|
|
|
Map of Asia |
Email this citation
The 1995 Financial Crisis in Japan
Working Paper 85
February 1996
The Berkeley Roundtable on the International Economy
Table of Contents
2. The Japanese System Of Financial Regulation
2.1. Where it came from...
2.2. ...and where it is now
3. The Bank Failures Of 1995
3.1. The Credit Cooperatives
3.2. The Regional Bank: Hyogo Bank
3.3. The J_sen (Housing Loan Companies)
4. The Process Of Bad Loan Accumulation
5. The Bad Loan Situation In 1995
5.1. Overview
5.2. Deposit Insurance
5.3. Tokyo Kyodo Bank - The Japanese RTC
During 1995, thirteen Japanese financial institutions went effectively bankrupt. With perhaps 70 trillion in bad loans hovering over the banking system, 14 of the top 21 banks are expected to post pre-tax losses for fiscal year 1995 (ending March 31, 1996). Non-disclosure and a lack of transparency resulted in domestic and international uncertainty about the soundness of the Japanese financial system. This led large U.S. rating companies to lower their evaluations of Japanese banks and resulted in a "Japan premium" in international financial markets. As of December 1994, Japanese banks had borrowed a total of $1,612 billion internationally. If they fail to resolve their domestic problems quickly, this crisis could have severe consequences for the international financial community.
Analysts disagree about the extent and the consequences of these problems 1 . While some have forecast a "Great Depression" scenario, complete with a meltdown of the entire financial system, others claim that the Ministry of Finance (MOF) will eventually manage to resolve the problems through an expensive orchestrated wave of mergers. Most commentators seem to agree, however, on the nature of the crisis: The "bubble" economy of the 1980s and the boom and bust of real estate prices is the primary reason for this crisis.
However, if Japan's bad loan problems were simply the result of real estate speculation and fraud, why is this crisis likely to lead to a major reconstruction of the banking system? In comparison, the Savings and Loan crisis in the U.S. did not have an immediate impact on the structure of the large money center banks and American banking as a whole. This paper argues that the "bubble" alone does not explain the Japanese crisis, the regulatory response, and the consequences that this crisis will have for the entire banking system. In order to evaluate the crisis and its ramifications, we need to examine what the structure of Japanese finance was when the bad loan problems occurred, and what exactly happened during this crisis. Why did the crisis occur? What explains the (non-) response to the emerging bad loan debacle in the early 1990s? And how does this crisis relate to the structure and evolution of the postwar Japanese financial system?
This paper argues that the crisis is more the result of problems that are structural (or the design of the regulatory system) rather than behavioral (behavior common during the bubble economy, e.g., speculation). The Japanese financial system and its regulatory structure did not evolve from the fundamental premises on which it operated during the period of rapid growth (1950-1973). The system is still characterized by collusive regulation (dango gy_sei), a lack of investment responsibility, and a predominance of extra-legal administrative guidance characterized by reciprocal obligations and "quid-pro-quo" deals. Ultimately, this results in an entanglement of regulators and regulatees that makes rule enforcement difficult. In this system of "collusive regulation" neither regulators nor regulatees were interested in disclosure and rule enforcement. To be sure, fraud during the bubble years added to the problem. But when the bad loans began to surface, no action was taken. By 1995, the problem had grown into a crisis. A reorganization of MOF, the primary regulator, is now being discussed in Japan as a countermeasure against future regulatory mistakes. However, as long as the structural features of collusive regulation remain untouched, an organizational change, such as a division of MOF into two agencies, is unlikely to sufficiently impact the environment of financial regulation in Japan and lead to a real change.
In order to evaluate the current events in the overall regulatory context, the paper begins with a brief review of the Japanese financial system in the period of rapid growth (1955-1975), and a characterization of the current regulatory system (Section 2). Section 3 presents a comprehensive portrait of "1995 bankruptcies": The five credit cooperatives, one regional bank, and seven housing loan companies 2 . Section 4 analyzes the approaches and tools employed by the regulators in response to the bad loan situation in 1995. Section 5 presents data and analysis of the extent of the bad loans, as well as the situation of the Deposit Insurance system. Section 6 concludes.
2. The Japanese System Of Financial Regulation
2.1. Where it came from...
In order to understand how and why the irresponsible banking practices of the 1990s could have happened, it is helpful to review briefly the Japanese financial system during the "period of rapid growth" (1950-1973). In this period, the primary policy goal of the Japanese government was to foster economic development. One of the policy tools was to channel low-cost funds from the savers to "designated" industries by inducing policy-conforming bank behavior. The financial structure in this period was characterized by four major structural features: (1) regulated interest rates; (2) indirect financing; (3) "window guidance" and administrative guidance; and (4) segmentation and strict demarcation of business areas 3 .
Interest Rate Regulation
Interest rate regulation, based on the "Temporary Interest Rate Adjustment Law" of 1947, aimed at controlling all interest rates in order to provide low-cost funds to designated industrial sectors. According to Royama (1986, p.234), this one single law is the basis and institutional framework for "cartellized" or "collusive cooperation" (karuteru-teki ky_ch_ k_i) in all layers of the financial system. Since interest rates were not responsive to supply and demand pressures, their low level resulted in excess demand for funds, which, in turn, had to be controlled through non-price measures, such as informal contacts and regulation 4 .
For the financial institutions, this policy had four major implications. First, profit margins were high and predictable, leading to a fairly stable hierarchy within the banking system. Second, there was little pressure to cut costs and increase efficiency. Third, the large banks were in constant contact with the regulators who aimed to influence the flow of funds for industrial policy purposes. Part of this process were the bailouts by mother banks of troubled industrial entities, which were usually negotiated with the involvement of the government. And fourth, the regulators allowed the banks to define their own disclosure requirements through their trade associations. Because many of these loan decisions were made based on criteria other than creditworthiness, the disclosure requirements remained rather flexible. In the period until 1973, this was not a problem because in an environment of steep economic growth, no one lost. Depositors did not question the business practices of the banks, while the major shareholders were usually the loan customers and related financial institutions, who were themselves not interested in more detailed disclosure.
Indirect Financing and the Dominance of Bank Loans 5
Between 1950 and 1973, the Japanese capital market was small and regulated, and corporations had no direct access to international finance 6 . Suzuki (1980) estimates that, as of 1974, 74.1% of all external financial sources raised by Japanese corporations were provided through banks (the corresponding number for the Federal Republic of Germany was 66.6%, and only 28% for the U.S.). For the policy makers, the dominance of banks facilitated the process of strategically funding economic growth. Therefore, the bank dominance in corporate funding was further supported by complementary policies to limit the role of the capital markets.
The bond market was suppressed from two sides: Government bonds were not issued on a large scale until 1973, and corporate bonds were relatively expensive because of regulation 7 . Capital increases through stock issues were also relatively expensive 8 . Trading volume on both stock and bond markets remained low until the 1980s. This benefitted the banks because there was no competition for loan customers from the capital markets. This also insured that high interest rate margins could be sustained.
Window Guidance
Facing an excess demand for funds under artificially low interest rates, banks soon found themselves in a financial position labelled "overloan" (oobaa roon). In an "overloan" situation, the banks' total capital and deposits are insufficient to meet the loan demand. This requires that the banks themselves borrow money from the central bank to furnish loans to the corporate sector. Therefore, the central bank could directly influence the credit behavior of the major banks by rationing credit. This happened in a process of window guidance 9 , a subset of administrative guidance specifically geared towards the lending behavior of commercial banks. As one effect of this guidance, the hierarchy of the large banks remained unchanged, because banks were apportioned central bank credit based on existing loan market share. This lowered the need for strategic management decisions, cost-cutting efforts, and competitive positioning vis-a-vis other large banks. Further, it was also unclear who would ultimately be responsible for failure in an environment where the financial authorities micro-guided the banks' lending behavior. The large banks did not have to develop a keen sense for the inherent risk of loans. Risk-insensitivity was furthered by the explicit policy of "convoy regulation" (gos_sendan sh_gi).
Segmentation of the Financial Industry
Until 1993 10 , the banking structure was segmented and categorized along various lines: (1) between commercial and investment banking (along this line, trust banks developed as a separate category; their main function is to manage funds entrusted to them on both corporate and private accounts); (2) between functions (by size, location and/or profession of the customer); and (3) between short-term versus long-term lending among large banks (city banks were legally allowed to furnish loans with maturities of up to three months, while specialized banks for long-term credits formed a category of their own). Within this multi-layered structure, each bank category was subject to special laws, detailed restrictions, and administrative guidance. The only common theme in this regulatory approach was the "convoy approach": All changes and deregulation in the system had to be considered with regard to how they would affect the weakest of the banks. If there was some danger of putting a small bank at risk, no change would occur.
Facing a large number of rules, city banks spun off business as nonbank subsidiaries in order to circumvent some of the rules. For instance, when MOF asked banks to cut their loans to the real estate area after the burst of the bubble in 1992, the banks obeyed on paper. In reality, however, many simply shifted this business to their nonbank affiliates. In this way, risky loans, possibly shady transactions, and illegal loans came to be concentrated in the nonbanks. Their high risk exposure was not a problem until 1995 when it coincided with the bankruptcies in other parts of the financial system.
The Process of Deregulation
After the first oil shock, some of these basic structural features began to unravel. First, the recession of the 1970s led the government to a large-scale issue of deficit-financing bonds. This triggered the development of the bond market, as the government had to lower the restrictions on dealings in these bonds in order to place them. The recession indirectly also led to a change in the flow of funds: corporations began to invest their funds at securities firms, which offered gensaki, a kind of repurchase agreements. Gensaki had not been included in the interest rate regulation law. Losing their corporate customers to the securities firms, the city banks began to lobby for a free-interest instrument themselves. This developed into a ping-pong game between city banks and securities firms of lobbying for yet more non-regulated instruments. Between 1979 and 1987, the open market (short-term financial market for banks and corporations) developed and grew rapidly 11 .
The features of indirect finance and bank dominance were also affected. The revision of the Foreign Trade Law in 1980 and a relaxation of restrictions on financial transactions in international financial markets led to an increase in bond issues by Japanese firms in Europe. This led to the phenomenon of "disintermediation"; i.e., a movement of the corporations away from financing through bank loans towards direct means of financing.
Interest rate deregulation slowly lowered the banks' profit margins. They now had to compete with both other banks and the capital markets for corporate clients. The city banks opened up to small corporations as well. Because smaller banks in general have higher refinancing costs, they moved to offer higher interest rates when this was allowed in 1990 in order to compete with the larger banks. To recover the cost that such a strategy entails, the smaller banks also had to extend loans with higher risk. All of this happened while the stock price and real estate market "bubble" was well under way (cf. MOF 1993).
The problem in the process of deregulation was that the regulators allowed the basic features of the financial setup - indirect finance, stable bank rankings and window guidance, business area segmentation, and interest rate regulation - to disappear without preparing the system for the forces of competition. The financial authorities decided to continue the "cartellized cooperation" and situational regulation based on (often oral) administrative guidance, rather than introducing more straightforward regulatory patterns based on increased disclosure. Interest rates were fully deregulated at a time when many of the smaller banks were suffering from the competitive pressure exerted by the large banks, in the face of a faltering "bubble economy", and before effective rules against fraud had been introduced. It was also unclear whether lenders (investors) would shoulder the losses or mother banks would assume responsibility for bailouts. While the financial system could no longer be guided by situational regulation, there was a emergent need for cooperation in order to solve this problem without major failures in the financial system.
