results are robust to the reduction in the number of observations. The second robustness check concerns the maturity-matching principle. This principle states that short-term assets are ?nanced with short-term

assets. In unreported regressions, we add the variable inventories, another short-term asset, to our speci?cations of Sect. 4.1. Previous studies such as Titman and Wessels (1988) and Michaelas et al. (1999) consider inventories as tangible ?xed assets, but if the maturity-matching principle is true, inventories should positively relate to short-term debt and have no signi?cant relationship with long-term debt, since inventories are a short-term asset. We ?nd that the coef?cients for inventories are signi?cant for both long- and short-term debt, which is not in line with the maturity-matching principle. Third, we replace asset growth by sales growth as proxy for growth opportunities. Our results are very similar to those reported in Sect. 4.1. Finally, to mitigate potential endogeneity issues, we ran regressions where we computed all explanatory variables using lagged values of total assets. Our results remain robust.

5 Concluding remarks We employed a large, proprietary panel dataset to study the impact of ?rm and industry characteristics on the capital structure decisions of Dutch small ?rms. Our results on the impacts of ?rm characteristics are mostly in line with the predictions of the pecking-order theory. SMEs use pro?ts to reduce their debt level, since they prefer internal funds over external funds. However, if a ?rm is growing, it increases its debt position because it needs more funds, and our results show that this happens according to the pecking-order theory. Furthermore, pro?ts particularly affect short-term debt, whereas asset growth only affects long-term debt. Therefore, this suggests that, after internal funds, long-term debt comes next in the pecking order for SMEs. Shortterm debt is more expensive and can be amortized easily. Our results also indicate that both inter- and intraindustry heterogeneity are important drivers of capital structure, in line with both pecking-order and trade-off theories of capital structure. Our analysis of inter-industry effects reveals that different industries exhibit different degrees of leverage, in line with the trade-off theory. The impact of ?rm characteristics for each industry is mostly in line with the peckingorder theory and this for almost all industries. Our intra-industry results indicate that ?rms display

123

H. Degryse et al.

considerable heterogeneity after controlling for ?rm characteristics. This suggests that the degree of industry competition, the degree of agency con?icts, and the heterogeneity in employed technology are also important drivers of capital structure. A more detailed investigation of this is left for future research.
Acknowledgements We thank two anonymous referees, Martin Brown, Fabiana Penas, Bert Sikken, and Willem van der Velden, for comments that improved the paper.

Open Access This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

Appendix See Table 6.

Table 6 Results from the industry-speci?c regressions Total debt Estimate Panel A: Manufacturing Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation R2 (?rm ?xed eff.) R2 (pooled) FF test (p-value) Panel B: Construction Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation R2 (?rm ?xed eff.) R2 (pooled) FF test (p-value) 0.080* 0.429* 0.449* 0.275* -0.138* 0.027* -0.021 0.288* 0.946 0.231 6.474 (0.000) 0.0243 0.0410 0.1724 0.0333 0.0277 0.0093 0.0157 0.1303 0.092* 0.529* 0.557* 0.055* -0.023 0.026* -0.016 0.081 0.959 0.410 6.498 (0.000) 0.0217 0.0374 0.1666 0.0254 0.0182 0.0091 0.0146 0.1094 -0.012 -0.100* -0.108 0.220* -0.115* 0.001 -0.005 0.208* 0.915 0.104 4.639 (0.000) 0.0144 0.0304 0.0850 0.0273 0.0205 0.0065 0.0140 0.0986 0.070* 0.390* 0.365 0.333 -0.186* 0.016* -0.003 0.112 0.948 0.172 8.346 (0.000) 0.0226 0.0508 0.2019 0.0479 0.0418 0.0096 0.0173 0.1487 0.096* 0.519* 0.418* 0.080 -0.069* 0.015 -0.013 -0.011 0.960 0.402 7.786 (0.000) 0.0206 0.0458 0.1855 0.0417 0.0320 0.0093 0.0148 0.1347 -0.026 -0.129* -0.115 0.252* -0.117* 0.001 0.011 0.123 0.924 0.150 5.764 (0.000) 0.0185 0.0395 0.1266 0.0386 0.0308 0.0080 0.0137 0.1253 Std. error Long-term debt Estimate Std. error Short-term debt Estimate Std. error

