Printer Friendly

Trade credits and bank loans for SMEs in the textile sector.

Introduction

The understanding of financing decisions in emerging markets is far less developed than in developed economies. Nevertheless, access to external finance is of particular importance for emerging markets since it is one of the key factors that contributes to the healthy development of real sector firms in these countries. (1) For firms in emerging markets, particularly small- and medium-sized enterprises (SMEs), access to bank loans is limited compared to firms in developed countries. Therefore, SMEs operating in emerging markets rely on trade credit to fulfill their need for additional finance. Moreover, Demirguc-Kunt and Maksimovic (2001) emphasize that trade credit is an answer for markets where alternative sources of finance are mostly unavailable and the development of the stock and bond market is modest.

However, there is an ongoing debate in the current finance literature if trade credit is a complement or substitute to bank credit. (2) In other words, does trade credit facilitate the access to bank credit or does it replace bank credit as quenching the need for external finance?

The purpose of this paper is to demonstrate the relationship between trade credits and bank loans through investigating SMEs operating in the textile sector in Turkey. The contribution of our work to the debate in the literature is at least threefold. First of all, one side of the debate focuses on the asymmetric information concept. Specifically, suppliers are thought to have a cost advantage over banks in acquiring information about the financial health of buyers. Small firms are known to lack the long-term relationship with providers of external finance and they are opaque in terms of their investment projects compared with large firms (Scherr and Hulburt, 2001). After studying small US firms, Petersen and Rajan (1995) reason that small firms consider borrowing from suppliers as a financing "lender of last resort" due to their limited access to bank loans. Moreover, Arslan (2003) indicates that SMEs in Turkey are more likely to be credit rationed than large firms due to the asymmetric information problem. Thus, focusing on small firms provides a better insight on the role of asymmetric information in determining the relationship between trade credits and bank loans. Besides, given that SMEs are firms within a financial growth cycle, as expressed by Berger and Udell (1998), they need more financial resources to fund their growth. (3)

Fisman and Love (2003) emphasize the industry-specific element to trade credit incentives. Therefore in this study, we concentrate on the textile sector to investigate the bank loans-trade credits relationship. The reason we only investigate SMEs in the textile industry is due to the fact that the representative sector is of particular importance for Turkish industrial production. Specifically, the Turkish textile sector accounts for almost 20% of Turkish Gross Domestic Product (GDP), 21% of total domestic industrial employment and 38% of total Turkish export earnings. As of 2002, Turkey is the second biggest supplier of textiles to the European Union after China. (4) Bagdadioglu and Ulucan (2005) indicate that almost 96% of textile firms in Turkey are classified as an SME and their biggest corporate concern is obtaining investment capital. The Halkbank Report (2003) shows that bank loan usage by Turkish SME textile firms only reach 10% of the level of that of SME textile firms in developed countries. Similarly, it is shown by Fisman and Love (2003) that trade credit usage in the textile sector in the US is above the median usage of the entire industries in the country. To sum up, investigating the relationship between trade credits and bank loans for such a strategic industry of Turkey is a crucial issue to keep track of their healthy financial stability.

Furthermore, we also investigate the role of financial crises on the relationship between trade credits and bank loans. Meltzer (1960) found that during the mid-1950's money tightening in the US, firms substituted bank loans with trade credits. Aligning with these findings, empirical results of Nilsen (1994) show that small firms react to economic recession by resorting to trade credits. Emerging markets are characterized as having large economic fluctuations in short-time intervals. Love, Preve and Sarria-Allende (2007) study the effect of financial crises on trade credit for a sample of 890 firms in six emerging economies. The authors found that the provision of trade credit increases right after a crisis but contracts in a crisis period because of a decline in trade credit due to a bank credit crunch, which the authors call the "redistribution view." In contrast, Miwa and Ramseyer (2006) find that small firms react to economic recessions by increasing their usage of trade credit. As for Turkey, the textile sector was the one most adversely affected among the industries in Turkey by the 1997 Asia crisis and 1998 Russian debt default due to the "contagion effect" (5) and the 2000-01 internal financial crises of Turkey. The largest number of closed factories and lost jobs in Turkey belonged to the textile sector owing to the indicated external and domestic crises (Halkbank, 2003). Consequently, investigating how financial crises influence the relationship between trade credits and banks loans for SMEs in the Turkish textile sector can help to find out the pattern of the change in the interaction between these concepts during economic constriction and expansion periods.

Last but not least, we investigate the accounting data of on average 493 SME textile firms operating in the period beginning in 1997 and ending in 2005 with a total of 1973 observations in our analyses. Thus, we provide an outstandingly large number of firm level data for a single sector in an emerging market. The source of our data is the database of The Central Bank of Turkey (TCMB--Turkiye Cumhuriyet Merkez Bankasi). If we had been obliged to work with textile firms that both fit into the SME definition and were publicly traded on the Istanbul Stock Exchange (ISE), we would have ended up with only 12 firms on average within the period of our analyses. Since the sample is made up of small sized firms, which make up more than 96% of the total textile industry in Turkey, our study expands the horizon of investigation on emerging markets by not being restricted only to publicly traded firms on stock exchanges.

Our results show that there is a non-linear relationship between trade credits and bank loans for textile SMEs in Turkey. In other words, trade credits initially encourage bank credits by complementing them. However, after a certain level of trade credit usage, trade credits curb bank loans and act as a substitute to them. Moreover, we do not detect that outside and domestic financial crises alter the pattern of the non-linear relationship. Finally, we also study the impact of specific firm characteristics on obtaining bank loans. We find a positive relationship between bank loans and firm size, firm age and having foreign sales. However, there is a modest negative relationship between bank loans and profitability. Nevertheless, the impact of interest coverage ratios on bank loans is not statistically different from zero. Finally, we find a slightly fast speed of adjustment to target bank loans for the firms.

