WPS4754 Policy ReseaRch WoRking PaPeR 4754 Market Power and the Matching of Trade Credit Terms Daniela Fabbri Leora Klapper The World Bank Development Research Group Finance and Private Sector Team October 2008 Policy ReseaRch WoRking PaPeR 4754 Abstract This paper studies the decision of firms to extend reveals that access to bank financing and profitability are trade credit to customers and its relation with their not significantly related to trade credit supply. Rather, financing decisions. The authors use a novel firm-level firms that receive trade credit from their own suppliers database of Chinese SMEs with unique information on are more likely to extend trade credit to their customers, market power in both output and input markets and and to "match maturity" between the contract terms of on the amount, terms, and payment history of trade payables and receivables. This matching practice is more credit simultaneously extended to customers (accounts likely used when firms face strong competition in the receivable) and received from suppliers (accounts product market (relative to their customers), and enjoy payable). The analysis shows that suppliers with relatively strong market power in the input market (relative to weaker market power are more likely to extend trade their suppliers). These results highlight the importance of credit and have a larger share of goods sold on credit. supply chain financing for market competition and risk Examination of the importance of financial constraints management in credit constrained firms. This paper--a product of the Finance and Private Sector Team, Development Research Group--is part of a larger effort in the department to study SME finance. Policy Research Working Papers are also posted on the Web at http://econ. worldbank.org. The author may be contacted at lklapper@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Market Power and the Matching of Trade Credit Terms Daniela Fabbri Amsterdam Business School University of Amsterdam Amsterdam, The Netherlands 31-20-525-4344 d.fabbri@uva.nl Leora F. Klapper* The World Bank Development Research Group Washington, DC 20433 1-202-473-8738 lklapper@worldbank.org * Corresponding author. We thank Mariassunta Giannetti, Vicente Cunat, Arturo Bris, Luc Laeven, Inessa Love, Max Macsimovic, Rohan Williamson, Chris Woodruff, and participants at the "Small Business Finance­What Works, What Doesn't?" Conference in Washington the Second Joint CAF-FIC-SIFR Conference on Emerging Market Finance in Stockholm and the 4th Csef-Igier Symposium on Economics and Institutions in Capri for valuable comments and Taras Chemsky for excellent research assistance. The opinions expressed do not necessarily represent the views of the World Bank, its Executive Directors, or the countries they represent. "Large, creditworthy buyers force longer payment terms on less creditworthy suppliers. Large creditworthy suppliers incent less credit worthy SME buyers to pay more quickly" CFO Magazine, April 2007 1. Introduction Trade credit is an important source of funds for both small and large firms around the world (Petersen and Rajan, 1997, Demirguc-Kunt and Maksimovic, 2002). Many firms use trade credit both to finance their input purchases (accounts payable) and offer financing to their customers (accounts receivable), even in the case of small firms with credit constraints (McMillan and Woodruff, 1999; Marotta, 2005; van Horen, 2005). In this sense, trade credit is an important capital structure decision linking the liability side to the asset side of the firm's balance sheet. A number of reasons could induce firms to extend and use trade credit simultaneously. For example, firms that need to extend credit to their customers ­ possibly as a competitive gesture ­ might be more likely to demand trade credit themselves. In other words, firms might leverage accounts payable received from suppliers in order to finance accounts receivable extended to customers. At the same time, the increased focus of firms and analysts on cash holdings might offer incentives to firms to extend payment terms in order to maintain higher cash balances (Bates, Kahle, and Stulz, 2008). Furthermore, the risk management practice of firms to match the maturity between their assets and liabilities suggests that firms might use trade finance to match their short-term payables and receivables. This might be true for both small firms 1 1For instance, changes in inventory practices have led to firms holding more inventory (short-term assets), which would increase the demand for payables (short-term liabilities). Additional information on supply chain management operations is available from the "The Accounts Payable Network" (theaccountspayablenetwork.com). 1 without access to short-term bank credit, as well as large firms that can negotiate trade credit terms cheaper than alternative sources of external financing. These financial and operational decisions introduce a strong link between the use and extension of trade credit and suggest that accounts receivable could be used to finance accounts payable. Furthermore, trade credit terms offered to customers should be set to match trade credit terms received from suppliers. Lastly, the timing of payments to suppliers should match the receipt of payments remitted from customers. If this was the case, it would suggest a symbiotic relationship between the decision to supply trade credit to customers and the demand of trade credit from suppliers. Studying the two sides of the trade credit relationship simultaneously could therefore illuminate some important questions, for instance, why both small and large firms, constrained and constrained, take and offer trade credit. However, because of data limitations, previous related literature focuses only on one-side of the trade credit relationship in isolation. In this paper, we use a novel data set which allows us to take a two-sided approach. We use a large firm-level survey of Chinese firms, which is a unique source of data for two main reasons. First, it provides detailed information on the market power and competition of both output and input market. Second, it contains information on the amount, terms, and payment history of both trade credit extended by firms to their customers (accounts receivable), as well as the receipt of trade credit by firms from their own suppliers (accounts payable). Third, since only about 30% of firms in our sample have a line of credit, we can discuss the unique role of trade credit for firms in an emerging market. 2 To our knowledge, this is the first firm-level data set providing detailed and rich information on both the market environment and contract features of supplier and customer "supply-chain" financing. For example, firms are surveyed on the percentage of sales and inputs financed with trade credit; the number of days offered (maturity); whether a discount is offered for early payment; and the number of days until payment is actually made /received. This allows us to look simultaneously at both sides of firm activity - the relation between a firm, its customers (product market), and its suppliers (input market) - and to precisely match these two sides. This unique perspective is crucial to highlight a direct and novel link between the supply and demand for trade credit. The design of our analysis is straight-forward. First, we document the importance of trade credit as a competitive gesture. Specifically, firms that face stronger competition in the product market are more likely to extend trade credit and have a larger share of goods sold on credit. We then investigate how these firms finance the provision of trade credit and we provide empirical support for a matching story. We document that firms are likely to depend on their own receipt of trade credit to finance the extension of trade credit and to match maturity between contract terms of payables and receivables. We find very large and significant relationships between the decision to offer and the use of trade credit; the percentage of inputs and the percentage of sales financed by trade credit; the number of days extended to customers and the maturity received from suppliers; and whether the firm is offered a discount by its suppliers and offers a discount to its customers. Furthermore, we find that firms match the ex-post timing of payments, i.e. firms that receive payments early from customers are significantly more likely to remit early to suppliers, and viceversa. 3 Finally, we investigate in which circumstances firms are more likely to adopt a matching strategy. This matching practice seems to be affected by the availability of internal resources and bank financing and the use of informal credit. Firms with positive retained earnings and bank credit are less likely to rely on accounts payable, while firms using more costly informal sources of financing are more dependent on their own receipt of supplier financing to extend credit to their own customers. In addition, this matching practice is more likely among firms that face stronger competition in the output market and enjoy stronger market power in the input market. These new results highlight the importance of supply chain financing for market competition and growth. Our paper contributes to the literature on trade credit along several dimensions. First, we provide a unified setting to analyze the importance of market structure and competition for the use of trade credit, from the perspective of the firm relative to both its customers and suppliers. The related literature finds contrasting results on this issue. For example, McMillan and Woodruff (1999) document that the supplier provides more trade credit when he has stronger market power, in line with the idea that strong market power gives the supplier an informal mechanism to enforce the repayment of the credit contract, through the threat of stopping the supply of the intermediate goods (Cunat, 2006). More recent papers (Fisman and Raturi, 2004, Van Horen, 2005 and Giannetti, Burkart and Ellingsen, 2006) document an opposite relationship, consistent with the Wilner's (2000) argument that a customer ­ in particular in financial distress ­ obtains more trade credit if he generates a large percentage of the supplier's profits (the supplier's bargaining power is low). 4 In line with this second view, we document that firms use the provision of trade credit to their customers as a competitive gesture when facing stronger competition in the product market. One key advantage of our paper is the detailed information about both the firm's market environment and the features of trade credit contracts simultaneously received and extended. We measure the degree of product market competition by exploiting information on the firm's marketing strategy (we know whether the firm has lowered prices or introduced a new product line in the past year), on the number of suppliers used by a given customer, on the market share of competitors, and on the importance of the supplier for a given customer.2 We relate all of these proxies for competition not only to the decision to offer trade credit and the amount of trade credit offered, but also to the specific credit terms offered to customers. We also exploit information on the structure of the input market (for example, we know whether the firm is the largest supplier's most important customer) to measure the firm's relative position in relation to its suppliers and its customers simultaneously. Given that firms use payables to finance receivables, and payables are likely to depend on the bargaining power of the firm towards its supplier, the main purpose is to investigate how the two market structures (down-stream and up-stream) affect the firm's ability to finance receivables. We document that firms are more likely to finance receivables with payables and to match the maturity of the two contracts in two situations: First, when they enjoy stronger market power in the input market - they can set their own credit conditions. Second, when firms face stronger competition in the output market - they are forced to offer trade credit to reduce competition. This last result shows that a two-sided 2 Our finding that firms introducing new products are more inclined to offer trade credit is also consistent with theories saying that suppliers extend trade credit as a quality guarantee because new products are more likely to be untested (Smith, 1987; Lee and Stowe, 1993; Long, Malitz and Ravid, 1993). 