Chapter 29 Banks and the Asian Crisis
- International Monetary Fund
- Published Date:
- January 2001
For me, the papers in this session skirt around the central lesson of the Asian crisis: that banks still matter, enormously. In the United States, the capital markets dominate the credit markets, and only twenty-two percent of the indebtedness of American enterprise is to banks. In Asia, eighty percent or more of the indebtedness of enterprise is to the banks. The United States suffered a recession when banks were afraid to lend in 1990–91; for Asia, the disabling of the banks means the dismantling of the economy.
The IMF Capital Markets report (Adams, Mathieson, Schinasi and Chadha, 1998, p. 79) argues that one of the weaknesses of the Asian economies was “excessive reliance on banks as the primary source of financial intermediation.” But it is difficult for a developing country to create financial markets that replace its banks as a primary source of financing, especially in an era of regulatory liberalization, when banks are both traders and the objects of trading. Neither Japan nor Germany has yet managed to make this work. In the United States, though the stocks of a few large financial corporations that owned banks were listed in the 1920s, bank stocks per se were not traded to a significant degree until the 1960s; until the 1930s, indeed, bank shares carried a liability for the remaining fifty percent of paid-in capital if anything awful happened to the bank.
It is all very well to talk about transparency, but the fact is that banks even in industrialized countries are blind pools of assets, and the certainty that those assets are sufficient to cover the liabilities is certified by the regulatory authorities under cover of bank secrecy. The world has changed in America because suppliers of funds to banks now feel, correctly or otherwise, that they have enough information to make their own judgment on the condition of the asset portfolio. So the need for the supervisors to certify that the bank can meet its obligations becomes an invitation, usually accepted, to moral hazard. In their hearts—and indeed in their heads, if not the mouthpart—banking supervisors agree with Lowell Bryan that “market discipline by depositors is another name for bank panics.” (Bryan, 1991, p. 225) On Monday we advocate transparency; on Tuesday we speak of bilateral netting of derivatives and we fight against mark-to-market. When I hear a bank supervisor talk about transparency, I fear, I think of John Randolph’s comment about a Senate colleague that the word honor in the mouth of Daniel Webster was like the word love in the mouth of a whore.
Moreover, if you can look through the mirror instead of into it, the question is not one of too-big-to-fail. LTCM is classic too-big-to-fail, moral hazard American style (the fear being not of a market collapse from liquidation, but of exposure of sloppy practice both by the banks and by their supervisor, which would really spook the markets). What happened in the South Korea, by contrast, is an expression of unavoidable if implicit claims on the taxpayer by creditors of failed banks chartered, regulated and supervised by the government that represents that taxpayer. Few economies can withstand the shock of a disruption to the payments system, and the money supply of the country is, after all, created mostly by the banks. To quote Andrew Sheng, “[A] central problem in the global banking system is that, irrespective of public or private ownership of banks, commercial bank losses in excess of capital have become de facto quasi-fiscal deficits.” (Sheng, 1996, p. 9) Then, if the bubble bursts, the capacity of the government and the central bank to direct lending by their banks is largely lost, because taxpayers recognize that additional loans are on their marker, and they don’t like it. Viz, not only Japan, but our own S&L crisis. I think Mr. Fons is right that a perception of transparency (mayhap, we should say “honesty”) leads to cheaper funding—but only when the truth is more encouraging than previously believed.
The genie of market pricing cannot be put back in the bottle, and it probably must be given more wishes to execute. And markets do demand transparency; indeed, it is the clash of tectonic plates between the information systems of the markets and the information systems of banking that makes the earthquake. In a world where people don’t believe central banks, and have reason not to believe central banks, commercial banks must be much more heavily capitalized: if the market demands twelve percent to fifteen percent from GE Capital in America, banks operating in less developed countries, multinational or local, must show at least that much. The cost to the borrower of reduced leverage in the lender should be less than the savings from the bank’s ability to fund itself at lower risk premia. (Though it is important right now to see the cognitive dissonance of appeals for the private sector to return to lending to developing and transition economies, while accepting in advance that haircuts to its positions and “precommitments” will be taken even in well-structured loans if the country involved gets in trouble.)
Moreover, as the reaction to the failure of Long Term Capital Management has shown, neither LDC banks nor multinational Western banks can afford to carry contingent liabilities of the size implied by “total return” swaps, long-dated currency options and other leveraged bespoke derivative contracts. The fact that the banking industry has sold its customers, its regulators and its academicians on the proposition that custom-made derivatives are risk-reducing does not mean that the proposition is true. Tony Terraciano asked me a few years ago why I thought banks stepped on a rake every few years, and I had to say I didn’t know the answer, and Alan Greenspan has now said that he doesn’t know the answer either. But turning over risk management to people who step on rakes is not good governance.
Because I think the banking system is the key, I agree that disaster forecasting is most likely to be accurate if it emphasizes too-rapid growth of bank assets, as in Calomiris and to a lesser degree in Kaminsky. Assuming that the banks are made to recognize bad loans, growth in scheduled assets, to use an old-fashioned phrase, should be an excellent predictor of trouble—currency trouble, too, because the issuance of bank liabilities to fund non-performing assets is the functional equivalent of printing money. That, it seems to me, is the real reason to demand good supervision: the good supervisor prevents evergreening and forces banks to admit that loans have gone sour. What reduces risk premia below the danger level is the willingness to roll over and even increase the loans of companies that are not earning their debt service—what Hy Minsky called Ponzi finance. This gives a record of greatly reduced loan losses, which was one of the things the Fed kept citing in its goldilocks economy evaluations, and permits lending officers to shave terms and conditions.
Understanding the disaster in Asia begins, it seems to me, with the realization that there were real losses to be allocated, and no system for allocating them. Everybody plays sauve que peut, and the banks have more peut, especially the foreign banks, so the market takes the worst hit. Because there is no system for protecting bondholders (let alone stockholders, who are of course residual claimants), and because derivative instruments link debt paper in different countries, both psychology and mathematics become vectors of contagion. The newly issued UNCTAD Trade and Development Report for 1998 contains a very intelligent summary of what happened in the Asian countries after confidence fled, but it is flawed by the hopeful assumption that if only everybody could have held on a little longer, the values represented by the paper would have risen enough to make all the lenders whole. This was not true; indeed, the banking systems in all three countries where the IMF intervened were so seriously undercapitalized that compound interest was working against them.
I have suggested that one prophylactic against the return of these conditions is the separation of the lending function from the payments system function, which will permit central banks to get out of the business of guaranteeing the par value of deposit liabilities. (Mayer, 1998, p. 34) But we need time for that. Time will be needed also for the creation of a viable standstill process. In the interim, the two most useful steps would be increasing the capital requirements for banks engaged in international lending or borrowing and a change in the law to treat cash-settled OTC derivative contracts as the gaming instruments they usually are, making them unenforceable in American courts.