The remainder of the paper is organized as follows. In section 1 we introduce the background of our empirical hypotheses. Section 2 presents the construction of the sample and the methodology. Section 3 includes the results of our analysis with discussions. Finally, section 4 includes concluding remarks.

Background

Trade credit arises when a customer is permitted by a supplier to delay payments for goods already delivered. As indicated in Cunat (2007), although its implicit interest rates are commonly very high as compared with the rates on bank loans, (6) trade credit is widely used in firms thanks to its specific advantages. Firstly, firms incur lower transaction costs from obtaining a trade credit than when they attempt to access institutional loans. Furthermore, it is less costly for firms to have their suppliers postpone trade credit payments than renegotiating bank loans. Finally, being in the same value chain, there is a mutual dependence between suppliers and firms. Suppliers may finance the growth of small customers to assure growth of their own sales and eventually capture future profitable business from the firms. Owing to the stronger communication between suppliers and the firms, trade credits allow the firms to match the timing of cash outlays and cash receipts from sales.

However, how do trade credits interact with bank loans? In other words, the question focuses on whether trade credit facilitates obtaining external finance from banks or alternatively hinders the access to bank loans by substituting for them. The findings in the literature are contradictory.

On the one hand, by approaching the matter from the perspective of the asymmetric information problem, a strand of the literature finds that the relationship is complementary. Moral hazard problems, which refer to the case when credit is diverted to non-value maximizing projects by borrowers, emanate from the asymmetric information between the creditors and borrowing firms. Trade credit is believed to decrease the moral hazard problem because, in case of a default, suppliers are entitled to cease the supply of goods to the firms. Besides, losing the current supplier is more costly in the presence of some kind of product specificity--a particularly common issue for textile firms--or when there is a certain link between the suppliers and customers of end products, which is an issue more common for emerging markets. In accord with this, Biais and Gollier (1997) show that trade credit increases the availability of bank credit, especially for small firms that are typically affected by asymmetric information problems. Cook (1999) also finds evidence for the facilitating role of trade credit, through diminishing the adverse selection problem. The author shows that trade credit provides a positive signal for firms, hence increasing the probability of acquiring bank credit. Love, Preve and Sarria-Allende (2007) confirm the complementary relationship for six emerging market countries from another aspect called the "redistribution view" through showing that a credit crunch affecting financial lenders also affects trade credit lenders. Consequently, on the one hand, we hypothesize that trade credit boosts the usage of bank credit for Turkish SMEs in the textile sector.

On the other hand, the prediction that trade credit is a substitute for insufficient bank financing is empirically confirmed by Atanasova and Wilson (2004) and Petersen and Rajan (1994, 1995) and Marotta (1997). The authors observe that credit-constrained firms make a larger use of trade credit when credit conditions are tighter. Moreover, Petersen and Rajan (1997) stress that excessive trade credit could give a negative signal to a bank that a firm cannot obtain credit at any bank. In accordance to this finding, Wilson and Summers (2002) provide evidence that to assure their financial health by maintaining a strong value chain, suppliers provide credit for firms that can no longer rely on banks for additional credit. Finally, another approval is reached by Huyghebaert, Van de Gucht and Van Hulle (2007)--in their study of business start-ups--who indicate that suppliers are more willing to renegotiate the outstanding debt or grant additional debt, whereas banks are more likely to liquidate borrowers upon default. Therefore, on the other hand it is also plausible to hypothesize a negative relationship between trade credit and bank credit.

We also consider the impacts of firm specific characteristics on bank loans as control variables in our analyses. Firstly, we consider the profitability of firms. According to the pecking order theory of Myers (1984), the more firms are profitable, the less external finance they resort to. Yet, the trade-off theory of capital structure refrains adopting an expectation as to the effect of profitability on obtaining an institutional loan. Furthermore, McConnell and Petit (1984) indicate for small firms that since the cost of debt decreases for firms as their size grow, smaller firms tend to use less bank debts. For this reason, size is the third firm specific characteristic that we include in our tests. Moreover, Milne (1991) and Guariglia (1999) express that if a firm has a large interest cover ratio then it is classified as credit constrained. Consequently, Mateut and Mizen (2002) highlight that a firm with higher interest cover ratio uses larger trade credits and less bank loans. Hence, our expectation is to reach a negative relationship between interest cover ratio and bank loans. Gonenc and Arslan (2003) study Turkish publicly traded firms and find that, when compared with firms having no foreign sales, those with foreign sales have a higher level of short term borrowing in their liabilities. In accord with this finding, we expect a positive relationship between banks loans and foreign sales. Finally, we control for the association between firm age and bank loans since Petersen and Rajan (1994) explain that young firms tend to use more external finance given that they are in the growth stage of their life cycle.

The Case for Turkey

Zhu and Xu (2007) and Demirguc-Kunt and Maksimovic (2001) highlight that a firm's usage of bank credit relative to trade credit is higher in countries that have efficient legal systems and large banks. Aligning with this, Fisman and Love (2003) find that firms in countries with less developed financial markets and weak legal enforcement appear to substitute bank loans with informal credit provided by their suppliers. A typical feature of an emerging market is poorly functioning financial markets with limited financial instruments. In accord with this, being an emerging market, in Turkey a private bond market is not available and company openness is far below 25% on average (Erol, 2005). Moreover, the legal infrastructure in Turkey relatively exacerbates the asymmetric information between creditors and borrowers compared with that of developed Anglo-Saxon states (La Porta, et al., 1998).

Turkey faced one of the most important events in its economic history within the period of our analysis, namely 1997-2005. The adverse influences of the Asian meltdown in 1997 and the Russian debt default in 1998 spread to real sector firms in Turkey and the textile sector was the worst affected among them. In 1999, Turkey signed up an economic stabilization agreement with the IMF. Therefore, Turkey benefited from a low level of inflation, interest and fixed exchange rates through the years 1999-2000. Henceforth, the years 1999-2000 are referred to as the pre-crisis period. However, due to poor public finance management and a fragile banking system, Turkey was struck by a financial crisis that has had a financial and economic impact in the years 2001-02. Henceforth, these years are referred to as the crisis period. These crises led the Turkish government to adopt new alternative economic policies under another IMF package and in turn during the following years, namely 2003, 2004 and 2005, the financial and economic indicators became normal again.