5 approach where input and output markets are simultaneously considered can shed a new light on trade credit patterns among firms. Our paper is also related to the literature on matching debt maturity. Previous literature offers theoretical explanations (Diamond, 1991 and Hart and Moore, 1991) and empirical evidence (Guedes and Opler, 1996; Stohs and Mauer 1996; Demirguc-Kunt and Maksimovic 1999) for firm's matching maturity of assets and liabilities. We bring forward the idea that firms with no or limited access to internal resources and formal external financing match the maturity of assets and liabilities by using account payables as a risk management tool, i.e. to hedge their short-term receivables risk. Some suggestive evidence of a relationship between payables and receivables has been documented in other papers, but the lack of data on trade credit contracts simultaneously received and extended precluded a compelling analysis. For example, Petersen and Rajan (1997) show that U.S. firms whose assets consist mainly of current assets (they do not use receivables) demand significantly more trade credit. Johnson, McMillan and Woodruff (2002) find that in European transition economies firms with a larger proportion of invoices paid by customers after delivery tend also to pay their suppliers late. Similarly, Boissay and Gropp (2007) document that French firms with late paying customers are more likely to default on their suppliers. While these last two papers show that firms involuntary pass the delay of payments from customers to their suppliers, we document that firms use payables to hedge their short-term receivables risk in a voluntary and consistent way. In particular, firms offer (ex-ante) to their customers amount and credit terms (discount and number of days) that are similar to the ones they have been offered from their suppliers. We also 6 find that firms having customers paying early are more likely to remit early to their suppliers. This result provides further support to our idea that firms do not simply try to increase as much as possible the amount of payables, but they time the payment structure in the down-stream and up-stream market. Our novel set of results is important to rationalize two stylized facts for which the extant literature still lacks a clear understanding. While most of the related theoretical and empirical literature has stressed the importance of credit constraints as a motivation to take credit (Biais and Gollier, 1997; Burkart and Ellingsen 2004; Petersen and Rajan, 1995; Burkart, Ellingsen, and Giannetti 2006 among others), the existence of credit constraints does not explain (a) the large use of trade credit by unconstrained firms (Marotta 2005) or by listed companies, which generally have access to both public and private financial markets (Demirguc-Kunt and Maksimovic 2002), and (b) that small or credit constrained firms are more likely to both grant and receive trade credit than large unconstrained firms (see McMillan and Woodruff, 1999 for evidence on Vietnamese firms and Marotta 2005 for the Italian manufacturing sector).3 Our argument is that the matching story would provide an explanation for these two stylized facts. Small and credit-constrained firms are able to offer trade credit (needed to compete in the product market) as long as they can use payables. Large and unconstrained firms might decide to take trade credit to optimally hedge receivables risk. Our paper is also related to the more recent literature analyzing the role of trade credit as an informal source of finance in developing economies where formal financing 3Using the complete World Bank Enterprise Surveys database, which includes a sample of over 40,000 mostly small and medium sized firms surveyed in over 50 mostly developing countries around the world, we find that 69% of all firms report selling goods on credit, while 51% use trade credit financing (compared to 62% of firms that have access to bank financing).These are authors' estimations. 7 channels are underdeveloped (McMillan and Woodruff, 1999; Johnson McMillan and Woodruff, 2004; Allen, Qian, and Qian, 2005 and Cull, Xu and Zhu, 2007). The general idea in this literature is that trade credit can represent a viable substitute for formal bank credit and therefore a crucial determinant of firm growth for a country at the earlier stages of its development or during the transition towards a market based economy. Cull, Xu and Zhu (2007) test whether Chinese firms have indeed used the trade credit channel to bypass or substitute for the lack of formal institutions to stimulate growth, but they do not find supporting evidence.4 Instead they argue that the Chinese growth is fueled by a growing private sector with increasing competitive pressure. Competition is then likely to be an important motivation for the use of trade credit in line with our results. Given that widespread competitive pressures across sectors is a common feature of developed and market-based economies, our findings are not specific to China but have more general implications. A similar conclusion also holds for the matching story ­ the credit constrained firms in our sample are more likely to match contract terms of payables and receivables ­ given that relatively illiquid firms with limited access to bank loans are also widespread in developed economies, mainly among small and medium-sized enterprises (see Berkovitz and White, 2004 and Fabbri and Klapper 2008 for evidence on credit constraints among U.S. firms).5 Finally, when we compare trade credit statistics in China and other countries, we find that the average use of trade credit in China is very similar to other developed and 4They focus on differences in ownership structures, since in China it is not so much the lack of a formal financial system but rather its institutional bias in favor of state-owned enterprises that could give rise to trade credit among viable firms with restricted or no access to credit from state-owned banks. 5 Berkovitz and White (2004) find that the probability of a firm being credit constrained is 29% for noncorporate U.S firms and 26% for U.S. corporations. Fabbri and Klapper (2008) document that less than half of SMEs in the U.S. have a line of credit. 8 developing countries. For example, using the complete Enterprise Survey database of over 70 countries, we find that the average number of firms using trade credit for working capital or investment purposes (the only comparative variable available across countries) is 47% in China, and 51% for the complete sample. Similarly, the average percentage of trade credit used for working capital purposes (averaged across firms that use trade credit) is 48% of total working capital financing in China, versus 37% for the complete sample. These figures provide further support to the generality of our findings, as well as recent institutional changes in China. For example, unique features of the Chinese economy ­ such as the bias towards state-owned banks and state-owned firms ­ have been decreasing since 2001 (two years before our survey takes place) and we carefully address related potential biases. Moreover, there are no country-specific regulations on inter-firm financing. Finally, when we replicate our main results using data for Brazil, we find additional support for both the market power and the matching story found in China.6 The remainder of the paper is organized as follows. Section 2 describes the data. Section 3 develops our main testable hypotheses. Section 4 presents empirical results. Section 5 discusses our results and addresses some related issues. Section 6 concludes. 2. Data and Summary Statistics We use firm-level data on about 2,500 Chinese firms, which was collected as part of the World Bank Enterprise Surveys conducted by the World Bank with partners in 76 6 Unfortunately data for Brazil does not include as detailed information on supply chain contracts; therefore, we use data for Brazil as only a further robustness check for the evidence found in China. 9 developed and developing countries.7 The dataset includes a large, random sample of firms across multiple manufacturing and service sectors. The surveys include both quantitative and qualitative information on barriers to growth, including sources of finance, regulatory burdens, innovations, access to infrastructure services, legal difficulties, and corruption. The 2003 World Bank Enterprise Survey in China asked additional questions on the market environment, such as the number and importance of supplier and customer relationships, and supplier and customer ("supply-chain") financing. For the purpose of our analysis, the key questions regard the extension and terms of trade credit. Importantly, the survey asks both (i) whether firms offer trade credit to customers and (ii) whether customers accept trade credit from the firm.8 This allows us to precisely identify the decision of firms to offer trade credit. From our sample of 2,400 firms, 2,295 firms report whether or not they extend trade credit to their customers.9 Table 1 shows variable names, definitions, and means for all variables. We include measures of trade credit, general firm characteristics, indicators of market power of the firm (relative to its customers and to its suppliers), financial characteristics, and indicators of the collateral value of goods sold and customer creditworthiness. Detailed summary statistics are shown in Table 2 (the full sample) and Table 3 (disaggregated by firms that do and do not offer trade credit). Table 4 shows a correlation matrix of our explanatory variables. Our main dependent variable is a dummy variable equal to one if the firm offers trade credit (accounts receivable), and zero otherwise (AR_d). We find that 39% of the firms in our sample offer trade credit, and that the average percentage of goods sold on 7The survey instrument and data are available at: www.enterprisesurveys.org. 8Detailed questions on supply chain terms were not asked in any other country survey. 9We exclude from our sample 157 firms that provide financial services. 