2.2. ...and where it is now
The "bubble economy" is not the only reason for the 1995 crisis. Nor is it a satisfactory explanation for the current difficulties. Next to behavioral, there were also structural factors that contributed to the bad loans problems and the ensuing financial crisis. Table 1 summarizes these factors.
| Structural | Behavioral |
|---|---|
|
|
Obviously, structural and behavioral factors interact in many ways. For instance, financial deregulation led to a partial breakdown of the main bank system through easy access to international financial markets, and also because companies were willing to jeopardize long-term relationships with one bank during the "bubble" period, when they felt invincible. Further, the partial breakdown obstructed the banks' monitoring ability, which in turn led to an increase in fraudulent behavior of both their customers and employees. However, the behavioral causes listed in Table 1 are not at all special to Japan; they could and did occur elsewhere, such as with the Savings and Loans crisis in the U.S. (see, e.g., Akerlof/Romer ). Instead, it is the structural causes that help us to fully understand what happened in Japanese finance in 1995. These structural factors also help explain why the bad loan problem triggered a system-wide crisis and an international Japan premium.
(1) Anachronistic Structure of the Banking System. The clear segmentation of the banking business into special categories of banks for special needs was addressed by MOF in the mid-1980s, and one bank category was dissolved. The Financial System Reform Act of 1993 paved the way for a piecemeal abolition of the remaining firewall between investment and commercial banking. But that was not enough. It took the failures of several credit cooperatives before a streamlining and merging of the financial cooperative system was addressed by local governments.
(2) Multi-layered Structure of Banking Supervision.Japan's multi-layered structure of banking supervision may not be appropriate for the financial markets of the 1990s. Commercial banks, large or small, are inspected regularly by both MOF (who looks after the legality of fiscal statements and transactions) and BOJ (who checks the soundness of loan positions). Smaller cooperatives, if they are inspected at all, are supervised by their local governments. Such a division of supervisory authority was also a problem in the U.S. S&L crisis. However, to make matters worse for Japan, there are only about 400 inspectors in Japan, as compared to 8000 in the U.S.. These 400 inspectors are in charge of 1250 financial institutions, not counting the nonbanks, i.e., there are 0.32 inspectors per institution, compared to the U.S. ratio of 0.8. Accordingly, while all U.S. banks are inspected once a year, Japanese bankers meet their regulators only every two to three years (Nikkei 11/21/1995, p.7). One of the reasons why there are not more inspectors in Japan is the "Total Staff Number Law" of 1967 which established a limit on the total number of bureaucrats in the government (AMA 1984). This cannot be the only reason, however, because if banking supervision was considered important enough, there would be ways to hire more inspectors. The political rationale behind the limited supervisory system seems to be that MOF is quite content with the close relationships with the regulatees. In contrast to the U.S., the functions of supervision, support, and protection of an industry are not separated across differing agencies; instead, these functions are all combined in MOF's Banking Bureau.
(3) "Yokonarabi" Policy, Convoy Protection, and "Wait-and-See Approach". The policy approach of MOF's Banking Bureau is in many ways still the same as in the period of rapid growth until 1973. Within one bank category, e.g., the city banks, MOF follows a "yokonarabi" strategy (lit.: "parallel, stagnating movement") which aims to maintain a stable ranking and hierarchy. This is enforced by permit and licensing requirements on new and original products which allow the regulator to suppress competitive moves by banks that are not supposed to move (Schaede 1994). One indicator of this approach is that, until 1995, MOF did not allow commercial banks to post pre-tax losses (NW 11/13/1995), because that would have undermined the "stability" of the system. Disallowing bad loan write-offs will eventually undermine the system in worse ways, because banks will engage in unofficial or illegal means to achieve the same end. Given the limited number and scope of inspections, false accounting was one way to disguise bad loans.
(4) Policy Tools: Administrative Guidance and Situational Regulation. The primary tools that the regulators use to implement their policies are administrative guidance and situational regulation 12 . In the banking industry, the patterns of administrative guidance are clearly observable. Often, when a large bank bails out a small, failing bank, it seems as if there was no rational economic justification involved, such as positive synergy effects. But a bail-out is often followed by a "favor" from MOF to the "white knight" supporting bank. Obviously, cooperation of this kind requires a constant interchange between the banks and the regulators. Every bank, therefore, has a special position called "MOF-tan", the designated MOF person, whose primary function it is to pay a visit to the MOF bureaucrat in charge and chat. Over time, this system has come to be called dang_ gy_sei (collusive regulation).
Administrative guidance is enforced with a carrot-and-stick mechanism: "If you are nice to me, I will be nice to you". The regulators strike deals with the regulatees in order to implement certain actions, such as stock "price keeping operations" in return for future support of the regulatee. Over time, the regulator gets entangled in a web of obligations and favors that render rule enforcement difficult.
"Entangled regulation" is related to the amakudari phenomenon of government officials who assume board positions in private banks and companies after early retirement at around age 55. The amakudari take on a double-sided role because they are paid by the private companies but remain dependent on the government personnel agency for further promotion (Schaede 1995). They often act as an intermediary and information channel between regulators and regulatees, but they can also serve functions as lobbyists on the one hand, and implementors of informal regulation on the other. Small Japanese banks have historically hired many second-tier retired MOF officials. In the case of the 1995 banking crisis, these amakudari are important because their existence makes it impossible for MOF to claim that it was unaware of the developments.
(5) Lack of Transparency and Collusion.The necessary consequences of a regulatory system that builds on administrative guidance are a lack of transparency and collusion. Hyogo Bank offers a good example (see below). When the banks faced problems in the 1990s, a retiring MOF bureaucrat was sent over to become the bank's president. Large city banks were asked to provide Hyogo Bank with low-interest rate loans. According to one rumor, Hyogo Bank and some other ailing financial institutions received emergency funds from BOJ in late 1992. MOF and BOJ stubbornly denied that this was the case, and BOJ supposedly demanded from the banks that this was not leaked to the mass media (Nikkei Kinyu 9/4/1995, p.1). Given the quid-pro-quo nature of administrative guidance, there must not and cannot be full disclosure or absolute standards and rules for who will be helped out. This lack of standards and rules, sometimes aptly labelled "go-tsug_ shugi" ("as it fits" or "according to circumstances" system) became MOF's biggest problem in the 1995 financial crisis and undermined its credibility in the eyes of many.
"By circumstance" regulation means that who is responsible in the event of a default may differ on a case-by-case basis. In order not to contradict themselves in public by ruling differently in similar cases, the general political solution for the bureaucrats is to negotiate a collusive outcome with all parties involved. The political drama surrounding the j_sen settlement is a fitting example.
These features of financial regulation created a system of "collusive regulation". This system helped the regulators to achieve the policy goals of stabilizing the financial system and channeling money into growth areas. The regulatees (banks) profited from the system through high profit margins and implicit bail-out insurances by the government. This system worked well until 1980, when the Japanese financial market was largely cut off from international influence. Internationalization in the 1980s, however, introduced outside forces into the system that slowly undermined its viability. When the bubble ended, MOF realized that collusive regulation did no longer achieve the goal of controlling bank behavior.
The developments in the 1990s have led some commentators to suggest that Japan is not prepared for a truly competitive financial market with unregulated interest rates. These commentators also argue that Japan will be better off without further financial deregulation. They claim that a country based on administrative guidance and long-term "quid-pro-pro" affiliations, where it is customary to ignore rules and were violations are met with leniency (amae), simply lacks the independent spirit necessary for true deregulation and competition (Kinyu zaisei jijo 10/2/1995, p.13).
3.1. The Credit Cooperatives
Credit cooperatives (Shiny_ kumiai, short: shinkumi or shinso) have their roots in the Meiji period but were reorganized in 1949 based on the "Law for Small Business Cooperatives etc." (Ch_sho kigy_-t_ ky_d_ kumiai h_). In 1951, many cooperatives were turned into non-for-profit Shinkin (Shiny_ kinko). Shinkumi, in contrast, are for-profit membership organizations for small- and medium-sized business entrepreneurs. They accept deposits and installment savings from, and provide loans to, members as well as national and local governments. Supervision was kept simple, allegedly to respect the cooperatives' independence. The establishment of a shinkumi, as long as business is confined to one prefecture, is automatically approved. Supervision is in the responsibility of the local government (see Suzuki 1987, BOJ 1995, Zenginkyo 1994).
There are only a few limitations on how surplus funds can be invested and on the size of loans. Most important is the "20% rule": Shinkumi are allowed to accept deposits from non-members of up to 20% of total loans or 400 million (increased to 800 million in 1989) 13 . This 20% rule created a major problem for the cooperatives under increased competition following interest rate deregulation. When the large banks moved towards smaller customers, the cooperatives could only compete by expanding beyond their traditional member customers. Further, because of their high cost structure, the small banks had to extend loans to risky projects for which they could demand higher interest rates, which in turn would enable them to offer higher savings rates.
The Two Tokyo Cooperatives
The first two major bank failures occurred in late 1994 when Tokyo Kyowa Credit Association and Anzen Credit were closed. Fraud was a major factor in these cases, and no "white knight" bank could be found to take over the two banks' business. The regulatory authorities funded the establishment of a new bank to assume all assets and liabilities, and closed the two shinkumi. Total bad loans of 150 billion had been accumulated by Takahashi Harunori, the chairman of Tokyo Kyowa since 1985 and president of EIE International Ltd since 1978. Through EIE, Takahashi made aggressive domestic and international real estate investments in hotels, golf courses, and office buildings. In its prime, EIE had 70 overseas subsidiaries and total assets of 1 trillion ($10 billion) (NW 3/13/1995). At the same time, Takahashi also began to "wine and dine" politicians and young MOF officials among whom he carefully picked those who were likely to make a career within MOF.
Takahashi's real estate investments were to a large extent financed by Long-Term Credit Bank of Japan (LTCB). In March 1992, LTCB's outstanding loans to EIE topped 380 billion. A year earlier, LTCB had imposed restructuring rules on EIE, but when the liabilities kept climbing, LTCB finally cut off all financial support in mid-1993. LTCB's president Horie Tetsuya resigned after Diet hearings in April 1995. Although he denied all allegations of reckless lending and any knowledge of the fraudulent lending schemes, LTCB agreed to support the bailout scheme for the two shinkumi with 27 billion (NW 3/20/1995; NW 4/3/1995).
Parliamentary hearings in spring 1995 revealed that Tokyo Kyowa had outstanding loans of about 13 billion to lower house politician Yamaguchi Toshio (who was arrested on December 7, 1995). Moreover, out of the 37.6 billion of loans from Tokyo Kyowa to companies controlled by Takahashi, 60% exceeded the legal maximum of 20% or 800 million to non-members. Between 1992 and 1994, Takahashi had arranged for loans of 18.2 billion from both Tokyo Kyowa and Anzen Credit to his group of companies, although he knew that these loans were uncollectible. Takahashi was accused of breach of trust and extending loans exceeding the legal limits for shinkumi. In 1995 he was indicted three times for a total of 19.4 billion of illegal loans to his affiliated companies (Mainichi 11/6/1995). In total, the "Tokyo Kyowa route" and the "Anzen Credit route" of illegal funding totalled more than 35 billion (Nikkei 11/10/1995). Entire bailout costs amounted to 186 billion ($1.86 billion). All assets and liabilities were taken over by Tokyo Kyodo Bank, established and funded for this very purpose by BOJ.
Fraud and real estate speculation, insider deals, and bribery were major factors in this case. Moreover, the multi-layered supervision structure created a problem. There was a lack of regulatory oversight by the Tokyo Metropolitan Government when the small cooperatives turned into major real estate investors overseas. Finally, the involvement of LTCB, a major bank with relatively strong monitoring capabilities, hints at a weakness of "relationship banking", which was an even more severe problem in the cases to follow in 1995.
Cosmo Credit Cooperative
The second shinkumi to falter in 1995 was Cosmo Credit, the largest credit cooperative in Tokyo 14 . In March 1989, Cosmo Credit's loans exceeded deposits for the first time; by April 1994, deposits were at 438.4 billion ($4.4 billion), and loans at 506.4 billion ($5 billion). Out of these loans, 60% were in real estate companies, and more than 70% were expected to be unrecoverable. The main regulator, the Tokyo Metropolitan Government, not only allowed Cosmo Credit to continue its business but also to attract new deposit with interest rates of up to three percentage points above average. Cosmo Credit also falsified its balance sheets and violated the 20% non-member loan limitation rule: 55% of all Cosmo Credit loans exceeded 800 million (Kinyu Bijinesu 10/1995, p.6).