Data and Methodology

We use a dataset of Turkish textile SME firms for the period beginning in 1997 and ending in 2005. Our sample is made up of 1973 observations and on average 493 firms. The data for this paper is obtained from the TCMB, which annually collects audited balance sheet and income statement data from the entire Turkish real sector firms so as to prepare internal analyses on industries. Firms voluntarily submit their data to the TCMB. In order to ensure reliability, the data is internally filtered by the Department of Statistics in the TCMB.

There is no consensus for the institutional definition of SMEs in Turkey. We based ourselves on the definition in Hall, Hutchinson and Michaelas (2004). Therefore, for our analyses, we picked the textile firms that have at least 10 employees and not more than 200. We were aware that the exclusion of micro firms, which employ between one and nine employees, and other firms with 200-250 employees could lead to potential biases, so we made robustness checks. However, the results of our robustness checks including both the micro firms and the ones with over 200 employees are very similar to those obtained in this paper. This result is not contrary to our expectations since the average number of employees in the Turkish textile sector for the years 1997-2005 is 105 (Turkish Statistical Institute, 2006). The results on our robustness checks are not reported for brevity and are available upon request.

We measured bank loan as the ratio of short-term bank loan to total assets. We did not use total bank loans because Turkish SMEs predominantly use short-term bank loans. Moreover, the ratio of short-term bank loans to the total bank loans for Turkish SMEs is 95%. (7) This also aligns with the indication of Chittenden and Bragg (1997) in their study of firms in UK, Germany and France that long term loans constitute a very small percentage of the liabilities of small firms. Furthermore, trade credit was measured as the ratio of accounts payable to total assets. Size was one of the control variables and it was measured as the inflation-adjusted natural logarithm of total assets and age of firms represented the number of years a firm had been operating. (8) Moreover, profitability was calculated as the ratio of earnings before interests and taxes to total assets whereas interest coverage ratio was calculated as the ratio of financial expenses to net sales. Finally, foreign sales was measured as the ratio of foreign sales to total sales.

We used the Generalized Method of Moments (Arrelano and Bond, 1991) methodology to estimate the relationship between bank loans and trade credits. Within this methodology, which avoids potential endogeneity problems, it is essential to take the adjustment process into account while developing the dynamic model. Therefore our estimation model was established within the following process:

Equation (1) is the static model:

[[gamma].sub.it] = [beta][x.sub.it] + [[xi].sub.n] + [[gamma].sub.t] + [u.sub.it] (1)

Where [beta] is a kx1 vector of constants, [x.sub.it] is a kx1 vector of explanatory variables, [[xi].sub.i] is the firm effect, which is assumed to be constant for firm i over t, [[gamma].sub.t] is time effect assumed constant for given t over i, is error term, and t=1,.... T and i=1, ...., N.

Let [y.sup.*] it be the target bank loans estimated from the equation (1), then

[y.sub.it] [-y.sub.it-1] = [lambda]([y.sup.*.sub.it-yit-1]) (2)

The coefficient is [lambda] is negatively associated with adjustment costs and has a value between 0 and 1. When [lambda]=1 then [y.sub.it-1], therefore adjustment occurs without any frictions. In other words, Ozkan (2001) states that this value approaches to 1 as the costs of being in disequilibrium exceed the adjustment costs. However, when [lambda] = 0 then = [y.sub.it] = [y.sup.*.sub.it] and in this case firms suffer less from disequilibrium and hence their slower adjustment process. After considering these issues, equation (2) was transformed into the following equation, which was estimated within our dynamic analysis.

[y.sub.it] = (1-[lambda])[y.sub.it-1] + [beta][x.sub.it] + [[xi].sub.i] + [[gamma].sub.t] + [u.sub.it]

Results

Table 1 provides descriptive statistics of the variables we used in our analyses. Panel A shows the descriptive of the whole period beginning in 1997 and ending in 2005. Panel B includes the 1997-98 period that reflects the outside shocks of the 1997 Asian crisis and the 1998 Russian crisis. Panel C exhibits the results for 1999-2000 pre-crisis period of Turkey, Panel D shows those for the 2001-02 financial crisis period and, finally, Panel E demonstrates the post-crisis period. The mean value of bank loans is 20.9% for the whole period and the variable displays a declining pattern when examining its mean value from Panel B to Panel E. It is remarkable that the level of bank loans of Turkish textile SMEs during the post-crisis period was less than that of the bank loans during the crisis period. The mean value of trade credit in the panels shows that its usage was higher than that of bank loans for the entire period except the pre-crisis period of 1999-2000, in which interest rates had abruptly and dramatically fell and loan terms became less strict. Nevertheless, the level of trade credits exceeded the level of bank loans during the post-crisis period. Our results show that the mean value of size does not show much difference during different periods. The most interesting result derived from the tables is that in the crisis period of 2001-02, the mean (median) value of profitability was 7.1% (4.4%) while the mean (median) value of the variable was 1.4% (1.8%) in the post-crisis period. Results in the table explain that the ratio of foreign sales to total sales was 37.7% on average. Another noticeable result in the panel is that the mean value of foreign sales was 40% during the crisis period while the mean value of the variable fell to 30.45% during the post-crisis period. These results on both foreign sales and profitability can be attributed to the devaluation of the domestic currency during the crisis period and the abrupt appreciation of the freely floating domestic currency during the post-crisis period. In contrast to the suggestion, the highest mean value of foreign sales ratio, which was 43.2% among the different periods, was reached during the pre-crisis period. During the pre-crisis period, Turkey adopted a fixed exchange rate regime, which enabled stability in prices of foreign sales, thus encouraging domestic exports. Finally, the firms in our sample were fairly young since the mean (median) age of our sample was 18.48 (18).

