Asian Financial crises

Chapter 6 Origins of the Asian Crisis: Discussion

International Monetary Fund
Published Date:
January 2001
  • ShareShare
Show Summary Details

The fundamental views on the causes of the recent Asian crisis have clearly emerged: a “bad policy” view and a “financial panic” view. I discuss these perspectives, their implications for public policy, and outline challenges for the bad policy camp.

In order to obtain a proper perspective on the papers in this session, it is useful to review the dominant views on emerging markets crises before and after the Mexican crisis of 1994. In so doing, my discussion will overlap somewhat with Reuven Glick’s. However, my emphasis will be different and will inevitably reflect my work with Andrés Velasco on the subject.


Before the 1994 Mexican crisis, a dominant view had arisen from the seminal paper by Paul Krugman (1979). Its central tenet was that a fixed exchange rate system eventually had to be abandoned if a government had limited reserves and ran a persistent fiscal deficit. Hence, this “first generation” view saw a balance of payments crisis as the predictable outcome of inconsistent macroeconomic policy.

While useful in other episodes, the “first generation” view is clearly inadequate to explain recent events. Both Mexico before 1994 and the Asian countries that went into crisis last year had been posting small fiscal deficits or even surpluses.

A “second generation” view was proposed by Maurice Obstfeld (1994). Obstfeld argued that governments may not have to abandon fixed exchange rates; however, they may choose to do so if the social cost of defending fixed rates, particularly in terms of unemployment, becomes too large. An interesting implication of Obstfeld’s hypothesis is that crises may be driven by market expectations. Expectations of devaluation may increase the interest rate and unemployment levels needed to defend an exchange rate. The government may decide to devalue as a consequence, confirming expectations.

Obstfeld’s work stimulated a new flurry of research on crises. Unfortunately for this view, neither the Mexico before its 1994 crisis nor Asian countries before their recent crisis exhibited a particularly weak macroeconomic picture. Hence the “second generation” view also seems inapplicable to recent crises.


Given this state of affairs, economists have been led to develop new approaches to financial crises that can explain what happened in Mexico and, more recently, in Asia. While hundreds of papers on the subject have been written recently, two camps have clearly emerged: a bad policy camp and a. financial panic camp.

Bad policy advocates follow the “first generation” view in asserting that, ultimately, crises are the inevitable outcome of misguided government policy. However, they need to solve the puzzle of why policy imbalances were not apparent from conventional measures of monetary or fiscal policy. Their answer has been to argue that Asian policies were “really” bad but, at the same time, were obscure enough so that their effects did not show up in conventional measures. While not manifest, the damage on the economy accumulated over time, which eventually led to a crisis and a policy reversal.

What kind of policies fit such a description? Implicit liability insurance is the leading candidate. It has been argued that Asian governments guaranteed domestic private liabilities, in spite of the fact that such guarantees encouraged domestic borrowers to take socially costly actions (such as investing in excessively risky projects or simply stealing funds). This strategy, it is said, “succeeded” for a while, because the government had enough funds to keep it going. The “insurance fund” had a limit, however, which was eventually reached. At that point, private agents understood that further borrowing would not be guaranteed and, in an “attack,” creditors exchanged existing private liabilities for the government insurance fund.

The story just described was developed first by Michael Dooley and is detailed in his contribution to this conference, which is an outgrowth of Dooley (1997). Other important papers in the bad policy camp are Mc Kinnon and Pill (1996), and Krugman (1998) although, in my opinion, Dooley (1997) is still the most compelling statement of that view. The contribution by Craig Burnside, Martin Eichenbaum, and Sergio Rebelo also belongs to this camp, although its primary focus is to explain when an exchange rate peg will be abandoned after a financial crisis, not to explain the crisis itself.

The opposite position is that the Asian crisis was the result of a financial panic. Those in this camp argue that economic fundamentals in Asia, including government policies, may not have been entirely satisfactory yet did not warrant a crisis. Instead, the cause was that international creditors, fearing a crisis, suddenly refused to roll over credit to Asian countries. As a consequence, these countries had to scramble for short term funds, which resulted in costly liquidations, asset price collapses, domestic bank runs, and credit crunches.

A key policy implication is that the Asian crisis did not need to happen: if foreign lenders had not panicked, Asian financial systems would not have had to endure the 1997 credit shock, and the costly disruption of the system would have been avoided, justifying the optimistic expectations. Hence, for financial panic advocates, market expectations were key to the Asian crisis.