The major reason for Cosmo Credit's default was rogue management by its president, Taido Sanpachi. His goal was to turn the cooperative into a commercial bank, and he expanded his business aggressively into real estate financing. All of his loans were based on physical (tangible) collateral (which is easier to evaluate that project loans). This collateral requirement meant that his loan customers, as well as depositors, were dispersed all over Tokyo. Because shinkumi, in principle, are meant to focus on a small local community, no "white knight" bank could be found, and cooperative had to be closed down.
Cosmo Credit had as one subsidiary a company group called SS Pharmaceutical Group, which served to guarantee Cosmo Credit's external funding (bank loans to increase the lending base). In 1991, with interest rate liberalization, Cosmo Credit introduced a high interest rate large-scale deposit called "mammoth", which carried a spread of 1.5-2 percentage points over similar instruments offered by city banks. In the three years between 1992 and 1994, the mammoth attracted 205 billion ($2 billion) in savings deposits. Taido postponed restructuring of the "mammoth" and got caught in falling interest rates in the early 1990s (Kinyu Zaisei Jij_ 7/17/1995, pp.34-35).
As early as in 1993, it was clear that Cosmo Credit was in serious trouble. The only reason Cosmo remained afloat was that it disguised its bad loans with more than 30 dummy companies. In so-called "self-auctions" (jik_ ky_raku) or tobashi, Cosmo Credit would sell a bad loan to an affiliated dummy company by having the dummy bid the highest price on the collateral concerned. The dummy in turn received a loan from Cosmo Credit for this purchase and for the future interest payments on the loan. On the book, a bad loan turned into an interest-earning loan.
After Cosmo Credit was closed, BOJ provided rescue loans and also forwarded unsecured funding of 20 billion to Tokyo Kyodo Bank, the institution originally set up to take over the business of Tokyo Kyowa and Anzen Credit. Seven banks with close business relations to SS Pharmaceutical were asked to contribute, and banks with outstanding loans to Cosmo Credit itself were requested to write off 63 billion ($630 million) or 60% of Cosmo's total loan balance. In contrast to the previous case, when the Tokyo Metropolitan Government balked, this time 20 billion were extended to support the bailout (Nikkei 7/4/1995). President Taido Sanpachi was held to be liable with personal assets. As was revealed in the fall of 1995, however, he mysteriously had no assets. The houses he lived in belonged either to one of his companies (organized as y_gen gaisha; i.e. with limited liability) or his brother-in-law (the president of SS Pharmaceuticals). Similar to Takahashi but not quite as organized, Taido had very close relations to high-ranking politicians and MOF officials. For instance, when he held a 40th anniversary party for Cosmo Credit on March 10, 1992, then Prime Minister Miyazawa sent a flower bouquet, and high-flying LDP politicians such as Kato K_ichi, and Yoshida Masateru, a former director general of MOF's Banking Bureau, attended personally (Kinyu Bijinesu 11/1995, p.33).
Next to fraud and reckless business expansion on the part of Mr. Sanpachi, this case reveals several regulatory mistakes. The first was the "convoy protection / "wait-and-see" approach: the regulators supported Cosmo's business expansion in the late 1980s in the face of a negative deposits to loan ratio. The second regulatory problem was that "self-auctions" are in fact a legal practice in Japan. Self-auctions are just one method of fake accounting, and by allowing this practice, the regulators invited non-disclosure. In combination, the "wait-and-see" attitude in the late 1980s, the deregulation of interest rates in the absence of strict disclosure rules, and a legal system that allows for accounting tricks led to the demise of this institution.
Kizu Credit Cooperative
The third failure occurred on August 30, when Osaka-based Kizu Credit Cooperative was closed at the same time as Hyogo Bank (see below). Kizu shinkumi was founded in 1953, and as of March 1995 had 27 branches and 674 employees. With deposits of 1.17 trillion ($11.7 billion) and loans of 1.98 trillion ($10.7 billion) in 1995, Kizu was the second largest credit cooperative in the country. By November 1995, the true extent of unrecoverable loans was estimated to be 960 billion. That is, more than 90% of Kizu's assets were unrecoverable, and total bailout and settlement costs was estimated to exceed 1.4 trillion ($14 billion) (Nikkei, 11/22/1995, p.1). While 500 billion of the settlement costs were to be covered by BOJ loans and Deposit Insurance funds, some 500 billion had to be furnished by large commercial banks. Some of these were accused of being at fault because of their policies of "referred deposits" (sh_kai yokin; lit: "introduced deposits"; see below).
Between 1989 and 1993, Kizu had built up a "deposit machine", with a monthly increase in deposits of 10 billion to a total of 840 billion in 1990. A significant portion of these were "referred deposits" from large city banks, in particular Sanwa Bank, LTCB, and Tokai Bank. Beginning with a mere 2 billion in 1987, at the peak in 1990 these accounted for 443 billion, or half of Kizu's total deposits (Yomiuri 9/22/1995). Over the course of five years, Sanwa Bank referred over 2 trillion ($20 billion) of deposits. In a typical "referral" transaction, the bank would talk its client into issuing low interest commercial paper (CP) and investing the proceeds in a large-scale high-interest savings deposit at Kizu Credit, thereby earning what appeared to be a free lunch 15 . The banks also profited because they earned a fee on the CP issue. Kizu, in turn, used these deposits as a lending base for high interest, high risk loans to real estate companies.
In 1991, MOF issued an administrative guidance requesting city banks to withdraw their "referred deposits". While this seemed to be a responsible regulation for the large banks, it turned out to be a regulatory blunder for the smaller institutions. Kizu's outstanding balance of referred deposits fell to zero by December 1992. To fill the gap, Kizu aggressively attracted new deposits by offering very high interest rates. The spread to average market rates peaked in 1992 with 2.6 percentage points (see Figure 1). On August 1, 1995 Kizu offered 3.5% on a 3-month deposit; this was 5-6 times higher than the average city bank rate of 0.6%. In order to recoup the deposit costs, Kizu had to make even riskier loans with high lending rates (Nikkei 9/14/1995).
88% of Kizu's loans were to real estate. Naturally, with falling land prices, Kizu accumulated non-performing loans. It employed three mechanisms to deal with these. The first was self-auctions similar to Cosmo Credit's. Between 1985 and 1995, there were 416 cases of "self auction", and auction income totalled 93 billion ($930 million) (Nikkei 9/30/1995). The second mechanism was to extend "bicycle loans" (oikashi; literally: over-loans) 16 . In order to deflect outright failures, Kizu offered defaulting customers a new loan with which to pay interest to Kizu. As of March 1995, Kizu had an estimated 60 billion ($600 million) of such bicycle loans. The third mechanism was simply that of false accounting at which Kizu excelled 17 . While the settlement procedures were discussed, the arrest of former professional boxer Watanabe Jiro highlighted a gangster side-story to the plot. Initially arrested for blackmailing, gambling with borrowed money, and some relation to a b_ryokudan (gang or mafia), a relation to Kizu Credit soon became clear. "Back-hand money" (ura kin) worth 38 billion ($380 million) related to the pro-boxer was deposited in large-scale deposits at Kizu through a real estate company, but withdrawn prematurely on August 29, the day before the bank was closed (Kinyu bijinesu 11/1995, p.32; TK, 9/16/1995, p.74).
As with the other cases, Kizu' story is one of fraud, reckless lending, high interest rates and fake accounting. However, the case also points at the regulatory problems of an anachronistic 20% rule, the "wait-and-see" approach, and lack of transparency and disclosure rules. In addition, the Kizu case highlights the problems of relationship banking and the question of who is responsible for a default when interbank relations are managed by administrative guidance. The guidance for large banks to limit their real estate exposure was a blunder not only because it jeopardized the small banks' lending base, but also because it was a situational regulation that attempted to remedy the consequences of risk shifting among banking categories from the large to the small banks.
Osaka Credit Cooperative
On December 7, the president of Osaka Credit, Kawase Tokunari, resigned and a plan was announced to continue operation as usual until Tokai Bank, one of the large city banks, would take over the cooperative's business in summer, 1996. In December 1995, Osaka Credit's deposits of 340 billion made it the fourth largest shinkumi in the country, while bad loans totalled 150 billion. The novel twist of this bailout was that Tokai Bank was to take over only the "good" business of Osaka Credit, while most of the bad loans were to be assumed by the RTC (the Japanese Resolution Trust Corporation, Tokyo Kyodo, see below). The cooperative had suffered from steep withdrawals of deposits after the failure of Cosmo Credit and Kizu Credit. In September 1995, the Osaka Credit received a non-collateralized loan from the National Federation of Credit Cooperatives (Zenkoku shiny_ ky_d_ kumiai reng_kai) for 50 billion, as well as a "support deposit" (ky_ryoku yokin) from Tokai Bank of 30 billion. The shinkumi was thus kept alive and running until the bailout plan was finalized. A bank run was averted.
Three problems led to Osaka Credit's demise. The first was that it was run in a "one-man" fashion without checks and balances. To expand his business, the president applied very lax standards to loan approvals during the bubble period. Second, Osaka Credit excelled in "bypass loans" (ukai yokin). These were intricate schemes involving affiliated nonbanks such as Osaka Mortgage (_saka teit_ sh_ken) to circumvent the 20% limit on loans to one customer. This was used particularly to extend loans to a crooked money broker, Nishiki Finance. Third, Osaka Credit had business relations with some of the outstanding notorieties of the bubble, such as Yokoi Hideki (Hotel New Japan) (Nikkei 12/8/1995).
When Osaka Credit went bankrupt, it had underwritten more than 5 billion of debt certificates issued by clients of a company called Nishiki Finance. This Kansai-based nonbank specialized in small business loans and declared bankruptcy on August 16, 1995. By the end of 1995, its failure had triggered more than 500 bankruptcies (and several suicides) of small businesses. In total, Nishiki had more than 2200 clients. To these, Nishiki provided loans based on "loan bills", which were simply IOUs. The trick was that Nishiki required up to 25 times the amount of the loan in bills; i.e., a company that needed 1 million had to write IOUs (debt bills) of up to 25 million. The understanding of the clients was that Nishiki would keep this paper in a vault and destroy it when the loan was repaid. Instead, Nishiki turned around and cashed in a total of 25.7 billion without authorization. Osaka Credit also extended loans to Nishiki through affiliated firms (the "bypass loans"). Osaka Credit claims that, rather than being at fault, it became itself a victim of Nishiki Finance (TK 9/16/1995, pp.78-79; Jitsugy_kai 11/1995, pp.73-77; Nikkei 11/19/1995, p.3).
The rescue plan for Osaka Credit is a possible model for future bailouts of failing cooperatives. The "white knight" bank takes over the business, but does not have to assume the bad loans. Instead, the bad loans are transferred to the RTC where they will eventually be settled with public funds. This makes it easier to find a rescue bank, and the authorities avoid having to shut the bank down. Also, settlement through the RTC uses public funds, rather than tapping into the depleted Deposit Insurance system (see below).
However, the scheme highlighted one problem. Tokai bank, the "white knight", volunteered to take the bank over because of a "strong request from the government and the City of Osaka, and for the stability of the financial system", not because it felt responsible for the bank (Nikkei 12/8/1995). Although the city bank had seconded young employees as well as an executive director to Osaka Credit regularly since the 1960s, it claims that it had no influence on management. This introduces the problem of how to define an "affiliated" bank, and what the responsibility and burden of such a bank should be.