We report the Pearson correlation matrix of our variables in Table 2. Our results in the correlation matrix show that correlations between the variables in our analyses did not exceed 50%. It should be noted from the table that the correlation coefficient between trade credits and bank loans was positive yet slim.

Univariate Results

Our objective in this part was to show the pattern of the relationship between trade credits and bank loans. Therefore Figure 1 demonstrates the mean values of the trade credits and banks loans beginning in 1997 and ending in 2005. The lines in the figure show that, except for the year 1997, the level of trade credit was always above that of bank loan. Aligning with our expectations, the gap between trade credit and bank loan was the biggest during the crisis period and it narrowed in the post-crisis period. However, the gap was narrower in 1997 and 1998, in which the Asian and Russian crises took place, than the pre-crisis period of 1999-2000. After 1998, trade credits and bank loans slightly moved in the same direction and this gives the first clue for the "complementing argument," which stipulates that usage of trade credit facilitates access to bank credit.

[FIGURE 1 OMITTED]

We also divided the observations of trade credits into four separate quartiles and report respective mean, median and standard deviation values for the usage of bank loans corresponding to the each of the quartiles. The objective of this task was to examine whether the usage of bank loans increase or decrease in accord with increase or decrease in usage of trade credits. The results for each of the years are reported separately in order to present the trend in the relationship throughout the different periods. We used t-statistics to verify the statistical significant difference of the mean values of the first and the last quartiles.

Our univariate results illustrate that the relationship between trade credits and bank loans is non-monotonic. Specifically, in the first quartile bank loans initially increased. Yet, after a certain level of trade credit usage, which corresponds either to the second or third quartiles, bank loans started to decrease. The differences between the mean value of the lowest and the highest quartiles are statistically significant in the years 1997, 1998, 2000, 2004 and 2005. The quartile analysis for the year 2005 is the sole exception to the general pattern since bank loans were found to increase monotonically with trade credits in that year. The univariate results obtained in Table 3 justify the need to make further analysis through multivariate analyses to control for the non-linear relationship between trade credits and bank loans.

Panel Results

Table 4 presents the results of our dynamic analyses. It should also be noted that we also took the squared term of the trade credit as an explanatory variable in order to test the non-monotonic relationship between bank loans and trade credits, which was captured during our univariate analysis.

Our results show for both models that while the trade credit is positively statistically significant, the squared term of trade credit, however, is negatively statistically significant. Moreover, both the economic and statistical significance levels of the variables are slightly close to each other. Our results suggest that at low levels of trade credits, banks are encouraged to supply loans to Turkish textile SMEs. However, after a certain level of trade credits, the supply of bank loans becomes adversely related with the trade credits.

Bearing in mind that the textile industry in Turkey is one of the worst affected by exogenous macroeconomic shocks, we focused on whether the relationship between trade credits and bank loans differed during the crisis periods of Turkey. Therefore, we interacted both the trade credit and the square of trade credit variable with the crisis dummy. Crisis dummy takes the value of unity if the observation year is 1997, 1998, 2001 and 2002 and zero otherwise. Similar to our previous findings, the interaction of trade credit variable with the crisis dummy is significantly and positively related to bank loans. Moreover, the interaction of squared term of trade credit variable with the crisis dummy is again significantly negatively associated with bank loans. Nevertheless, we observed a slight decline of economic and statistical significance of both of the variables during the crises years.

In Model 2, we incorporated control variables explaining the bank loan usage of Turkish textile SMEs. As expected, the size of firms is positively related with bank loans, but with a very low economic and statistical significance. Furthermore, we found a very modest negative relationship between the profitability of firms and the usage of bank loans given that both the economic and statistical significance of the profitable variable is at low levels. Nevertheless, our results demonstrate that the impact of interest coverage ratio on bank loans is not statistically significantly different from zero. Age of firms is found to be strongly positively associated with bank loans. Lastly, we found that foreign sales are positively related with bank loans despite a low economic significance.

Finally, the results on the lagged bank loan variable display the speed of adjustment of Turkish textile SMEs to their target bank loans. The speed of adjustment for the firms is slightly closer to one since it is around 0.60. Therefore, we deduce for the Turkish textile SMEs that the cost of being in disequilibrium is slightly greater than the costs of adjustment to the target bank loan level.

Discussions

Our results provide a mixed support for the two distinct arguments regarding the relationship between bank loans and trade credits. In other words, at low levels of trade credit our results confirm our hypothesis on the complementary argument that was also supported by the findings of Biais and Gollier (1997) and Love, Preve and Sarria-Allende (2007). Given that trade credit signals that a firm is creditworthy, initially, the eagerness of banks to lend Turkish textile SMEs increases. At this stage the firms are relatively at the advantageous situation for the supply of external funds since both trade credits and bank loans go hand in hand. However, at the higher levels of usage of trade credits its relationship with bank loan is found to be confirming the substitution argument of Wilson and Summers (2002) and Atanasova and Wilson (2004). We understand from our results that at a higher level of trade credits, banks begin the credit rationing of textile SMEs. This is caused by the fear of banks that the firms become more likely to be unable to service their further financial obligations considering the burden caused by their higher level of trade credit usage. Interestingly, the non-linear pattern of the relationship between bank loans and trade credits is not affected by crises. Put differently, the attitude of suppliers of bank loans to the level of trade credits used by the firms does not change during the economic upswings or downturns. Therefore, our findings for crisis periods contradict those of Meltzer (1960) and Nilsen (1994).