123

Firm and industry characteristics Table 6 continued Total debt Estimate Panel C: Wholesale trade Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation R2 (?rm ?xed eff.) R2 (pooled) FF test (p-value) Panel D: Retail trade food Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation R2 (?rm ?xed eff.) R2 (pooled) FF test (P-value) Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation R2 (?rm ?xed eff.) R2 (pooled) FF test (p-value) 0.062 0.445* 0.290 0.465* -0.151 0.026 0.004 0.189 0.953 0.261 8.069 0.074* 0.312* 0.321* 0.430 -0.249* 0.027* -0.016 0.396* 0.955 0.140 10.570 (0.000) (0.000) 0.0188 0.0424 0.1496 0.0454 0.0398 0.0072 0.0156 0.1364 0.0556 0.0813 0.2517 0.1186 0.0791 0.0224 0.0377 0.2495 0.123* 0.494* 0.433 0.186 -0.042 0.012 0.015 0.069 0.955 0.384 6.933 0.083* 0.495* 0.521* 0.154* -0.095* 0.028* -0.018 0.188 0.962 0.359 9.267 (0.000) (0.000) 0.0177 0.0400 0.1570 0.0386 0.0297 0.0073 0.0152 0.1342 0.0510 0.0914 0.2534 0.1043 0.0646 0.0244 0.0374 0.2567 -0.061* -0.049 -0.142 0.279* -0.109* 0.014 -0.011 0.120 0.908 0.093 4.819 -0.009 -0.182* -0.200 0.277* -0.154* 0.000 0.001 0.208 0.930 0.148 6.579 (0.000) (0.000) 0.0144 0.0306 0.1054 0.0395 0.0299 0.0057 0.0144 0.1142 0.0290 0.0647 0.1153 0.0900 0.0551 0.0164 0.0316 0.1686 0.066* 0.344* 0.491* 0.372* -0.241* 0.019* 0.006 0.154 0.950 0.141 8.857 (0.000) 0.0231 0.0579 0.1234 0.0403 0.0353 0.0088 0.0162 0.1517 0.086* 0.494* 0.689* 0.069* -0.092* 0.011 0.010 0.066 0.957 0.398 7.234 (0.000) 0.0181 0.0487 0.1327 0.0309 0.0281 0.0081 0.0149 0.1197 -0.019 -0.150* -0.198 0.303* -0.149* 0.008 -0.003 0.088 0.929 0.141 6.108 (0.000) 0.0161 0.0387 0.1242 0.0377 0.0325 0.0076 0.0144 0.1313 Std. error Long-term debt Estimate Std. error Short-term debt Estimate Std. error

