Stop Rolling over the Bad Loan Problems

TSURU Kotaro
Senior Fellow, RIETI

The resolution of the non-performing loans (NPLs) problem must be the top priority on the policy agenda of Prime Minister Koizumi's cabinet, which has committed to "reform without sanctuary." If the NPLs were removed completely from banks' balance sheet within two to three years, as requested by the recent economic package, it would mark a major step toward solving the NPL problems, especially for major banks. Whether this cabinet can successfully deal with the problems, however, will depend on properly understanding the delay of the NPL resolution and its influence on the economy. Without this understanding, the history of the NPL problem might repeat itself.

A key phrase in explaining the lending behavior of Japanese banks in the 1990s is "soft budget constraints" (the refinancing of loss-making enterprises). Originally referring to the situation where governments loosen financial discipline on state owned enterprises in socialist economies, a similar phenomenon has also been observed in bank-firm relationships in transition economies.

Why do soft budget constraints occur? Banks have already incurred ample sunk costs, including monitoring costs in past lending relationships. They therefore have an incentive to extend loans to these unprofitable projects and recover their loss if refinancing is expected to yield any profit whatsoever.

In Japan's case, soft budget constraints have been concentrated mainly in the real estate sector. Bank loans to real estate have steadily increased, even in the 1990s after the bubble economy burst and most real estate projects went sour. Even if real estate loans were not performing, banks tended to build up "loan-loss provisions" for them, while waiting for land prices to recover, rather than carry out "loan-loss write-offs" and remove NPLs from their balance sheet. Collateral value of real estate related loans is strongly correlated with land price movements and there is an incentive to liquidate collateral when prices are high.

This wait-and-see approach toward dealing with the NPL problems further promoted the inefficient refinancing of this sector. In extreme cases, weakly capitalized banks might be willing to refinance risky projects and, thus, "gamble for resurrection," hoping that the government would save the insolvent bank with a capital injection. Contrary to expectations, land prices have kept falling over the past decade. A forbearance approach of building up provisions and refinancing poor projects, which seemed to be a rational choice by banks, has subsequently produced new NPLs.

Given the above story, it would be difficult to conceive that the effect of NPLs on the economy was a simple credit crunch or "disorganization," which implies shrinkage of economic transactions due to mutual distrust among trading partners. Instead, the refinancing of inefficient real estate, construction, and retailing sectors has crowded out new loans to profitable manufacturing sectors.

There is, therefore, a coexistence of a credit crunch and soft budget constraints. The credit crunch that occurred at the end of 1997 should be considered a process of normalization: banks, under pressure from the introduction of comprehensive disclosure rules and the Prompt Corrective Action procedure, hesitated to refinance inefficient firms, thereby strengthening the perception of a credit crunch. The government, however, made the serious policy mistake of expanding its loan guarantees for small and medium sized companies, which further loosened financial discipline on these entities.

Along with the mechanism of soft budget constraints, "reputational concerns" have attracted much attention in recent academic literature, as did the mechanism of forbearance of the NPL problems. For example, bank managers may not want to write off NPLs immediately, since such write-offs prove their mismanagement of the loans. A bank's reluctance to accept capital injection is also associated with their reputational concerns. Bank regulators, who want good reputations as capable monitors, may postpone the closure of insolvent banks and delay the drastic resolution of the NPL problems.

Moreover, a case-by-case approach is important for dealing with NPL problems, since the magnitude of the problem differs from bank to bank. Once authorities announce a specific measure on an individual bank, financial markets revise their expectations on the financial conditions of the bank, and if they overreact, the authorities may have to consider systemic risks. To reduce these risks, the authorities had no option but to maintain so-called convoy system.

Thus, the fundamental resolution of the NPL problems must involve an end to the soft budget constraints and a clarification of the final responsibility of lenders, borrowers, and regulators. Regulatory authorities should not give banks free hand in the disposal of NPLs. Instead, by standardizing their classification and provision rate to a specific borrower (which may differ among banks), the authorities should mandate that banks resolve the NPLs. More generally, structural reform should be aimed at breaking past "non-performing" relationships.

Long-term relationships built in business, government, and politics, supported by a network of relation-specific investment, have now become an obstacle to Schumpeterian creative destruction. If the resolution of the NPL problems advances further, a credit crunch, due to the end of inefficient refinancing, and, in the worst case, disorganization, may arise. The government should overcome the pain of structural reform and promote the creation of new economic networks, by facilitating resource reallocation and enhancing labor mobility.

June 5, 2001

June 5, 2001