Finally, our results lend a weak support to the prediction that larger textile SMEs are more eligible for bank loans due to the decreased asymmetric information between their financial quality and supplier of funds. For this reason our results fairly confirm those of McConnell and Petit (1984) on the impact of size on acquiring bank loans. Moreover, given that the profitability variable is only just negatively related with bank loans, our results provide a modest support to the pecking order hypothesis, which suggests that profitability of firms diminish their resorting to bank loans. However, our results regarding interest coverage variable do not go hand in hand with those regarding profitability, since the former is found to be statistically insignificant. The results on the impact of firm age on bank loan usage are the contrary to the findings of Petersen and Rajan (1994), which indicate that young firms tend to use more external finance given that they are in the growth stage of their life cycle. Instead, we reason that older Turkish textile SMEs are more attractive for the supply of bank loans since they are more transparent and, hence, better known by the financial institutions given that they have a longer track of financial records and activities to be judged upon. In other words, in the market, older firms have an established reputation, which facilitates their access to bank loans. Finally, our findings on the foreign sales of Turkish textile SMEs are in agreement with those of Gonenc and Arslan (2003) regardiing Turkish publicly traded firms. Specifically, textile SMEs with export activities have a higher ability to acquire bank loans.

Conclusion

Acquiring external funds for an SME is of particular importance to textile firms in Turkey. Therefore, we studied two alternatives of external finance: trade credits and bank loans. In other words, the objective of this study was to find the relationship between trade credits and bank loans of Turkish SMEs operating in the textile sector. Moreover, the objective of the study also encompassed observing if the association between trade credits and bank loans differed during crisis periods. Our data was obtained from TCMB, hence we were not limited to the few textile SMEs that are publicly traded on the ISE. Our analyses included 1973 observations from an average 493 firms covering the period beginning in 1997 and ending in 2005. We conducted our analyses through GMM, incorporating a dummy variable to control for crises.

Our univariate analyses indicate a non-linear relationship between trade credits and bank loans. Our regression analyses also confirm that bank loans initially increase with trade credits. However, after a certain level of trade credit usage, supply of bank loans starts to decrease. Moreover, this relationship pattern is not affected by economic downturns and upswings. Therefore, at low levels of trade credit usage we support the complementing argument for Turkish textile SMEs, but at higher levels of trade credit usage we confirm the substitution argument.

We also included control variables in our analyses. We find that, on the one hand, firm size, firm age and foreign sales have a positive impact on bank loans. On the other hand, profitability has a fairly adverse influence on bank loans of Turkish textile SMEs. Given that we conducted our analyses on a partial adjustment model we find a modest speed of adjustment to target bank loans for the Turkish textile SMEs.

The main limitation of this study was focusing on only a single sector in one country. The issue of substitution or complementarity of trade credits and bank loans is important and until now there is insufficient knowledge on the relationship, especially in emerging economies. Therefore, further research could follow the same path for other sectors, and even for other countries, especially emerging economies.

Note

The earlier version of this paper was presented at the 14th annual conference of the Multinational Finance Society in July 2007 in Greece. We give our thanks to Serap Celen and Gulin Kucuay for their valuable help in managing the data in the Central Bank of Turkey. The authors are especially grateful for the insightful and detailed comments of Mehmet Baha Karan, Manuel Rocha Armada, Edgar Ortiz, Lakshman Alles, and Elisabeth O'Dowd. The usual disclaimer regarding errors applies.

Contact

For further information on this article, contact:

Ozgur Arslan, Department of Business Administration, Hacettepe University, Beytepe, Ankara 06800, Turkey

Tel.: 00 90 312 297 8700/Fax: 00 90 312 299 2055

E-mail: arslan@hacettepe.edu.tr

References

Arellano, M. and S. Bond. 1991. "Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations," Review of Economic Studies 58: 277-92.

Arslan, O. 2003. "Kucuk ve orta olcekli isletmelerde calysma sermayesi ve bazy finansal yonetim uygulamalary," C.U. Iktisadi ve Idari Bilimler Dergisi 1: 121-35.

Arslan, O., C. Florackis and A. Ozkan. 2006. "The Role of Cash Holdings in Reducing Investment-Cash Flow Sensitivity: Evidence from a Financial Crisis Period in an Emerging Market," Emerging Markets Review 7, no. 4: 320-38.

Atanasova, C.V. and N. Wilson. 2004. "Disequilibrium in the UK Corporate Loan Market," Journal of Banking & Finance 28, no. 3: 595-614.

Bagdadioglu, N. and A. Ulucan. 2005. "Financial Efficiency of the Turkish Textile and Clothing Industry Before the 2005 Trade Liberalization," Eurasian Review of Economics and Finance 1, no. 4: 1-10.

Berger, A.N. and G.F. Udell. 1998. "The Economics of Small Business Finance: The Roles of Private Equity and Debt Markets in Financial Growth Cycle," Journal of Banking and Finance 22, nos. 6-8: 613-73.

Biais, B. and C. Gollier. 1997. "Trade Credit and Credit Rationing," Review of Financial Studies 10, no. 4: 903-37.

Burkart, M. and T. Ellingsen. 2004. "In-Kind Finance: A Theory of Trade Credit," American Economic Review 94, no. 3: 569-90.

Chittenden, F. and R. Bragg. 1997. "Trade Credit, Cash-flow and SMEs in the UK, Germany and France," International Small Business Journal 16: 14-22.

Cook, L.D. 1999. "Trade Credit and Bank Finance: Financing Small Firms in Russia," Journal of Business Venturing 14, nos. 5-6: 493-518.

Cunat, V. 2007. "Trade Credit: Suppliers as Debt Collectors and Insurance Providers," Review of Financial Studies 20, no. 2: 491-527.

De Blasio, G. 2005. "Does Trade Credit Substitute Bank Credit? Evidence from Firm-level Data," Economic Notes 34, no. 1: 85-112.

Demirguc-Kunt, A. and V. Maksimovic. 2001. Firms as Financial Intermediaries: Evidence from Trade Credit Data. World Bank Working Paper.

Eiteman, D.K., A.I. Stonehill and M.H. Moffett. 2006. Multinational Business Finance. N.p.: The Addison Wesley Series in Finance.