Panel E: Retail trade nonfood

123

H. Degryse et al. Table 6 continued Total debt Estimate Panel F: Catering and leisure Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation R2 (?rm ?xed eff.) R2 (pooled) FF test (p-value) Panel G: Transport Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation R2 (?rm ?xed eff.) R2 (pooled) FF test (p-value) Panel H: Services Size (log) Tangible assets Intangible assets Net debtors ROA Growth (assets) Tax rate Depreciation R2 (?rm ?xed eff.) R2 (pooled) FF test (p-value) 0.057* 0.390* 0.279* 0.258* -0.098* 0.022* -0.026 0.104 0.954 0.176 7.134 (0.000) 0.0193 0.0444 0.1170 0.0442 0.0352 0.0075 0.0149 0.1111 0.069* 0.471* 0.355* 0.046 -0.042 0.015* -0.025 -0.090 0.967 0.402 7.194 (0.000) 0.0171 0.0424 0.1026 0.0331 0.0243 0.0069 0.0140 0.0937 -0.011 -0.081* -0.076 0.212* -0.056* 0.007 -0.001 0.193* 0.945 0.192 5.801 (0.000) 0.0143 0.0281 0.0675 0.0360 0.0273 0.0056 0.0123 0.0848 0.092* 0.393* 0.252 0.383* -0.248* 0.024* -0.037 0.310* 0.959 0.264 10.568 (0.000) 0.0213 0.0666 0.3891 0.0714 0.0638 0.0083 0.0253 0.1374 0.114* 0.511* 0.108 0.151 -0.105* 0.017 -0.040 0.111 0.958 0.419 8.083 (0.000) 0.0228 0.0707 0.3385 0.0838 0.0531 0.0097 0.0251 0.1634 -0.023 -0.118* 0.143 0.233* -0.143* 0.007 0.003 0.198 0.916 0.191 5.378 (0.000) 0.0175 0.0444 0.2491 0.0514 0.0433 0.0080 0.0213 0.1136 0.118* 0.339* 0.353* 0.344* -0.128* 0.014 -0.029 0.317* 0.963 0.228 9.731 (0.000) 0.0283 0.0537 0.1716 0.1001 0.0493 0.0085 0.0261 0.1104 0.164* 0.350* 0.426* 0.252* -0.067* 0.016* -0.028 0.232* 0.968 0.335 9.869 (0.000) 0.0266 0.0513 0.1502 0.0975 0.0339 0.0079 0.0238 0.0870 -0.046* -0.011 -0.073 0.092 -0.061 -0.002 -0.001 0.085 0.932 0.117 5.927 (0.000) 0.0163 0.0343 0.1012 0.0771 0.0390 0.0048 0.0186 0.0942 Std. error Long-term debt Estimate Std. error Short-term debt Estimate Std. error

Notes: * Signi?cant at the 5% level

References
Ang, J. S. (1992). On the theory of ?nance for privately held ?rms. Journal of Small Business Finance, 1(3), 185–203. Baker, M., & Wurgler, J. (2002). Market timing and capital structure. Journal of Finance, 57(1), 1–32.

Balakrishnan, S., & Fox, I. (1993). Asset speci?city, ?rm heterogeneity and capital structure. Strategic Management Journal, 14(1), 3–16. Berger, A. N., & Udell, G. F. (2006). A more conceptual framework for SME ?nance. Journal of Banking and Finance, 30(11), 2945–2966.