10 credit is 14% (AR_per); within the subsample of firms that offer trade credit, the average volume of credit extended represents over 35% of sales. On average, firms that extend trade credit offer customers about one month to pay (the median value of AR_days is 30 days). Finally, we find that 20% of firms that offer trade credit offer a prepayment discount on credit to its customers (AR_discount). The use of trade credit in China is comparable to similar developing countries. Although China is known to be a country with a high involvement of the public sector in the economic activity, there are no specific policies that regulate trade credit (Cull et al., 2007). Our dataset also allows us to examine the payment performance of firms' customers (i.e. the collection of accounts receivables). First, we include the percent of total sales received by customers due to the overdue penalty (AR_overdue). We find that 75% of firms receive overdue fines, with an average fee of 19% (and a median of 10%) of total sales. We also construct a multivariate dummy variable (AR_gap) equal to one if the difference between the number of days offered to customers and the number of days before the payment of receivables from customers actually takes place is greater than zero (i.e. customers prepay their receivables); equal to 0 if the difference equals zero (i.e. customers pay on time); and equal to negative one if the difference is less than zero (i.e. the customers pays late). In our sample, 35% of firms receive early payments of receivables, 37% of firms receive on-time payments, and 27% receive late payments. Next, we examine firms' use of trade credit from their own suppliers, accounts payable (AP). We find that 45% of firms use AP (AP_d), and the average percentage of supplies financed with credit is about 10% (AP_per); within the sample of firms that use AP, credit used equals about 20% of input purchases. Similar to accounts receivable, the 11 median term of payables is approximately one month (AP_days). We also find that about 7% of firms that use trade credit are offered a prepayment discount on credit from their suppliers (AP_discount). Table 3 shows that 62% of firms that extend trade credit to their customers receive credit from their suppliers, while only 34% of firms that do not extend credit use payables; this difference is significant at 1%. Furthermore, a first glance at the data also shows significant differences in payment terms. For example, firms that extend trade credit to their customers are offered 46 days before its supplier imposes penalties, while firms that do not extend credit receive a shorter offers (only 29 days); this difference is significant at 1%. We also include the percent of total input costs paid to suppliers due to overdue penalties (AP_overdue). About 25% of firms pay overdue fines, at an average of 1.67% of input costs. We also construct a multivariate dummy variable (AP_gap), equal to one if the number of days received from suppliers is less than the number of days until the firm pays its suppliers (i.e. the firm pays its payables to suppliers early); equal to 0 if the difference equals zero (i.e. the firm pays its suppliers on time); and equal to negative one if the difference is less than zero (i.e. the firm pays its suppliers late). In our sample, 29% of firms make early payments of receivables, 51% of firms make on-time payments, and 20% of firms pay late. Additional preliminary support of the interdependence between the use and extension of trade credit is shown in Table 5. Panel A summarizes the main sources of firm financing: Banks, Suppliers (trade credit), Family and Informal Sources, and Retained Earnings and Equity. Data is summarized by age, size, and ownership. For instance, we find that larger firms use more trade credit than smaller firms ­ that are more 12 likely to be credit constrained ­ and that foreign-owned firms use more trade credit than state-owned firms. Panel B shows summary statistics disaggregated by firms that offer trade credit (AR_d=1) and firms that do not offer trade credit (AR_d=0). We find significant differences in the use of trade credit among these two groups. For instance, young firms and small and medium sized enterprises (with less than 250 employees) that extend trade credit use significantly more trade credit from suppliers. We include in all regressions some general firm characteristics, which are likely to be associated with trade credit. First, the log number of years since the firm was established (L_Age).10 Second, we use as a proxy for firm size the log number of total employees (including contractual employees) (L_Emp). All our empirical results are robust to using alternative measures of firm size, such as dummies indicating small, medium, and large firms. Third, we include a dummy variable equal to one if the percentage of the firm owned by foreign individuals, foreign investors, foreign firms, and foreign banks is greater than 50%, and equal to zero otherwise (D_Foreign). Forth, we include a dummy variable equal to one if the percentage of the firm owned by the government (national, state, and local, and cooperative/collective enterprises) is greater than 50% (D_State). We include these ownership dummy variables to control for possible preferential access to financing from foreign and state-owned banks, respectively. It might also be the case that foreign and state-owned firms have preferential foreign and government product markets, respectively, and are not as sensitive to market competition. In our sample, 7% of firms are foreign owned, while 23% are state owned. Fifth, we include a dummy variable equal to one if the firm sells its products abroad, and equal to zero otherwise (D_Export). We include this variable to control for possible differences in 10The minimum age of firms in the sample was restricted to four. 13 trade credit use among national and foreign customers. In our sample, 9% of firms are identified as exporters. We also include in all regressions 17 city dummies. Our next set of variables measure market power and competition. First, we measure the importance of the firm's largest customer with a dummy variable equal to one if the percent of total sales that normally goes to the firm's largest customer is greater than 5% (the median), and 0 otherwise (Saleslargestcust_5). A value equal to one suggests that the firm's largest customer is important to its overall revenue and that the firm's market power is weak, relative to its customers. Second, we measure the importance of the firm for its largest customer with a dummy variable equal to one if the number of suppliers used by the firm's largest customer is greater than 5 (the median), and 0 otherwise (Suplcust_5). In other words, a value equal to one implies that the customer is less dependent on the firm ­ i.e. ending the relationship poses less of a risk of a holdup problem ­ and consequently less market power for the firm, relative to its customers. Third, we measure the importance of the firm's main competitor in the product market with a dummy variable equal to one if the firm's main competitor's share in the domestic market for the firm's most important product is greater than 1% (the median), and 0 otherwise (Compmktshare_1). This value proxies for market competitiveness; a value of one connotes weaker market power, relative to the firm's competitors (and buyers). Moreover, we proxy for a more competitive environment with a dummy equal to one if on average, and relative to the average of the last year, the firm has lowered prices on its main business line, which we assume was done as a competitive gesture (Lowered_prices). We also include a dummy variable if the firm has introduced a new 14 product (or service) or business line in the past year, assuming that this would require the firm to compete with a new product (New_product). As shown in Table 3, in bivariate tests, firms operating in more competitive environments ­ using all measures of market power and competition ­ are more likely to extend trade credit to customers. We also construct a new dummy variable (Bi_mktpower) that measures simultaneously the market structure in the input and output markets. Our unique dataset includes information on the market power of manufacturing firms in relation to both their customers and their suppliers. First, we construct a dummy variable that measures the bargaining power of a firm relative to its suppliers (Main_customer), which is equal to one if the firm is the most important customer of its main supplier and zero otherwise. We compare this variable to Saleslargest_5, which is a measure of the market power of a firm relative to its customers. In our sample, 751 of 1762 firms (43%) have weak bargaining power relative to their customers (Saleslargest_5 equals one), and 637 of 1514 firms (42%) are in a strong position relative to their suppliers (Main_customer equals one). Firms with available information on both sides of the markets are 1205: 29% of firms have weak bargaining power relative to their customers and strong market power relative to their suppliers; 37% of these firms have weak bargaining power relative to both their customers and their suppliers; 14% of firms have strong bargaining power relative to their customers and strong market power relative to their suppliers; finally, 20% of these firms have strong bargaining power relative to their customers and weak market power relative to their suppliers. The correlation between the market power in the input and output markets is slightly negative (­0.0289) but not significantly different from zero, suggesting that there 15 is no correlation between the strength of the contractual position of a firm in the input and output market. We hypothesize that the most likely scenario for a firm to use its own payables to finance its receivables ­ and match payment terms ­ is in the case where a firm is in the strong position to demand trade credit from its suppliers, but must offer trade credit from a weak position relative to its customers. Hence, Bi_mktpower takes value equal to one if two conditions are satisfied: First, the proportion of total sales that normally goes to the firm's largest customer is greater than 5%, which indicates that the firm has weak bargaining power towards its customer (Saleslargest_5 equals one). Second, the firm is the most important customer of its main supplier, i.e. the firm has strong bargaining power towards its supplier. In the remaining cases, the variable is assumed to be zero. Next, we include various measures of financial liquidity. We use the percentage of unused line of credit, equal to zero if the firm does not have a line of bank credit (LC_unused), which is 7% on average. Less than 30% of firms have access to a line of credit from the banking sector, and on average, firms that have a line of credit have 26% unused. The low-level of financial access to formal credit market might suggest that many Chinese firms are credit constrained. We also include two dummy variables equal to one if the firm uses local or foreign bank financing (D_Bank) and family or informal credit (D_Fam_Informal). Finally, since we do not have balance sheet information on cash holdings, we measure the availability of internal resources by using a dummy variable equal to one if the firm uses retained earnings to financing working capital or investment (D_RE). Bivariate tests find that firms that extend trade credit are 16 significantly more likely to have a larger unused line of credit and positive retained earning. Related literature has also documented that weak legal institutions constrain the ability of firms to access external financing such as long term debt or equity (Demirguc- Kunt and Maksimovic, 1998; 1999), and reduce firm growth opportunities (Demirguc- Kunt and Maksimovic, 2006). In addition, Johnson, McMillan