Trade Payables and Shareholder Wealth: Evidence from North Korean Shock on South Korean Companies*

Hocheol Nam Graduate School of Economics, Kyushu University 6-19-1, Hakozaki, Higashiku Fukuoka 812-8581 JAPAN

Konari Uchida** Faculty of Economics, Kyushu University 6-19-1, Hakozaki, Higashiku Fukuoka 812-8581 JAPAN

Abstract We find that the portfolio of Korean firms with small accounts payable and large accounts receivable experiences negative excess returns when investors learn that North Korea sank a South Korean warship in May 2010. The negative effect of small accounts payable (large accounts receivable) is especially evident for small and highly leveraged companies as well as for non-subsidiaries (parent companies). The results suggest that trade credits provide financially constrained and unhealthy firms with insurance against negative shocks. To the best of our knowledge, this is the first research to provide evidence that trade payables have favorable effects on shareholder value. Key Words: Trade credits; Shareholder wealth; Event study; Korea JEL Classification: G14; G32

*

An early version of this paper was presented at the Japan Finance Association conference and the 12th International Conference on Asian Financial Markets and Economic Development. We thank Jungwook Sim for his helpful comments. We are grateful for financial support provided by JSPS KAHENHI Grant Number 15H03367. ** Corresponding author. Faculty of Economics, Kyushu University 6-19-1, Hakozaki, Higashiku, Fukuoka 8128581 JAPAN. Tel.: +81-92-642-2463 E-mail: [email protected] 1

This paper investigates the relation between trade payables and stock price reaction to news of a negative external shock. Hill, Kelly, and Lockhart (2012) argue that accounts receivables, which represent 18% of total assets of US manufacturing companies, are significantly related to shareholder wealth. As for the borrower (customer) side, previous studies stress that trade credits are one of many financing measures rather than a simple tactic to delay payment (Meltzer, 1960; Atanasova, 2007). An important feature of trade credits is that lenders (suppliers) can closely monitor borrowers (customers) over the course of business, and thus information asymmetry between them is significantly reduced (Biais and Gollier, 1997; Petersen and Rajan, 1997). Accordingly, trade payables serve as an important financing source for financially constrained firms. In addition, trade credits tend to build on long-term relations between suppliers and clients, and suppliers have an incentive to rescue financially distressed clients to prevent violation of valuable relationships (Cuñat, 2007). These ideas give rise to the prediction that accounts payables serve as a substitute for bank debt (Meltzer, 1960). In fact, Atanasova (2007) and Cuñat (2007) show evidence that lowcredit-quality companies rely on trade credits, especially when they cannot access institutional loans. Financially constrained companies are likely to favor trade credits when the government tightens monetary policy because banks tend to significantly decrease the loan supply during tight monetary periods (Kashyap, Stein, and Wilcox, 1993). Indeed, many previous studies show evidence to support this idea (Nilsen, 2002; Choi and Kim, 2005; De Blasio, 2005; Mateut, Bougheas, and Mizen, 2006; Atanasova, 2007). Previous studies also suggest that trade credit increases during liquidity shocks and financial crises (Cuñat, 2007; Garcia-Appendini and Montoriol-Garriga, 2013; Carbó-Valverde, Rodríguez-Fernández, and Udell, 2016). These previous studies commonly imply that accounts payables create significant value for financially constrained companies through information production and insurance. However, most empirical studies focus on determinants of firms’ reliance on accounts payable. To the best of our knowledge, only a few studies have reported evidence that trade payables increase shareholder wealth, while Hill, Kelly, and Lockhart (2012) find a positive relation between stock returns and the change in accounts receivable. This research attempts to fill this gap. A potential reason for the lack of previous studies is that an inverse relation likely exists between firm value and accounts payable since financially constrained companies tend to rely on trade credits. In addition, previous studies commonly suggest that trade credits are more expensive than other financing sources and, therefore, may offset their positive impacts on firm value (Ng, 2

Smith, and Smith, 1999). To overcome these problems, we examine daily stock price reactions to news of a negative external shock. Given that trade credits create additional value when the firm falls into constrained and/or distressed situations, stock price responses to unexpected negative shocks should be able to capture the value relevance of trade payables. We can also mitigate endogeneity concerns by using an unexpected event, which does not affect the preevent level of accounts payable. Specifically, we employ a North Korean shock which potentially harms Korean corporate performance as research material. On March 26, 2010, the Korean navy warship Cheonan went down in the West Sea, killing 46 sailors (see Table I for a series of events). Although the cause of the sinking was not identified in the following several months, the international joint investigation team announced on May 20 that a torpedo fired from a North Korean submarine sank the Cheonan. Then, the Korean government declared on May 24 that North Korea was responsible for sinking the Cheonan and announced economic sanctions against North Korea. The news of North Korea’s attack and economic sanctions likely increased investors’ fears about future war and retaliatory attacks, which potentially deteriorated Korean firm performance. In addition, investors were likely to require a higher risk premium for Korean firms’ securities due to increased uncertainty. Those potential negative impacts should have been incorporated in stock prices immediately after the news. Indeed, the Korean stock price index (KOSPI) declined by about 4.3% during the three trading days from May 20 to May 25, 2010 (May 21 is a holiday in Korea, and May 22 and 23 were weekend days). Using the North Korean shock for our research has several advantages. It offers a clear event day and enables us to examine the event’s impacts on shareholder value by tracing a few days’ stock returns. In contrast, research on financial crisis uses relatively long-term (e.g., one-year) returns to measure firm performance (Mitton, 2002; Lemmon and Lins, 2003). The North Korean shock was also unexpected, and Korean firms were less likely to take the shock into consideration when they determined the level of trade credits before the news. We can substantially mitigate endogeneity concerns by using this geopolitical event in our study. Furthermore, compared to other developed countries such as the US, UK, and Japan, the stock market and long-term firmbank relationships are not well developed in Korea. Instead, it is well known that family business groups are dominant in the Korean economy. Korean data are also advantageous for our research because trade credits serve as an important financing channel for Korean companies (Love, Preve, and Sarria-Allende, 2007). 3

[Insert Table I about here] We examine performance of portfolios formed by the pre-shock level of accounts payable by using the Fama-French three-factor model and Fama-French-Carhart four-factor model. Results suggest that the portfolio investing in firms with small accounts payable experiences a significant stock price (alpha) reduction from May 20 to May 25, 2010. This evidence highlights the uniqueness of trade credits since we find the opposite result for total liabilities. Specifically, the portfolio of firms with large total liabilities experiences significantly negative excess returns during the event period, probably due to high bankruptcy costs. Previous studies also argue that banks are informed investors that can monitor borrowers (Diamond, 1984; James, 1987). However, we do not find a significant relation between the level of bank debt and excess returns during the event period. The negative effect of small accounts payable is especially evident for small and highly leveraged companies. The result supports the view that trade payables are advantageous for financially constrained companies. We also find that the portfolio of firms with large accounts receivable significantly underperforms during the three-day event period. This negative effect of accounts receivable is also significant for small and highly leveraged firms. In addition, stock prices of firms with large accounts receivable do not significantly decline if the firms have large accounts payable. Taken together, our results show clear evidence that business suppliers provide insurance to financially constrained and unhealthy clients. Finally, we find that the effect of accounts payable (accounts receivable) is especially evident for non-subsidiaries (parent firms). Parent companies are likely to incur costs to extend trade credits to subsidiaries when negative external shocks occur. Meanwhile, we find no evidence that the results are mainly driven by business group companies. This research makes significant contributions to the literature. Although previous studies argue that trade payables provide an important financing channel to financially constrained firms, most empirical analyses are limited to determinants of firms’ reliance on accounts payable (Nilsen, 2002; Choi and Kim, 2005; De Blasio, 2005; Mateut, Bougheas, and Mizen, 2006; Atanasova, 2007; Cuñat, 2007; Garcia-Appendini and Montoriol-Garriga, 2013; CarbóValverde, Rodríguez-Fernández, and Udell, 2016). By using data of US companies during 1973 to 2006, Hill, Kelly, and Lockhart (2012) show evidence that stock returns are positively associated with the change in accounts receivable. To the best of our knowledge, this paper is the first to show direct evidence that trade payables affect value (avoid stock price reduction) 4

for constrained companies. In particular, our findings provide direct support for the view that trade payables serve as insurance for client companies (Wilner, 2000; Cuñat, 2007) by taking advantage of an unexpected negative shock. Paying attention to the stock price reaction to negative news, this research also finds new evidence that accounts receivable decrease shareholder wealth when negative shocks occur. Previous studies suggest that trade payables are advantageous for constrained companies during tight money periods and financial crises. We show evidence that trade payables also become beneficial when geopolitical risk is evident. Finally, our research adds to the literature of political risk by showing that the use of relationship-based financing mitigates the negative economic impacts of political risk (Chan and Wei, 1996; Chan, Chui, and Kwok, 2001; Amihud and Wohl, 2004). The remainder of this paper is organized as follows. Section I describes previous studies and our hypothesis. Section II presents our empirical methodology and data. Section III shows our main empirical results. Additional analyses are presented in Section IV. Finally, Section V offers a brief summary and the conclusion of our research.

I.