Erol, T. 2005. "Corporate Debt & Output Pricing in Developing Countries: Industry-level Evidence from Turkey," Journal of Development Economics 76, no. 2: 503-20.

Fabbri, D. and L. Klapper. 2008. Trade Credit Supply, Market Power and the Matching of Trade Credit Terms. Policy Research Working Paper, World Bank, No 4754.

Fisman, R. and I. Love. 2003. "Trade Credit, Financial Intermediary Development, and Industry Growth," Journal of Finance 58, no. 1: 353-74.

Ge, Y. and J. Qiu. 2007. "Financial Development, Bank Discrimination and Trade Credit," Journal of Banking & Finance 31, no. 2: 513-30.

Gonenc, H. and O. Arslan. 2003. "Capital Structure of the Turkish Domestic and International Real Sector Firms," ISE Review 7, nos. 25-26: 41-63.

Guariglia, A. 1999. "The Effects of Financial Constraints on Inventory Investment: Evidence from a Panel of UK Firms," Economica 66, no. 261: 43-62.

Guney, Y. and A. Ozkan. 2005. "New Insights on the Importance of Agency Costs for Corporate Debt Maturity Decisions," Applied Financial Economics Letters 1, no. 4: 233-38.

Halkbank. 2003. Annual Report.

Hall, G.C., P.J. Hutchinson and N. Michaelas. 2004. "Determinants of the Capital Structures of European SMEs," Journal of Business Finance & Accounting 31, nos. 5-6: 711-28.

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, no. 4: 435-55.

IFM & Partners. 2004. Consolidated Report, Study on the Implications of the 2005 Trade Liberalisation in the Textile and Clothing Sector. Institut Francais de la Mode (IFM), Centre d'Etudes Prospectives et d'Informations Internationales (Paris), European Public Advisers (Brussels), MAIA (Brussels), Pole Europe Conseil (Brussels).

La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R.W. Vishny. 1998. "Law & Finance," Journal of Political Economy 106, no. 6: 1113-55.

Love, I. 2003. "Financial Development and Financing Constraints: International Evidence from the Structural Investment Model," Review of Financial Studies 16, no. 3: 765-91.

Love, I., L.A. Preve and V. Sarria-Allende. 2007. "Trade Credit and Bank Credit: Evidence from Recent Financial Crises," Journal of Financial Economics 83, no. 2: 453-69.

Marotta, G. 1997. "Does Trade Credit Redistribution Thwart Monetary Policy? Evidence from Italy," Applied Economics 29, no. 1: 1619-29.

Mateut, S. and P. Mizen. 2002. Trade Credit and Bank Lending: An Investigation into the Determinants of UK Manufacturing Firms' Cccess to Trade Credit. Nottingham, UK Experian Centre for Economic Modeling (ExCEM), University of Nottingham.

McConnell, J.J. and R.R. Petit. 1984. "Application of the Theory of Finance to Small Business Firms." Part A in P.M. Horvitz and R.R. Petit (eds.), Small Business Finance: Problems in Financing of Small Business. Greenwich, CT: JAI Press.

Meltzer, A.H. 1960. "Mercantile Credit, Monetary Policy and the Size of Firms," Review of Economics and Statistics 42, no. 4: 429-36.

Milne, A. 1991. "Financial Effects on Inventory Investment." Mimeo, London Business School.

Miwa, Y. and J.M. Ramseyer. 2006. "Japanese Industrial Finance at the Close of the 19th Century: Trade Credit and Financial Intermediation," Explorations in Economic History 43, no. 1: 94-118.

Myers, S.C. 1984. "The Capital Structure Puzzle," Journal of Finance 39, no. 3: 575-92.

Nilsen, J.H. 1994. "Trade Credit and the Bank Lending Channel of Monetary Policy Transmission." Working paper, Department of Economics, Princeton University, March.

Ozkan, A. 2001 "Determinants of Capital Structure and Adjustment to Long Run Targets: Evidence from UK Company Panel Data." Journal of Business Finance and Accounting 28: 175-98.

Petersen, M.A. and R.G. Rajan. 1994. "The Benefits of Lending Relationship: Evidence from Small Business Data," Journal of Finance 49, no. 1: 3-37.

--. 1995. "The Effect of Credit Market Competition on Lending Relationships," Quarterly Journal of Economics 110, no. 2: 407-44.

--. 1997. "Trade Credit: Theories and Evidence," The Review of Financial Studies 10, no. 3: 661-91.

Scherr, F.C., and H.M. Hulburt. 2001. "The Debt Maturity Structure of Small Firms," Financial Management 30, no. 1: 85-111.

Turkish Statistical Institute. 2006. "Statistical Indicators 1923-2005." Prime Ministry Republic of Turkey, http://www.bilkent.edu.tr/~zaim/Yillikgostergeler.pdf.

Wilson, N. and B. Summers. 2002. "Trade Credit Terms Offered by Small Firms: Survey Evidence and Empirical Analysis," Journal of Business Finance & Accounting 29, nos. 3-4: 317-51.

Zhu, T. and L.C. Xu. 2007. Formal Finance and Trade Credit During China's Transition. Policy Research Working Paper Series from the World Bank, no. 4204.

Ozgur Arslan, Department of Business Administration, Hacettepe University, Beytepe, Ankara, Turkey

Goknur Umutlu, Department of Business Administration, Hacettepe University, Beytepe, Ankara, Turkey

(1.) See; Arslan, Florackis and Ozkan (2006).

(2.) See among others; Burkart and Ellingsen (2004), De Blasio (2005), Love, Preve and Sarria-Allende (2007), Ge and Qui (2007), Fabbri and Klapper (2008).

(3.) Finally, it should be noted that Fisman and Love (2003) suggest the necessity of studying usage patterns of trade credits and bank loans of smaller firms since observation for the largest publicly traded firms in studies can hardly be generalized to the rest of the population.

