10. A New Look at Bank-Firm Relationships and the Use of Collateral in Japan: Evidence from Teikoku Databank Data

A detailed description of the basic characteristics of firms in the TDB database and their inter-firm relationships, with special emphasis on the use of trade credit, can be found in Uchida et al. (2010).

Note that since the aim of the analysis is to examine firm–bank relationships, financial firms such as banks, securities firms, insurance companies, and finance companies were excluded in the compilation of the dataset.

See Uchida et al. (2010) for further details.

In Japan, SMEs are officially defined in Article 2, Paragraph 1 of the Small and Medium-sized Enterprise Basic Act based on the amount of capital (up to 300 million yen) or the number of regular employees (up to 300). Lower limits apply to enterprises in the wholesale, retail, and services industries.

Whether a firm is owner-managed is determined by checking the surname of the owner of the majority of the capital stock of the firm and the surname of the CEO or firm representative. If the two are the same, it is assumed that the firm is owner-managed.

There already exist many studies on main bank relationships in Japan, especially on large, listed firms (see, e.g., Aoki et al. 1994; Uchida and Udell 2010).

A detailed discussion of the different bank types in Japan can be found in Uchida and Udell (2010).

The TDB score is a metric that evaluates the creditworthiness of a firm. The score ranges from 1 to 100 points, and a higher score indicates higher creditworthiness. See Uchida et al. (2010) for more on TDB scores.

To mitigate the financing difficulties of SMEs, the Japanese government employs credit guarantee programs that ensure the repayment of defaulted loans. See Uesugi et al. (2010) and Ono et al. (2013) for details on credit guarantee programs in Japan.

However, (and as will be seen below), firms whose main bank is a government FI are not necessarily riskier.

More precisely, a main bank is a bank with which a firm has a deposit account and obtains loans (short-term and/or long-term loans). When a firm has a deposit relationship with one bank and a lending relationship with another, TDB identifies the latter as the main bank, even though the firm is allowed to list multiple banks as its main banks. Similarly, when a firm obtains short-term loans from one bank and long-term loans from another, then the former is labeled as the main bank.

In the empirical literature, the main bank is typically defined as the private FI with which a firm has the largest amount of outstanding loans (Kawai et al. 1996; Sheard 1989); however, other definitions are also used (see Uchida and Udell 2010).

The TDB database also contains information regarding discounted bills (tegata-waribiki), a traditional method for short-term financing in Japan. See Ono et al. (2010) for more on discounted bills.

Mester et al. (2007) provide empirical evidence on the importance of deposit relationships (transaction accounts) in monitoring borrowers.

The percentage of firms with paid-in capital of more than 100 million yen in the real estate registry data is approximately 3 %, whereas it is 8 % in the full sample used here. Similarly, the percentage of firms with more than 100 employees in the real estate registry data is approximately 6 %, while it is 8 % for the full sample.

Ono and Uesugi (2009) showed that 66.7 % of Japanese SMEs use personal guarantees.

In Japan, most firms open a checking account with their main bank, because promissory notes or checks are accepted and discounted only when drawn against a checking account with a bank and produced by using the uniform format promulgated by the bankers’ association. Once such a checking account is opened, a large part of the firm’s cash flow goes through the account. Therefore, taking deposits in the checking account as collateral could be seen as equivalent to collateralizing the cash flow of the firm.

In Fig. 12.3, account receivables and inventories are included in the item “other.”

Following the tradition of the Civil Law system, Japanese law does not allow non-possessory security interests in movables. The Civil Code stipulates that movables cannot be the subject of a mortgage, but can only be the subject of a pledge, which requires that the debtor (pledgor) gives possession of the movable to the creditor (pledgee). However, in order to enable the debtor to utilize their property as a security for their debt and still use the same property for the business, the practice of “transfer by way of security” has developed. This is the legal practice of transferring the ownership of the movable to the creditor based on the agreement that the transfer is only for the sake of securing the credit and that ownership will revert to the debtor when the latter has repaid the debt. In 1998, a registration (filing) system for account receivables was introduced with the expectation that registration enhances the effectiveness of transfers by way of security. Under the new registration system, the security interest is “perfected,” i.e. becomes valid vis-à-vis the third party (another creditor). In 2004, the registration system was extended to movables (inventories and equipment). Notwithstanding these reforms, the system has several limitations. In particular, registration is not the exclusive means for the perfection of the interest in a transfer by way of security; another means of perfecting the creditor’s right remains taking possession of the property. Therefore, registration does not preclude the possibility that another creditor claims an interest that has priority over the registered creditor based on the fact that the unregistered creditor had taken possession of the property (typically through the “fictitious possession” by the debtor’s declaring that the debtor possesses the movable for the sake of the unregistered creditor) prior to the registration. As a result, a creditor cannot be sure whether they can rely fully on a registration to secure the loan (Kozuka and Fujisawa 2009). The low level of account receivables and inventories registered as collateral may partly be the result of such shortcomings of the registration system.

In cases where land and building are not separable (as in the case of an apartment), this is counted as one piece of real estate.

Note, however, that the share of firms with four or fewer properties is probably overestimated, since it is possible that TDB does not investigate whether a firm has other properties.

Although the official real estate registry provides information about seniority among secured loans, the TDB does not collect such information.

Ono et al. (2014) study the cyclicality of LTV ratios and whether LTV ratios are good predictors of firms’ ex-post performance.