Osaka Credit, again, is a story of fraud and imprudent banking. But it is also a story of relationship banking, risk shifting from large banks to small banks, and collusive regulation. The takeover by Tokai bank was made possible by a MOF promise to take over Osaka Credit's bad loans. These will eventually be financed by public monies through the Japanese RTC (see below). The responsibility for the bad loans, and Tokai's role by sending employees to Osaka Credit, was not discussed.
3.2. The Regional Bank: Hyogo Bank
Regional banks are local versions of city banks. Historically, there were 64 regional banks. Until the 1980s, there were also so-called 69 savings banks (s_g_ gink_) that were subject to a special law. As part of financial liberalization, the s_g_ bank category was dissolved in the late 1980s. Some of the former savings banks turned into credit unions, a few merged with other banks, and the rest were subsumed into a new category called "second-tier regional banks" (dai-ni chih_ gink_), subject to the same law as the large regional banks and city banks. Traditionally, the customers of the former second-tier regional banks were small- and medium-sized businesses and individuals who paid their debt regularly. However, beginning in the early 1980s, large city banks began to compete for smaller customers. Similar to the credit cooperatives, in 1990 the small banks began to offer higher interest rates on large deposits, and, in turn, had to advance loans to projects with higher risks. As of March 1994, the 64 second-tier regional banks had a total loan volume of 51.4 trillion ($514 billion), out of which, at that time, 14% (7.2 trillion) was deemed noncollectible (Iso 1995).
Hyogo Bank was established in 1944 in Kobe. As of March 1995, the bank had 147 branches, 2747 employees, deposits of 2.5 trillion ($25 billion), and loans outstanding of 2.77 trillion ($28 billion). This made it the country's largest second-tier regional bank. The bank's shares were listed in the First Sections of the Tokyo and Osaka stock exchanges 18 . Hyogo Bank had been hobbled by non-performing loans and a lack of leadership since at least 1992 when MOF sent Yoshida Masateru, a former director general of MOF's Banking Bureau, to initiate a restructuring program (this is the same Mr. Yoshida who attended Cosmo Credit's festive party in 1992). MOF requested the large city banks to provide Hyogo with very low interest rate loans. Nevertheless, after the Kobe earthquake of January 1995, the bank was too short of funds to compete with the larger banks for reconstruction loans. Worse, the local governments withdrew a major portion of their deposits to fund reconstruction projects. These were originally inserted to support the bank's recovery (Nikkei 8/31/1995, p.1). On August 31, Hyogo Bank was restructured: It remained in business, but all operations were transferred to a new bank called Midori Bank on January 26, 1996 19 . No change occurred for the customers in daily banking matters, and a run on deposits was averted.
Hyogo Bank faced two major problems when it defaulted. The first was its aggressive real estate lending during the bubble period and heavy "bicycle business" (loans to defaulting customers) since 1992. The second problem was its 20 affiliated nonbanks. A nonbank is an institution that offers financial services but does not engage in banking business in the narrow sense (intermediation between depositors and borrowers). Examples include credit card companies, housing loan companies, mortgage companies, and leasing firms. There are more than 3500 such institutions in Japan. Regulatory supervision of nonbanks was initially in the hands of MITI (Ministry for International Trade and Industry), but when the nonbanks acquired increasingly bank-like characteristics, it was eventually shifted to MOF in 1992. As nonbanks were supervised by MITI during the bubble years, banks and securities firms set up numerous nonbanks to circumvent MOF regulation.
On the day Hyogo Bank was closed, its bad loans were estimated at about 1.5 trillion ($15 billion, or 55% of total loans). On September 1, 1995, it was revealed that Hyogo Bank's 20 affiliated nonbanks held additional bad loans of 1.75 trillion ($17.5 billion). The biggest lender to these 20 affiliated nonbanks was Hy_go Bank itself with 381 billion, or 25% of the total (Nikkei 9/1/1995). Because Hyogo Bank was itself the mother bank, which was now restructured, it was unable to assume the losses. In contrast to the prevalent Japanese custom of "mother bank responsibility", and a bail-out by the mother bank, a "lender responsibility system" was invoked here with rigor for the first time in the postwar period (Nikkei 9/1/1995).
Commentators quickly pointed out that the Hyogo Bank disaster was a political failure and that the lack of clear rules for how to share the burden led to a collusive outcome (Nikkei 9/4/1995, Nikkei Kinyu 8/31/1995, p.1). Also, in treating Hyogo Bank in the same non-transparent manner as the credit cooperatives, MOF had violated the basic rules of the stockmarket. On August 18, 1995, the stock price of Hyogo Bank was at 248; it fell to 7 on September 6. The total number of issued and fully paid shares was 330 million, held by 11,000 shareholders. An audit report of May, 1995, estimated Hyogo Bank's bad loans to be a mere 61 billion. Three months later, bad loans of 790 billion were revealed. In May 1995, auditors had even issued a positive earning forecast. Commentators regarded this to be disingenious at best, and questioned the role of auditors in this case, and in Japan in general (Kiny_ zaisei jij_ 10/2/1995, p.12).
The Hyogo Bank case highlights five of the eight features of Japanese financial regulation: multi-layered supervision, the convoy protection approach, administrative guidance, lack of transparency and collusion, and amakudari regulation. As a second-tier regional bank, Hyogo Bank was subject to MOF supervision, but it also played an important local function. It shifted most of its bad business to nonbanks which were outside MOF supervision. When it ran into problem in the early 1990s, MOF requested the local authorities and large city banks to deposit "support monies" at Hyogo Bank, thereby diluting the issue of lender responsibility. MOF also sent a retiring official to take over the presidency of the bank and to implement an improvement plan. This amakudari official largely assumed the function of a permanent on-site inspector and advised MOF not to formally inspect the bank, unless it wanted to take action. Because of the convoy approach, the underlying assumption that Japanese commercial banks should not go bankrupt, and the "wait-and-see" approach, MOF did not take any further action. The 1995 earthquake undermined all MOF efforts to cover up problems, and it was forced to address the issue when Kizu Bank was impacted by the defaulting credit cooperatives in Tokyo. As a result of the entanglement that administrative guidance created, no large bank could be held responsible for the problem, and the city bank closest to Hyogo refused to bail the bank out.
3.3. The J_sen (Housing Loan Companies)
The j_sen have become a leading symbol of Japan's bad loan problem, although their bad loans account for less than one sixth of the total bad loans. The j_sen combine all the elements that characterize the crisis: Irresponsible lending, regulatory lapses, fraud, poor regulatory oversight, "favor banking" among financial institutions that belong to the same networks, and a lack of clear responsibility between lenders, owners, and parent firms.
There were a total of eight j_sen (J_taku kiny_ sen mon kaisha, literally: "Special housing finance companies"). Of these, one belongs to the agricultural sector and is still afloat, while the other seven were dissolved beginning in March 1996. The history of individual housing loans is rather short in Japan. City banks and regional banks only entered the business in the 1960s, and at the time treated housing loans as consumer loans. With the trend towards the core family, urbanization, and economic growth, the housing loan market grew in the 1970s. J_sen were established by different groups of financial institutions as special subsidiaries to circumvent interest rate regulation. While under direct supervision by the Ministry of Finance, the j_sen, as nonbanks, could charge higher interest rates. Their conditions were less strict, and they also provided loans for the construction of apartments, stores, and office buildings. J_sen were refinanced to more than 90% through long-term loans, of which roughly one third were loans from the parent institutions. The jusen lent money with a spread of 1.02% points over their borrowings (MOF 1992, p.180). In 1976, j_sen accounted for 3.7% of all housing loans, but their share had increased to 12% in 1991.
In the second half of the 1980s, the housing loan environment changed dramatically. The market in total grew, while corporate loans were decreasing and interest rates were successively deregulated. The large banks now began to compete in this market, and the j_sen were driven into real estate loans and loans to developers, just as the bubble economy took off. By the time the real estate market collapsed in 1990, the j_sen had turned into "special real estate lenders". Neither the parent institutions nor the regulators intervened, and the amount of non-performing loans increased sharply through the 1990s recession. In 1994, total outstanding loans of j_sen were 12.8 trillion, but only 2.8 trillion were housing loans (BOJ 1995, p.391; SD 9/16/1995). In 1995, MOF estimated that 7 trillion of the j_sen loans were irrecoverable.
One of the largest groups of lenders for the j_sen were agricultural cooperatives (n_ky_), federations of these at the prefecture level (shinren) 20 , and their "central bank" N_rin Ch_kin. In 1994, there were 7350 agricultural cooperatives nationwide, of which 3373 also engaged in complete financial activities. As of March 1995, these 3373 n_ky_ were estimated to hold deposits of 74 trillion, with outstanding loans of 18.8 trillion. Because agricultural cooperatives, until 1992, faced the same 20% rule as the shinkumi, they entrusted about two thirds of their deposits to one of the 47 prefectural financial federations. These federations, also subject to the 20% rule, invested part of their surplus funds in Norin Chukin (which had total deposits of 28.4 trillion in 1994) (MOF 1992, Zenginkyo 1994, NW 12/25/1995).
In October 1980, the Ministry of Agriculture, Fishery, and Forestry (MAFF) and MOF made a special exception to the 20% rule on loans to non-members for the prefectural financial federations for loans to j_sen. Immediately, agriculture savings began to flow to the j_sen. Twelve years later, in June 1992, the restrictions on loan guarantees and loans to non-members of 20% to total loans were completely abolished for all agricultural cooperatives, including the local n_ky_ (MOF 1992, p.162). In the meantime, in March, 1990, MOF introduced a first countermeasure against the emerging real estate debacle: A loan limit on real estate was introduced for commercial banks. This led to a sharp drop of bank loans to the j_sen. In combination, these policies turned the agricultural cooperatives into the largest lenders, with total loans to j_sen of 5.6 trillion in 1993. With more than 50% of their loans extended to the j_sen, the federations were highly exposed. While the banks claimed that these loans were obviously risky, the n_ky_ insisted that there was an implicit understanding with MOF and MAFF that these loans were safe 21 (SD 9/16/1995; MOF 1992, 163; NW 12/25/1995, Nikkei Kinyu 10/3/1995).
Table 2 lists the bad loans, mother banks, and major lenders of the seven j_sen that were dissolved in March 1996. The seven companies are quite different in character, both in terms of size and dependence on the mother banks. Nippon Housing Loan, until 1992 considered the largest and best of the j_sen, and Dai-Ichi Housing Loan were publicly traded on the Tokyo Stock Exchange. Nippon Housing Loan was run by a retired MOF bureaucrat for 16 years. In total, there were 12 former MOF bureaucrats in leading positions at the 7 j_sen in 1995. At Housing Loan Service, the president was recruited in alternating order from three of the city banks, and the mother bank portion in both loans and referred deposits was higher than at the other j_sen. At Juso, fraud played a major role. Juso had extended loans on land collateral with zero value, such as forests, and employees reportedly were ordered to change the pictures and the text on the loan approval forms prior to MOF inspections (TK 8/26/1995, p.56; TK 9/16/1995, p.72). There is some evidence suggesting that Juso, Chigin-Seiho, and S_g_, the three j_sen with the worst bad loan ratios, were forced into risky loans by their financial suppliers (mother banks or other lenders) (Asahi 8/22/1995). Part of the referred loans may have been aimed to support the mother bank by lending money to a troubled client so that the client could pay interest to the bank. Other risky loans were an extension of "relationship banking" triggered by the MOF guidance that stopped banks from increasing loans to the real estate sector after 1990. Banks therefore requested the subsidiary to take over, regardless of the creditworthiness of the customer or the collateral.
Of the total loans outstanding to the j_sen of 13 trillion in March 1995, 5 trillion were furnished by the top 21 banks (city banks, trust banks, banks for long-term credit), 2 trillion by smaller banks and life insurance companies, and 6 trillion by the agricultural cooperatives (NW 11/27/1995). This helps explain the fierce fight that developed in fall, 1995, between the two major lender groups. The banks favored a "lender responsibility" approach which would have left the n_ky_ with a major portion of the bill, while the n_ky_ insisted on a "mother bank" approach which would have freed them from all responsibility. The result of this fight will have critical long-term ramifications for the contractual and institutional structure of Japanese finance. (See Appendix 1 for a summary of the argument).
