Nonbank financing and performance of informationally opaque businesses

Abstract

Previous studies argue that banks offer loans to informationally opaque businesses using relationship lending technology. Using survey data of small businesses in Japan, we show that informationally opaque and financially weak firms that do not have lending relationships use high interest rate nonbank loans because of low availability of bank credit. Furthermore, we show that nonbank loan applicants are likely to incur operating losses and default. These results imply that nonbanks have difficulty avoiding the information problem because borrowers have uninformative financial statements and weak financial conditions.

Acknowledgements

This study is supported by a Grant-in-Aid for Scientific Research, Japan Society for the Promotion of Science, Tokyo. I thank the seminar participants at the Research Institute of Economy, Trade, and Industry (RIETI) for their helpful comments and discussions. Any remaining errors are mine.

Notes

¹ See Bester (1994) for more detail.

² In Japan, Uchida et al. (2007) also showed that small banks offer loans to informationally opaque firms using relationship lending technology. Scott (2004) and Cole et al. (2004) also focused on bank size and relationship lending.

³ Kang and Stultz (2000), examining listed firms, use stock returns as a proxy of firm performance. Our sample contains data on unlisted small businesses; however, data on stock returns or Tobin's q are unavailable. Thus, we investigate firm performance using accounting profit.

⁴ See also Yabushita and Bushimata (2006) for more detailed information about nonbanks.

⁵ Outstanding bank loans also contracted after 1996.

⁶ Some registered nonbank lenders do not have branches because they do not offer credit. Hence, there are more registered nonbank lenders than branches.

⁷ See the website of the Japanese Bankers Association: http://www.zenginkyo.or.jp/en/stats/. Banks are defined as city banks, regional banks, trust banks and long-term credit banks.

⁸ Large firms comprise about 10% of the combined sample; the remainder are SMEs.

⁹ Berger et al. (2005) used ‘firm size’ as a proxy for informational transparency. However, as Uchida et al. (2007) argued, ‘audit’ is a more suitable proxy for information verifiability. Therefore, we use ‘audit’, rather than ‘firm size’, as a proxy.

¹⁰ Petersen and Rajan (1997) investigated demand for trade credit, which is a higher interest rate type of debt.

¹¹ See Belsley et al. (1980) for more detailed explanations about the VIF collinearity test.

¹² The Small and Medium Enterprise Agency (2003) showed that firms in poor financial health, firms lacking good operating relationships with their main bank and firms in the real estate and service sectors use nonbank loans. However, the agency does not show the effects of firm size, firm age, collateral assets and auditing.

¹³ In Asquith et al. (1994), the definition of financial distress is based on interest coverage ratios.

¹⁴ See Wooldridge (2001) for a more detailed discussion about selectivity bias.

¹⁵ See Greene (2002) for a detailed explanation of bivariate probit estimation.

¹⁶ Our results show that the estimated correlation between u_{i , t} and v_{i , t} is statistically negative. If we regress ROA as a dependent variable, this correlation is positive. These results suggest that firms that are more likely to borrow from nonbanks are less likely to incur losses, after we control for other variables.

¹⁷ However, some studies (for example, Marotta, 2005) argue that there is no evidence that trade credit is more expensive than loans.

¹⁸ However, as some trade creditors have long-term relationships with small businesses, they might offer credit with the relationship lending technology.