(4.) For further information on the importance of the textile sector for the Turkish economy, see IFM & Partners (2004).

(5.) See; Eiteman, Stonehill and Moffett (2006).

(6.) The authors outline that the interest rate of trade credit is more expensive than that of bank credit due to its higher insurance premium and default premium.

(7.) Besides, due to the missing variables in the database, we would have lost 67% of our observations if we had attempted to include long term bank loans.

(8.) Previously we had also included the squared term of variable size and age in the regression. However, the squared terms were found to be statistically insignifant and, therefore, excluded from the original analyses. The results are available upon request.
Table 1. Descriptive Statistics for Key Variables

                                               Standard
                       Mean          Median    Deviation

Panel A: Whole period

Bank Loan              0.209          0.161      0.205
Trade Credit           0.254          0.221      0.180
Size                  14.402         14.795      2.143
Profitability          0.037          0.028      0.160
Interest Coverage     -0.079         -0.036      0.188
Foreign Sales          0.377          0.254      0.367
Age                   18.477         18.000      8.778

Panel B: 1997-98 (Outside Shocks)

Bank Loan              0.248          0.219      0.213
Trade Credit           0.263          0.227      0.186
Size                  13.136         13.122      1.051
Profitability          0.027          0.031      0.140
Interest Coverage     -0.091         -0.065      0.104
Foreign Sales          0.410          0.303      0.383
Age                   20.488         19.000      7.819

Panel C: 1999-2000 (The Pre-Crisis Period)

Bank Loan              0.289          0.251      0.259
Trade Credit           0.269          0.242      0.176
Size                  14.212         14.161      1.093
Profitability          0.037          0.036      0.167
Interest Coverage     -0.099         -0.055      0.221
Foreign Sales          0.432          0.336      0.373
Age                   20.223         19.000      8.375

Panel D: 2001-02 (The Crisis Period)

Bank Loan              0.172          0.122      0.185
Trade Credit           0.291          0.250      0.198
Size                  15.169         15.195      1.005
Profitability          0.071          0.044      0.191
Interest Coverage     -0.105         -0.055      0.156
Foreign Sales          0.400          0.276      0.379
Age                   18.777         18.000      8.923

Panel E: 2003-04-05 (The Post-Crisis Period)

Bank Loan              0.168          0.142      0.153
Trade Credit           0.210          0.178      0.151
Size                  14.654         15.876      3.198
Profitability          0.014          0.018      0.131
Interest Coverage     -0.039         -0.018      0.219
Foreign Sales          0.304          0.167      0.330
Age                   15.930         14.000      8.847

                      Minimum       Maximum

Panel A: Whole period

Bank Loan              0.000          0.991
Trade Credit           0.000          0.889
Size                   6.709         18.291
Profitability         -2.406          0.877
Interest Coverage     -5.509          0.000
Foreign Sales          0.000          1.149
Age                    0.000         82.000

Panel B: 1997-98 (Outside Shocks)

Bank Loan              0.000          0.877
Trade Credit           0.000          0.833
Size                  10.190         17.545
Profitability         -1.159          0.652
Interest Coverage     -1.047          0.000
Foreign Sales          0.000          1.032
Age                    0.000         61.000

Panel C: 1999-2000 (The Pre-Crisis Period)

Bank Loan              0.000          0.991
Trade Credit           0.000          0.846
Size                  10.497         17.967
Profitability         -1.033          0.877
Interest Coverage     -3.922          0.000
Foreign Sales          0.000          1.060
Age                    0.000         82.000

Panel D: 2001-02 (The Crisis Period)

Bank Loan              0.000          0.955
Trade Credit           0.000          0.889
Size                  11.233         17.976
Profitability         -0.904          0.785
Interest Coverage     -1.856          0.000
Foreign Sales          0.000          1.149
Age                   0.00082.000

Panel E: 2003-04-05 (The Post-Crisis Period)

Bank Loan              0.000          0.965
Trade Credit           0.000          0.832
Size                   6.709         18.291
Profitability         -2.406          0.516
Interest Coverage     -5.509          0.000
Foreign Sales          0.000          1.065
Age                    0.000         82.000

Note: This table provides descriptive statistics for the main
variables used in our analyses. Bank loan is measured as the ratio of
short-term bank loan to total assets. Trade credit is measured as the
ratio of accounts payable to total assets. Size is the
inflation-adjusted natural logarithm of total assets. Profitability is
the ratio of earnings before interest and taxes (EBIT) to total assets.
Interest coverage is the ratio of financial expenses to net sales.
Foreign sales is the ratio of foreign sales to net sales. Finally age
represents the number of years the firm has been operating.

Table 2. Pearson Correlation Matrix (VV=1973)

                     Size     Age    Bank Loan   Trade    Profit
                                                 Credit

Size                 1.00
Age                  0.01    1.00
Bank Loan           -0.08    0.00      1.00
Trade Credit        -0.18   -0.01      0.05       1.00
Profit              -0.01    0.00     -0.33      -0.05     1.00
Foreign Sales       -0.03    0.06      0.23       0.07     0.07
Interest Coverage    0.00    0.00      0.00       0.02     0.03

                     Foreign    Interest
                      Sales     Coverage

Size
Age
Bank Loan
Trade Credit
Profit
Foreign Sales          1.00
Interest Coverage      0.01       1.00

Note: This table presents the Pearson's Correlation matrix for the
variables used in our analyses of 1973 observations. Size is the
inflation adjusted natural logarithm of total assets. Age represents
the number of years the firm has been operating. Bank loan is measured
as the ratio of short-term bank loan to total assets. Trade credit
is measured as the ratio of accounts payable to total assets. Profit
is the ratio of earnings before interest and taxes (EBIT) to total
assets. Interest Cov. is the ratio of financial expenses to net sales.
Foreign sales is the ratio of foreign sales to net sales.