Eventually, a "revised mother bank system" was agreed upon. This meant that the mother banks had to shoulder the major portion of the bad loans but received some support. A plan proposed on December 25, 1995, devised a two-phase scheme which included more than 1 trillion of public fund support. The mother banks were to write off completely their total loans to j_sen of 3.5 trillion, while the agricultural cooperatives were to keep all their principal, but supported the write-off with a contribution of 530 billion (less than 10% of the amount of their total loans of 5.5 trillion). By dividing the scheme into two phases, the regulators tried to conceal the fact that total irrecoverable loans exceeded 9.9 trillion, and total public funds exceeded 1.3 trillion. A political battle ensued over the allocation of these amounts, and the size of public funds used. The politicians offered what amounted to a horse trade: they would only support the use of public funds if those responsible in the banks and the ministries were removed and punished. To save the deal, MOF's administrative vice minister Kyosuke Shinozawa offered his resignation to the Finance Minister on December 26, 1995.
MOF was questioned in public in various ways. The first problem was that of former MOF bureaucrats serving as presidents at some of the j_sen. These connections leave no doubt that MOF knew the details of the problems for many years. Further, MOF shared supervision of the n_ky_ with MAFF, and supposedly promised in 1992 that the agricultural financial institutions would not suffer from this disaster (see appendix 1 for details).
Moreover, there had been no investigation of the j_sen since 1992. One possible reason is that MOF, knowing about the dilemma, did not investigate because it was hoping for a real estate price recovery which would have solved part of the problem without any action or public attention.
The j_sen problem features all the characteristics of the Japanese regulatory process. Institutionally, the j_sen were a leftover from the period of rapid growth and interest rate regulation and should have been abolished when the banks entered the housing loan market in the 1980s. In the same way, the agricultural cooperatives had outlived their raison d'etre when large banks entered the regional markets. Second, because these institutions served special functions, they were regulated by various agencies: The nonbanks, until 1992, by MITI, and the agricultural cooperatives by MAFF, and both also by MOF. When problems occurred in the early 1990s, MAFF and MOF were unable to find solutions that pleased all constituencies, and therefore adopted a "wait-and-see" attitude. Third, Regulation of both jusen and agricultural cooperatives was almost exclusively based on administrative guidance and "by circumstance". When MOF wanted to curtail real estate financing of city banks, it lifted the 20% rule specifically for j_sen loans by agricultural cooperatives to ensure further funding of the j_sen. This was a regulatory mistake, and it helps explain why the agricultural cooperatives were not held liable. Fourth, the amakudari (retired MOF officials) play an important role. It needs to be noted that the quality of the amakudari depends on the quality of his work and career while at MOF; j_sen were not among the high-ranking amakudari positions. However, the amakudari certainly functioned to inform the ministry of the severity of the situation. Therefore, the fact that MOF did not inspect the j_sen between 1991 and 1995 does not mean that MOF was unaware of the situation. Rather, MOF was too entangled, both with the large mother banks and with other ministries and agencies, to take actions, and therefore chose to pretend not to know. As shown in Table 2, total bad loans of the seven j_sen almost doubled between 1991 and 1995. Taken together, these regulatory mistakes cost the Japanese system more than 3.48 trillion ($35 billion).
The j_sen affair has had a major impact on both the structure of the banking system and regulatory oversight. On the regulatory side, the discussion of "mother bank responsibility" has jolted a pillar of the postwar Japanese system of corporate finance. This system operated on the tacit assumption that there is always a "main bank" that assumes a bail-out function should problems arise. Over time, this assumption, paired with encompassing regulation of bond issues and loan collateral, led to a system in which investments were considered to be risk-free. Deregulation took away the regulatory protection, but it became obvious that only a disaster would change the deeply ingrained risk-free investment mentality. The j_sen have provided this disaster.
In terms of the structure of the financial system, the bankruptcies of credit cooperatives have triggered a reorganization of the cooperative sector. Because of obsolete business practices and increasing competition from commercial banks, many of the agricultural cooperatives and federations are likely to disappear. The huge amounts of bad loans, to j_sen and other nonbanks will also undermine the viability of some of the large banks, and major mergers are to be expected over the next five years. The fierce, and often ugly, discussion during the j_sen debate may have introduced a further rift into the once quite cozy, and certainly mutually polite, world of banking and finance. And finally, the ways in which the bad loans are written off and digested by the major banks will bifurcate the banking system into strong and weak banks. The old, stable, and predictable large bank hierarchy was destroyed.
4. The Process Of Bad Loan Accumulation
With this factual background, we can now analyze the major patterns of bad loan accumulation. The mechanisms employed can be grouped in three categories:
(1)Improper bank management and the circumvention of rules (above limit loans to one customer; reckless real estate lending; super-high interest rates; self-auctions; bicycle loans; non-disclosure, false accounting);
(2)"Relationship business" (referred deposits; referred loans; risk shifting to nonbanks); and (3)Fraud (insider trading; mafia connections; bribery).
Table 3 summarizes these categories for the six failure cases of 1995. A pattern emerges. All banks are characterized by reckless real estate lending, non-disclosure, and false accounting. Most of them also offered above market rates in an effort to stay afloat 22 .
Improper Bank Management
The activities grouped in category 1 are violations of rules. There are two possible reasons for non-enforcement. First, the regulators may not have known about the violations, because of non-disclosure or sloppy investigation. Given the size of these activities, and teh presence of retired MOF officials as amakudari, this is unlikely. Second, and more likely, the regulators did know, but did not enforce the rules for fear of having to address the underlying structures of the banking system. Most of these activities are an outgrowth of former mistakes or old regulation that proved to be problematic in an environment of competition with large banks. Consider the 20% rule for credit cooperatives. This rule limits the competitiveness of small banks at a time when the raison d'etre for these cooperatives - to take care of small business that is ignored by large banks - has disappeared. Cooperatives that obey the 20% rule will necessarily go out of business.
The second problem area is reckless real estate lending. At one level, this is simply an outgrowth of the bubble economy. At another level, however, this is a problem of the introduction of competition into a system not based on strict disclosure rules and monitoring but on "cartellized cooperation", and, therefore, provides temptation and opportunities for shady deals. All institutions that failed in 1995 had ratios of bad loans to total loans of more than 50%. This is striking in light of the original spirit of cooperatives of mutual help and trust. Under the old regulatory approach of "convoy" protection, this was not a problem, because everyone made money. The regulatory mistake was to allow competition into a system that was not prepared to sufficiently supervise the actors.
This is related to the premium interest rates offered by the failing cooperatives. The coincidence of deposit and loan rate deregulation and the MOF guidance to large banks to withdraw their referred deposits forced the cooperatives to compete on price. This led them to engage in loan projects with higher risks. When these soured during the recession of the 1990s, it led to a vicious cycle of tobashi and bicycle loans.
Both tobashi and bicycle loans aim to achieve the same end. A tobashi (lit: "something that is flying"; a sailing line item) is an accounting gimmick. Bad loans are transferred to another company, often a dummy, with the sole objective of having it temporarily disappear from the books. In relation with the real estate debacle, tobashi were used for "bicycle loans" and "self auctions" (jik_ ky_raku). The problem was that both are legal practice in Japan. Tobashi, therefore, fall into the category of regulatory mistakes.
Relationship Business
The virtues of "alliance capitalism", long-term stable trade relations, mutual shareholdings, and a network of business groups have been pointed out 23 . The vices of the keiretsu system are usually considered in terms of closed markets and limited foreign access to Japanese trade circles. However, the 1995 banking crisis highlights a distinct downside to this system for the domestic market. This is revealed when the leading firm or bank misuses its power and exploits affiliated companies by shifting to them especially risky business, regardless of the desires of the affiliate. In the 1995 financial crisis, the mother banks refused to assume responsibility for the ensuing management difficulties of their affiliates. That is, the affiliated nonbanks were used for purposes of rule circumvention and transfer of risky projects but were left unsupported when their loans became irrecoverable.
As Table 3 reveals, Kizu Credit is the blackest sheep of all 1995 failures, with record bad loans, self-auctions, and bicycle loans. Not coincidentally, Kizu also was the largest recipient of "referred deposits". Large city banks have two motivations to refer both deposits and loans to smaller banks. First, these enhance business ties with large corporations by introducing them to high interest rate deposits. In the period when parts of the market were interest-deregulated while other were still regulated, even a simple arbitrage scheme resulted in a "free lunch" for the corporation and fee for the large bank. Second, referring loan business to smaller institutions served the purposes of relationship banking. By referring a client to a smaller institution, banks could circumvent the Banking Law regulation that prohibits them from extending disproportionately high loans to one single customer 24 .
Referred deposits as such are legal, but if they induce a cooperative to exceed the 20% limit, they become illegal. Further, a large portion of the referrals fell in the category of "covert guided deposits" (d_ny_ yokin magai), in which the referring city bank designated both the recipient of the loan as well as the interest rate. This constitutes a direct intrusion by the large bank into the business decision of the cooperative. Usually, such tight business relations were accompanied by personnel exchange (shukk_) and the sending of a managing director from the large bank to the small bank 25 . Nevertheless, Sanwa Bank denied any involvement in Kizu Credit and initially refused to support the cooperative's bailout. One banker commented: "For the big banks, these small institutions are like toys. After the fact, they claim they didn't know it" (TK9/16/1995, pp.74-77).
Fraud
Next to patterns of behavior that are best explained by exploitation of situational regulation and regulatory negligence, there are those that simply constitute fraud. An argument can be made that fraud will always happen, no matter how the regulatory system is designed: If someone wants to cheat, he will. The more important issue is how fraud, once it has occurred, is dealt with. The Japanese default mechanism seems to be not to deal with it at all.
The first item in this category is "insider deals", such as the withdrawal of large amounts of deposits one day before Kizu Credit was closed, or the triangle loan relation involving the two Tokyo cooperatives and their affiliated pharmaceutical firm. Most likely, the full extent of these transactions will never be revealed because they also involve politician and bureaucrats 26 .
In addition to systematic deception, there were also individual cases of fraud that added to the overall picture. The business relations between Osaka Credit and Nishiki Finance (see section 2.1.) are a case in point. Complicating everything is the possible involvement of criminal gangs (yakuza). One way in which the yakuza is involved are direct bank loans, particularly from small banks and nonbanks. One indicator of the yakuza role in the j_sen affair is that none of the companies that owed money to the jusen has been shut down, even after the j_sen themselves were dissolved 27 .
5. The Bad Loan Situation in 1995
5.1. Overview
Table 4 summarizes the bad loan situation in 1995 by bank categories according to MOF announcements. However, when MOF announced in March 1995 that total bad loans of all banks were 40 trillion, no one in the banking industry believed that (TK 6/24/1995). Rather, Toyo Keizai, a weekly economic magazine, kept publishing numbers that contradicted and embarrassed MOF. For instance, in June 1995, Toyo Keizai published an estimate that the true number of bad loans of the top 21 banks was 24.6 trillion, rather than 12.5 trillion (TK 6/24/1995). As Table 3 indicates, by September 1995, MOF had revised its numbers to 23.4 trillion. In June 1995, Toyo Keizai also estimated that the smaller banks faced bad loans of about 27 trillion, so that total bad loans in the banking system were 51.5 trillion. The magazine was also consistent in pointing to the problem of nonbank bad loans. These will eventually turn into bad loans for the lender banks and are estimated those to reach 20 trillion, so that the total amount of bad loans reached 71.5 trillion.