While the emphasis on the role of market expectations resembles Obstfeld’s (1994) views, the financial panic view postulates very different mechanisms. In particular, Chang and Velasco (1998a, b) have shown that the key condition for a small country to be prone to financial panics is international illiquidity. A country’s financial system is internationally illiquid if its potential short-term obligations in hard currency exceed the hard currency liquidation value of its assets. If holders of the short-term liabilities of the financial system lose confidence and attempt to redeem their holdings, the system will become bankrupt, making the confidence loss self fulfilling, if and only if it is internationally illiquid.

In other words, financial panic advocates argue that the root of the Asian crisis was a maturity mismatch: short-term international liabilities were far greater than short-term assets. Evidence in favor of this view has been provided in Chang and Velasco (1998c) and Radelet and Sachs (1998). To illustrate, consider Table 1, taken from Chang and Velasco (1998c). In the table, the short-term liabilities of Korea, Indonesia, Malaysia, the Philippines, and Thailand are proxied by their short-term borrowing from BIS reporting banks; their short-term assets are proxied by the level of their international reserves. The bottom panel of the table shows that the ratio of short-term debt to reserves far exceeded unity in June 1997 in Korea, Indonesia, and Thailand. This means that if, as it was the case, international bankers had refused to roll over credit, these countries would have not had enough reserves to meet their immediate obligations. Also, while the debt-reserves ratio was below one in Malaysia and the Philippines, it had more than doubled since 1994.

Table 1.Asian-5: Short-Term Debt vs Reserves

Debt (millions of









(millions of US$)






Debt to Reserves







Source: BIS, IMF and Chang and Velasco (1998c).
Source: BIS, IMF and Chang and Velasco (1998c).

A crucial question for the financial panic camp is then: how did Asian countries become internationally illiquid? Financial liberalization appears as the main culprit. As documented by Chang and Velasco (1998c), the Asian countries affected by the recent crisis had enacted sweeping reforms designed to make their financial systems more market oriented. These reforms are beneficial in principle, but may also have exacerbated international illiquidity; this is indeed one consequence of the theory put forward in Chang and Velasco (1998b).

The distinction between the bad policy view and the financial panic view is not only an academic matter. It is in fact a crucial issue because the two perspectives have very different implications for public policy. To be concrete, consider the issue of whether there should be an international lender of last resort. If we believe the bad policy camp, the answer is clearly negative. In fact, if we take Dooley’s analysis literally, such a facility only makes problems worse, since access to international credit would only induce governments to keep lending for ill-fated purposes. On the other hand, if one believes that crises are caused by financial panics, one must conclude that the existence of an international lender of last resort would be of great help. Depositors and international creditors would be reassured that their short-term claims would be ultimately honored, and hence they would have no reason to panic.

Needless to say, I believe that the financial panic view is more promising than the bad policy view. The former is consistent with a wide array of evidence, and its theory has been developed in a much greater level of detail; I refer you to my papers with Andrés Velasco for demonstration of these assertions. But my role here is to discuss the papers in this session, and hence I will conclude with some further comments on them and on the bad policy view.


What obstacles must advocates of the bad policy view overcome if they are to gain the upper hand in the current debate? I think that the main challenge is to make that view more believable both in the theoretical sense and in the empirical sense. Theories developed by the bad policy camp often assume policies that are too contrived to be believable. The model in the paper by Burnside, Eichenbaum, and Rebelo, for example, does not “work” unless very strong and arguably unreasonable assumptions are imposed on the availability and timing of taxes, and the actions of monetary authorities.

The policies postulated in Dooley’s contributions are more intuitively plausible. However, it is unclear why a government would stick to those policies. More generally, the bad policy camp needs to find some explanation of why the bad policies are implemented. Is it because of politics? Concerns about redistribution? Time inconsistency? Answers are needed to lend credibility to the bad policy position. In addition, without them we can say little or nothing about the desirability or possibility of reforms.

Finally, advocates of the bad policy view should bite the bullet and present supporting empirical evidence. In spite of the attention this view has received, I have yet to see one compelling piece of data supporting it. It is not enough to show that countries were corrupted and that their financial system were distorted. A lot of emerging economies suffer from comparable levels of corruption and financial distortions, but were not attacked recently as one would expect if the bad policy evidence were valid. Also, while corruption and financial distortions in Asia have presumably been present for decades, the Asian crisis countries had performed remarkably well up to last year. If corruption and financial distortions are the cause, why then did it take so long for these countries to go into crises?

Ultimately, the bad policy view can succeed only if can present valid evidence that matches the predictions of its theory. I believe this will be a formidable task, partly because bad policy theories are constructed so that policy effects are not manifest in standard measures of monetary or fiscal imbalances. But I want to encourage “bad policy” believers to try. This effort can only contribute to a richer debate, the outcome of which has key implications for world welfare.


    Other Resources Citing This Publication