and Woodruff (2002) document that entrepreneurs who report that courts are effective grant 5% more trade credit on average, but this effect is significant only for new-relationships. In a developing country such as China, legal contracts and confidence in the judicial system to enforce contracts could be all important indicators in the decision to extend credit. Legal institutions matter when contracts are written and disputes between parties arise. We thus include a dummy variable equal to one if the firm generally does not enter written contracts with clients, and equal to zero if the firm generally does use written contracts (Contracts). In our sample, 88% of firms ­ 1950 out of a total of 2216 firms ­ enter into written contracts with customers, and 82% do the same with suppliers of raw materials. Thus, almost all firms in our sample use written contracts. Furthermore, 29% of the sample report disputes with customers; 23% of the sample report disputes with suppliers. These figures are relatively low if compared with the average (58%) of firms located in the formerly planned economies of Eastern Europe and former Soviet Union (Johnson, McMillan and Woodruff, 2002). In case of disputes, however, the use of court action is also quite low. Among the 572 firms having disputes with customers, 47% of them use court action, 8% use 17 arbitration, and 74% negotiation. In the case of negotiation, 80% of firms recover the full value, while in case of court action 84% of the firms do not recover anything. Similar figures arise when we look at disputes with suppliers: among the 435 firms reporting at least one dispute with suppliers, only 23% rely on court action; 5% firms use arbitration, and 65% use negotiation. Overall, this evidence seems to suggest that when firms have disputes, they prefer to negotiate and to avoid relying on the assistance of third-parties, in line with Johnson, McMillan and Woodruff (2002). This could suggest that the quality of legal institutions is not crucial given that only few firms rely on court action. However, it could also be that firms rely seldom on courts since they anticipate that the cost will be too high. To gain a better understanding of the role of legal enforcement, we use firm-level survey information on the relation between the firm and the local judiciary. Firms are asked to evaluate the likelihood that the legal system will uphold contracts and property rights in business disputes in a scale ranging from zero to one. We call this variable (Property_right); the median value is 80%, suggesting that most firms in our sample do not consider the legal system as a major constraint to doing business.11 This variable is defined at the firm-level and therefore reflects how the firm perceives the quality of legal institutions. This is a nice feature of our data since it is likely that the effect of institutions on firm activity also depends on the characteristics of the firm (i.e., size). Lastly, we include a set of variables to test the robustness of previous theories, in particular, that trade credit is related to collateral values and customer creditworthiness. We include two direct measures of collateral value: First, the percent of sales made to 11Our results are robust to the inclusion of Law_predictability, which measures the predictability of laws or regulations that materially affect the operation and growth of business; however, this variable is missing for about 1/3 of firm observations. 18 clients' unique specification i.e. that cannot be sold to other clients (Uniqueness), which is about 40% of goods, on average. Second, the approximate percent of goods that are certified (Certified_products), which is about 47% of goods, on average. Finally, we are concerned that trade credit patterns within supply chains might be endogenous to industry characteristics. For instance, there might be "industry standards" which set the percentage and terms of trade credit. We control for this by including 12 "industry" dummies (2-digit NACE codes) in all regressions, which is the finest level of sector classification available. Although additional detailed information is available on the firm's "main business" line, this includes 1,818 descriptions and 99% of classifications describe only one firm. Nevertheless, we studied the trade credit patters of firms within a few classifications with more than 10 firms ­ and found no "systematic" patterns. For instance, 33 firms are classified as "dress manufacturing." On average, 31% of firms extend trade credit and 53% of firms receive trade credit. For the 11 firms that extend trade credit to their customers, the percentage of sales offered ranges from 10% to 100% (the median is 20%); terms offered include 3, 4, 10, 30, 40, and 90 days; and three firms offer a discount. Examinations of additional narrow industry classifications found similar disparities across firms.12 3. Hypotheses 3.1 The Market Power Hypothesis First, we test whether the decision to extend trade credit to customers depends on the market structure where the firm operates: 12Additional industries available upon request. 19 Accounts Receivable (AR) indicators = f {Firm characteristics, Market Power indicators} {1} 3.2 The Financial Constraint and Supply Chain Hypotheses The next step is to understand how firms finance the supply of trade credit. In principle, firms can use internal resources, like retained earnings. Alternatively, if they have access to external finance, they can use bank credit. According to previous literature (Frank and Maksimovic 2005, Burkart, Ellingsen, and Giannetti 2006), firms that are financially constrained on bank credit should extend less trade credit. An extension of this argument is that firms that are financially constrained should extend shorter terms. Finally, firms could also rely on family and informal loans. We provide an alternative hypothesis ­ that firms finance their extension of trade credit (accounts receivables) with access to trade credit from their own suppliers (accounts payables). An extension of this argument is our "matching maturity hypothesis," which assumes that firms aim to match the maturity of their assets and liabilities also in trade credit decisions. It is well known in the capital structure literature that firms use long-term debt to finance their long-term assets as a sound risk management practice. We assume that firms have only one short-term asset ­ accounts receivables ­ and one short-term liability ­ accounts payables. We thus estimate the following hypothesis: AR indicators = f {Firm characteristics, Financial characteristics, Accounts Payable (AP) indicators} (2) If it is indeed the case that firms finance their extension of accounts receivables with their receipt of accounts payables, firms would need to match terms to avoid late 20 charges, i.e. a firm would need to use money remitted from customers to pay its suppliers. We use interaction terms to test whether credit constrained firms or firms with less access to internal and external financing depend more on matching the terms of their accounts receivable and accounts payable. In summary, we estimate the following hypothesis: AR indicators = f {Firm characteristics, Financial characteristics, AP indicators, Interaction of Financial and AP indicators} (3) We also use interaction terms of market structure and accounts payable usage to test whether firms operating in a more competitive market are more likely to use their receipts of accounts payable to finance their extension of trade credit and to match credit terms. We use two indicators of market power: First, a measure of the percent of total sales that normally goes to the firm's largest customer (Saleslargest_5). We expect a positive correlation between this variable and accounts receivable indicators. Second, an index of competition in the output and input markets, Bi_mktpower, as defined in Section 2. We expect that firms are more likely to match receivables with payables (both the amount and credit terms) when they (i) face stronger competition in the output market, i.e. they are forced to offer trade credit to their customers, and (ii) command stronger market power in the input market, i.e. they can set their own credit conditions to suppliers. We thus test the following hypothesis: AR indicators = f {Firm characteristics, Market Indicators, AP indicators, Interaction of Market and AP indicators} (4) An alternative way to test our matching maturity story is to look at actual trade credit terms used, rather than the terms offered. Our data set provides unique information 21 on the percentage of owed accounts payable and accounts receivable that are overdue and on the number of days before accounts receivable are actually received and accounts payable are actually paid. If firms aim to match the maturity of their assets and liabilities ­ and/or use remitted receivables to finance their payment of payables ­ we should find that firms have a larger share of accounts payable overdue when their customers delay the repayment of a larger share of receivables. Similarly, we expect firms to pay accounts payable relatively quicker to their suppliers if their customers pay faster as well. Our model is then: AP indicators = f {Firm characteristics, AR indicators, Financial indicators} (5) 4. Results Regressions are shown in Tables 6 to 11. All regressions control for general firm characteristics. Consistently, we find that larger firms are more likely to extend trade credit, which might be related to their longer and more established customer and supplier relationships. Moreover, younger firms are more likely to offer trade credit, which can be due to the fact that new firms face stronger competition when entering the product market. We find no consistent significant relationships, however, with foreign or state ownership or exports. Table 6 shows that various measures of weaker market power and competition have a positive and highly statistical significant effect on the decision to offer trade credit and the percentage of sales financed with credit. For instance, the larger the number of suppliers of the firm's most important customer, the larger the market share of competitors and the larger the percentage of sales to the largest customer, the more likely 22 are firms to extend trade credit. In addition, firms that have introduced new products or lowered prices in the past year are more likely to extend trade credit, presumably as a competitive gesture. This suggests that when firms face an increase of competition in the product market, they are more likely to offer trade credit to their customers and allow customers to pay a larger share of sales on account. In this case, trade credit might be used as a competitive device to reduce actual competition or to prevent entry. In both cases, trade credit becomes crucial for the survival of the firm. A discrete change in product market competition (measured by saleslargest_5) increases the likelihood to offer trade credit (AR_d) by about 5 percentage points, which corresponds to 12.5% of the average likelihood. The same shock also increases the share of goods sold on credit (AR_per) by 68 percentage points. Both figures suggest that the economic effect of competition on trade credit supply is economically relevant. This would explain why even small firms without access to bank credit might still want to extend trade credit to their customers.13 Firms have different channels to finance the supply of trade credit, such as external financing (bank credit or informal sources), internal resources (retained earnings), or alternatively, credit from suppliers (accounts payable). Next, we examine the importance of a firm's access to finance from its own suppliers on its decision to extend credit to its customers, after controlling for other potential sources of accounts receivable financing. Table 7 documents the relevance of each source of financing. As shown in Panel A, unused bank credit lines (LC_unused) does not appear to have an 13Our results hold after controlling for the number of customers. The coefficient of this variable is positive and significant, suggesting that firms with a larger number of customers are more likely to offer trade credit or sell a larger percentage of their goods on credit. We do not show these regressions since the number of customers also appears to proxy for firm size and is highly correlated with our control variables. 