123

Firm and industry characteristics Bradley, M., Jarrell, G., & Kim, E. (1984). On the existence of an optimal capital structure: Theory and evidence. Journal of Finance, 39(3), 857–878. Brennan, M. J., & Schwartz, E. S. (1978). Corporate income taxes, valuation and the problem of optimal capital structure. Journal of Business, 51, 103–114. Cassar, G., & Holmes, S. (2003). Capital structure and ?nancing of SMEs: Australian evidence. Accounting and Finance, 43(2), 123–147. Cetorelli, N., & Gambera, M. (2001). Banking market structure, ?nancial dependence and growth: International evidence from industry. Journal of Finance, 56(2), 617–648. Chen, L. H., Lensink, R., & Sterken, E. (1999). The determinants of capital structure: Evidence from Dutch panel data, working paper University of Groningen. Cole, R. (2008). What do we know about the capital structure of privately held ?rms? Evidence from the Surveys of Small Business Finance. http://www.sba.gov/advo/ research/rs324tot.pdf. De Jong, A. (1999). An empirical analysis of capital structure decisions in Dutch ?rms. PhD. dissertation. CentER, Tilburg University. De Jong, A., Kabir, R., & Nguyen, T. T. (2008). Capital structure around the world: The roles of ?rm- and countryspeci?c determinants. Journal of Banking and Finance, 32, 1954–1969. DeAngelo, H., & Masulis, R. W. (1980). Optimal capital structure under corporate and personal taxation. Journal of Financial Economics, 8, 3–29. Fama, E. F. (1980). Agency problems and the theory of the ?rm. Journal of Political Economy, 88, 288–307. Fama, E. F., & French, K. R. (2002). Testing trade-off and pecking order predictions about dividends and debt. Review of Financial Studies, 15(1), 1–33. Frank, M., & Goyal, V. K. (2008). Trade-off and pecking order theories of debt. In E. Eckbo (Ed.), The handbook of empirical corporate ?nance. North Holland: Elsevier. Graham, J. R., & Harvey, C. R. (2001). The theory and practice of corporate ?nance: Evidence from the ?eld. Journal of Financial Economics, 60, 187–243. Hall, G. C., Hutchinson, P. J., & Nicos, M. (2004). Determinants of the capital Structures of European SMEs. Journal of Business Finance and Accounting, 31(5&6), 711–728. Harris, M., & Raviv, A. (1991). The theory of capital structure. Journal of Finance, 46(1), 297–355. Heyman, D., Deloof, M., & Ooghe, H. (2008). The ?nancial structure of privately held Belgian ?rms. Small Business Economics, 30, 301–313. Jensen, M. C. (1986). Agency costs of free cash ?ow, corporate ?nance and takeovers. The American Economic Review, Papers and Proceedings, 76, 323–329. Jordan, J., Lowe, J., & Taylor, P. (1998). Strategy and ?nancial policy in UK small ?rms. Journal of Business Finance and Accounting, 25(1&2), 1–27. Leibenstein, H. (1966). Allocative Ef?ciency vs. ‘X-Ef?ciency’. The American Economic Review, 56, 392–415. MacKay, P., & Phillips, G. (2005). How does industry affect ?rm ?nancial structure? The Review of Financial Studies, 18(4), 1433–1466. Maksimovic, V., & Zechner, J. (1991). Debt, agency costs and industry equilibrium. Journal of Finance, 54, 1619–1645. Michaelas, N., Chittenden, F., & Poutziouris, P. (1999). Financial policy and capital structure choice in U.K. SMEs: Empirical evidence from company panel data. Small Business Economics, 12(2), 113–130. Mitchel, K. (1991). The call, sinking fund and term-to-maturity features of corporate bonds: An empirical investigation. Journal of Financial and Quantitative Analysis, 26, 201– 222. Modigliani, F., & Miller, M. H. (1958). The cost of capital, corporation ?nance and the theory of investment. The American Economic Review, 48(3), 261–297. Modigliani, F., & Miller, M. H. (1963). Corporate income taxes and the cost of capital: A correction. The American Economic Review, 53(3), 433–443. Myers, S. C. (1977). Determinants of corporate borrowing. Journal of Financial Economics, 5, 147–175. Myers, S. C. (1984). The capital structure puzzle. Journal of Finance, 39, 575–592. Myers, S. C., & Majluf, N. S. (1984). Corporate ?nancing and investment decisions when ?rms have information that investors do not have. Journal of Financial Economics, 13, 187–221. Ortiz-Molina, H., & Penas, M. (2006). Lending to small business: The role of loan maturity in addressing information problems. Small Business Economics, 30, 361– 383. Sogorb-Mira, F. (2005). How SME uniqueness affects capital structure: Evidence from a 1994–1998 Spanish data panel. Small Business Economics, 25(5), 447–457. Stulz, R. (1988). Managerial control of voting rights: Financing policies and the market for corporate control. Journal of Financial Economics, 20, 25–54. Titman, S., & Wessels, R. (1988). The determinants of capital structure choice. Journal of Finance, 43(1), 1–19. Van der Wijst, N., & Thurik, R. (1993). Determinants of small ?rm debt ratios—an analysis of retail panel data. Small Business Economics, 5(1), 55–65. Van Dijk, R. (1997). Corporate ?nance and equity investment: Panel data analyses. PhD. dissertation. Tinbergen Institute Research Series no. 168, Erasmus University Rotterdam. Verbeek, M. (2008). A guide to modern econometrics (3rd ed.). Chichester: Wiley. Vos, E., Yeh, A., Carter, S., & Tagg, S. (2007). The happy story of small business ?nancing. Journal of Banking and Finance, 31, 2648–2672.

123




友情链接: