Financial heterogeneity and the dynamics of credit rationing in Japan

ABSTRACT

A perceptual gap between banks and firms exists in Japan, preventing the credit channel of monetary policy. Banks believe that bankable customers are scarce, while firms believe that banks do not issue loans without collateral or guarantees. To explain this gap, I focus on the dispersion in the degree of financial constraints across listed Japanese firms from FY1991 to FY2019. I construct a firm-specific and time-varying measure of financial constraints through the structural estimation, and investigate its distribution over time. The results reveal a right-skewed distribution for the index of financial constraints, indicating that many firms face minor financial constraints, while a few face severe financial constraints. The spread between the 75th and 25th percentiles of the index of financial constraints increased after the bubble burst, indicating that Japan’s financial heterogeneity has recently become outstanding. Finally, decomposing financial heterogeneity into within – and between-industry effects shows that the observed financial inequality is due to the increase in inequality among firms within narrowly defined industries.

JEL CLASSIFICATION:

• D24
• D25
• G31

KEYWORDS:

• Between-industry effects
• credit rationing
• financial inequality
• firm investment
• within-industry effects

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ Studies that use the same approach in measuring financial constraints include Hubbard and Kashyap (1992), Whited (1992), Hubbard et al. (1995), Ogawa et al. (1996), Barran and Peeters (1998), Ogawa and Suzuki (1998), Love (2003), and Whited and Wu (2006).

² For the sake of clarity, I have omitted taxes. The profit, firm’s discount rate, and price of capital goods are all appropriately tax-adjusted in the estimation. Tax adjustment is essential because our sample period is longer than the quarter-century.

³ I also use the same definition as Whited (1992) to estimate the investment Euler equation. I can obtain qualitatively similar results. However, the over-identification test rejects the null hypothesis that the model specification is valid at the 5% significance level.

⁴ Lian and Ma (2021) emphasize that the Japanese firms primarily use asset-based lending contrary to the US’s earnings-based lending and explain the background behind this. Ono et al. (2021) describe this background in more detail.

⁵ Whited (1992), Hubbard et al. (1995), and Whited and Wu (2006) use the same instrumental variables. As tax payments include the presence of firms with tax losses and the use of investment tax credit, their use as an instrument minimizes the importance of the model’s misspecification problem. Similarly, as the depreciation charges must follow the firm’s accounting rules and the interest expenses are determined by the past debt contracts, the use of these variables as instruments also minimizes the model’s misspecification problem.

⁶ The scale is close to Whited’s (1992) result; see p.1,447.

⁷ One reason for this difference may be due to the Japan-specific traditions of debt contracts. As Lian and Ma (2021) point out, Japanese firms primarily use asset-based lending, contrary to the US’s earnings-based lending. Such a practice emerged, because Japanese bankruptcy courts were largely dysfunctional before a major reform around 2000.

⁸ Since the investment Euler equation includes the one-period lead variable, we can obtain the time-series variations of Θ until FY2018.

⁹ For the literature reporting a positive and significant relationship between financial access and productivity, see Gatti and Love (2008) for Bulgarian firms, Butler and Cornaggia (2011) for farmers from US midwestern states, Chen and Guariglia (2013) for Chinese manufacturing firms, and, more recently, Ferrado and Rugierri (2018) and Levine and Warusawitharana (2019) for European countries.

10. The use of listed company data has advantages. For example, detailed information such as depreciation expense, interest expense, and the value of land stock is available only for listed companies.

Additional information

Funding

This work was supported by the Japan Society for the Promotion of Science (JSPS) [KAKENHI grant number 17K13767]; Kyoto University [The Joint Research Program of KIER].