23 effect on the likelihood to offer trade credit, the percentage of sales financed by trade credit, the length of the payment period, or the offer of a pre-payment discount.14 We also include dummy variables indicating that the firm uses retained earnings, bank financing, and informal or family sources of financing. Panel B shows consistently that after even including these variables, only matching accounts payable terms are significant.15 Our results are also robust to the inclusion of a dummy indicating positive profitability (not shown). The most intriguing finding of Table 7 is that firms use accounts payable to finance the provision of accounts receivable and that firms "match maturity" of trade credit received from their own suppliers with the one offered to their customers. We find very large and significant relationships between the decision to offer and the use of trade credit; the percentage of inputs purchased on account and the percentage of goods sold on credit; the number of days extended to customers and the ones received from suppliers; and whether the firm is offered a discount by its suppliers and offers a discount to its customers. A discrete change in the decision to take trade credit (AP_d) increases the firm's likelihood to offer trade credit (AR_d) by about 33 percentage points. A one standard deviation increment in the percentage of inputs purchased on credit (AP_per) increases the percentage of goods sold on credit by 87 percentage points. A discrete change in the likelihood to receive a pre-payment discount from suppliers (AP_discount) increases the firm's likelihood to offer a pre-payment discount to customers (AR_discount) by about 22 percentage points. Finally, a one standard deviation increment in the number of days the firm is allowed to use trade credit (AP_days) 14Note that the regressions using trade credit terms ­ AR_days and AR_discount ­ only include firms that offer trade credit (i.e AR_d=0). 15All results in Table 7 are robust to the exclusion of accounts payable terms. 24 increases by 3 days the payment period offered to customers (AR_days). This increment corresponds to the 8% of the average payment period offered. These figures suggest that changes in the contract terms received by suppliers have a significant economic impact on the contract terms offered by firms to their customers. A related question is whether the use of accounts payable to finance accounts receivable is more likely among credit-constrained firms or among firms lacking internal resources. Table 8 shows our results. The interaction terms between the use of accounts payables and LC_unused and our D_bank dummy are never statistically significant, suggesting that the decision to use accounts payable to finance accounts receivables does not depend on access to bank financing. The same results are also found if we include the smaller sample of trade credit terms (maturity and prepayment discounts, not shown). However, we find a significant effect of the interaction of the percentage of goods sold on credit and our bank dummy, suggesting that firms with access to bank financing rely less on accounts payable to finance the provision of trade credit to their own customers. Table 8 also includes the interaction of accounts payable terms and dummies indicating the use of retained earnings and informal sources of financing. In this case, we find some interesting significant terms: the coefficient on the interaction of accounts payable and bank financing is significantly negative, while the interaction of accounts payable and informal financing is significantly positive. These results suggest that whereas retained earnings might be used to finance the extension of trade credit, firms that use more costly informal sources of financing are more dependent on their own receipt of supplier financing to extend credit to their own customers 25 Table 9, however, shows that the extent to which firms match accounts payable with accounts receivable does depend on the bargaining power enjoyed simultaneously in the input and output markets. In particular, the coefficients of the interaction term between the index of market power - Bi_mktpower - and the decision to offer trade credit or the share of sales financed by trade credit ­ AR_d and AR_per ­ are positive and statistically significant. These findings suggest that our supply chain hypothesis is most likely to hold when firms need to offer trade credit to their customers (as a competitive gesture), but have enough bargaining power with suppliers to set their own credit conditions.16 Table 10 uses ex-post information (for the sample of firms that have positive accounts payable) on effective payment terms ­ the spread between number of days offered to customers and the number of days before customers remit (AR_gap) and the spread between the number of days offered by suppliers and the number of days before the firm remits (AP_gap) ­ to test our matching maturity story. We construct two dummy variables (D1_AR_gap) and (D2_AR_gap) which take values equal to one if AR_gap is equal to one (customers pay in advance) and negative one (customers pay late), respectively, and zero otherwise. We also control for external and internal resources through LC_unused and retained earnings (D_RE). The share of unused bank credit and having internal funds (retained earnings) are never significant in determining early or late payments. However, we find strong statistical significance for our matching maturity story. Columns 1-3 show that firms are significantly more likely to pay their suppliers late if their customers remit late and, 16Notice that the variable Bi-mktpower is only available for the sub-sample of manufacturing firms. All results for all tables are robust for the subsample of manufacturing firms. 26 similarly, Columns 4-6 show that firms are significantly more likely to pay their suppliers early if their customers remit early. In Column 6, we control for whether the firm is offered a discount from its suppliers for early payment. We find that firms offered discounts are significantly more likely to pay early (AP_discount), although the interaction with AR_gap is insignificant (not shown). We also test the determinants of overdue payables to suppliers (as a percentage of input costs) (AP_overdue). In general, the share of overdue payments from customers (as a percentage of sales, AR_overdue) is positive but insignificant (not shown). Overall, this evidence suggests that firms use the proceeds from remitted receivables to pay outstanding payables. Finally, Table 11 tests a series of variables found to be significant indicators of trade credit supply in previous literature. We find that after including accounts payable terms, having a larger share of goods made to clients' unique specification ­ i.e. outputs with low collateral value ­ does not effect the use or share of trade credit (Uniqueness). This variable might also proxy for customer opportunism (see Burkart and Ellingsen 2004, Burkart et al. 2006 and Cunat 2006), for instance, the lower diversion value of the transacted goods. But our findings do not provide empirical support for either of these alternative interpretations. However, entering into written contracts with clients (Contracts=1) increases the likelihood to offer trade credit. We also add our measure of firm's perceptions of property rights to our previous models. Our results (in particular the ones predicting the use of accounts payable) do not change. In addition, the likelihood that the legal system will uphold property rights (Property_rights) seems to have no significant impact on trade credit supply. Our results are consistent with the previous 27 evidence that most firms in our sample do not have disputes with trading partners and any disputes that do occur seldom rely on court action. Finally, we perform a series of robustness tests. We find that our results still hold if we restrict the sample to only manufacturing firms, profitable firms (Profit_d=1), to firms with a state ownership (national, state, and local, and cooperative/collective enterprises) lower than 50% (D_State=0), or to non-exporter firms (D_Export=0). Similarly, we get similar results if we replace the log of total employment with the log of total sales, although we are less comfortable using accounting data of the large number of unaudited firms in our sample. Our results are also robust to the inclusion of a dummy variable equal to one if the firm belongs to a government sponsored industrial park, science park, or Export Promotion Zone (EPZ).17 Next, although our data set generally provides only cross-sectional firm-level data for one year, some questions in the survey refer to past firm activity. These questions allow us to control for some changes in firm policy that occurred in the past. For example, the variable New_product included in our regressions reflects whether a firm has introduced new products (or services) in the past year and therefore this variable also controls for changes in the firm's investment policy. In addition, limited accounting information is available, both for the current and previous years. We use this information to control for potential idiosyncratic shocks at the firm level. For example, we construct a set of dummy variables to control for whether the firm increased sales or fixed assets in the past three years. These dummies are insignificant and do not affect our main results. 5. Discussion 17Regression results available upon request. 28 The evidence presented in the previous section raises some important questions. Why, for example, contrary to previous literature, our measure of credit constraints (LC_unused) does not seem to affect the decision to offer trade credit both directly or indirectly through the interactions with accounts payable? A possible explanation could be that the percentage of unused credit lines does not necessarily capture the tightness of credit constraints. Firms have to pay fees on the proportion of unused credit lines and therefore have incentives to reduce the unused portion. It follows that a fully used credit line does not necessarily identify a credit constrained firm. However, when instead we include a dummy indicating the availability of bank credit (both from local and foreign commercial banks), we find that firms with external bank finance are more likely to sell a larger percentage of goods on credit, but again no significant effect is found for the decision to extend credit. Another possible explanation could be that trade credit is not necessarily more expensive than bank credit. The central question then becomes how costly trade credit is, relative to bank financing. Conventional wisdom is that trade credit is primarily a financing of "last resort" for firms that have exhausted or unable to access bank credit. However, some more recent papers challenge this view. For example, Giannetti, Burkart and Ellingsen (2006) document that a majority of the U.S. frms in their sample appear to receive cheap trade credit. Similarly, Miwa and Ramseyer (2005) argue that there is no evidence that sellers use "extravagant cash discount' in Japan, while Marotta (2005) document that trade credit provided by Italian manufacturing firms is only slightly, if at all, more expensive than bank credit. Our findings are in line with this recent evidence. 29 Furthermore, our survey allows us to calculate the implicit cost of trade credit offered (AR_cost) using information on the discount offered for cash payment and the number of days before the supplier imposes penalties. Table 12 provides summary statistics on this rate and on the interest rate charged by banks (Bank_cost). The evidence is intriguing for several reasons. We find that 10% of firms in our sample receive an implicit rate lower than 4% and 25% of firms receive an implicit rate lower than 10%; the median implicit rate in our sample is 27%. Next, we examine the spread between the cost of bank financing and trade credit, accounting for annual fees that banks charge borrowers (Diff_AR_Bank_fee). We find that for over 20% of firms in our sample trade credit is indeed cheaper than bank loan. In addition, for over 40% of firms in our sample, the difference between the two rates is below 10 percentage points. Although we are unable to precisely measure the cost of trade credit, this evidence suggests that large firms might use trade credit as a relatively cheap source of working capital financing. A second important issue is that our analysis does not address the causality issue between payables and receivable, i.e. which of the two decisions ­ the offer of trade credit to customers or the demand of trade credit from suppliers ­ do firms make first? We take a first step in addressing this issue by estimating a simultaneous-equation-model where we let accounts receivable (AR_d and AR_per) depend on accounts payable (AP_d and AP_Per) and vice versa (after controlling for the usual firm characteristics, sector and city heterogeneity).18 The results show that accounts payable depend significantly and positively on accounts receivables, while accounts receivable are significantly related only to the degree of competition in the product market ­ but not on accounts payable (not shown). This would suggest that firms first decide to offer customers delayed 18We use as identification condition the market structure of the input and output markets. 30 payments, and then decide to purchase inputs on credit, perhaps in order to use their payables to finance the provision of receivables. We interpret this evidence as further support to our competition-matching story. Our result on the role of market competition raises another question: if competition is a main driver behind the decision to offer trade credit, why do firms not simply reduce the product price? In many developed countries, laws prohibit firms from selling identical products to customers at different prices (i.e. the Clay Act in the U.S.). However, China has no such laws. Alternatively, if firms are able to finance their extension of accounts receivable with accounts payable, firms might be able to both increase their total assets (relative to a price discount), and reduce their leverage ratio (i.e. total debt), which might be more favorable for loan officers and equity analysts. We leave a more in depth discussion of these institutional questions to future research. 6. Conclusions This paper uses firm-level data on about 2,500 Chinese firms to study the decision to extend trade credit. Supplier financing is often overlooked in the capital structure literature, although it is arguably the most important source of small and medium sized financial intermediation ­ particularly in countries with less developed financial and information systems. We show that firms are likely to offer trade credit as a competitive gesture. We also find that firms are likely to depend on credit from their own suppliers to finance the extension of trade credit to their customers and to match credit terms between accounts payable and accounts receivable; in particular, firms with stronger market power in the input market and facing strong competition in the output market. Furthermore, we find evidence that firms match their ex-post payment decisions. Finally, we find that for 31 over 20% of firms in our sample trade credit is indeed cheaper than bank financing, which may explain why firms without credit constraints choose to use trade financing. These results highlight the importance of supply chain financing for market competition and growth. 32 References: Allen, Franklin, Jun Qian, and Meijaun Quian, 2005. Law, Finance and Economic Growth in China. Journal of Financial Economics 77, 57-116. Bates, Thomas, Kathleen Kahle, and Rene Stulz, 2008, Why do U.S. Firms Hold so Much More Cash than They Used to?, working paper. Beck, Thorsten , Asli Demirguc-Kunt and Vojislav Maksimovic 2006. Financial and Legal Constraints to Growth: Does Firm Size Matter. Journal of Finance 1, 137-177. Berkowitz Jeremy and Michelle J. White, 2004. Bankruptcy and Small Firms' Access to Credit. Rand Journal of Economics 35, 69-84. Biais, Bruno and Christian Gollier, 1997. Trade Credit and Credit Rationing. Review of Financial Studies 10, 903-937. Boissay, Frederic and Reint Gropp, 2007. Trade Credit Defaults and Liquidity Provision by Firms, European Central Bank Working Paper n. 753. Brennan, Michael, Vojislav Maksimovic and Josef Zechner, 1988. Vendor Financing. Journal of Finance 43, 1127-1141. Burkart, Mike and Tore Ellingsen, 2004. In-Kind Finance: A Theory of Trade Credit. American Economic Review 94, 569-590. Burkart, Mike, Tore Ellingsen, Mariassunta Giannetti, 2006. What You Sell is What You Lend? Explaining Trade Credit Contracts. ECGI Finance Research Paper No. 71/2005. Calomiris, Charles., Charles. Himmelberg, and Paul Wachtel, 1995. Commercial Paper, Corporate Finance and the Business Cycle: A Microeconomic Perspective, Carnegie- Rochester Series on Public Policy 42, 203-50. Cull, Robert, Lixin Colin Xu, and Tian Zhu, 2007. Formal Finance and Trade Credit During China's Transition. Journal of Financial Intermediation, forthcoming. Cunat, Vicente, 2006. "Trade Credit: Suppliers and Debt Collectors as Insurance Providers," Review of Financial Studies, 20. Demirguc-Kunt, Asli, and Vojislav Maksimovic, 1998. Law, Finance, and Firm Growth. Journal of Finance 6, 2107-2137. Demirguc-Kunt, Asli, and Vojislav Maksimovic, 1999. Institutions, Financial Markets, and Firms Debt Maturity. Journal of Financial Economics 54, 295-336. 33 Demirguc-Kunt, Asli, and Vojislav Maksimovic, 2002. Firms as Financial Intermediaries: Evidence from Trade Credit Data, World Bank Working Paper. Diamond D., 1991. Debt Maturity Structure and Liquidity Risk. Quarterly Journal of Economics, 56. 709-738. Fabbri Daniela and Leora Klapper, 2008. The Choice to Collateralize. Working Paper. Fabbri, Daniela and Annamaria Menichini, 2007. Trade Credit, Collateral Liquidation and Borrowing Constraints. Swiss Finance Institute Working Paper. Ferris, J. Stephen, 1981. A Transaction Theory of Trade Credit Use. The Quarterly Journal of Economics 96, 247-270. Fisman, Raymond and Inessa Love, 2003. Trade Credit, Financial Intermediary Development, and Industry Growth, Journal of Finance 58, 353-374. Fisman, R., Raturi, M., 2004. Does Competition Encourage Credit Provision? Evidence from African Trade Credit Relationships. Review of Economics and Statistics 86, 345- 52. Frank, Murray and Vojislav Maksimovic, 2005. Trade credit, Collateral, and Adverse Selection. University of Maryland Working Paper. Giannetti, Mariassunta, Burkart, Mike and Tore Ellingsen, 2006. What You Sell is What You Lend? Explaining Trade Credit Contracts. Review of Financial Studies, forthcoming Guedes, Jose and Tim Opler, 1996. The determinants of the Maturity of Corporate Debt Issues. The Journal of Finance 51, 1809-1833 Hart, Oliver and John Moore, 1991. A Theory of Debt Based on the Inalienability of Human Capital. The Quarterly Journal of Economics 109, 841-879. Johnson, Simon, John McMillan and Christopher Woodruff, 2002. Courts and Relational Contracts. Journal of Law, Economics and Organization18, 221-277. Lee, Yul W. and John D. Stowe (1993), "Product Risk, Asymmetric Information, and Trade Credit," Journal of Financial and Quantitative Analysis 28, 285­300. Lee, Hau L. and Chung-Yee Lee Springer, 2007, Building Supply Chain Excellence in Emerging Economies, Science+Business Media: United Kingdom. Long, Michael S., Ileen B. Malitz and S. Abraham Ravid, 1993. Trade Credit, Quality Guarantees, and Product Marketability. Financial Management 22, 117-127. 34 Love, Inessa, Lorenzo A. Preve and Virginia Sartia-Allende, 2006. Trade Credit and Bank Credit: Evidence from Recent Financial Crisis. World Bank Working paper. Marotta, Giuseppe, 2005. Is Trade Credit More Expensive than Bank Credit Loans? Evidence from Italian Firm-Level Data. Applied Economics 37(4), 403-416. McMillan, John and Christopher Woodruff, 1999. Interfirm Relationships and Informal Credit in Vietnam, The Quarterly Journal of Economics 114, 1285-1320. Mian, Shehzad L. and Clifford W. Smith, 1992. Accounts Receivable Management Policy: Theory and Evidence. Journal of Finance 47, 169-200. Petersen, Mitchell and Raghuram G. Rajan, 1995. The Effect of Credit Market Competition on Lending Relationships. Quarterly Journal of Economics 110, 407-443. Petersen, Mitchell A. and Raghuram G. Rajan, 1997. Trade Credit: Theory and Evidence. Review of Financial Studies, 10 (3), 661-691. Smith, J. K., 1987. Trade Credit and Informational Asymmetry. Journal of Finance, 42, 863­872. Stohs, M., and Mauer, D., 1996. The Determinants of Corporate Debt Maturity Structure. Journal of Business, 69, 279-312. Summers, Barbara and Nicholas Wilson, 2002. Trade Credit Terms Offered by Small Firms: Survey Evidence and Empirical Analysis. Journal of Business and Finance Accounting 29, 317-335. Van Horen, N., 2005. Do firms use trade credit as a competitiveness tool? Evidence from Developing Countries, World Bank Working Paper. Wilner, Benjamin S. 2000. The Exploitation of Relationships in Financial Distress: The Case of Trade. Credit. The Journal of Finance 55, 153-178. 35 Table 1: Variable Definitions and Mean Statistics Variable Name Definition Mean Measures of Trade Credit AR_d Dummy (0/1), =1 if the firm offers credit to its customers (i.e. accounts receivable), =0 if the firm does not offer trade credit 0.39 AR_per The percent of monthly sales sold on credit 14.02 AR_days Average number of days customers are allowed to use the credit before the firm imposes penalties 60.15 AR_discount Dummy (0/1), =1 if the firm offers a pre-payment discount on credit to its customers, =0 otherwise (and = . if AR_d is =0) 0.20 Dummy variable =1 if the difference between the number of days offered to customers less the number of days until receivable AR_gap payments are received from customers is greater than zero (i.e. customers prepay their receivables); =0 if the difference equals 0.08 zero (i.e. customers pay on time); and =-1 if the difference is less than zero (i.e. the customers pays late) AP_d Dummy (0/1), =1 if the firm uses supplier credit (i.e. accounts payable) to purchase inputs, =0 otherwise 0.45 AP_per The percent of inputs purchased on credit (based on period averages), = 0 if the firm does not use trade credit 9.58 AP_days Average number of days the firm is allowed to use the credit before its suppliers imposes penalties 31.67 AP_discount Dummy (0/1), =1 if the firm received a pre-payment discount on credit from its suppliers, =0 otherwise (and = . if AR_d is =0) 0.07 Dummy variable =1 if the difference between the number of days offered to the firm by its suppliers less the number of days until AP_gap the firm pays its suppliers is greater than zero (i.e. the firm prepays its payables); =0 if the difference equals zero (i.e. the firm paysits 0.09 suppliers on time); and =-1 if the difference is less than zero (i.e. the firms pays its suppliers late) General Firm Characteristics L_Age Log number of years (+1) since the firm was established 2.57 L_Emp Log average number of total employees (including contractual employees) 4.94 Dummy (0/1), =1 if the percentage of the firm owned by foreign D_Foreign individuals, foreign institutional investors, foreign firms, and 0.07 foreign banks is greater than 50, =0 otherwise Dummy (0/1), =1 if the percentage of the firm owned by the D_State government (federal, state, local, and collective/cooperative 0.23 enterprises) is greater than 50, =0 otherwise D_Export Dummy (0/1), =1 if the firm is exporting, =0 otherwise 0.09 36 Indicators of (Weaker) Market Power of the Seller (relative to its Customers) Saleslargest_5 Dummy (0/1), =1 if the percent of total sales that normally goes to the firm's largest customer is greater than 5%, =0 otherwise 0.57 Dummy (0/1), =1 if on average, and relative to the average of the Lowered_prices last year, the firm has lowered prices on its main business line, =0 0.48 otherwise Dummy (0/1), =1 if the firm's main competitor's share in the Compmktshare_1 domestic market for the firm's most important product is more 0.28 than one percent, =0 otherwise New_product Dummy (0/1), =1 if the firm has introduced new products (or services) in the past year, =0 otherwise 0.42 Suplcust_5 Dummy (0/1), =1 if the number of suppliers used by the firm's largest customer is greater than 5, and 0 otherwise 0.44 Dummy (0/1), =1 if the percent of total sales that normally goes to Bi_mktpower the firm's largest customer is greater than 5% (i.e Saleslargest_5 = 1), and the firm is its largest supplier's most important customer, 0.29 =0 otherwise. Financial Characteristics The percent of the firm's line of credit or overdraft facility that is LC_unused currently unused (=0 if the firm does not have a line of credit or 0.07 overdraft facility) D_Bank Dummy (0/1), = 1 if the firm uses local or foreign bank financing for working capital or investment 0.49 D_RE Dummy (0/1), = 1 if the firm uses retained earnings for working capital or investment 0.36 D_Fam_Informal Dummy (0/1), = 1 if the firm uses family or informal financing for working capital or investment 0.20 Other Firm Characteristics Property_rights The likelihood (%) that the legal system will uphold contracts and property rights in business disputes 63.20 Uniqueness The percent of sales made to clients' unique specification (i.e. that cannot be sold to other clients) 37.53 Certified_dum Dummy (0/1), =1 if the firm has profits greater than zero, =0 otherwise 0.57 Contracts Dummy (0/1), =1 if the firm generally enters into written contracts with clients, =0 otherwise 0.88 37 Table 2: Summary Statistics See Table 1 for variable definitions. Variable Name Obs. Mean Std. Dev. Min Max AR_d 2,157 0.39 0.49 0.00 1.00 AR_per 2,184 14.02 27.97 0.00 100.00 AR_days 818 60.15 47.41 1 180 AR_discount 823 0.20 0.40 0.00 1.00 AR_gap 809 0.08 0.79 -1.00 1.00 AP_d 2,100 0.45 0.50 0.00 1.00 AP_per 2,069 9.58 21.41 0.00 100.00 AP_days 656 31.67 25.77 1 90 AP_discount 829 0.07 0.26 0.00 1.00 AP_gap 656 0.09 0.70 -1.00 1.00 L_Age 2,243 2.57 0.74 1.39 3.99 L_Emp 2,239 4.94 1.48 0.00 11.16 D_Foreign 2,242 0.07 0.26 0.00 1.00 D_State 2,242 0.23 0.42 0.00 1.00 D_Export 2,265 0.09 0.28 0.00 1.00 Saleslargest_5 1,762 0.57 0.49 0.00 1.00 Compmktshare_1 2,180 0.28 0.45 0.00 1.00 Lower_price 2,222 0.48 0.50 0.00 1.00 New_product 2,223 0.42 0.49 0.00 1.00 Suplcust_5 1,646 0.44 0.50 0.00 1.00 Bi_mktpower 1,205 0.29 0.45 0.00 1.00 LC_unused 2,152 0.07 0.21 0.00 1.00 D_Bank 1,549 0.49 0.50 0.00 1.00 D_RE 1,457 0.36 0.48 0.00 1.00 D_Fam_Informal 1,421 0.20 0.40 0.00 1.00 Property_rights 1,935 64.42 38.52 0.00 100.00 Uniqueness 2,047 37.53 42.05 0.00 100.00 Contract 2,216 0.88 0.33 0.00 1.00 38 Table 3: Mean Differences, by Trade Credit Supply See Table 1 for variable definitions. t-statistics show the mean difference of firms that offer trade credit to customers versus firms that do not offer trade credit to customers. ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively. Variable Name AR_d=0 AR_d=1 Sig. AP_d 0.34 0.62 *** AP_per 4.44 17.51 *** AP_days 28.98 46.05 *** AP_discount 0.05 0.09 ** AP_gap 0.07 0.12 L_Age 2.60 2.51 *** L_Emp 4.84 5.11 *** D_Foreign 0.06 0.10 *** D_State 0.24 0.19 *** D_Export 0.09 0.11 * Saleslargest_5 0.54 0.64 *** Bi_mktpower 0.26 0.33 *** Compmktshare_1 0.23 0.36 *** New_product 0.36 0.51 *** Lower_price 0.41 0.60 *** Suplcust_5 0.42 0.56 *** LC_unused 0.06 0.09 *** D_Bank 0.49 0.52 D_RE 0.33 0.42 *** D_Fam_Informal 0.20 0.21 Property_rights 63.20 66.76 ** Uniqueness 37.45 37.60 Contract 0.85 0.93 *** 39 Table 4: Correlation Matrix See Table 1 for variable definitions ***, ** and * indicate significance at the 1%, 5%, and 10% levels, respectively. Panel A: Explanatory Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) AP_d (1) 1.00 AP_per (2) 0.51** 1.00 AP_days (3) . 0.25** 1.00 AP_discount (4) . 0.02 0.09* 1.00 L_Age (5) -0.07** -0.03 0.00 0.04 1.00 L_Emp (6) 0.11** 0.14** 0.16** 0.03 0.29** 1.00 D_Foreign (7) 0.05* 0.12** 0.03 0.01 -0.11** 0.09** 1.00 D_State (8) -0.12** -0.08** -0.04 0.01 0.40** 0.21** -0.15** 1.00 D_Export (9) 0.04* 0.07* 0.06 0.02 -0.08* 0.21* 0.36* -0.12* 1.00 Saleslargest _5 (10) 0.03 0.12** 0.18** -0.01 -0.05* 0.02* 0.10** -0.08** 0.12* 1.00 New_product (11) 0.12** 0.14** 0.15** 0.02 -0.06** 0.23** 0.03 -0.03 0.06* 0.14** 1.00 Lower_price (12) 0.09** 0.11** 0.15** 0.01 -0.05** 0.06** 0.01 -0.07** 0.06* 0.18** 0.22** 1.00 Compmktshare_1 (13) 0.13** 0.11** 0.06 0.04 -0.05** 0.27** 0.10** -0.02 0.06* 0.13** 0.31** 0.16** 1.00 Bi_mktpower (14) 0.08*** 0.14*** 0.11** -0.05 -0.06** 0.17*** 0.08*** 0.00 0.13* 0.46*** 0.09*** 0.13*** 0.19*** 1.00 LC_unused (15) 0.10** 0.16** 0.09** 0.01 -0.02** 0.20** 0.08** 0.00 0.10* 0.08** 0.16** 0.08** 0.17** 0.08** 1.00 Panel B: Dependent Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) AR_d 0.27** 0.30** 0.23** 0.08** -0.06** 0.09** 0.08** -0.06** 0.04* 0.10** 0.15** 0.18** 0.13** 0.08*** 0.07** AR_per 0.23** 0.35** 0.20** 0.00 -0.06** 0.10** 0.09** -0.08** 0.12* 0.16** 0.14** 0.17** 0.14** 0.11*** 0.08** AR_days 0.09** 0.07 0.14** -0.02 -0.02 0.09** 0.02 0.02 0.05 0.11** 0.12** 0.16** 0.16** 0.03*** 0.08* AR_discount 0.06 -0.02 -0.08 0.13** -0.03 -0.05 0.00 0.00 -0.06 -0.09** 0.01 0.00 -0.01 -0.03*** 0.00 40 Table 5: Types of Financing, by Firm Type, Means (Percentages) See Table 1 for variable definitions. The minimum age of firms in our sample is 4 years. Bank financing includes both domestic and foreign banks. "Other" includes and other sources of financing. In Panel B, ***, ** and * indicate significance of the mean difference at the 1%, 5%, and 10% level, respectively. Panel A: All Firms Obs. Bank Trade Family & Informal Retained Earnings Credit Sources & Equity Other All firms 1,664 29.16 2.43 7.98 17.13 43.30 By age 4-7 452 24.31 1.61 13.08 18.73 42.28 8-13 597 25.80 3.82 7.33 20.20 42.85 >13 615 35.99 1.68 4.87 12.98 44.48 By size < 50 478 14.94 1.17 12.29 17.32 54.28 51-100 280 21.08 1.84 12.20 18.60 46.28 101-250 353 32.82 3.29 5.91 19.09 38.90 > 250 553 43.21 3.27 3.45 14.97 35.10 By Ownership Foreign 125 26.38 8.30 2.42 16.29 46.61 State 362 39.11 0.25 2.94 9.32 48.39 Domestic private 1,177 26.40 2.48 10.13 19.62 41.38 Panel B: by Extension of Trade Credit AR_d = 0 AR_d = 1 Trade Family & Family & Obs. Bank Trade Credit informal Obs. Bank informal sources Credit sources All firms 969 27.97 1.14 8.19 626 30.37 4.43*** 7.88 By age 4-7 260 25.15 0.82 11.81 170 23.16 3.02** 15.25 8-13 322 23.68 1.51 8.77 256 28.45 6.34*** 5.67 >13 387 33.43 1.03 5.27 200 38.96 3.18** 4.44 By size < 50 311 14.85 0.43 12.21 137 14.80 2.58** 13.13 51-100 162 21.46 1.01 10.50 109 19.83 3.23 14.48 101-250 199 32.00 1.84 7.33 142 33.83 4.89** 4.41 > 250 297 42.56 1.47 3.29 238 42.10 5.77*** 3.91 By Ownership Foreign 52 22.00 1.60 4.81 70 29.57 13.64*** 0.74 State 226 36.71 0.20 2.26 114 39.37 0.39 4.84 Domestic private 691 25.56 1.41 10.38 442 28.18 4.01*** 9.79 41 Table 6: The Relationship between Trade Credit Supply and Market Power and Competition The reported estimates are from logit (Colums 1-5) and OLS (Columns 6-10) regressions. AR_d is a dummy indicating the use of accounts receivable. AR_per is the percent of monthly sales sold on credit. We transform AR_per in a continuous variable by using the inverse function of the logistic model. See Table 1 for variable definitions. All regressions include 12 sector dummies and 17 city dummies. Robust p-values are shown in parentheses, ***, **, * indicate significance at the 1%, 5%, and 10% level, respectively. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) AR_d AR_d AR_d AR_d AR_d AR_per AR_per AR_per AR_per AR_per L_Age -0.24 -0.19 -0.14 -0.17 -0.24 -0.85 -0.67 -0.67 -0.6476 -0.5279 [0.01]*** [0.02]** [0.00]*** [0.03]** [0.01]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.06]* L_emp 0.16 0.10 0.10 0.09 0.16 0.51 0.36 0.41 0.3216 0.3123 [0.00]*** [0.01]** [0.00]*** [0.02]** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.03]** D_Foreign 0.55 0.34 0.14 0.34 0.55 1.49 0.78 1.08 0.8942 0.7311 [0.03]** [0.10]* [0.24] [0.09]* [0.03]** [0.08]* [0.27] [0.13] [0.21] [0.37] D_State -0.11 -0.10 -0.02 -0.12 -0.11 0.01 0.01 0.01 -0.1076 0.0142 [0.48] [0.49] [0.78] [0.39] [0.48] [0.98] [0.98] [0.99] [0.77] [0.98] D_Export 0.33 0.35 0.39 0.34 -0.329 0.04 0.06 0.07 0.1368 0.1318 [0.16] [0.07]* [0.04]** [0.08]* [0.16] [0.96] [0.93] [0.91] [0.84] [0.86] Saleslargest_5 0.21 0.77 [0.08]* [0.02]** Compmktshare_1 0.21 0.86 [0.07]* [0.02]** Lower_price 0.56 1.55 [0.00]*** [0.00]*** New_product 0.32 0.95 [0.00]*** [0.00]*** Suplcust_5 0.21 0.51 [0.08]* [0.15] Constant -2.94 -1.92 -2.13 -1.92 -2.94 -11.94 -11.37 -11.89 -10.99 -12.27 [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** Observations 1,611 1,974 2,011 2,016 1,611 1,623 2,000 2,038 2,040 1,524 Pseudo R-squared 0.09 0.08 0.09 0.08 0.07 0.13 0.12 0.12 0.12 0.11 42 Table 7 Panel A: The Relationship between Trade Credit Demand and Trade Credit Supply The reported estimates are from logit in Columns (1), (3), (5), (7), OLS in Columns (2) and (6) and tobit regressions in Columns (4) and (8). AR_d is a dummy indicating the use of accounts receivable. AR_per is the percent of monthly sales sold on credit We transform AR_per in a continuous variable by using the inverse function of the logistic model. AR_discount is a dummy indicating if the firm offers a pre-payment discount on credit to its customers; and AR_days