Table 5 shows a discrepancy of 11 trillion in the MOF numbers for the top 21 banks between March and September 1995. The MOF numbers were based on a survey of the top 21 banks. The discrepancy between March and August 1995 can be explained by the fact that new bad loans surfaced, while land prices and the stock index kept falling. In addition, MOF's definition of a "non-performing loan" does not reflect the full picture and was revised in fall 1995 28 . Disclosure rules are vague, and banks have external and internal bad loan numbers, but the MOF survey quoted only the external figures. Finally, the MOF bad loan figures exclude all nonbanks. It is in this category are that most of the "bad surprise" loans are hidden. Tables 6 and 7 shed some light on the nonbank loan situation of the largest commercial banks. Table 6 shows the portion of the largest twelve loans by city banks and banks for long-term credit that were directed to nonbanks in the areas of real estate, mortgage lending, leasing, and developing/home construction. On average, 57% of the largest loans of the top 14 Japanese banks went to risky borrowers. If only the largest 12 loans are considered, the top 14 banks furnished nonbanks with an average of 715 billion of loans. Table 7 draws a picture of nonbank and real estate loans. As of March 1995, the largest 61 nonbanks (including the j_sen at the time) had total loans outstanding of 42.2 trillion, or 69 billion on average. Out of these loans, 47.5% came from the top 21 banks. No information is available on the amount of additional loans to these nonbanks from bank subsidiaries. At the same time, the largest 30 real estate companies had total loans outstanding of 9.3 trillion, of which 52% were provided directly by the top 21 banks. Again, no figures are available for loans by the banks' subsidiaries to this sector. That is, nonbank bad loans are huge and 50% of them come from the very large banks. Table 3 strongly suggests that the official MOF figures, even the revised set of September 1995, fall short of reality by an order of magnitude of 20 trillion.
In the final analysis, one of the biggest changes that the bad loan crisis will bring for Japanese finance may be that it undermined the strict hierarchy among banks that characterized the financial system in the postwar period. The top 21 banks are now clearly divided into strong and weak banks, and some of these may not survive. Also, the large banks adopted different strategies and time frames for dealing with their bad loans, so the bifurcation of the banking structure is bound to become more pronounced over the coming decade.
Table 6 outlines strength and bad loan exposure for the top 21 banks as of March 1995, by indicating operating profits and securities reserves on the one hand, and total bad loans, as made public at that point, on the other hand. The ratio of profits and liquid reserves to bad loans is presented in column 8. In only five cases was this ratio positive. Mitsubishi Bank was in best shape, followed by Asahi Bank, IBJ, Sanwa Bank, and Bank of Tokyo. At the other end of the spectrum, we find all trust banks, two banks for long-term credit, and Hokkaido Takugin. Five of these banks received a de facto "junk" rating from Moody's rating agency in August 1995. The table, however, may be misleading as to the strength of even the five banks with positive ratios. Again, the "bad loans" include neither direct loans to nonbanks nor bad loans held by affiliated nonbank subsidiaries (on average, the large banks have 2.4 affiliated nonbanks; cf. Table 5, last column).
5.2. Deposit Insurance
The Deposit Insurance Corporation (DIC) was established in 1971 to insure deposits of up to 10 million ($100,000). The senior deputy governor of the Bank of Japan ex officio also serves as DIC President. In 1986, the DIC's charter was extended to allow for cash support of takeovers of failed institutions. As Table 9 shows, until 1995, this was the DIC's primary activity. DIC supported five "white knights" that took over failing smaller institutions. Even with the failures in 1995, the basic logic remained to transfer business to another bank and support the activities of this new bank with DIC funds (NW 2/27/1995). As of March 1995, the DIC fund totalled 876 billion ($8.8 billion, as compared to $13.1 billion in the U.S. at the time). This covered about 0.15% of total deposits of 541 trillion at the time. The DIC has an additional credit line of 500 billion from the BOJ.
In 1995, bailouts exceeded these amounts. After the DIC had supported Tokyo Kyowa to assume the business of the two Tokyo credit cooperatives and Cosmo Credit with a total of 140 billion, Kizu Credit and Hyogo Bank required a total of 800 billion in support. The DIC fund was depleted, and, for the first time, the DIC had to utilize its BOJ credit line. By August, the biggest worry was that if there were a further bank failure in fiscal year 1995 (which lasts until March 31, 1996), the DIC might be pathetically underfunded 29 (Nikkei Kinyu 8/31/1995,p.1).
In December, 1995, MOF decided on a sharp increase in the premium from 0.012% to 0.048% of the total balance for all types of deposits (in comparison, the U.S. FDIC rate varies by risk and can be as high as 0.27%). This raised the costs of banks' contributions from 70 billion to 490 billion a year, or 8% of projected net earnings. In addition, DIC's credit line at BOJ was extended from 500 million to 1000 billion (FT12/23/1995; NW 11/13/1995). These system changes suggest that the financial authorities were preparing for further bankruptcies in 1996.
It is important to note that the Japanese logic for the DIC system differs from that in the U.S.. The basic Japanese idea is to use these funds to make sure that no bankruptcy will occur, so that the DIC's primary role is to support the "white knights". Tokyo Kyowa Bank was initially designed as a bank, albeit with the function to assume the credit cooperatives' business and sell off their bad loans. This system favors large depositors who otherwise would not be insured. By building on rescue institutions, it also creates routes for disguising the insertion of public funds into the system.
In November 1995, MOF created an additional special bailout fund (tokubetsu kikin) supported by the government, regional local governments, and the private sector. Initial funding was 200 billion, and this fund was to play the role of a second, private/public deposit insurance. The insurance premium for the large banks was an additional 0.048%, on top of the DIC's 0.048%. MOF's argument to convince the large banks was that the two Tokyo Credit and Cosmo Credit cases had shown how difficult it was to obtain support from city governments (i.e. local politicians). Therefore, the MOF bureaucrats designed an administrative solution to circumvent politicians altogether (Nikkei 9/30/1995, p.1). The organization of this fund as a public/private kikin ("fund") allowed MOF to channel public monies without attracting the attention of the public. In effect, this MOF scheme designed a way to prop up failing banks without political checks and balances 30 .
5.3. Tokyo Kyodo Bank - The Japanese RTC
With the failure of the two Tokyo Cooperatives, the financial authorities for the first time in the postwar period faced the situation of a bankruptcy where no bank could be found that was willing to take over. However, rather than simply paying the depositors off from the Deposit Insurance system and closing the business down completely, MOF and BOJ founded Tokyo Kyodo Bank. While it looked like a bank, its function was to assume the business of the defaulted cooperatives and to sell off their bad loans in a period of five years. Initial capital of the bank was 40 billion, of which 50% was provided by BOJ and the remainder by commercial banks. The 15 executives of the bank were recruited from MOF, BOJ, and the large commercial banks (Ito 1995).
By September 1995, Tokyo Kyodo faced overdue loans of 58 billion, or 90% of the bank's total assets of 65 billion. In addition, it had been asked to also assume the business of the failed Cosmo Credit and Kizu Credit. The authorities decided to turn Tokyo Kyodo into a Japanese version of the U.S. RTC (Resolution Trust Corporation) which had assumed and then sold off the assets and bad loans of the failing Savings and Loans in the 1980s. As a specialized organization for dissolution management, the Japanese RTC is eligible for support through funds from the general budget and the special bankruptcy fund created under the Deposit Insurance system. It may also request support from politicians, the courts, and the police. The RTC's function is to assume a defaulting bank's business, if no other bank is willing to do so, to sell off the bad loans, and to repay all deposits. As became clear in the case of Osaka Credit, an additional task is to support a "white knight" bank by taking over some of the defaulting bank's bad loans. This, over time, may become the most important function of the Japanese RTC (Nikkei 11/22/1995, p.1.).
At the end of 1995, the Japanese financial system was under a storm cloud of possibly more than 70 trillion in bad loans. Two thirds of the top 21 banks (out of which eleven are among the 20 largest banks in the world) were to post pre-tax losses for FY 1995. It can be expected that orchestrated mergers on a large scale will lead to significant changes in the banking landscape. For the smaller banks, the failures of seven cooperatives are only a leading indicator of the real problem. Many of the bank categories will disappear, leading to further changes for the large banks.
The 1995 financial crisis will enter Japanese history books as a crucial point in the development of the postwar financial system and banking structure. The 1995 crisis has changed the face of banking as it brings the end of the yokonarabi (in parallel) system. It has also touched upon one of the basic features of postwar Japanese financial system, corporate finance, and corporate governance: The mother bank responsibility system. While the j_sen solution still sustains the mother bank logic, the public discussion about this logic and the problems associated with a lack of lenders' responsibility (such as collusion, unfairness, and deception of the public) have seriously undermined the assumption that all investments are based on mother bank guarantees. If the lenders' responsibility becomes part of the investment decision, the close contact with the regulator becomes less important. In the long run, a lender responsibility approach undercuts the importance of administrative guidance. The twist is, however, that neither the regulators nor the banking industry seem to want this to disappear.
The failures of 1995 were all related to speculative real estate financing and fraud related to the bubble. But while the crisis would have been much less pronounced without the bubble, there were other problems that are likely to have triggered a crisis. This analysis of the 1995 crisis has highlighted five areas of major structural problems and policy weakness.
(1) Lax disclosure rules allowed banks, large and small alike, to hide or disguise losses and shady deals. When the regulators found some of the fake accounting entries, they responded with leniency in the hope that restructuring would avert bankruptcies.
(2) In 1990, regulators adopted a "wait-and-see" attitude in the hope that the market would recover and bad loans would heal themselves. When the market did not recover by 1994, it became clear that this had been a policy mistake. The regulators tried to cover it up, thereby adding to the overall bail-out bill. It is this "wait-and-see" regulatory attitude that explains why disclosure rules were kept flexible: Fake accounting suited the regulators' strategy of betting on the market recovery.
(3) There was a problem of "out-of-synch deregulation". In 1979, MOF initiated interest rate deregulation very slowly and carefully, in an attempt not to undermine the banking hierarchy. However, MOF failed to introduce regulatory measures that were needed to balance market forces, such as standard disclosure rules to enhance information openness. By not introducing such rules, MOF kept alive the basic logic of "mother bank responsibility" in an environment where interest rates were not managed in a way that would allow the mother bank to assume responsibility.
(4) Related to this is the heavy reliance on administrative guidance in banking regulation. In the period of rapid growth, when all interest rates were regulated, the financial authorities had to rely on administrative guidance because the price mechanism did not work. Extra-legal administrative guidance, especially the ad-hoc "situational" part, gives bureaucrats leverage over the industry because it creates uncertainty for the regulatees, and there is usually no recourse against it; i.e., administrative guidance means power for the regulators. Accordingly, when the price mechanism was slowly introduced into the financial markets in the 1980s, the bureaucrats still clung to this regulatory tool. This explains why MOF kept the "mother bank responsibility" system alive. In an environment of administrative guidance, the logic of the lender's responsibility does not work.
(5) As a result of all of the above, there is little policy transparency, and solutions to problems tend to be collusive based on the "quid-pro-quo" logic of administrative guidance. This worked well in the "period rapid growth" (1950-1973), because there was little interdependence with other financial markets, both within and outside Japan. For instance, one of the problems in 1995 that would not have occurred in the 1950s is the treatment of the general shareholders of Hyogo Bank. In the restructuring of Hyogo Bank, the role, rights, and obligations of shareholders were completely ignored.
The major reason why the financial crisis occurred, and why it was not handled efficiently, is that the financial system and its regulatory structure did not evolve from the 1950s system. The financial structure never shed its "rapid growth" characteristics. The pursuit of interest rate deregulation in this environment only made matters worse. The problem with this non-evolution is a lack of tools to effectively deal with the problems of the deregulated system.