Literature review and hypotheses Trade credits have been viewed as one of firms’ financing sources (non-bank debt) (Meltzer,

1960; Biais and Gollier, 1997; Burkart and Ellingsen, 2004; Atanasova, 2007). An important feature of trade credit financing is that it builds on the relationship between suppliers and clients, which significantly decreases problems arising from information asymmetry. Dass, Kale, and Nanda (2015) find that relationship-specific investments (proxied by research and development (R&D) expenditures) of upstream firms are positively associated with trade credits. Given that suppliers can closely monitor clients over the course of business, trade credits serve as an important financing method, especially for financially constrained firms that do not have access to bank debt (substitution view) (Petersen and Rajan, 1995; Biais and Gollier, 1997). Atanasova (2007) provides evidence that financially constrained companies rely on trade credits when they cannot access institutional loans. Cuñat (2007) finds that firms without collateralized assets and with less liquidity use more trade credits. Once constrained companies receive trade credits, suppliers’ information is transmitted to banks, and those firms may get access to bank loans (Biais and Gollier, 1997).1 Thus, the nature of trade credits should affect the investment 1

This theoretical argument explains the fact that many companies use both bank debt and trade credits. 5

behaviors of constrained companies. In fact, Guariglia and Mateut (2006) find that internal funds (proxied by coverage ratio) do not affect inventory investments by UK financially constrained firms when those firms have large trade credits, although inventory investments of the average constrained firm show significant sensitivity to internal funds. These results suggest that trade credits support the financing of constrained companies. Generally, monetary tightening decreases the bank loan supply, especially to financially constrained companies. The literature has investigated whether trade credits absorb the reduction in bank loan supply during monetary tightening (Meltzer, 1960). Nilsen (2002) argues that small and large firms without bond ratings increase trade credits when the government tightens monetary policy. Choi and Kim (2005) find that accounts payable and receivable increase during monetary tightening. By using UK data, Mateut, Bougheas, and Mizen (2006) show evidence that bank loans decrease during a tight monetary policy period (1990-1992) and instead trade credits increase. Atanasova (2007) also finds that financially constrained UK firms rely more on trade credits during periods of tight money. Although financially constrained firms are generally forced to curtail investments by monetary tightening, the substitution role of trade credits absorbs the negative impact (Biais and Gollier, 1997). De Blasio (2005) reports evidence that trade credits are positively associated with investments in Italy when the government tightens monetary policy. Financial crisis also shrinks the monetary supply. Garcia-Appendini and Montoriol-Garriga (2013) and Carbó-Valverde, Rodríguez-Fernández, and Udell (2016) show evidence that creditconstrained firms tend to increase trade credits, especially during a financial crisis, while less constrained firms use bank debt. For firms in emerging markets, Love, Preve, and SarriaAllende (2007) find that the provision of trade credit increases right after a crisis, although it contracts in the following months and years.2 Love, Preve, and Sarria-Allende (2007) attribute the result to the accumulation of unpaid credit until suppliers take write-downs (or buyers resume payments). Given that it takes time to build a long-term business relationship, both creditors and suppliers want to maintain their relationship once it is established. Wilner (2000) postulates that trade credit suppliers can renegotiate with lenders with less cost and, thus, suppliers are likely to provide financially distressed clients with a moratorium to avoid

Burkart and Ellingsen (2004) theoretically argue that banks are willing to lend to firms that receive trade credits since availability of trade credits boosts firms’ investments rather than diversion. 2 Love, Preve, and Sarria-Allende (2007) find that bank debt also declines after financial crisis. 6

violation of their relationships. 3 Cuñat (2007) argues that suppliers provide clients with insurance against liquidity shocks. These facts suggest that trade credits are valuable for client firms, especially during a financial crisis.4 In sum, trade payables are likely to generate benefits (information production and insurance) to financially constrained companies, especially when liquidity shocks occur. To the best of our knowledge, however, only a few studies show that accounts payable increase shareholder wealth, while Hill, Kelly, and Lockhart (2012) show evidence that stock returns are positively related to the change in accounts receivable. There are two potential reasons for the lack of previous findings. First, a reverse causality problem exists in that poorly performing (and thus financially distressed) firms rely on accounts payable. Second, trade credits are generally considered more costly than bank debt for borrowing companies (Petersen and Rajan, 1994). For example, a common term of trade credits in Ng, Smith, and Smith’s (1999) sample is “2/10 net 30,” which combines a 2% discount for payment within 10 days and a net period ending on day 30 (implicit interest rate is 43.9%). Put differently, firms receiving trade credits incur high costs in exchange for the monitoring and insurance effects, which offset positive effects on shareholder value. In a similar vein, Wilner (2000) theoretically argues that trade credits are associated with low costs of renegotiation, and thus firms are willing to pay high interest rates on trade credits. We predict that trade payables are value-enhancing, especially when firms fall into a difficult situation. This research attempts to examine the relation between trade payables and the stock price response to an unpredicted negative external shock to uncover the value-creating effects of trade credits. Given the nature of trade credits, low-value companies are likely to rely on trade payables. In addition, firm characteristics (including unobservable characteristics) associated with the use of trade payables also affect firm value. These facts generate typical endogeneity problems when we implement cross-sectional analyses of firm value and accounts payable. An effective way to avoid the endogeneity concern is to trace stock price responses during a few days when unexpected negative shocks are announced. We propose the following

In addition, trade credits are less risky for suppliers than bank debt because suppliers can threaten to cut off future supplies to enforce repayment and easily repossess goods in case of failed payment (Petersen and Rajan, 1997; Kohler, Britton, and Yates, 2000). 4 Cuñat (2007) also finds that trade credits tend to increase when firms encounter unexpected liquidity shocks. By using survey data, Ng, Smith, and Smith (1999) find that firms adopting trade credits generally do not respond to fluctuations in market demands and interest rates. 3

7

hypothesis: Hypothesis: Trade payables are positively associated with stock price reactions to unpredicted negative shocks.

II. Methodology and data To capture value effects of trade payables, this research examines stock price reactions to negative external shocks. Numerous studies have examined few-day stock price reactions to specific corporate announcements, given the premise that stock prices immediately incorporate the value of the new information. In this research, we adopt a macro-level shock rather than firm-level event since firms announcing a specific event (in this case, an event related to a negative shock) are likely to share common characteristics, which may generate sample selection biases. Many previous studies focus on monetary tightening policy and financial crisis and attempt to examine the substitution relation between trade credits and bank debt. Given that we have interest in insurance effects as well as liquidity supply effects, we adopt an unexpected shock which increases the uncertainty of corporate economic performance. Specifically, we adopt the North Korean shock on South Korean companies. It is commonly recognized that South Korea is exposed to geopolitical risk. On March 26, 2010, the Korean navy warship Cheonan went down in the West Sea, killing 46 sailors. The cause of the sinking was not identified during the several following months, and then the international joint investigation team officially announced on May 20 that a torpedo fired from a North Korean submarine sank the Cheonan. In response to the announcement, the North Korean government declared on May 20 that the North would start all-out war against the South if the South imposed economic sanctions. Then, the South Korean government declared on May 24 that North Korea was responsible for sinking the Cheonan and announced economic sanctions against North Korea. This incident was likely to increase the uncertainty, at least temporarily, which Korean companies encountered. Korean firms might have had to stop (or decrease) production, logistics, and sales activities if North Korea launched a missile against South Korea and, thus, might have encountered a substantial decline in economic performance. The South Korea government was also concerned about an economic slowdown, which is illustrated by the government and Ministry of Strategy and Finance launching a special joint economic and 8

finance response team on May 21, 2010, to mitigate the shock to the South Korean market. In fact, tension between the two nations significantly increased after the sinking of the Cheonan. In November 2010, for example, North and South Korea fired at each other for about one hour on an island that sits off a disputed border. The North Korean shock has an important advantage in this research. This event was likely unpredicted, and investors suddenly became reminded of the geopolitical risk which Korean firms encounter and feared the possibility of future war. Indeed, the Korean currency (won) was depreciated to 1 USD = 1194.10 won on May 20, which is the highest level since October 29, 2009. Also, South Korea’s credit default swap premium recorded its highest level of the year after the announcement of the Cheonan investigation results (The Hankyoreh, May 26, 2010). The unpredicted nature significantly mitigates endogeneity problems because Korean firms are less likely to determine the level of trade credits in consideration of the shock. In addition, the North Korean shock offers a clear event day and enables us to examine its impacts on shareholder value by tracing a few days’ stock returns. The short-term event window is advantageous to capture pure effects of crisis, in contrast to financial crisis research which uses relatively long-term (e.g., one-year) returns (Mitton, 2002; Lemmon and Lins, 2003). We also stress that trade credits are likely to play an important role for Korean corporate finance. Compared to other developed economies such as Japan, the UK, and the US, the stock market and long-term bank-firm relationships are not well developed in Korea. Instead, business groups which have complex ownership structures, called chaebols, are dominant in the Korean economy. Business groups generally have the internal capital market, where trade credits may serve as an important financing channel. Given that the stock market incorporates the value of potential future events, we trace stock price reactions from May 20 to May 25, when investors initially learned that North Korea had sunk the Cheonan. Although the sinking occurred on March 26, KOSPI showed a 0.7% increase during the three days from March 26 to March 30, suggesting that investors were less likely to fear future wars and uncertainty at that time. In contrast, KOSPI declined by about 4.3% during the three trading days from May 20 to May 25, 2010. We argue that the value of trade credits should be priced during this period when investors recognize that the sinking was caused by North Korea. Since all firms share the event window, abnormal returns across firms are likely correlated. The conventional event study methodology may understate the standard error and lead to biased 9

statistical inference. Schwert (1981) and Campbell, Lo, MacKinlay, Adamek, and Viceira (1997) suggest examining returns of portfolios during event windows which invest in firms with specific characteristics to diversify away from this cross-sectional correlation. Recent studies on stock price impacts of a macro-level event commonly employ this approach.5 We adopt this approach and form portfolios by accounts payable over assets or a specific variable of interest at the end of June, year t, by using financial data during January to December, year t – 1. We invest in those portfolios until June of year t + 1, then rebalance the portfolio at the end of June, year t + 1. To capture the effect of the North Korean shock on stock price, we implement calendar-time portfolio regressions based on the Fama and French three-factor model and Fama-French-Carhart four-factor model, with an event window dummy (Cai and Walkling, 2011):

𝑹𝒑,𝒕 − 𝑹𝒇,𝒕 = 𝜶 + 𝜷𝟏 (𝑹𝒎,𝒕 − 𝑹𝒇,𝒕 ) + 𝜷𝟐 𝑺𝑴𝑩𝒕 + 𝜷𝟑 𝑯𝑴𝑳𝒕 + 𝜷𝟓 𝑬𝑽𝑬𝑵𝑻𝑫𝒕 + 𝒆𝒕 , 𝑹𝒑,𝒕 − 𝑹𝒇,𝒕 = 𝜶 + 𝜷𝟏 (𝑹𝒎,𝒕 − 𝑹𝒇,𝒕 ) + 𝜷𝟐 𝑺𝑴𝑩𝒕 + 𝜷𝟑 𝑯𝑴𝑳𝒕 + 𝜷𝟒 𝑴𝑶𝑴𝒕 + 𝜷𝟓 𝑬𝑽𝑬𝑵𝑻𝑫𝒕 + 𝒆𝒕 ,

where 𝑅𝑝,𝑡 is the value-weighted return of the portfolio, 𝑅𝑓,𝑡 is the risk-free rate, 𝑅𝑚,𝑡 is the market return, 𝑆𝑀𝐵 is the size factor return, 𝐻𝑀𝐿 is the book-to-market factor return, and MOM is the momentum factor return. The dummy variable 𝐸𝑉𝐸𝑁𝑇𝐷 equals 1 for the three trading days between May 20 and May 25, 2010, and 0 for all other dates (event dummy). We estimate the models by using daily stock return data covering from November 25, 2009, to November 25, 2010. We pay attention to the coefficient 𝛽5, which is the average daily excess return during the event window. We predict that the portfolios investing in firms with high accounts payable have larger 𝛽5 than the portfolios investing in low accounts payable companies. In the following part of the paper, we present only the results of the four-factor model because the three-factor model generates qualitatively the same results. We obtain daily stock price data as well as annual financial data from OSIRIS, provided by Bureau van Dijk. Financial institutions are not included in the analysis due to the different format of financial information. Firms are removed from the portfolio formation if stock return data are not available during the investment period. As a result, 1,219 (1,186) companies are adopted for portfolio formation at June 2009 (June 2010) and included in the portfolio from For instance, Cai and Walkling (2011) adopt this method to examine the effect of the Sarbanes-Oxley Act on stock prices. 5

10

July 2009 to June 2010 (from July 2010 to June 2011). Table II presents summary statistics of companies included in the portfolio formation at June 2009 (year 2008 financial data). Accounts payables occupy about 9.2% of total assets while accounts receivables account for 20% of assets. Accounts receivables are larger than payables probably because our sample consists of listed companies. The average (median) sample firm has five (two) subsidiaries. By using OSIRIS data, we define as subsidiaries firms with a corporate shareholder that owns more than 20% of the firm’s shares.6 About 22% of sample companies are identified as a subsidiary. These figures suggest that the parent-subsidiary relation is common among Korean listed companies. To identify affiliation with family business groups, we access the Online Provision of Enterprises Information System (OPNI), through which the South Korean government provides information on Korean family business groups.7 The OPNI identifies affiliations with family business groups by the existence of cross-shareholdings with affiliated companies. Approximately 12% of sample companies (145 firms) are identified as family business group companies. [Insert Table II about here]

III. Empirical results A. Accounts payable and stock price reaction to the North Korean shock To examine whether trade credits reduce negative impacts of the North Korean shock, we first investigate the excess returns of portfolios formed by accounts payable over total assets. Companies in our dataset are equally divided into three groups based on accounts payable over assets at the end of June 2009 and 2010, respectively, and included in the assigned portfolio from July of a year to June of the next year. Year 2008 (2009) accounting information is used for portfolio construction at June 2009 (2010). We conduct calendar-time portfolio regressions for those portfolios by using data from November 25, 2009, to November 25, 2010. Results are presented in Table III. Panel A indicates that the portfolio investing in firms with high accounts payable (High) does not experience a significant stock price reduction during the negative shock (the coefficient of EVENTD is insignificant). In contrast, the portfolio

Shareholders classified as industrial company in OSIRIS shareholder data are identified as corporate shareholders. 7 See http://groupopni.ftc.go.kr/ 6

11

consisting of firms with low accounts payable experiences a significant reduction of alpha when the North Korean shock is released to the public. Panel A also presents results for the difference portfolio (High minus Low), suggesting that the coefficient of the event dummy is significantly different between High and Low portfolios. This result supports our hypothesis that trade credits absorb damage from negative external shocks. [Insert Table III about here] To highlight the uniqueness of trade credits, we form portfolios by leverage (total liabilities over assets). Panel B of Table III suggests that the portfolio of highly leveraged firms (High) experiences significantly negative excess returns when news of the North Korea shock is released. The difference portfolio (High minus Low) has a negative and significant coefficient on EVENTD, suggesting that the stock market significantly depreciates highly leveraged companies. There are several interpretations of this result. The North Korean shock raises the probability that Korean firms encounter operating performance declines, which increases expected bankruptcy costs. The increased uncertainty may worsen firms’ financing conditions because investors may charge a high risk premium, and thus firms may be forced to curtail investments. This negative effect is likely serious for highly leveraged companies which do not have sufficient internal funds. These arguments highlight the positive aspect of trade credits. Although trade credits are a component of liabilities, they have the opposite effects on shareholder wealth as other liabilities. Hoshi, Kashyap, and Scharfstein (1990) suggest that Japanese firms with close relations with a bank decrease investments less than those without such relations. Gilson, John, and Lang (1990) show evidence that financially distressed firms with more bank debt are more likely to restructure debt privately (less likely to choose in-court bankruptcy). These facts motivate us to build portfolios formed by bank loans (scaled by assets). Panel C of Table III shows that bank debt is not a substitute for trade credits in terms of shareholder wealth effects. All three portfolios formed by bank loans carry an insignificant coefficient on EVENTD. Generally, Korean banks do not maintain close relationships with non-financial companies and, thus, do not behave like Japanese main banks. Table III suggests that trade credits significantly mitigate the impacts of negative external shocks on Korean companies. To further examine the effect of trade credits against a negative external shock, we formulate 2*2 portfolios which adopt accounts payable as one formation variable. Specifically, firms are equally divided into two groups based on a specific variable, and then each portfolio is further 12

divided into two portfolios based on accounts payable. Table III suggests that leveraged companies tend to experience significant stock price reductions at the release of news on the North Korean shock. We posit that trade payables mitigate the negative leverage effects because trade credits are a stable financing source for borrowers (Ng, Smith, and Smith, 1999). Panel A of Table IV indicates that the portfolio investing in firms with high leverage and low accounts payable (High-Low) experiences a significant reduction in alpha during the event period. This result suggests that non-trade credit liabilities harm shareholder wealth when negative external shocks occur. Meanwhile, stock prices of highly leveraged firms do not show a significant stock price reduction if the trade credits account for a significant portion of liabilities (High-High). This result provides clear evidence that trade payables significantly decrease costs associated with leverage. We also form portfolios by debt ratio (debt over assets) and trade credits. Again, Panel B suggests that trade payables mitigate the costs associated with debt financing. [Insert Table IV about here] Suppliers of trade credits can reduce information asymmetry about borrowers (Biais and Gollier, 1997; Petersen and Rajan, 1997). Accordingly, previous studies commonly suggest that small companies tend to rely on trade payables (Carbó-Valverde, Rodríguez-Fernández, and Udell, 2016). Given that small companies are generally subject to financing constraints as well as high default probability, trade payables should have large value impacts for small companies. Carbó-Valverde, Rodríguez-Fernández, and Udell (2016) find that small firms rely on trade credits during a financial crisis. We construct 2*2 portfolios by firm size (total assets) and accounts payable. Panel C indicates that the negative effect of small trade payables is more evident for small firms. Specifically, large companies do not experience a reduction of alpha at the event period, regardless of the level of trade payables. However, small companies with low accounts payable show a significant stock price decline at the time of the North Korean shock. Panel C also suggests that stock prices of small companies do not significantly decline if the companies have large accounts payable. Generally, firms with large intangible assets are subject to information asymmetry as well as to a lack of collateral for loans (Long and Malitz, 1985; Jensen, Solberg, and Zorn, 1992; Healy and Palepu, 2001; Anderson, Mansi, and Reeb, 2003). With a premise that those firms suffer from financing constraints, we construct 2*2 portfolios using intangibles assets (scaled by total assets) and accounts payable. Consistent with the view that trade credits are beneficial to 13

financially constrained companies, Panel D indicates that the portfolio of firms with large intangible assets and small accounts payable experience a significant stock price reduction. Taken together, the results support our hypothesis that trade credits generate benefits to distressed and/or constrained companies. As an additional test, we construct portfolios by using alternative measures of financial constraints and accounts payable: (a) cash and equivalents, (b) earnings after taxes, depreciation, and amortization, and (c) short-term debts (all variables are scaled by total assets). Untabulated results support our prediction that trade credits absorb the negative impacts of the North Korean shock on the shareholder value of distressed firms. B. Accounts receivable and stock price reaction to the North Korean shock We have shown evidence that firms relying on trade payables can mitigate stock price reductions due to negative external shocks. Cuñat (2007) indicates that suppliers have an incentive to provide financially distressed clients with liquidity to avoid violation of the business relationship. Meanwhile, negative external shocks may also harm suppliers’ liquidity status and their access to bank debt. Trade credit provision should be costly for suppliers at that time. Indeed, Love, Preve, and Sarria-Allende (2007) find that after the East Asian financial crisis, firms constrained in bank finance reduced credit extension to their customers in terms of quantity and length of time. We investigate whether firms with large accounts receivable experience stock price reductions during the event period to highlight characteristics of trade credits from the supplier side. First, we construct three portfolios solely by accounts receivable over assets. Panel A of Table V suggests that the portfolio investing in firms with large accounts receivable (High) experiences a significant decline in alpha during the event period (the EVENTD coefficient is significantly negative). Meanwhile, the portfolio of firms with small accounts receivable (Low) shows insignificant excess returns during the event period. The negative and significant coefficient on EVENTD for the difference portfolio indicates that firms with large accounts receivable underperform those with small trade receivables at the time of the North Korean shock. Trade credit suppliers incur costs when negative shocks come to the market. [Insert Table V about here] Panel B of Table V presents results of 2*2 portfolios of accounts receivable and payable. The result shows clear contrasts between accounts payable and receivables. Only the portfolio investing in firms with large accounts receivable and small payables (High-Low) shows a 14

significant stock price reduction during the event period. Even though firms have large accounts receivable (small accounts payable), stock prices of those firms do not show a negative reaction if the firms have large accounts payable (small accounts receivable). Put differently, the result suggests that wealth transfer exists from suppliers to customers when negative external shocks occur. Those results provide evidence that trade credits serve as insurance against negative shocks. Trade credit provision should be costly, especially for suppliers with poor liquidity status. Love, Preve, and Sarria-Allende (2007) find that firms with high short-term debt reduce the provision of trade credits when the aggregate bank credit supply shrinks due to financial crisis. Molina and Preve (2009) find that firms with financial difficulties decrease their trade receivables. With respect to stock price performance, Hill, Kelly, and Lockhart (2012) show evidence that leverage weakens the positive association between excess returns and increases of trade receivables. These findings motivate us to predict that the negative effect of large accounts receivable at the North Korean shock is especially evident for highly leveraged companies. Panel C of Table V presents regression results for the portfolio formed by firm size (leverage for Panel D) and accounts receivable. Small firms and highly leveraged companies are likely to suffer from poor liquidity and contracted access to bank debt when negative shocks occur. Panels C and D clearly show that small and highly leveraged firms incur significant costs associated with trade credit provision at the time of the North Korean shock. Specifically, the Low-High portfolio in Panel C and the High-High portfolio in Panel D have a significantly negative coefficient on EVENTD. This result is consistent with the findings of Love, Preve, and Sarria-Allende (2007), Molina and Preve (2009), and Hill, Kelly, and Lockhart (2012), suggesting that the insurance nature of trade credits creates a significant burden for small and highly leveraged suppliers. IV. Additional analysis A. Subsidiaries and trade credits Previous analyses suggest that trade credit providers experienced significant stock price reductions when the news of the North Korean shock hit the market. Generally, companies maintain long-term business relations with their subsidiaries, and parent companies likely 15

provide trade credits to subsidiaries. Parent companies are also less likely to rely on trade payables because they are large and less subject to information asymmetry, while it is difficult for subsidiaries to obtain access to public finance (e.g., bond issue) and even to bank debt. Once negative shocks occur, investors can expect companies to extend trade credits to their subsidiaries. To test the idea, we create 2*2 portfolios based on the number of subsidiaries and accounts receivable. Panel A of Table VI suggests that firms with many subsidiaries show significant reductions in alpha during the event period if they have large accounts receivable before the North Korean shock (High-High). We also construct 2*2 portfolios by investments (in balance sheet) over assets and accounts receivable (Panel B) to address the concern that the number of subsidiaries does not take the size of subsidiaries into consideration. Since investments include shareholdings of subsidiaries, we premise that firms with large investments are expected to provide financial supports to their subsidiaries. Consistent with this notion, Panel B of Table VI indicates that the portfolio of firms with large investments and large accounts receivable (High-High) shows significant underperformance during the event period, while the other three portfolios do not experience significant stock price reductions. Overall, our results support the idea that parent companies are expected to extend trade credits to subsidiaries when negative shocks occur. [Insert Table VI about here] Next, we examine the issue from the subsidiary side. Given the premise that subsidiaries receive trade credits from their parent companies, we construct portfolios based on the existence of a parent company and accounts payable. As mentioned, we identify as subsidiaries 268 sample firms (Table II). Panel C of Table VI indicates that the portfolio of independent firms (firms without a parent company) with small trade credits (Independent-Low) experiences significantly negative excess returns at the North Korean shock. In contrast, subsidiaries do not experience significant stock price reductions regardless of the level of trade credits. These results are consistent with our prediction that parent companies are likely to provide trade credits to subsidiaries when negative shocks occur. B. Business group It is well documented that business groups prevail in the Korean economy, in which affiliated firms have long-term business relationships. Corporate groups generally have internal capital 16

markets, in which trade credits are extensively provided. Meanwhile, business groups may provide affiliated companies with insurance in various forms against negative shocks. This fact raises a concern that our results on trade credits are biased by unobserved characteristics of business groups. To address this concern, we construct portfolios by family business group affiliation, which is available from OPNI data. Of 1,219 (1,186) companies included in portfolio formation at the end of June 2009 (2010), 145 (156) are identified as family group companies (Table II). Although we find no significant differences in accounts payable and accounts receivable (over assets) between family- and non-family group companies, family group companies are significantly larger than non-family firms (untabulated). In Panel A of Table VII, we construct two portfolios, one of which invests in family group companies and the other in non-family group companies. Panel A suggests that non-family business companies experience significant stock price reductions at the news of the North Korean shock, while stock prices of family business companies show insignificant responses. Although the result is consistent with the view that business groups provide affiliated companies with special insurance mechanisms (other than trade credits extension), we cannot rule out the possibility that family group companies do not suffer from significant stock price reduction since they are significantly larger than non-family group companies. Meanwhile, this result does not support the idea that expropriation of minority shareholder wealth, which is evident in business groups with complex ownership structures, becomes serious during a crisis period (Mitton, 2002; Lemmon and Lins, 2003; Baek, Kang, and Park, 2004).8 [Insert Table VII about here] Panel B of Table VII creates portfolios by family group affiliation and accounts payable. Results suggest that family group companies do not experience significant stock price reductions regardless of the level of trade credits. In contrast, the portfolio investing in nonfamily group companies with small accounts payable generates a negative and significant excess return during the event period. Panel C of Table VII presents results for portfolios formulated by family group affiliation and accounts receivable. Again, family group firms do

Baek, Kang, and Park (2004) show evidence that Korean firms with higher ownership concentration by unaffiliated foreign investors suffered less from deteriorating stock performance during the financial crisis of 1997. In a similar vein, Mitton (2002) finds that firms with higher outside ownership concentration showed better stock price performance in emerging markets during the crisis. Using a sample of 800 firms in eastern Asian countries, Lemmon and Lins (2003) find that crisis-period stock returns are 10-20 percentage points lower for firms in which managers had high levels of control rights but had separated their control and cash flow ownership. 8

17

not experience a significant stock price reduction during the crisis regardless of the level of accounts receivable. These results rule out the concern that our main findings are attributable to extensive use of trade credits by business group companies. C. Regression of Tobin’s Q We have so far argued that trade payables have positive effects on borrowing firms’ value when the North Korean shock occurs. To further examine the relation between trade payables and firm value, we implement regressions of Tobin’s Q by using 8,973 observations from 2007 to 2013, for which necessary data are available. Tobin’s Q is computed by the sum of market value of equity and book value of total liabilities divided by total assets. We premise that trade payables are positively associated with firm value for the year 2010 when the North Korean shock occurred and employ accounts payable (scaled by assets) and the interaction term with the year 2010 dummy as key independent variables. Given that firm characteristics associated with trade credits usage (e.g., financing constraints) are likely related to firm value, we employ the firm fixed effects model with year dummies to mitigate the endogeneity problem. We also add accounts receivable, firm size (natural logarithm of total assets), R&D expenses, return on assets (ROA) (earnings before interest and tax scaled by assets), and debt ratio (debt over assets) as well as cash holdings (cash and equivalents) as control variables. R&D expenses and cash holdings are scaled by assets. [Insert Table VIII about here] Models (1) through (3) of Table VIII indicate that accounts payables have a positive and significant coefficient. Although previous studies suggest that trade credits are more costly than bank debt (Petersen and Rajan, 1994; Ng, Smith, and Smith, 1999), the result suggests that trade credits enhance Korean firm value during the period under investigation. Importantly, the interaction term of accounts payable and the year 2010 dummy has a positive and significant coefficient in those models, suggesting that accounts payable incrementally create value for the year when the North Korean shock occurred. As a further analysis, model (4) adds the interaction term of accounts payable and the year 2008 dummy because the global financial crisis occurred in the analysis period. Model (4) carries a positive coefficient on the interaction term with the year 2008 dummy, although the significance level is marginal. Consistent with our main argument, this result suggests that accounts payable create additional value during the global financial crisis, when many companies are likely to suffer from liquidity constraints. 18

Importantly, model (4) also carries a positive and significant coefficient on the interaction term of accounts payable and the year 2010 dummy. To further examine the trade payables’ effects on firm value, we divide the sample equally into two groups based on asset size (Models (5) and (6)) and debt over assets (Models (7) and (8)), and then implement the regression analyses separately for the subsamples.9 Trade credits are generally beneficial, especially for small companies which suffer from financial constraints (Nilsen, 2002; Carbó-Valverde, Rodríguez-Fernández, and Udell, 2016). The regression for small companies (Model (5)) yields a positive and significant coefficient on the interaction term of accounts payable with the 2010 dummy, which is consistent with our hypothesis. Meanwhile, accounts payable do not significantly increase the value of large companies for the year 2010 (Model (6)). We also premise that trade payables are beneficial for highly leveraged companies given that insurance effects of trade credits decrease bankruptcy costs. However, the regression for high debt ratio companies (Model (8)) carries an insignificant coefficient on accounts payable while the variable has a positive and significant coefficient in the regression for low debt ratio companies (Model (7)). A potential interpretation is that the high interest rates charged on trade credits place a heavy burden on highly leveraged companies in normal situations (Ng, Smith, and Smith, 1999). Importantly, Model (8) carries a positive and significant coefficient on the interaction term of accounts payable and the year 2010 dummy, suggesting that trade credits become value-enhancing for highly leveraged companies when firms encounter a negative shock. We do not find such an effect for low-leveraged companies. Meanwhile, Table VIII does not show a significant coefficient on accounts receivable and the interaction term with the year 2010 dummy. With respect to other control variables, ROA has a positive and significant coefficient. Cash holdings are positively associated with Tobin’s Q. Liquidity is important for Korean firms during the period, while free cash flow problems are less serious. We do not find a significant coefficient on other control variables. D. Other negative shocks We have so far examined stock price reactions from May 20 to 25, 2010, when investors recognized that North Korea had sunk the Cheonan. As a further analysis, we implement the

We do not use leverage (liabilities over assets) as a sample division variable because it is potentially correlated with accounts payable over assets. 9

19

same analyses by using other news on North Korea’s provocative behaviors. On May 25, 2009, North Korea announced that it had successfully conducted an underground nuclear test.10 For the three days starting on that day, KOSPI declined by 2.99%. We formulate three portfolios by trade payables (scaled by assets) and implement calendar-time portfolio regressions (fourfactor model) during the period from November 27, 2008, to November 27, 2009. The regression carries a negative coefficient on EVENTD for the low accounts payable portfolio, while the high accounts payable portfolio has a positive coefficient on EVENTD. Although those coefficients are statistically insignificant, the EVENTD coefficient for the difference portfolio is positive and significant at the 10% level. Also, the three-factor model regression carries a significant coefficient on EVENTD for the low accounts payable portfolio and the difference portfolio. These results are consistent with our argument that trade payables are valuable for borrowers when negative external shocks hit the market. We also conduct the calendar-time portfolio regression for the following provocative events by North Korea: (a) firing shells at the South Korean island of Yeonpyeong on November 23, 2010, which killed two Korean soldiers, (b) firing a long-range rocket (internationally seen as a disguised test of long-range missile technology banned under United Nations resolutions) on April 13, 2012, (c) nuclear test on February 12, 2013, in defiance of international bans, and (d) missile test on May 20, 2013. One-year data surrounding the event day are used for estimation. Most events carry a negative coefficient on EVENTD for the low accounts payable portfolio while the high accounts payable portfolio has a positive coefficient on EVENTD. Although the coefficients are not significantly different from zero, the signs are consistent with our main finding. Insignificant coefficients are attributable to the fact that North Korea’s serial provocations became less surprising for investors. Indeed, KOSPI did not decline substantially during the three days starting with the new release day. The maximum of the three-day KOSPI reduction for those events is 0.83% for firing shells at Yeonpyeong (from November 23 to 25, 2010). For the last two events, KOSPI increased over the three days. V. Conclusion This paper investigates the relation between trade payables and stock price reactions to news of a negative external shock. Previous studies suggest that trade credits build on long-term 10

It was the second nuclear test. The first test was conducted in 2006. 20

relationships between suppliers and customers, and problems arising from information asymmetry are substantially mitigated (Biais and Gollier, 1997; Petersen and Rajan, 1997). In addition, suppliers have an incentive to renegotiate with distressed clients to prevent violations of valuable relationships. However, most empirical studies focus on determinants of firms’ reliance on trade payables and, to the best of our knowledge, only a few studies have shown evidence that accounts payable affect shareholder wealth. This research attempts to fill this void. Specifically, we investigate excess returns of portfolios, formed by the level of accounts payable, when news of the North Korean shock was released in May 2010. This approach is advantageous in mitigating potential endogeneity problems. Results suggest that the portfolio investing in firms with small accounts payable experiences a significant stock price (alpha) reduction from May 20 to May 25, 2010. We also find that the portfolio of firms with large accounts receivable experiences a significant stock price reduction during the three-day event period. The negative effect of small accounts payable (large accounts receivable) is especially evident for small and highly leveraged companies as well as for non-subsidiaries (parent companies). Firms with large accounts receivable (small accounts payable) do not experience negative excess returns if they have large accounts payable (small accounts receivable). Taken together, our results show clear evidence that business suppliers provide insurance to financially constrained and unhealthy clients. This research makes significant contributions to the literature. To the best of our knowledge, this paper is the first to provide direct evidence that trade payables affect shareholder value (avoid stock price reduction) for constrained companies. Specifically, our findings provide direct support for the view that trade credits serve as insurance for client companies (Wilner, 2000; Cuñat, 2007). We also provide new evidence that accounts receivables decrease shareholder wealth when negative external shocks occur. Previous studies suggest that trade credits provision is costly for distressed companies (Love, Preve, and Sarria-Allende, 2007; Molina and Preve, 2009; Hill, Kelly, and Lockhart, 2012). We reinforce this argument by showing evidence that highly leveraged companies providing substantial trade credits suffer from significant stock price reductions when negative shocks hit the market. Our results also suggest that trade payables become beneficial when geopolitical risk is evident. Finally, our research adds to the literature of political risk by showing that the use of relationship-based financing mitigates negative economic impacts of political risk (Chan and Wei, 1996; Chan, Chui, and Kwok, 2001; Amihud and Wohl, 2004). 21

References Amihud, Y. and A. Wohl, 2004, “Political News and Stock Prices: The Case of Saddam Hussein Contracts,” Journal of Banking and Finance 28(5), 1185-1200. Anderson, R. C., S.A. Mansi, and D. M. Reeb, 2003, “Founding Family Ownership and the Agency Cost of Debt,” Journal of Financial Economics 68(2), 263-285. Atanasova, C., 2007, “Access to Institutional Finance and the Use of Trade Credit,” Financial Management 36(1), 49-67. Baek, J., J. Kang, and K. Park, 2004, “Corporate Governance and Firm Value: Evidence from the Korean Financial Crisis,” Journal of Financial Economics 71(2), 265-313. Biais, B. and C. Gollier, 1997, “Trade Credit and Credit Rationing,” Review of Financial Studies 10(4), 903-937. Burkart, M. and T. Ellingsen, 2004, “In-kind Finance: A Theory of Trade Credit,” American Economic Review 94(3), 569-590. Cai, J. and R. A. Walkling, 2011, “Shareholders' Say on Pay: Does It Create Value?” Journal of Financial and Quantitative Analysis 46(2), 299-339. Campbell, J. Y., A. W. Lo, C. A. MacKinlay, P. Adamek, and L. M. Viceira (Eds.), 1997, The Econometrics of Financial Markets, Princeton, NJ, Princeton University Press. Carbó-Valverde, S., F. Rodríguez-Fernández, and G. F. Udell, 2016, “Trade Credit, the Financial Crisis, and SME Access to Finance,” Journal of Money, Credit and Banking 48(1), 113-143. Chan, Y., C. W. Chui, and C. Y. Kwok, 2001, “The Impact of Salient Political and Economic News on the Trading Activity,” Pacific-Basin Finance Journal 9(3), 195-217. Chan, Y. and K. C. J. Wei, 1996, “Political Risk and Stock Price Volatility: The Case of Hong Kong,” Pacific-Basin Finance Journal 4(2-3), 259-275. Choi, W. and Y. Kim, 2005, “Trade Credit and the Effect of Macro-financial Shocks: Evidence from U.S. Panel Data,” Journal of Financial and Quantitative Analysis 40(4), 897-925. Cuñat, V., 2007, “Trade Credit: Suppliers as Debt Collectors and Insurance Providers,” Review of Financial Studies 20(2), 491-527. Dass, N., J. R. Kale, and V. Nanda, 2015, “Trade Credit, Relationship-specific Investment, and Product Market Power,” Review of Finance 19(5), 1867-1923. De Blasio, G., 2005, “Does Trade Credit Substitute Bank Credit? Evidence from Firm-level Data,” Economic Notes: Review of Banking, Finance and Monetary Economics 34(1), 8522

112. Diamond, D. W., 1984, “Financial Intermediation and Delegated Monitoring,” Review of Economic Studies 51(3), 393-414. Garcia-Appendini, E. and J. Montoriol-Garriga, 2013, “Firms as Liquidity Providers: Evidence from the 2007-2008 Financial Crisis,” Journal of Financial Economics 109(1), 272-291. Gilson, S. C., K. John, and L. H. P. Lang, 1990, “Troubled Debt Restructurings: An Empirical Study of Private Reorganization of Firms in Default,” Journal of Financial Economics 27(2), 315-353. Guariglia, A. and S. Mateut, 2006, “Credit Channel, Trade Credit Channel, and Inventory Investment: Evidence from a Panel of UK Firms,” Journal of Banking and Finance 30(10), 2835-2856. Healy, P. M. and K. G. Palepu, 2001, “Information Asymmetry, Corporate Disclosure, and the Capital Markets: A Review of the Empirical Disclosure Literature,” Journal of Accounting and Economics 31(1-3), 405-440. Hill, M. D., G. W. Kelly, and G. B. Lockhart, 2012, “Shareholder Returns from Supplying Trade Credit,” Financial Management 41(1), 255-280. Hoshi, T., A. Kashyap, and D. Scharfstein, 1990, “The Role of Banks in Reducing the Costs of Financial Distress in Japan,” Journal of Financial Economics 27(1), 67-88. James, C., 1987, “Some Evidence on the Uniqueness of Bank Loans,” Journal of Financial Economics 19(2), 217-235. Jensen, G. R., D. P. Solberg, and T. S. Zorn, 1992, “Simultaneous Determination of Insider Ownership, Debt, and Dividend Policies,” Journal of Financial and Quantitative Analysis 27(2), 247-263. Kashyap, A. K., J. C. Stein, and D. W. Wilcox, 1993, “Monetary Policy and Credit Conditions: Evidence from the Composition of External Finance,” American Economic Review 83(1), 78-98. Kohler, M., E. Britton, and T. Yates, 2000, “Trade Credit and the Monetary Transmission Mechanism,”

Bank

of

England

Working

Paper

115.

Available

at

SSRN:

http://ssrn.com/abstract=234693. Lemmon, M. L. and K. V. Lins, 2003, “Ownership Structure, Corporate Governance, and Firm Value: Evidence from the East Asian Financial Crisis,” Journal of Finance 58(4), 1445-1468. Long, M. S. and I. B. Malitz, 1985, “Investment Patterns and Financial Leverage,” in B.F. 23

Friedman, Ed., Corporate Capital Structures in the United States, Chicago, IL, University of Chicago Press. Love, I., L. A. Preve, and V. Sarria-Allende, 2007, “Trade Credit and Bank Credit: Evidence from Recent Financial Crises,” Journal of Financial Economics 83(2), 453-469. Mateut, S., S. Bougheas, and P. Mizen, 2006, “Trade Credit, Bank Lending and Monetary Policy Transmission,” European Economic Review 50(3), 603-629. Meltzer, A. H., 1960, “Mercantile Credit, Monetary Policy, and Size of Firms,” Review of Economics and Statistics 42(4), 429-437. Mitton, T., 2002, “A Cross-firm Analysis of the Impact of Corporate Governance on the East Asian Financial Crisis,” Journal of Financial Economics 64(2), 215-241. Molina, C. A. and L. A. Preve, 2009, “Trade Receivables Policy of Distressed Firms and Its Effect on the Costs of Financial Distress,” Financial Management 38(3), 663-686. Ng, C. K., J. K. Smith, and R. L. Smith, 1999, “Evidence on the Determinants of Credit Terms Used in Interfirm Trade,” Journal of Finance 54(3), 1109-1129. Nilsen, J. H., 2002, “Trade Credit and the Bank Lending Channel,” Journal of Money, Credit and Banking 34(1), 226-253. Petersen, M. A. and R. G. Rajan, 1994, “The Benefits of Lending Relationships: Evidence from Small Business Data,” Journal of Finance 49(1), 3-37. Petersen, M. A. and R. G. Rajan, 1995, “The Effect of Credit Market Competition on Lending Relationships,” Quarterly Journal of Economics 110(2), 407-443. Petersen, M. A. and R. G. Rajan, 1997, “Trade Credit: Theories and Evidence,” Review of Financial Studies 10(3), 661-691. Schwert, G. W., 1981, “The Adjustment of Stock Prices to Information about Inflation,” Journal of Finance 36(1), 15-29. Wilner, B. S., 2000, “The Exploitation of Relationships in Financial Distress: The Case of Trade Credit,” Journal of Finance 55(1), 153-178.

24

Table I A series of events associated with the North Korean shock This table presents a series of events associated with the North Korean shock. The information is available from the News Library of the Ministry of National Defense Republic of Korea website (http://www.mnd.go.kr/mbshome/mbs/mnd_eng/) and Ministry of Unification website (http://eng.unikorea.go.kr/), Ministry of Foreign Affairs Republic of Korea (http://www.mofa.go.kr/ENG/), and Ministry of Strategy and Finance (http://english.mosf.go.kr/). March 26, 2010 Korean navy warship Cheonan went down in the West Sea (37°55'45"N, 124°36'02"E (southwest of Baengnyeong Island)); 46 South Korean sailors died. May 20, 2010 The international joint investigation team (consisting of 24 members from South Korea, Australia, US, Sweden, and England) officially announced that a torpedo fired from a North Korean submarine sank the Cheonan. May 21, 2010 The South Korea government and Ministry of Strategy and Finance launched a special joint response team for economics and finance to mitigate the shock in the South Korean market from the Cheonan incident (http://english.mosf.go.kr/). The North Korean government declared that it would start all-out war against South Korea if the South imposed economic sanctions. May 24, 2010 President Lee Myung-bak released a statement to the nation, declaring that North Korea was responsible for sinking the Cheonan and announced the following economic sanctions against North Korea: (1) Prohibit North Korean ships from sailing in South Korean territorial waters (2) Cease all North-South trade (3) Ban South Koreans’ visits to North Korea (4) Prohibit investment by South Koreans in North Korea (5) Cease aid projects The Ministry of Unification South Korea announced that the South Korean government has reported the Cheonan incident to the United Nations Security Council. The Ministry of National Defense Republic of Korea declared that the ministry will undertake the following acts: (1) Broadcasting propaganda and anti-Kim Jong-il messages to North Korea (2) Blocking sea lanes to North Korea (3) Holding large-scale military drills with US forces in the West Sea in June 2010

25

Table II Summary statistics This table shows summary statistics for 1,219 companies which are included in our portfolio formation at the end of June 2009. Year 2008 financial data are presented. Accounts payable, bank loans, accounts receivable, and investments are scaled by total assets. Debt ratio is defined as the sum of loans, other short-term debt, and total long-term interest-bearing debt divided by the book value of total assets. Leverage is computed by the book value of total liabilities divided by the book value of total assets. Subsidiary dummy is a binary variable which takes the value of 1 for companies with a corporate shareholder who owns more than 20% of shares of the firm. Family business dummy is a binary variable which takes the value of 1 for companies affiliated with a family business group. We obtain family group affiliation for individual companies from the Online Provision of Enterprises Information System (OPNI).

Accounts payable/assets Bank loans/assets Debt ratio Accounts receivable/assets Leverage Total assets (billion won) Number of subsidiaries Investments Subsidiary dummy Family business dummy

Mean

Std. Dev.

Minimum

Median

Maximum

N

0.092 0.049 0.240 0.200 0.486 1214.442 4.335 0.035 0.221 0.119

0.080 0.070 0.184 0.121 0.247 6585.309 8.792 0.062 0.415 0.324

0 0 0 0.002 0.009 5.020 0 0 0 0

0.073 0.018 0.233 0.178 0.485 97.348 2 0.014 0 0

0.547 0.477 0.846 0.906 3.908 105829.3 173 0.732 1 1

1219 1219 1219 1218 1219 1219 1218 1218 1214 1219

26

Table III Calendar time portfolio regression: Leverage variable portfolios This table presents calendar-time portfolio regression results for portfolios formed by a leverage variable. Firms in our dataset are divided into three portfolios based on accounts payable over assets (Panel A), leverage (total liabilities over assets) (Panel B), and bank loans over assets (Panel C) at the end of June 2009 and 2010, respectively. These firms are included in the assigned portfolio from July of a first year to June of the next year. Year 2008 (2009) accounting information is used for portfolio construction at June 2009 (2010). The regression uses portfolio data from November 25, 2009, to November 25, 2010 (251 trading days), and this table presents results from the Fama-French-Carhart four-factor model, with an event dummy (EVENTD). EVENTD takes a value of 1 for the three trading days from May 20 to May 25, 2010, and 0 for all other dates. Difference indicates the difference portfolio whose return is High portfolio return minus Low portfolio return. T-statistics are reported in parentheses. Rm –Rf

SMB

HML

Mom

EVENTD

Alpha

R2

N

0.546 -0.572 (11.36)*** (-8.44)*** Middle 1.043 0.012 (31.60)*** (0.26) High 0.973 0.161 (29.75)*** (3.50)*** Difference 0.427 0.733 (7.09)*** (8.63)*** Panel B: Leverage portfolio

-1.582 (-34.43)*** -0.041 (-1.31) -0.114 (-3.66)*** 1.468 (25.49)***

-0.055 (-1.61) -0.114 (-4.88)*** 0.025 (1.06) 0.079 (1.86)*

-0.024 (-5.42)*** -0.003 (-0.99) -0.000 (-0.13) 0.024 (4.26)***

-0.008 (-7.63)*** 0.000 (0.64) -0.000 (-0.49) 0.008 (5.82)***

0.975

251

0.818

251

0.792

251

0.959

251

Low

-0.216 (-8.24)*** 0.035 (1.42) -1.549 (-36.22)*** -1.333 (-28.20)***

-0.017 (-0.90) 0.004 (0.24) -0.036 (-1.15) -0.019 (-0.54)

-0.005 (-1.80)* -0.000 (-0.01) -0.025 (-5.93)*** -0.020 (-4.36)***

-0.006 (-9.30)*** -0.001 (-1.72)* -0.006 (-6.27)*** -0.001 (-0.51)

0.758

251

0.860

251

0.978

251

0.972

251

-1.707 (-6.07)*** 0.005 (0.27) -0.097 (-2.36)** 1.610 (5.63)***

0.521 (2.50)** -0.045 (-3.04)*** -0.095 (-3.12)*** -0.616 (-2.91)***

-0.030 (-1.07) -0.000 (-0.05) -0.005 (-1.22) 0.025 (0.87)

-0.013 (-1.92)* -0.001 (-1.51) -0.000 (-0.33) 0.012 (1.83)*

0.493

251

0.906

251

0.707

251

0.476

251

Panel A: Accounts payable portfolio Low

0.724 0.301 (26.42)*** (7.78)*** Middle 0.959 -0.077 (36.82)*** (-2.08)** High 0.618 -0.627 (13.82)*** (-9.95)*** Difference -0.106 -0.928 (-2.15)** (-13.32)*** Panel C: Bank loan portfolio Low 0.145 -0.287 (0.49) (-0.69) Middle 0.972 -0.032 (46.38)*** (-1.10) High 1.001 0.093 (23.28)*** (1.53) Difference 0.856 0.380 (2.86)*** (0.90)

27

Table IV Calendar time portfolio regression: 2*2 portfolios by accounts payable This table presents calendar-time portfolio regression results for 2*2 portfolios. Firms in our dataset are divided into two portfolios based on leverage (total liabilities over assets) (Panel A), debt ratio (debt over assets) (Panel B), total assets (Panel C), and intangible assets over assets (Panel D) at the end of June 2009 and 2010, respectively. Then, each portfolio is further divided into two portfolios based on accounts payable over assets. These firms are included in the assigned portfolio from July of a year to June of the next year. Year 2008 (2009) accounting information is used for portfolio construction at June 2009 (2010). The regression uses portfolio data from November 25, 2009, to November 25, 2010 (251 trading days), and this table presents results from the Fama-French-Carhart four-factor model, with an event dummy (EVENTD). EVENTD takes a value of 1 for the three trading days from May 20 to May 25, 2010, and 0 for all other dates. T-statistics are reported in parentheses. Rm –Rf

SMB

HML

Mom

EVENTD

Alpha

R2

N

0.869 0.026 -0.018 -0.017 (22.47)*** (0.47) (-0.50) (-0.63) Low-High 1.010 0.040 -0.045 -0.022 (25.00)*** (0.70) (-1.15) (-0.78) High-Low 0.610 -0.621 -1.554 -0.038 (13.96)*** (-10.09)*** (-37.19)*** (-1.23) High-High 1.032 0.202 -0.128 -0.122 (19.19)*** (2.66)*** (-2.48)** (-3.19)*** Panel B: Debt ratio and accounts payable portfolio (2*2 portfolio)

-0.002 (-0.58) 0.004 (1.00) -0.025 (-6.00)*** -0.000 (-0.10)

-0.003 (-3.35)*** -0.000 (-0.10) -0.007 (-6.83)*** 0.001 (0.57)

0.689

251

0.729

251

0.979

251

0.609

251

Low-Low

-0.001 (-0.36) -0.000 (-0.07) -0.024 (-5.49)*** -0.004 (-0.87)

-0.001 (-1.71)* 0.000 (0.37) -0.008 (-7.29)*** 0.000 (0.37)

0.772

251

0.771

251

0.976

251

0.668

251

-1.755 -0.042 -0.020 (-34.92)*** (-1.13) (-3.98)*** -0.620 0.005 -0.003 (-13.50)*** (0.16) (-0.60) -0.002 -0.055 -0.003 (-0.07) (-3.10)*** (-1.27) -0.105 -0.096 -0.002 (-2.54)** (-3.14)*** (-0.39) payable portfolio (2*2 portfolio)

-0.009 (-7.39)*** -0.003 (-2.33)** -0.002 (-3.07)*** 0.001 (0.62)

0.970

251

0.677

251

0.865

251

0.704

251

-0.002 -0.000 0.750 (-0.60) (-0.54) -0.004 -0.001 0.567 (-0.76) (-0.63) -0.025 -0.008 0.975 (-5.58)*** (-7.09)*** -0.005 0.001 0.635 (-0.96) (1.18) Significant at the 10% level

251

Panel A: Leverage and accounts payable portfolio (2*2 portfolio) Low-Low

0.926 -0.019 0.011 -0.033 (27.64)*** (-0.40) (0.33) (-1.38) Low-High 1.003 0.157 -0.124 -0.035 (28.07)*** (3.12)*** (-3.62)*** (-1.39) High-Low 0.581 -0.592 -1.572 -0.036 (12.44)*** (-8.99)*** (-35.18)*** (-1.08) High-High 1.009 0.086 -0.056 -0.082 (21.49)*** (1.30) (-1.25) (-2.46)** Panel C: Size and accounts payable portfolio (2*2 portfolio) Low-Low

0.525 (9.99)*** Low-High 0.921 (19.17)*** High-Low 0.933 (37.37)*** High-High 1.023 (23.61)*** Panel D: Intangible assets

-0.301 (-4.07)*** 0.910 (13.44)*** -0.052 (-1.48) 0.161 (2.64)*** and accounts

Low-Low

-0.001 -0.003 -0.030 (-0.03) (-0.07) (-1.11) 0.271 -0.196 -0.129 (3.38)*** (-3.59)*** (-3.21)*** -0.618 -1.553 -0.049 (-9.16)*** (-33.93)*** (-1.45) 0.102 -0.068 -0.082 (1.39) (-1.37) (-2.21)** level; **: Significant at the 5% level; *:

0.987 (26.06)*** Low-High 0.990 (17.37)*** High-Low 0.571 (11.92)*** High-High 1.041 (19.99)*** ***: Significant at the 1%

28

251 251 251

Table V Calendar time portfolio regression: Portfolios by accounts receivable This table presents calendar-time portfolio regression results. In Panel A, firms in our dataset are divided into three portfolios based on accounts receivable over assets at the end of June 2009 and 2010, respectively, and these firms are included in the assigned portfolio from July of a year to June of the next year. Year 2008 (2009) accounting information is used for portfolio construction at June 2009 (2010). Difference indicates the difference portfolio whose return is High portfolio return minus Low portfolio return. In Panel B, firms are divided into two portfolios based on accounts receivable over assets and each portfolio is further divided into two portfolios based on accounts payable over assets. In the following panels, firms are divided into two portfolios based on total assets (Panel C) or total liabilities over assets (Panel D), and then each portfolio is further divided into two portfolios based on accounts receivable over assets. The regression uses portfolio data from November 25, 2009, to November 25, 2010 (251 trading days), and this table presents results from the Fama-French-Carhart four-factor model, with an event dummy (EVENTD). EVENTD takes a value of 1 for the three trading days from May 20 to May 25, 2010, and 0 for all other dates. T-statistics are reported in parentheses. Rm –Rf

SMB

HML

Mom

EVENTD

Alpha

R2

N

0.917 -0.095 0.036 -0.047 -0.004 (26.28)*** (-1.94)* (1.08) (-1.90)* (-1.32) Middle 1.000 0.136 -0.113 -0.089 -0.001 (30.64)*** (2.96)*** (-3.63)*** (-3.83)*** (-0.38) High 0.621 -0.556 -1.593 -0.035 -0.024 (12.18)*** (-7.72)*** (-32.65)*** (-0.97) (-4.93)*** Difference -0.296 -0.460 -1.629 0.012 -0.019 (-5.15)*** (-5.69)*** (-29.68)*** (0.29) (-3.59)*** Panel B: Accounts receivable and accounts payable portfolio (2*2 portfolio)

-0.002 (-2.47)** -0.000 (-0.14) -0.006 (-5.34)*** -0.004 (-3.24)***

0.763

251

0.803

251

0.972

251

0.962

251

Low-Low

-0.005 (-1.29) 0.001 (0.44) -0.026 (-5.10)*** 0.001 (0.19)

-0.007 (-6.53)*** 0.001 (1.91)* -0.007 (-5.66)*** -0.001 (-0.62)

0.553

251

0.852

251

0.969

251

0.741

251

0.001 (0.26) -0.019 (-3.61)*** -0.001 (-0.52) -0.005 (-1.43)

-0.005 (-3.93)*** -0.009 (-7.01)*** -0.001 (-1.84)* -0.001 (-0.62)

0.479

251

0.967

251

0.853

251

0.747

251

-0.002 0.778 (-2.38)** -0.001 0.680 (-1.52) -0.001 0.729 (-1.31) -0.006 0.971 (-5.19)*** the 10% level

251

Panel A: Accounts receivable portfolio Low

0.714 -0.022 (15.96)*** (-0.34) Low-High 1.059 0.018 (36.33)*** (0.43) High-Low 0.592 -0.628 (11.12)*** (-8.37)*** High-High 0.984 0.162 (25.91)*** (3.03)*** Panel C: Size and accounts receivable

-0.057 -0.073 (-1.34) (-2.31)** -0.016 -0.056 (-0.59) (-2.72)*** -1.544 -0.068 (-30.36)*** (-1.82)* -0.119 0.016 (-3.29)*** (0.59) portfolio (2*2 portfolio)

Low-Low

0.790 0.671 -0.457 -0.017 (13.60)*** (8.19)*** (-8.24)*** (-0.42) Low-High 0.528 -0.262 -1.779 -0.020 (9.52)*** (-3.35)*** (-33.51)*** (-0.50) High-Low 0.956 -0.031 -0.011 -0.070 (35.87)*** (-0.81) (-0.45) (-3.69)*** High-High 0.980 0.121 -0.087 -0.066 (26.03)*** (2.27)** (-2.41)** (-2.46)** Panel D: Leverage and accounts receivable portfolio (2*2 portfolio) Low-Low

0.927 (27.88)*** Low-High 0.934 (21.98)*** High-Low 0.955 (24.35)*** High-High 0.618 (11.85)*** ***: Significant at the 1%

-0.071 0.043 (-1.51) (1.34) 0.341 -0.257 (5.70)*** (-6.32)*** 0.032 -0.059 (0.58) (-1.56) -0.591 -0.567 (-8.04)*** (-31.44)*** level; **: Significant at the

-0.014 (-0.58) -0.006 (-0.21) -0.105 (-3.79)*** -0.065 (-1.76)* 5% level; *:

29

-0.000 (-0.01) -0.003 (-0.75) -0.005 (-1.27) -0.025 (-5.01)*** Significant at

251 251 251

Table VI Calendar time portfolio regression: Portfolios by subsidiaries and largest shareholder This table presents calendar-time portfolio regression results for 2*2 portfolios. Firms in our dataset are divided into two portfolios based on the number of subsidiaries (Panel A) and investments over assets (Panel B) at the end of June 2009 and 2010, respectively. Then, each portfolio is further divided into two portfolios based on accounts receivable over assets. In Panel C, firms in our dataset are divided into independent companies and subsidiaries and then further divided into two portfolios based on accounts payable over assets. We identify as subsidiaries companies with a corporate shareholder who owns more than 20% of the firm’s shares. These firms are included in the assigned portfolio from July of a year to June of the next year. Year 2008 (2009) accounting information is used for portfolio construction at June 2009 (2010). The regression uses portfolio data from November 25, 2009, to November 25, 2010 (251 trading days). Each panel includes the results from the Fama-French-Carhart four-factor model, with an event dummy (EVENTD). EVENTD takes a value of 1 for the three trading days from May 20 to May 25, 2010, and 0 for all other dates. T-statistics are reported in parentheses. Rm –Rf

SMB

HML

Mom

EVENTD

Alpha

R2

N

-0.006 (-4.33)*** -0.001 (-1.39) -0.001 (-1.82)* -0.006 (-5.69)***

0.373

251

0.820

251

0.870

251

0.977

251

-0.001 (-2.32)** -0.001 (-0.65) -0.001 (-1.13) -0.007 (-5.47)***

0.856

251

0.752

251

0.704

251

0.969

251

-0.007 0.976 (-6.82)*** -0.001 0.630 (-0.47) 0.000 0.632 (0.23) 0.001 0.790 (1.81)* the 10% level

251

Panel A: Number of subsidiaries and accounts receivable portfolio (2*2 portfolio) Low-Low

0.739 0.507 -0.336 -0.006 (11.55)*** (5.62)*** (-5.50)*** (-0.14) Low-High 0.965 0.794 -0.541 -0.030 (29.78)*** (17.39)*** (-17.49)*** (-1.33) High-Low 0.959 -0.013 -0.019 -0.063 (38.63)*** (-0.37) (-0.80) (-3.58)*** High-High 0.631 -0.647 -1.536 -0.033 (13.75)*** (-10.01)*** (-35.03)*** (-1.03) Panel B: Investments and accounts receivable portfolio (2*2 portfolio) Low-Low

0.002 (0.40) -0.003 (-1.00) -0.001 (-0.43) -0.025 (-5.86)***

0.917 0.000 -0.014 0.031 -0.001 (36.42)*** (0.01) (-0.57) (1.72)* (-0.59) Low-High 0.950 0.129 -0.102 0.006 -0.004 (26.25)*** (2.54)** (-2.96)*** (0.24) (-1.25) High-Low 0.977 -0.003 -0.035 -0.124 -0.001 (22.98)*** (-0.04) (-0.85) (-4.13)*** (-0.18) High-High 0.593 -0.605 -1.557 -0.060 -0.025 (11.01)*** (-7.96)*** (-30.23)*** (-1.58) (-4.97)*** Panel C: Independent/subsidiary and accounts payable (2*2 portfolio) Independen 0.588 -0.641 -1.540 -0.048 -0.025 t-Low (12.41)*** (-9.59)*** (-33.98)*** (-1.43) (-5.67)*** Independen 0.973 0.211 -0.147 -0.112 -0.001 t-High (20.00)*** (3.08)*** (-3.16)*** (-3.25)*** (-0.20) Subsidiary- 1.019 0.124 -0.097 -0.056 -0.006 Low (19.79)*** (1.71)* (-1.98)** (-1.53) (-1.30) Subsidiary- 1.050 0.146 -0.092 -0.020 -0.003 High (29.60)*** (2.92)*** (-2.71)*** (-0.81) (-0.79) ***: Significant at the 1% level; **: Significant at the 5% level; *: Significant at

30

251 251 251

Table VII Calendar time portfolio regression: Portfolios with family business affiliation This table presents calendar-time portfolio regression results. In Panel A, firms in our dataset are divided into two portfolios based on affiliation with family business group. We obtain family group affiliation for individual companies from the Online Provision of Enterprises Information System (OPNI). Difference indicates the difference portfolio whose return is Family portfolio return minus Non-Family portfolio return. In Panels B and C, the family group firms and non-family group firms are further divided into two portfolios based on accounts payable over assets (Panel B) and accounts receivable over assets (Panel C) at the end of June 2009 and 2010, respectively, and then these firms are included in the assigned portfolio from July of a year to June of the next year. Year 2008 (2009) accounting information is used for portfolio construction at June 2009 (2010). The regression uses portfolio data from November 25, 2009, to November 25, 2010 (251 trading days), and this table presents results from the Fama-French-Carhart four-factor model, with an event dummy (EVENTD). EVENTD takes a value of 1 for the three trading days from May 20 to May 25, 2010, and 0 for all other dates. T-statistics are reported in parentheses. Rm –Rf

SMB

HML

Mom

EVENTD

Alpha

R2

N

0.892

251

0.973

251

0.964

251

0.960 -0.073 0.005 -0.072 -0.003 -0.001 0.829 (32.39)*** (-1.74)* (0.17) (-3.45)*** (-0.93) (-2.01)** 1.072 0.043 -0.027 -0.095 -0.002 0.002 0.702 (23.23)*** (0.67) (-0.60) (-2.90)*** (-0.47) (1.96)** 0.528 -0.482 -1.640 -0.049 -0.023 -0.008 0.972 (10.44)*** (-6.76)*** (-33.92)*** (-1.35) (-4.88)*** (-7.48)*** 0.851 0.494 -0.341 -0.002 -0.002 -0.004 0.745 (25.21)*** (10.40)*** (-10.57)*** (-0.09) (-0.77) (-4.82)*** Panel C: Portfolio by family business group and accounts receivable (2*2 portfolio) Family 0.965 -0.098 0.030 -0.074 -0.001 -0.001 0.799 -Low (29.34)*** (-2.11)** (0.95) (-3.16)*** (-0.30) (-1.41) Family 1.066 0.091 -0.072 -0.094 -0.005 0.002 0.727 -High (24.57)*** (1.49) (-1.74)* (-3.05)*** (-1.27) (1.64) Non-Family 0.733 0.268 -0.187 0.001 -0.003 -0.006 0.714 -Low (23.83)*** (6.19)*** (-6.35)*** (0.05) (-1.19) (-8.20)*** Non-Family 0.568 -0.358 -1.717 -0.014 -0.020 -0.008 0.968 -High (10.40)*** (-4.65)*** (-32.90)*** (-0.35) (-3.94)*** (-6.35)*** ***: Significant at the 1% level; **: Significant at the 5% level; *: Significant at the 10% level

251

Panel A: Portfolio by family business group 1.000 -0.030 -0.007 -0.081 (42.66)*** (-0.92) (-0.31) (-4.86)*** Non-Family 0.547 -0.447 -1.661 -0.030 (11.02)*** (-6.40)*** (-35.01)*** (-0.85) Difference 0.453 0.417 1.654 -0.051 (8.05)*** (5.26)*** (30.76)*** (-1.27) Panel B: Portfolio by family business group and accounts payable Family

Family -Low Family -High Non-Family -Low Non-Family -High

31

-0.002 -0.000 (-1.08) (-0.30) -0.022 -0.008 (-4.76)*** (-7.31)*** 0.020 0.008 (3.75)*** (6.32)*** (2*2 portfolio)

251 251 251

251 251 251 251

Table VIII Regression of Tobin’s Q This table reports regression results of Tobin’s Q by using 8,973 firm-year data from 2007 to 2013. Tobin’s Q is computed as the sum of market value of equity and book value of total liabilities divided by total assets. Accounts payable and accounts receivable are scaled by total assets. Size is the natural logarithm of total assets. R&D is R&D expenses over total assets. ROA is the earnings before interest and taxes scaled by total assets. Debt ratio is debt over total assets. Cash holdings is the firm’s cash and equivalents scaled by assets. All variables are winsorized at the top and bottom 1 percentage values. Models (1) through (4) use the entire sample whereas models (5) through (8) are results for subsamples. In models (5) and (6), we divide our sample into two groups by firm size (logarithm of total assets) (model (5) is for small companies). In models (7) and (8), sample firms are separated into two groups by debt ratio (model (7) is for low-leveraged companies). A firm fixed effects model with year dummy is used for estimation. T-statistics are reported in parentheses. Model (1) Sample Accounts payable Accounts payable *2010 year dummy Accounts payable *2008 year dummy Accounts receivable

Entire

R&D ROA Debt ratio Cash holdings Constant.

Entire

Model (3) Entire

0.690 (2.20)**

0.604 (1.88)* 0.585 (2.29)**

-0.030 (-0.11)

-0.035 (-0.13) 0.051 (0.30) 0.049 (0.92) 2.772 (1.49) 0.489 (2.12)** -0.203 (-1.22) 0.627 (2.31)** 1.269 (5.24)*** Yes 0.056 8973

-0.012 (-0.04) -0.161 (-0.74) 0.050 (0.94) 2.765 (1.48) 0.490 (2.12)** -0.203 (-1.22) 0.626 (2.31)** 1.266 (5.23)*** Yes 0.057 8973

0.050 (0.93) 2.761 (1.48) 0.490 (2.12)** -0.203 (-1.22) 0.628 (2.32)** 1.271 (5.25)*** Yes 0.057 8973

Model (4) Entire

0.616 (1.91)** 0.470 (2.45)**

Accounts receivable *2010 year dummy Size

Model (2)

0.552 (1.67)* 0.525 (2.73)*** 0.417 (1.83)* -0.020 (-0.07)

0.050 (0.94) 2.769 (1.48) 0.490 (2.12)** -0.204 (-1.23) 0.625 (2.31)** 1.274 (5.26)*** Yes 0.057 8973

Model (5)

Model (6)

Model (7)

Model (8)

Small firms

Large firms

Lowleveraged

Highleveraged

0.697 (1.44) 1.113 (2.35)**

0.014 (0.03) 0.188 (0.90)

1.654 (4.02)*** -0.046 (-0.18)

-0.688 (-1.43) 1.624 (3.19)***

-0.132 (-0.31) -0.381 (-0.90) 0.043 (0.38) 3.387 (1.27) 0.156 (0.51) -0.124 (-0.41) 0.615 (1.60) 1.522 (3.84)*** Yes 0.051 4233

0.382 (1.58) 0.007 (0.04) 0.056 (0.81) 0.764 (0.38) 1.235 (4.44)*** -0.171 (-1.21) 0.277 (1.21) 1.013 (2.71)*** Yes 0.098 4740

-0.565 (-1.15) 0.060 (0.22) 0.043 (0.44) -0.373 (-0.17) 0.565 (1.73)* -0.666 (-1.88)* 0.447 (1.54) 1.321 (3.29)*** Yes 0.066 4522

0.395 (1.35) -0.529 (-1.46) -0.004 (-0.05) 5.271 (1.56) 0.557 (1.95)* 0.070 (0.29) 1.371 (2.50)** 1.437 (3.82)*** Yes 0.056 4451

Year dummies R2 N ***: Significant at the 1% level; **: Significant at the 5% level; *: Significant at the 10% level

32