Corporate Debt Mix and Long-term Firm Growth in Japan

ABSTRACT

This paper examines how firms change debt financing channels in line with the development of financial markets. In this aim, a data set of Japanese listed firms from 1965 (with more than 10% of annual GDP growth) to 2015 (almost 0% GDP growth) is used. We find a long-term change in the debt mix from internal debt financing (e.g., trade credits) to external debt financing (e.g., public bonds). Furthermore, we document that the firm growth rate is positively related to bond financing and negatively associated with trade credits. These associations are not conditional on interlocking business relationships with Keiretsu. The findings imply that the role of established firms’ internal financing channels diminishes as financial markets develop along with economic growth.

Notes

1. One exception is Ang and McKibbin (2007), who use the macro data of the Malaysian economy from 1960 to 2001. Their analysis relies on aggregated data, which makes it difficult to analyze the influence of firm heterogeneity on growth. Indeed, the benefit of the financial system’s development would differ across firms.

2. The data are from the World Bank (https://www.worldbank.org/ja/publication/global-economic-prospects). Note that the data in the World Bank constitute the real GDP growth rate, whereas the growth rate in Japan has a nominal value.

3. While it is the common standard to define a fiscal year over 12 months, most listed firms in Japan followed a 6-month reporting practice until 1975. Among the 1,332 firms in the cleansed sample as of 1965, in the first year of the data set, 937 firms followed a 6-month reporting practice, 379 firms compiled their financial statements on a yearly basis, the rest changed the accounting period during 1965. The Japanese government changed accounting regulations in 1974, making it easier to adopt a 12-month accounting period. To keep consistency in our sample about the length of accounting period, we compute annual sales as the sum of two 6-month sales quantities – i.e., one from the previous term and the other from the current term. We restrict observations to those companies whose accounting period is exactly 12 months.

4. This approach is also employed in DeAngelo and Roll (2015), who examine the instability of the capital structure of US firms using a sample from 1950 to the 2000s.

5. Oztekin and Flannery (2012) also show the high financial leverage in bank-oriented countries.

6. we identify business group-affiliation firms from the Kigyou Keiretsu Soran (or Soran) data, which was published by Toyo-Keizai Publishing. Then, a value of one for the dummy variable is assigned to six-largest business group-affiliation by searching for the information contained in Soran since 1979 due to the data availability. We assume that on or before 1978, each firm’s keiretsu affiliation was the same as that in 1979. We believe this assumption is reasonable. The business group relationship is, usually, stable and does not change over time. Toyo-Keizai Publishing stopped publishing the Soran data in 2000. As a result, each company’s business-group affiliation after 2001 is assumed to be the same as that in 2000.

7. We also conduct an analysis with stricter restriction with the caliper as 0.001 and the result is shown in the appendix file. The result are almost similar to that in Figure 2.

8. We also define the long-term change of debt in several ways such as 5, 7, and 10 years differences. However the results are similar to the one with the growth ratio from 1965.

9. We also estimate the sample with allowing the exit during the sample period. But results are quite similar to the one we show. Please see Tables B2 to B4 in Appendix file.

Additional information

Funding

This study was supported by JSPS (19H01507; 20K01781).