The reason why the financial structure did not evolve is that a regulatory system of administrative guidance and collusive regulation is beneficial to both regulators and regulatees: It gives leverage to the bureaucrats and it allows the banking industry to exploit the flexibility in the regulation to their own advantage. This system may have some positive consequences. Regulatory cartels, like other cartels, can serve to lower total costs for the cartel members. Given the extent of bad loans, this could benefit the economy as a whole. At the same time, cartels also often work to raise the costs for those who are not members, in this case the Japanese taxpayers. MOF's initial attitude to avoid problems and wait in the hope that real estate prices would recover soon has already made the bill much higher than it would have been if the problem had been addressed in 1992. Public funds were needed to solve the major bankruptcy problem. As long as the economy grew, the costs of collusive regulation were of secondary importance. A crisis was needed to highlight the deficiencies and costs of a system of collusive regulation.
The author is grateful for valuable comments from Marie Anchordoguy, Mary Yoko Brannen, Takeo Hoshi, Thomas Lifson, Leonard Lynn, John McMillan, Masao Nakamura, John Ries, Tom Roehl, and the members of the PRISM group. Financial support by the Abe Fellowship Program is gratefully acknowledged.
References
(a) Newspapers and Magazines
Aera (weekly general magazine)
Asahi = Asahi Shinbun (daily general newspaper)
FT = Financial Times (daily)
Kinyu Bijinesu (monthly financial magazine)
Kinyu Zaisei Jij_ (bi-weekly financial and fiscal magazine)
Nikkei = Nihon Keizai Shinbun (daily economic newspaper)
Nikkei Kinyu = Nikkei Kinyu Shinbun (daily specialized financial newspaper)
NW = Nikkei Weekly (English language weekly digest of Nikkei)
SD = Sh_kan Daiyamondo (weekly economic magazine)
TK = Toyo Keizai (weekly economic magazine)
Yomiuri = Yomiuri Shinbun (daily general newspaper)
(b) Articles and Monographs
Akerlof, G. / Romer,...
AMA (Administrative Management Agency), 1984, "Staff Number Control", in: Tsuji, Kiyoaki (ed.), Public Administration in Japan; Tokyo: University of Tokyo Press, pp.71-86.
BOJ (Bank of Japan) (ed.), 1995, Waga-kuni no kiny_ seido (shinpan) (The Japanese Financial System; revised edition); Tokyo: Nihon Gink_ kiny_ kenky_jo.
Dattel, Eugene R., 1994, The Sun that Never Rose - The Inside Story of Japan's Failed Attempt at Global Financial Dominance; Chicago: Probus.
Eguchi Hidekazu / Hamada Koichi, 1978, "Banking Behavior under Constraints: Credit Rationing and Monetary Mechanism in Japan", in: Japanese Economic Studies Winter 1977/78, Vol.6 No.2, pp.3-41.
Gerlach, Michael, 1992, Alliance Capitalism; Berkeley: University of California Press.
Horiuchi Akiyoshi, 1984, Nihon no kiny_ seisaku - kiny_ mekanizumu no jissh_ bunseki (Monetary Policy in Japan - An Empirical Analysis); Tokyo: Toyo keizai shinp_-sha.
Iso Y_suke, 1995, "Nihon no kiny_ ga abunai", in: Aera, September 11, 1995, pp.16-17.
Ito Hirotoshi, 1995, T_ky_ Ky_d_ gink_ no kyok_ - _kurash_ no yami (Nihon musekinin jidai Part 5) (The fabrication of Tokyo Kyodo Bank - A MOF hush-up (Japan in the time of no responsibility, Part 5); Tokyo: Kadensha.
Kure Bunji, 1973, Kiny_ seisaku (Monetary Policy); Tokyo: Toyo keizai shinp_-sha.
MOF (Ministry of Finance)(ed.), various issues, _kurash_ Gink_-kyoku kiny_ nenp_, Tokyo.
MOF (Ministry of Finance, Institute of Fiscal and Monetary Policy) (Ed.), 1993, The Mechanism and Economic Effects of Asset Price Fluctuations -- A Report of the Research Committee, Tokyo: MOF.
Nanto, Dick K. / Jackson, William D. / Wells, F. Jean, 1995, "The Daiwa Bank Problem: Background and Policy Issues", CRS Report for Congress 95-1164 E, The Library of Congress, Washington D.C..
_kurash_ (MOF) (ed.), 1993, _kurash_ meikan 1993, Tokyo: _kurash_ insatsukyoku.
Patrick, Hugh, 1962, Monetary Policy and Central Banking in Contemporary Japan; Bombay.
Royama Shoichi, 1986, Nihon no kiny_ shisutemu (The Japanese Financial System); Tokyo: Toyo keizai shinpo-sha.
Schaede, Ulrike, 1989, Geldpolitik in Japan 1950-1985 (Monetary Policy in Japan 1950-1985); Marburg: Marburger Japan-Reihe Bd.1.
Schaede, Ulrike, 1990, "The Introduction of Commercial Paper: A Case Study in the Liberalization of the Japanese Financial Markets", in: Japan Forum Vol.2 No.2, October 1990, pp.215-234.
Schaede, Ulrike, 1994, "Finanzinnovation und Regulierung in Japan", in: von Stein, Heinrich (ed.), Banken in Japan heute; Frankfurt: Fritz Knapp, pp.331-366.
Schaede, Ulrike, 1995, "The 'Old Boy' Network and Government-Business Relationships in Japan", Journal of Japanese Studies Vol.21 No.2, Summer 1995, pp.293-317.
Suzuki Yoshio, 1980, Money and Banking in Contemporary Japan; New Haven:
Suzuki Yoshio (ed.), 1987, The Japanese Financial System; Oxford: Oxford UP.
Teranishi Juro, 1982, Nihon no keizai hatten to kiny_ (Japan's economic development and finance); Tokyo: Iwanami Shoten.
Wood, Christopher, 1992, The Bubble Economy - Japan's Extraordinary Speculative Boom of the '80s and the Dramatic Bust of the '90s; New York: The Atlantic Monthly Press.
Zenginkyo (Federation of Bankers Associations of Japan) (ed.), 1994, The Banking System in Japan;Tokyo: IBI Inc.
Appendix 1: The J_sen Debate
When the j_sen problem was first discussed in 1993, the mother banks of Nippon Housing Loan (i.e., major city banks) requested a dissolution of the troubled subsidiary. The agriculture-related cooperatives opposed a dissolution because of their high loan exposure. The MOF, hiding behind an argument of "maintaining financial stability" designed a 10 year plan for restructuring (Yomiuri 9/22/1995, Nikkei Kinyu 10/3/1995). However, when land prices continued to fall the situation turned out to be much worse than MOF had expected. In 1995 the issue was discussed abroad, and large Japanese banks had to pay the "Japan premium" in international financial markets. In September 1995, the agricultural cooperatives agreed to dissolve the j_sen, but only on the condition that they would not lose any of their principal of 5.5 trillion. The banks retorted that the "mother bank principle" was a leftover from the days of strict regulation. As part of the deregulation process in the early 1980s, MOF had issued an administrative guidance to all commercial banks that they should lower their portion of collateralized loans; i.e. increase loans without security. The banks argued that from the day they lowered their collateral standards, following the MOF guidance, the "lender responsibility" of the agricultural institutions, and any other lenders, began (SK 9/16/1995).
For the financial institutions, this controversy was one of life and death. If the "mother bank" approach had been adopted, the 11 city banks would have had to shoulder 1.8 trillion of bad loans, as opposed to only 770 billion under the "lender responsibility" system. In contrast, the agricultural banks would have had to carry 2.9 trillion under the "lender responsibility" as compared to zero with the mother bank approach (SK 9/16/1995, TK 9/16/1995). Even worse, among the three banks for long-term credit and seven trust banks, eight banks had twice as much in bad loans as in funds. The "mother bank" approach would have led to the failure of at least three banks within this group 31 (Nikkei Kinyu 9/4/1995).
The ministerial discussion behind this "mother banks" vs. "lender responsibility" controversy unfolded as follows. MOF early on opposed the "mother bank responsibility" approach. If only the banks were to carry the burden, at least some of them would have needed public funds to survive the disaster. The use of public funds would immediately have raised the question of MOF's regulatory responsibility. On the other hand, the "lender responsibility" would have required a bailout of the agricultural cooperatives through the agriculture-related deposit insurance system (N_rinkei kiny_ kikan no yokin hoken kik_), which is different from the banks' deposit insurance corporation. Banks would have quietly written off their loan losses over five years, and no public funds would be needed.
At this point, the famous "memo" surfaced. It contained a promise by MOF made in 1993 that the agricultural cooperatives would not have to pay for the j_sen mess. Commentators assume that this memo was leaked by MAFF in order to prevent MOF from bullying through its plan. From MAFF's perspective, if only the agricultural cooperatives had received public funds for the j_sen, MAFF's, rather than MOF's, responsibility would have been scrutinized by the public (TK 9/16/1995).
By October, 1995, it had become quite obvious that the bureaucrats' agenda in the j_sen dissolution was not so much to maintain order in the banking system as to save their own reputation. The deadlock was broken by Mitsubishi Bank that proposed a "revised mother bank approach". This was a compromise between the two extremes: the mother banks could write off all their bad loans and the agricultural cooperatives could write off some of theirs 32 . At this point, the politicians were called in. Under the new single-seat constituency system for the lower house, candidates need at least 51% of the vote (rather than 10-15% under the old system). Agricultural cooperatives have strong vote-gathering capabilities and can, therefore, easily activate politicians, especially those with rural constituencies. On the other hand, politicians with a relation to the financial interests, in particular former MOF bureaucrats, had to publicly support the MOF plan.
In typical fashion, two advisory councils were then established, each with the task to argue for a compromise but each with special interests; one for the coalition government and one for the Ministry of Finance 33 . While differing in detail, both councils pressed for a solution of the problem before the end of 1995, and suggested the establishment of a special "dissolution management institution" for the j_sen which would be financially supported by the mother banks and, possibly, public funds (Yomiuri 9/22/1995, Nikkei 9/23/1995, Nikkei Weekly 10/16/1995). This was an ideal solution, especially for the MOF bureaucrats, because if public funds were not used to directly bail out one, or a group of, financial institutions, but rather to provide capital for a dissolution institution, then this money could be easily disguised. In fact, it might not even have to come out of the general budget 34 .
One major problem for the "revised mother bank" agreement was "referred loans", and "loans with altered conditions" which total 570 billion. The "altered condition" could indicate that the loan was transferred from the mother bank to the j_sen, but no paperwork existed. The agricultural banks claimed they had tapes and copies of how the employees of some j_sen locked themselves into the company on a Sunday to make over some of the loan approval forms (TK 9/16/1995). If the "reversed" and "altered" loans are included in the mother bank responsibility, the burden of the agricultural cooperatives would decrease substantially.
By early 1996, a solution had not yet been reached.
Footnotes
Note 1: For example, John Reed, Citibank, on CNBC Asia News, September 1995; also Wood (1992) and Dattel (1994). Back.
Note 2: As this paper focuses in on the domestic crisis, the Daiwa Bank case is not included in this analysis. Daiwa Bank's New York branch had lost $1.1 billion, allegedly in bond trading, over a period of 11 years without informing the authorities. When the Japanese MOF was finally told about the problem, it chose not to report to the U.S. authorities for another month. At the time of writing, the case was still unravelling. For a detailed analysis, see Nanto et al., 1995. However, the arguments made in this paper on MOF regulation apply for the Daiwa case as well. Back.
Note 3: The following discussion builds on a vast body of literature, including BOJ (1995); Eguchi/Hamada (1978); Horiuchi (1984); Kure (1973); Patrick (1962); Royama (1986); Suzuki (1980); Teranishi (1982); and Schaede (1989). Back.
Note 4: This policy was possible over a long period of time because private household savings had a low price elasticity; i.e. household accumulated savings regardless of the level of interest rates. Reasons for this savings behavior include the extremely low level of social security benefits in this period and the absence of alternative investment opportunities. Back.
Note 5: The term "indirect financing" describes the route in which savings in an economy are channelled to investment. "Direct" means that savers own shares or hold corporate bonds. "Indirect" means that a financial institution (bank) enters the process as an intermediary who accepts and pools deposits and finances the corporations, either through loans or share- and bondholdings. The degree to which this process is direct is related to (a) the development and role of the capital market within the national economy, and (b) the access of corporations to international financial (loan) and capital (stocks and bonds) markets. Back.
Note 6: The Foreign Exchange and Foreign Trade Control Law (Gaikoku kawase oyobi gaikoku b_eki h_) of 1949, until its revision in 1980, did not allow for a free flow of capital. All major transactions had to be reported to the authorities, and raising funds abroad was generally not approved. Back.
Note 7: Based on the "balanced budget" policy line introduced by Joseph Dodge in 1949, the issue of government deficit financing bonds (akaji kokusai) requires the passing of an authorizing law by the Diet. Until the oil shock 1973, adherence to this policy line restricted the market for government bonds. Strict regulation on dealing in government bonds by banks also suppressed trading in government bonds.
The market for industrial bonds was regulated through (a) interest rate restrictions, and (b) collateral requirements, which together constituted de facto regulation on bond issues. All bond issues were subject to standardized guidelines on interest rates, which moved parallel with the discount rate. The low made industrial bonds an unattractive investment. Further, until 1975 all bond issues required physical collateral of 100%. This increased both the costs of a bond issue and disadvantaged some companies, such as trading companies, which did not hold physical assets with high collateral value. Back.
Note 8: Until the 1970s, it was customary to issue shares at face value, which was prescribed by law to be either 50 or 500. The face value was usually lower than the market value, but shareholders had preferred rights on capital increase issues at face value. In addition, dividends were customarily set at 10% of face value, regardless of the business results over the fiscal year, so that dividend payments became a burden during recessions. Back.
Note 9: See Patrick (1962) for a complete description of this process. Back.
Note 10: The Financial System Reform Act of 1993 removed the legal barriers to change; thereafter, moving into new business areas was allowed in a slow process. Back.
Note 11: CD were introduced in 1979; the government added financing bills (FBs) to Treasury Bills (TBs) in 1985, bankers' acceptances (BA), although never a large market, followed in 1985, and Commercial Paper (CP) was allowed in 1987. Back.
Note 12: The Japanese regulatory structure consists of four layers: the law; cabinet ordinances that accompany the law; ministerial written notifications that supplement the legal framework; and oral "invitations" to certain actions, addressed to an individual or a group of market participants. Notifications and invitations are referred to as "administrative guidance". Such guidance is not based on any law, but aims to realize an administrative goal through industry cooperation. The process is not transparent and often involves delicate conversations between ministry officials and corporations. Because rules can be implemented or revoked at the discretion of the ministry in charge, and without prior notice, this regulation is highly situational (ad hoc). The enforcement is based on a carrot-and-stick principle: following "advice" may reap rewards later. Back.
Note 13: More precisely, lending is permitted to non-members who have a deposit with the cooperative (which is used as collateral against the loan) and to local government bodies, as long as such lending does not exceed 20% of total lending. Back.
Note 14: The Mainichi newspaper reported about problems on Friday, July 28, 1995. On the following Monday, July 31, the cooperative experienced a run when 73 billion ($730 million, or 17% of total deposits) were withdrawn on one day. The shinkumi was closed down that evening. The large newspapers were under a MOF request not to report on problems of individual financial institutions. The City of Tokyo had originally proposed July 28 as "X-Day", but then changed its mind. Mainichi did not hear the latest news and printed the story on July 28. Cosmo claimed that without the newspaper's report, they would have survived. While this is highly unlikely, given Cosmo's loan situation, the authorities then issued a second request to the newspapers not to report on immediate problems before the closure, to avoid a run. When Kizu shinkumi and Hy_go Bank failed on August 30 (see below), newspapers and TV stations withheld the news until the banks were actually closed, and then issued an additional one-leaf special to the evening paper (Kinyu Bijinesu, 10/1995, pp.36-37). Back.
Note 15: The CP, commercial paper, market was opened in November 1987. In the first two years after its inception, this market was often referred to as a "Crazy Paper", because the interest rates to be paid by the issuer were lower than the interest rates paid by banks on large-scale deposits. On the competition in this new market segment, and the distortions in the interest rate structure it created, see Schaede 1990. Back.
Note 16: I adopt the term "bicycle loan" from a mechanism first observed in the Japanese credit card business: Because credit cards in Japan have low credit lines, consumers often have several cards and pay one off with the help of another. In a vicious cycle, consumers have to take on more and more credit cards to pay their debt; i.e. they have to keep going in order not to topple. Back.
Note 17: This explains the sudden huge increase in Kizu's accounts in 1995. In January 1992, Kizu reported 3.4 billion of uncollectible loans; in September 1992, this number had climbed to 23.8 billion; in July 1993, to 70.8 billion; and in October 1994, to 370 billion. In November 1995, 960 billion of nonrecoverable loans were revealed. Back.
Note 18: Major shareholders were Sumitomo Bank (4.99%; with 5% being the legal limit for shareholdings by financial institutions); LTCB (Long-Term Credit Bank of Japan, 4.2%); IBJ (Industrial Bank of Japan, 3.88%); Nippon Credit Bank (3.44%); Mitsubishi Bank (3.36%); and Sanwa Bank (2.31%). Back.
Note 19: Capital for Midori Bank was decided to be 80 million, but it was not to pay dividends for 10 years; after that period, the outstanding balance of bad loans will be settled by deposit insurance and the bank is expected to operate profitably. The deposit insurance system contributed 440 billion to pay off a major portion of the uncollectible loans (Nikkei November 23, 1995). The bank's starting capital was provided by the largest shareholders, the Kansai-based industry, and the bank's trade association.
In a close parallel to Sanwa Bank which denied all responsibility for Kizu Credit's problem, Sumitomo Bank, which was the largest shareholder of Hy_go Bank, declared that it had no management relationships, and although it had financed the non-banks with 60 billion ($600 million), it did not participate in their business (Toyo Keizai 9/16/1995, pp.74-77). Back.
Note 20: The 47 Shin y_ n_gy_ ky_do kumiai ren g_kai, or shinren, are the middle layer of the system. They are the prefecture-level federations for the agricultural cooperatives, and function as banks for their members, the local n_ky_. Back.
Note 21: According to one source, the ministries even encouraged these loans; cf. NW 12/25/1995. This relates to the famous memo of February 1993 between MAFF and MOF, in which MOF agrees to special treatment in terms of interest payment and preferential treatment for the agriculture cooperatives over other j_sen lenders. In 1995, MOF denied having made such concessions, but the cooperatives claim they built their loans on trust in MOF. Back.
Note 22: The strategy of offering above market rates to attract new savings to be able to pay interest on the existing accounts was also prevalent in the U.S. S&L crisis. A second strategy that the S&L bankers who faced bankruptcy (so-called "zombie banks") employed was to aggressively bet on derivatives (futures and options) because they had nothing to lose and all to win. Such speculative derivative trading has not been reported in detail for the Japanese "zombie" cooperatives and banks, although both bankers and regulators in Tokyo suspected in summer 1995 that there was speculative trading which remained unreported (interviews, Tokyo, summer 1995). Back.
Note 23: See Gerlach 1992 for a detailed study. Back.
Note 24: The large-scale finance rule" (_guchi y_sei kisei) is contained in Art.13 of the Banking Law. It prohibits a bank from giving a large-scale loan (in % of total loans) to one single customer. The definition of "large-scale" is contained in an accompanying ordinance. Back.
Note 25: For instance, in the case of Kizu Credit, Sanwa Bank in the peak year of 1990 shifted 317 billion to Kizu Credit, while LTCB transferred 85 billion, and Tokai Bank 37.4 billion. Sanwa Bank claimed that the shukk_ (temporary secondment of a Sanwa Bank employee to the cooperative) occurred at Kizu's request and did not result in an involvement in management decisions. However, since the establishment of relation with Kizu Credit in 1978, Sanwa Bank had sent a total of 14 managing directors to Kizu. Back.
Note 26: On December 7, 1995, Yamaguchi Toshio, Lower House LDP politician and former Minister of Labor, was arrested. He is believed to have arranged for Takahashi, the major manipulator of the two Tokyo cooperatives, to finance a special corporation related to the Ministry of Health and Welfare. In March 1995, Nakajima Yoshio and Taya Hiroaki, were reprimanded for dining and golfing with Takahashi. Nakajima had also been a very active real estate and stock market investor during the bubble period, and he was dismissed from the ministry in summer 1995. MOF adopted a new moral code which stipulates that "any relationships between MOF officials and members of the private sector should be avoided if they raise public criticism" (NW 5/29/1995). In September, the government requested that high-ranking civil servants practice self-discipline on stockholdings, especially in companies on which they have inside information (Nikkei 9/14/1995, p.1). Back.
Note 27: There are also the so-called jisageya ("price deflation players"). These are the "deflation equivalent" to the jiageya ("price increase players"), who during the bubble years browbeat small landowners into selling to developers. Jisageya work to prevent the sale of a property unless they are paid off. One mechanism is to approach a company about to go bankrupt and offer a loan or "buy" rental rights for cash. Once the jisageya have moved into the building, renter protection legislation makes it impossible to sell (NW 11/27/1995). Back.
Note 28: The official definition is "loans to bankrupt borrowers, loans with payments more than six months in arrears, and restructured loans with interest rates below the official discount rate". This excludes loans with altered conditions such as interest rates below market or "bicycle loans". Back.
Note 29: Note that all these troubles coincided with the Daiwa Bank problem in New York. From a purely domestic perspective, it is somewhat understandable the MOF did not worry about the New York problem enough to even report it. There were other, potentially more life-threatening issues to be dealt with at home. Back.
Note 30: This scheme was triggered by the political haggling surrounding the election of new city mayors in Tokyo and Osaka in April 1995. Tokyo's new mayor Aoshima Kunio had promised in his campaign that he would not support the failed banks. The City Council then agreed to support Cosmo Credit after all. The haggling further added to the bailout costs of Cosmo. Back.
Note 31: Toyo Keizai (9/16/1995, p.70) provided data to show that three banks would need bailout support under either approach: Nippon Credit Bank, Chuo Trust, and Nihon Trust. Nihon Trust was later acquired by Mitsubishi Bank, and an acquisition of Chuo Trust by Norin Chukin was temporarily discussed. Analysts assume that Nippon Credit Bank will be merged with one of the large banks, possibly IBJ, before long. Back.
Note 32: Mitsubishi Bank's motif was not pure altruism. The bank itself had only limited exposure as a mother bank of Housing Loan Service, but Bank of Tokyo, with which Mitsubishi Bank was planning to merge in 4/1996, had a much larger exposure in loans to the j_sen. Therefore, Mitsubishi Bank was better off under the mother bank responsibility, and therefore had an incentive to devise a scheme that would shift some of the burden to the mother bank. Back.
Note 33: These were the Kiny_ sh_ken projekuto chiimu (Finance and Securities Project Team) and the Kiny_ seido ch_sa kai Antei iin-ka (Stabilization Subcommittee of the Financial System Advisory Council). Both councils issued an intermediate report in mid-September, and a final report in December. Back.
Note 34: The so-called "second budget", FILP (Fiscal Investment and Loan Program, Zaisei t_y_shi) allows for preferential funding, and even subsidies, for public-private institutions. Although this was never said in public, this is the best way to disguise government's financial maneuvers. Sure enough, in November it was proposed that the management dissolution institution should be organized as a tokush_ h_jin, a special corporation (Nikkei 11/19/1995), which can be easily funded through FILP. This, however, was politically not feasible, because the ongoing administrative reform asks for the abolition of tokush_ h_jin. Back.