Table 3. Trade Credit Quartiles

Years   Bank Loan            1st Quartile   2nd Quartile

1997    Mean                    0.2750         0.3219
        Median                  0.2450         0.3311
        Standard deviation      0.2623         0.2317
        n                        151            151

1998    Mean                    0.2445         0.2608
        Median                  0.1934         0.2345
        Standard deviation      0.2446         0.2265
        n                        128            127

1999    Mean                    0.2053         0.2765
        Median                  0.1288         0.2521
        Standard deviation      0.2291         0.2207
        n                        135            134

2000    Mean                    0.2983         0.3189
        Median                  0.2127         0.2988
        Standard deviation      0.2972         0.2587
        n                        137            137

2001    Mean                    0.1454         0.1974
        Median                  0.0370         0.1153
        Standard deviation      0.2163         0.2157
        n                        154            153

2002    Mean                    0.1334         0.1489
        Median                  0.0173         0.1226
        Standard deviation      0.2012         0.2113
        n                        211            211

2003    Mean                    0.1670         0.2189
        Median                  0.1243         0.1840
        Standard deviation      0.1858         0.2064
        n                        200            200

2004    Mean                    0.1306         0.1677
        Median                  0.0708         0.1231
        Standard deviation      0.1678         0.1636
        n                        186            185

2005    Mean                    0.1275         0.1732
        Median                  0.0813         0.1503
        Standard deviation      0.1474         0.1529
        n                        172            172

Years   3rd Quartile   4th Quartile     t-test

1997       0.3105         0.1906      3.339 ***
           0.3187         0.1624
           0.2286         0.1670
            151            151

1998       0.2619         0.1927       1.975 **
           0.2651         0.1714
           0.2072         0.1681
            127            128

1999       0.2405         0.2030        0.092
           0.2304         0.1900
           0.1906         0.1756
            134            134

2000       0.2669         0.1819      3.868 ***
           0.2649         0.1520
           0.2259         0.1891
            137            137

2001       0.1769         0.1210        1.133
           0.1456         0.0601
           0.1772         0.1571
            153            154

2002       0.1686         0.1300        0.204
           0.1357         0.1010
           0.1632         0.1350
            211            211

2003       0.2051         0.1730        -0.356
           0.2060         0.1546
           0.1672         0.1490
            200            200

2004       0.1933         0.1675      -2.306 **
           0.1852         0.1580
           0.1465         0.1391
            185            185

2005       0.1796         0.1879      -3.72 ***
           0.1530         0.1672
           0.1500         0.1537
            172            172

Note: This table provides univariate mean comparisons of bank loan for
trade credit quartiles from lowest to highest. It also provides median
comparisons, standard deviation comparisons and number of observations
denoted as "n." Bank loan is measured as the ratio of short-term bank
loan to total assets. Trade credit is measured as the ratio of short
term accounts payable to total assets. ***, ** and * indicate
significance levels at 1%, 5% and 10% respectively.

Table 4. Relationship Between Bank Loans and Trade Credits

Dependent Variable = Bank Loan

Independent Variables                   Model 1              Model 2

[Bank Loan.sub.t-1]                    0.375 ***           0.394 ***
                                         (3.02)              (3.51)

Trade Credit                           0.523 ***           0.415 ***
                                         (4.01)              (4.26)

Trade Credit * (Crisis Dummy)          0.311 ***           0.206 ***
                                        -(2.99)              (3.17)

(Trade Credit)2                       -0.642 ***           -0.718 ***
                                        (-3.25)              (-3.01)

(Trade Credit)2 *                     -0.405 ***           -0.462 ***
(Crisis Dummy)                          (-2.74)              (-3.22)

Size                                                        0.016 *
                                                             (1.79)

Profitability                                               -0.009 *
                                                             (1.82)

Interest Coverage                                            -0.122
                                                             (-0.92)

Age                                                        0.283 ***
                                                             (3.11)

Foreign Sales                                               0.065 **
                                                             (1.99)

Wald 1                                (22.05) ***          (25.02) ***

Wald 2                                 (17.24) **          (32.15) ***

Sargan                                   115.3                65.14

[lambda]                                 0.625                0.606

AR 1                                  (-3.172) ***        (-3.826) ***

AR 2                                  (-2.624) ***           0.4371

Note: This table presents GMM estimations for the years 1997-2005.
Asymptotic standard errors are robust to heteroscedasticity. Bank
loan is measured as the ratio of short-term bank loan to total
assets. Trade credit is measured as the ratio of accounts payable
to total assets. Crisis is a dummy variable taking the value of
unity if the observations are at the years 1997, 1998, 2001 and
2002 and zero otherwise. Size is the inflation adjusted natural
logarithm of total assets. Profitability is the ratio of earnings
before interest and taxes (EBIT) to total assets. Interest coverage
is the ratio of financial expenses to net sales. Foreign sales is
the ratio of foreign sales to net sales. Finally age represents the
number of years the firm has been operating. Wald 1 tests
statistical significance of estimated coefficients that are
asymptotically distributed under the hypothesis of
non-relationship. Wald 2 tests statistical significance of time
related variables. Ho hypothesis of the Sargan test is the
orthogonality condition in terms of errors and has a asymptotic
distribution. AR1 and AR2 tests are respectively first and second
order autocorrelations of the residuals. [lambda] is the adjustment
factor and calculated through the estimated coefficient of the lagged
dependant variable. ***, ** and * indicate significance levels at
1%, 5% and 10% respectively.
COPYRIGHT 2009 Canadian Council for Small Business and Entrepreneurship
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Small and medium enterprises
Author:Arslan, Ozgur; Umutlu, Goknur
Publication:Journal of Small Business and Entrepreneurship
Geographic Code:1USA
Date:Sep 22, 2009
Words:8361
Previous Article:Typologie de dirigeants de tres petite entreprise.
Next Article:The relationship between owner characteristics and use of bootstrap financing methods.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters