Ebook Bad Loans and Accounting Discretion
Why did Japan’s non-performing loan problems and the banking crisis last more than a decade? According to Hutchison and McDill (1999), the average duration of a banking crisis is 3.9 years that is smaller than the Japan’s case by 10 years ! Finland and Sweden went through more serious asset market collapses than Japan and yet needed only four years to solve the bad loan problems. Korea were hit by a serious Asian currency crisis but also solved the bad loan problem in four years.
Our argument is that an inadequate enforcement of capital adequacy requirements was one of important sources of the long-lasting non-performing loan problem in Japan. Capital adequacy requirements constitute a core element of prudential regulation that is intended to induce sound risk management of bank asset portfolios. Nonetheless, the regulatory authorities allowed banks to engage in accounting discretion and to overstate bank capital. In response, banks had a strong incentive to conceal non-performing loans by extending additional loans to almost insolvent firms. It is obvious that the extended loans became further non-performing loans.
There are a number of studies that found the evidence of the accounting discretion under capital regulations by Japanese banks, including Shimizu and Horiuchi (1998), Ito and Sasaki (2002), Fukao (2002), and Shrieves and Dahl (2003). There are also a number of studies that found the evidence of bad loans made by Japanese banks, including Sekine et al. (2002), Hori and Osano (2002), Caballero et al. (2003), Peek and Rosengren (2005), and others. The former literature revealed the evidence of accounting discretion that was rooted on capital requirements by Japanese banks, while the latter found the evidences of the continuation of bad loans. To our knowledge, however, there are few studies that investigated the relationship between accounting discretion and bad loans.
The purpose of this paper is to provide theoretical and empirical analyses that investigate whether a lax enforcement of capital adequacy requirements, accompanied with discretionary accounting practices, promoted banks to supply loans to almost insolvent firms and consequently prolonged the non-performing loan problem in Japan.
As stated above, bad loans are often associated with accounting discretion and difficult to detect because reliable data is limited. Data on accounting discretion is not available, either. In order to work out this difficulty, we construct a theoretical model in an attempt to derive testable implications from observable data on the relationship among bad loans, capital adequacy requirements, and accounting discretion.
The capital requirements, when it is subject to accounting discretion, provide banks with the perverse incentive to extend bad loans to almost insolvent firms. Continuing loans to bad firms becomes a tool for inflating regulatory capital although it finally makes damages to the bank. The market evaluation to the bank declines, while the bank has ample regulatory capital that is sufficient to meet the minimum capital requirements. Recapitalization to poorly-capitalized banks also provides them with the perverse incentive. Recapitalization by issuing subordinated debt, if it is issued to insiders, may not play a disciplinary role, but rather may work as a tool of softening bank’s capital constraint. Consequently, subordinated debt and the continuation of bad loans can serve as complementary device for capital watering.
Following the theoretical findings, we construct a number of measures from accounting variables that are expected to be correlated with bad loans. The first measure is the market-valued capital as a proportion of bank assets that is expected to be correlated with the bank asset quality to the extent that the stock market incorporates the information on financial conditions of individual banks accurately. The second measure, which we believe to be most appropriate for account discretion, is the difference between the risk-based capital ratio (RBC) and the market-valued capital as a proportion of bank assets.
The theoretical finding predicts that the divergence between the two ratios reflects the perverse incentive of banks to extend bad loans. The third measure is subordinated debt as a proportion to risk-weighted bank asset. The theoretical finding predicts that recapitalization by issuing subordinated debts to insiders works as a tool of softening the bank capital constraint. The fourth measure is the distance of the reported RBC ratio from the minimum requirement level that is motivated by the idea that more severely capital-constrained banks have a greater incentive of extending bad loans.
Using these measures of accounting discretion as explanatory variables, we estimate the share of bad loans in total loans, because we can safely regard that the share of bad loans is controlled by banks and thus we can work out the identification problem associated with the loan supply function. Given the assumption that accounting variables are “manipulated”, banks choose the loan portfolio and accounting variables simultaneously [e.g., Shrieves and Dahl (2003)]. In order to avoid such an endogeneity problem, we estimate coefficients not only by OLS but also by GMM.
We find that the RBC ratio is not correlated with the banks’ bad loans. Capital requirements do not seem to have worked well to discipline banks’ behavior. On the other hand, the market-valued capital ratio is correlated with the banks’ bad loans. The stock market seems to have incorporated the information on financial conditions of individual banks to some degree. Our proposed accounting measures, the difference between the risk-based capital ratio (RBC) and the market-valued capital ratio, and the subordinated debt ratio, are found to be useful in uncovering evidence supporting that banks had the perverse incentive of extending bad loans in an attempt to inflate regulatory capital.
The estimation using the distance of the reported RBC ratio from the minimum requirement level suggests evidence that banks extended bad loans beyond the minimum requirement. Furthermore, we find a different result between major banks and regional banks, suggesting the too-big-to-fail policy with regard to the regulation and supervision in Japan.
The remainder of this paper is organized as follows. Section 2 reviews the banking problem in the 1990s. Section 3 sets up the model. Section 4 conducts the theoretical analysis. Section 5 reports our empirical method and results. Section 6 concludes.
Posted in :