is the average number of days customers are allowed to use the credit before the firm imposes penalties. Columns (3-4) and (7-8) include only the subsample of firms that use accounts receivable. See Table 1 for variable definitions. All regressions include 12 sector dummies and 17 city dummies. Robust p-values are shown in parentheses, ***, **, * indicate significance at the 1%, 5%, and 10% level, respectively. (1) (2) (3) (4) (5) (6) (7) (8) Unused Line of Credit (LC_Unused) Bank Financing (D_Bank) AR_d AR_per AR_discount AR_days AR_d AR_per AR_discount AR_days L_Age -0.13 -0.6859 -0.06 -2.09 -0.25 -0.67 0.14 -2.62 [0.11] [0.00]*** [0.76] [0.50] [0.01]** [0.02]** [0.54] [0.47] L_emp 0.06 0.2743 -0.03 0.56 0.08 0.28 -0.11 1.82 [0.15] [0.02]** [0.76] [0.73] [0.11] [0.05]** [0.36] [0.33] D_Foreign 0.39 0.1974 0.03 -4.06 0.47 0.39 0.34 -3.69 [0.07]* [0.77] [0.93] [0.54] [0.05]* [0.66] [0.44] [0.61] D_State -0.03 0.1048 -0.15 -5.78 -0.00 0.033 -0.17 -14.09 [0.85] [0.78] [0.63] [0.28] [0.98] [0.94] [0.69] [0.03]** D_Export -0.33 0.2945 -0.87 9.87 -0.52 -0.42 -0.45 7.48 [0.10]* [0.64] [0.06]* [0.18] [0.02]** [0.56] [0.34] [0.35] LC_unused 0.16 0.68 -0.13 5.80 [0.51] [0.41] [0.77] [0.50] D_Bank -0.08 -0.40 0.45 7.96 [0.55] [0.32] [0.11] [0.09]* AP_d 1.41 1.36 [0.00]*** [0.00]*** AP_per 0.09 0.08 [0.00]*** [0.00]*** AP_discount 1.30 0.59 [0.00]*** [0.17] AP_days 0.10 0.11 [0.02]** [0.03]** Constant -2.96 -13.25 -19.37 64.48 -3.16 -13.18 -0.41 73.42 [0.00]*** [0.00]*** [0.00]*** [0.01]*** [0.00]*** [0.00]*** [0.83] [0.00]*** Observations 1,881 1,869 646 556 1299 1,291 477 409 Pseudo R-squared 0.13 0.18 0.13 0.01 0.15 0.18 0.17 0.01 43 Table 7: Panel B: The Relationship between Trade Credit Demand and Trade Credit Supply The reported estimates are from logit in Columns (1), (3), (5), (7), OLS in Columns (2) and (6) and tobit regressions in Columns (4) and (8) AR_d is a dummy indicating the use of accounts receivable; AR_per is the percent of monthly sales sold on credit. We transform AR_per in a continuous variable by using the inverse function of the logistic model. AR_discount is a dummy indicating if the firm offers a pre-payment discount on credit to its customers; and AR_days the average number of days customers are allowed to use the credit before the firm imposes penalties. Columns (3-4) and (7-8) include only the subsample of firms that use accounts receivable. See Table 1 for variable definitions. All regressions include 12 sector dummies and 17 city dummies. Robust p-values are shown in parentheses, ***, **, * indicate significance at the 1%, 5%, and 10% level, respectively. (1) (2) (3) (4) (5) (6) (7) (8) Retained Earnings (D_RE) Family & Informal Financing (D_Fam_Inf) AR_d AR_per AR_discount AR_days AR_d AR_per AR_discount AR_days L_Age -0.19 -0.56 0.18 -4.62 -0.22 -0.59 0.13 -2.02 [0.08]* [0.06]* [0.49] [0.22] [0.04]** [0.05]** [0.61] [0.59] L_emp 0.06 0.20 -0.08 2.26 0.08 0.25 -0.07 2.46 [0.27] [0.16] [0.54] [0.22] [0.16] [0.08]* [0.57] [0.18] D_Foreign 0.60 0.75 0.29 -8.56 0.51 0.27 0.36 -6.55 [0.02]** [0.42] [0.52] [0.23] [0.06]* [0.77] [0.45] [0.37] D_State 0.02 0.40 -0.41 -15.03 -0.03 0.35 -0.51 -18.89 [0.91] [0.40] [0.39] [0.02]** [0.86] [0.47] [0.29] [0.00]*** D_Export -0.40 -0.30 -0.42 5.67 -0.47 -0.30 -0.28 8.38 [0.10] [0.70] [0.41] [0.46] [0.05]* [0.70] [0.58] [0.29] D_RE 0.14 0.24 0.29 -2.08 [0.32] [0.57] [0.31] [0.65] D_Fam_Inf -0.13 -0.05 -0.36 6.47 [0.47] [0.92] [0.34] [0.25] AP_d 1.50 1.44 [0.00]*** [0.00]*** AP_per 0.09 0.09 [0.00]*** [0.00]*** AP_discount 0.54 0.64 [0.25] [0.16] AP_days 0.13 0.14 [0.01]*** [0.00]*** Constant -3.26 -13.59 -1.14 40.54 -2.90 -13.48 0.44 52.36 [0.00]*** [0.00]*** [0.54] [0.15] [0.00]*** [0.00]*** [0.77] [0.06]* Observations 1,207 1,203 453 388 1,187 1,182 445 381 Pseudo R-squared 0.16 0.2 0.18 0.01 0.15 0.2 0.17 0.02 44 Table 8: The Financing of Trade Credit Supply: Is Trade Credit Demand More Important for Credit Constrained Firms? The reported estimates are from logit (Colums 1-4) and OLS (Columns 5-8) regressions. AR_d is a dummy indicating the use of accounts receivable and AR_per is the percent of monthly sales sold on credit. We transform AR_per in a continuous variable by using the inverse function of the logistic model. See Table 1 for variable definitions. All regressions include 12 sector dummies and 17 city dummies. Robust p-values are shown in parentheses, ***, **, * indicate significance at the 1%, 5%, and 10% level, respectively. (1) (2) (3) (4) (5) (6) (7) (8) AR_d AR_d AR_d AR_d AR_per AR_per AR_per AR_per L_age -0.13 -0.25 -0.18 -0.22 -0.69 -0.66 -0.54 -0.59 [0.11] [0.01]** [0.09]* [0.04]** [0.00]*** [0.02]** [0.06]* [0.05]* L_emp 0.06 0.08 0.05 0.08 0.27 0.28 0.19 0.25 [0.15] [0.11] [0.30] [0.16] [0.02]** [0.05]** [0.18] [0.09]* D_Foreign 0.39 0.49 0.61 0.51 0.21 0.45 0.78 0.36 [0.07]* [0.05]** [0.02]** [0.06]* [0.76] [0.59] [0.40] [0.70] D_State 0.03 -0.02 0.01 -0.03 0.10 -0.03 0.37 0.32 [0.85] [0.93] [0.94] [0.86] [0.79] [0.95] [0.43] [0.52] D_Export -0.33 -0.53 -0.39 -0.47 0.29 -0.51 -0.31 -0.30 [0.10]* [0.02]** [0.11] [0.05]* [0.65] [0.47] [0.69] [0.70] AP_d 1.40 1.53 1.72 1.44 [0.00]*** [0.00]*** [0.00]*** [0.00]*** AP_per 0.09 0.12 0.10 0.08 [0.00]*** [0.00]*** [0.00]*** [0.00]*** LC_unused 0.11 0.84 [0.78] [0.33] LC*AP_d 0.10 [0.84] LC*AP_per -0.01 [0.79] D_Bank 0.07 0.26 [0.68] [0.53] D_Bank *AP_d -0.33 [0.20] D_Bank *AP_per -0.06 [0.00]*** D_RE 0.38 0.59 [0.04]** [0.18] D_RE*AP_d -0.54 [0.05]** D_RE*AP_per -0.03 [0.16] D_Fam_Informal -0.11 -0.53 [0.66] [0.30] D_Fam_Inf*AP_d -0.03 [0.93] D_Fam_Inf*AP_per 0.06 [0.05]** Constant -2.97 -3.27 -3.44 -2.90 -13.2556 -13.5008 -13.606 -13.387 [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** Observations 1,881 1,299 1,207 1,187 1,869 1,291 1,203 1,182 Pseudo R-squared 0.13 0.15 0.17 0.15 0.18 0.19 0.2 0.2 45 Table 9: The Financing of Trade Credit Supply: Is Trade Credit Demand More Important in Competitive Markets? The reported estimates are from logit (Colums 1-2) and OLS (Columns 3-4) regressions. AR_d is a dummy indicating the use of accounts receivable and AR_per is the percent of monthly sales sold on credit. We transform AR_per in a continuous variable by using the inverse function of the logistic model. See Table 1 for variable definitions. All regressions include 12 sector dummies and 17 city dummies. Robust p- values are shown in parentheses, ***, **, * indicate significance at the 1%, 5%, and 10% level, respectively. (1) (2) (3) (4) AR_d AR_d AR_per AR_per L_Age -0.23 -0.3 -0.86 -0.72 [0.02]** [0.01]*** [0.00]*** [0.00]*** L_emp 0.09 0.09 0.38 0.30 [0.06]* [0.14]* [0.00]*** [0.01]*** D_Foreign 0.54 0.6 0.90 0.28 [0.04]** [0.03]** [0.26] [0.68] D_State -0.04 0.13 0.12 0.22 [0.82] [0.53] [0.77] [0.58] D_Export -0.23 -0.34 0.25 0.16 [0.32] [0.19] [0.73] [0.81] AP_d 1.22 1.34 [0.00]*** [0.00]*** AP_per 0.06 0.08 [0.00]*** [0.00]*** Saleslargest_5 0.03 0.15 [0.85] [0.66] Sales*AP_d 0.47 [0.05]* Sales*AP_per 0.03 [0.18] Bi_mktpower -0.43 -0.26 [0.07]* [0.59] Bi*AP_d 0.75 [0.02]** Bi*AP_per 0.04 [0.03]** Constant -2.77 -2.78 -11.58 -11.42 [0.00]*** [0.00]*** [0.00]*** [0.00]*** Observations 1,547 1,080 1,529 1,803 Pseudo R-squared 0.16 0.15 0.19 0.19 46 Table 10: Ex-post Matching of Trade Credit Terms The reported estimates are from multinomial logit regressions (AR_gap=0 is the base outcome). AR_gap is a dummy variable equal to 1 if the difference between the number of days offered to the firm by its suppliers less the number of days until the firm pays its suppliers is greater than zero (i.e. the firm prepays its payables); equal to 0 if the difference equals zero (i.e. the firm pays its suppliers on time); and equal to - 1 if the difference is less than zero (i.e. the firms pays its suppliers late). See Table 1 for variable definitions. All regressions include 12 sector dummies and 17 city dummies. Robust p-values are shown in parentheses, ***, **, * indicate significance at the 1%, 5%, and 10% level, respectively. (1) (2) (3) (4) (5) (6) AP_gap = ­ 1 = + 1 L_Age -0.05 -0.22 -0.23 -0.41 -0.62 -0.45 [0.82] [0.41] [0.29] [0.03]** [0.02]** [0.02]** L_emp 0.14 0.22 0.11 0.13 0.24 0.13 [0.19] [0.10] [0.31] [0.17] [0.06]* [0.16] D_Foreign 0.29 -0.23 0.41 -0.23 0.05 -0.39 [0.50] [0.68] [0.37] [0.58] [0.93] [0.37] D_State -0.13 -0.52 0.02 -0.39 -1.11 -0.34 [0.73] [0.27] [0.97] [0.27] [0.02]** [0.34] D_Export 0.19 0.27 0.06 -0.60 -1.59 -0.72 [0.70] [0.65] [0.91] [0.20] [0.01]*** [0.13] D1_AR_gap -0.23 -0.12 -0.23 0.56 0.99 0.41 [0.55] [0.81] [0.55] [0.06]* [0.01]** [0.18] D2_AR_gap -0.89 -1.28 -1.40 -0.48 -0.50 -0.62 [0.03]** [0.02]** [0.00]*** [0.17] [0.29] [0.08]* LC_unused 0.25 -0.41 [0.67] [0.45] D_RE -0.38 -0.56 [0.31] [0.12] AP_discount 0.40 0.89 [0.46] [0.04]** Constant -22.32 -0.43 -0.87 0.30 0.94 -0.91 [0.00]*** [0.80] [0.55] [0.79] [0.58] [0.56] Observations 583 374 569 583 374 569 Pseudo R-squared 0.23 0.28 0.23 0.23 0.28 0.23 47 Table 11: Other Determinants of Trade Credit Supply: Collateral Value, Customer Creditworthiness and Legal Institutions The reported estimates are from logit (Colums 1-3) and OLS (Columns 4-6) regressions. AR_d is a dummy indicating the use of accounts receivable and AR_per is the percent of monthly sales sold on credit. We transform AR_per in a continuous variable by using the inverse function of the logistic model. See Table 1 for all variable definitions. All regressions include 12 sector dummies and 17 city dummies. Robust p-values are shown in parentheses, ***, **, * indicate significance at the 1%, 5%, and 10% level, respectively. (1) (2) (3) (4) (5) (6) AR_d AR_d AR_d AR_per AR_per AR_per L_Age -0.13 -0.11 -0.14 -0.67 -0.65 -0.67 [0.14] [0.18] [0.13] [0.01]*** [0.00]*** [0.01]*** L_emp 0.04 0.05 0.07 0.25 0.25 0.33 [0.35] [0.28] [0.12] [0.04]** [0.03]** [0.01]** D_Foreign 0.43 0.39 0.34 0.34 0.20 0.39 [0.05]** [0.07]* [0.13] [0.63] [0.77] [0.60] D_State 0.01 0.01 0.00 0.18 0.14 0.17 [0.97] [0.96] [1.00] [0.66] [0.72] [0.68] D_Export -0.35 -0.31 -0.31 0.28 0.35 0.20 [0.09]* [012] [0.14] [0.67] [0.59] [0.76] AP_d 1.42 1.41 1.29 [0.00]** [0.00]*** [0.00]*** AP_per 0.09 0.091 0.08 [0.00]*** [0.00]*** [0.00]*** LC_unused 0.28 0.20 0.28 0.87 0.72 0.97 [0.30] [0.45] [0.31] [0.33] [0.39] [0.28] Uniqueness 0.00 -0.00 [0.68] [0.88] Contracts 0.43 0.53 [0.02]*** [0.23] Property_rights -0.00 -0.00 [0.57] [0.83] Constant -2.49 -3.28 -2.89 -12.60 -12.07 -11.62 [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** [0.00]*** Observations 1,766 1,865 1655 1,756 1,853 1,640 Pseudo R-squared 0.13 0.13 0.12 0.18 0.18 0.18 48 Table 12: Summary Statistic on The Implicit Cost of Trade Credit The table presents summary statistics for the implicit rate of trade credit (AR_cost), the bank interest rate (Bank_cost), the sum of bank interest rate and annual fee (Bank_cost_fee) and the difference between the rate of trade credit and the two bank interest rates, namely (Diff_AR_Bank) and (Diff_AR_Bank_fee). The numbers are percentages. To calculate (AR_cost) we use the following expression: 360 /(#daysnet-#daysdiscount) AR _ cost = 100 - 100 -1 discount% 75%of Obs Mean Std. 5% of the 10% of 25% of the 50% of the Dev. sample the sample sample sample the Min Max sample AR_cost 133 167.74 540.65 2.03 4.10 10.16 27.43 64.91 1.20 3600.00 Bank_cost 81 6.30 7.79 0.60 4.00 5.00 5.80 6.31 0.01 74.00 Bank_cost_fee 53 10.90 15.30 2.76 5.24 5.85 7.00 9.00 1.60 104.00 Diff_AR_Bank 81 215.19 668.98 -4.45 -0.19 6.10 23.43 78.06 -5.36 3624.00 Diff_AR_Bank_fee 53 139.84 528.40 -12.45 -4.98 0.45 21.42 57.91 -30.18 3619.00 49