Most international private and official institutions and professional observers were surprised by the abruptness of the 1997-98 Asian financial crisis and its subsequent contamination. There is also a broad consensus that neither multilateral agencies nor international investors anticipated the full scope of the crisis.
In this paper we use standard options pricing techniques and a country's macroeconomic value estimation method to estimate investors risk perceptions concerning five countries--Indonesia, Malaysia, the Philippines, Korea and Thailand--before, during and after the Asian crisis. We find that implied volatilities were generally high for all five countries over the whole period. By the end of 1996, just before the crisis, implied volatilities were at levels comparable to those at the height of the Mexican peso crisis at the end of 1994. However, the market seemed to discriminate shrewdly among the five countries. Indonesia had the highest implied volatility, which was also rising precipitously before the crisis, and was the hardest hit by the crisis. The Philippines, which was only marginally affected by the crisis, had the lowest implied volatility and it was falling before, during and after the crisis. Malaysia had the second lowest implied volatility and it was affected less than either Korea or Thailand, which had higher volatilities. We conclude that the market was sensitive to the countries' growing economic, financial and political problems and was able to judiciously discriminate among them. In anticipation of potential problems, the market had positioned itself at the short end of the maturity spectrum. In this context, the crisis looks more like a long anticipated rational portfolio rebalancing due to the evolving economic situation in each of the five countries than a mindless, indiscriminate panic. In fact, it was having anticipated the risk and prepared for it that made the reaction to Thailand's problem so rapid and so spectacular.
Our analysis differs from mainstream research conclusions. We look at three key variables, namely the country risk premium and maturity on international borrowings and the implied volatility of the country's capacity to generate the foreign exchange necessary to service that debt. Along with International Securities Market Association data, one can picture the observed risk premium as follows:
Based on the risk premiums, it looks like by the end of 1996, the markets had begun to anticipate financial problems for Indonesia and Malaysia. Indonesia experienced an increase of its risk premium from 436 to 633 basis points between 1995 and 1996 and Malaysia's rose by 58% over the same period to a higher level than 1994 in the wake of the Mexican peso crisis. The risk premium for the Philippines, Thailand and Korea, however, was falling. At the end of 1997, the risk premium was obviously strongly impacted by the events of mid-1997 although the Philippines seems to have been less exposed than its neighbors and better able to manage the crisis. In any case, in 1997 all five countries suffered a dramatic rise in their premium. Risk premiums, however, are very crude instruments for measuring risk. The first problem with rate spreads and secondary market discounts on bank debt as financial risk proxies is that they reflect the risk of the individual instruments rather than the overall risk of the issuer. Disparities in these measures exist even intra-country, depending on: fees; the timing of the loan due to the demand and supply conditions of bank credit; tax treatment of the bonds, and loan features such as the length and size of the loan. We have compensated for this problem by taking a weighted average of all the available instruments. The second problem is that they assume an exclusive, linear relationship with the benchmark. However, if the relationship between volatility and the spread is non-exclusive--that is it depends on other variables--or non-linear, it is possible for the spread to fall when the volatility is rising. To overcome this problem, we measure volatility directly in the following section.
We use the Clark (1991) model of the value of a country's net export capacity to estimate the implied volatility of the country's capacity to generate the foreign exchange necessary to service its foreign debt. This model measures the dollar value of a country's capacity to generate net exports. This quantity is analogous to a company's market value in corporate finance and has been shown to be useful in determining country creditworthiness and in forecasting sovereign debt defaults and reschedules. These values are then applied in the Black-Scholes option pricing formula to compute the implied volatility of the five economies for each of the five years. The resulting instantaneous measure of risk represents the market's perception of the risk of each economy at each point in time.
From these results, we can observe that the market considered these countries as extremely risky as far back as 1993 with implied volatility ranging from 41% to 78% between 1993 and 1996. Comparable figures for Mexico were 28.2% at the end of 1993, the year before the peso crisis, and 36.2% at the end of the first quarter of 1995, the height of the peso crisis. Furthermore, in 1996 implied volatility for all the countries but Korea stayed at or above the levels reached during 1995, the year of the Mexican peso crisis. For Korea, it had fallen only slightly and was still higher than that of Malaysia, the Philippines and Thailand. Over the whole period from 1993 to 1996, implied volatility rose for three countries (Indonesia, Malaysia and the Philippines) and fell for two (Korea and Thailand). These two countries that experienced a fall in their implied volatility were the two that had the highest implied volatilities in 1993. The reduction only brought them to the same level as the other three countries. All this suggests that as early as 1993 the market was anticipating the potential difficulties that would eventually materialize in 1997. It is interesting to note that Indonesia had the highest implied volatility on the eve of the crisis and was the country that suffered most when it hit. On the other hand, the Philippines had the lowest implied volatility and was the least affected. Malaysia had the second lowest and was the second least affected.
In light of the actual 1997 events, the market's perception of the Philippines as the lowest risk and Malaysia as the second lowest was a shrewd assessment. Indeed, even though the Philippines had not been able to achieve its neighbors' economic performance over the last decade, it did not suffer their specific weaknesses to the same extent. One of the main causes of the Asian problems arose from its financial sector's vulnerability. Another was the heavy reliance on short-term debt. In graph 1 we see that the Philippines had the lowest amount of short-term debt and in graph 2 we see that it represented a relatively small percentage of overall debt compared to the other countries. The low implied volatility indicates that the markets spotted this and acted on it.
Conclusion
Based on this evidence and contrary to most previous statements on the topic, we conclude that capital markets were sensitive to the five countries' mounting economic and political problems. Investors and creditors were able to discriminate between each country's specific features as well. Furthermore, the markets anticipated the upcoming crisis by positioning itself at the short end of the maturity spectrum. Thus, when the crisis did come, investors were able to get out fast. In this sense, it was having anticipated the risk and prepared for it that made the reaction to Thailand's problem so rapid. Thus, we conclude that while the rating agencies, international institutions and other professional observers failed to adequately assess the region's economic vulnerabilities, the markets themselves were able to accurately assess the situation for the region in general and for individual countries in particular. The spectacular scale of the crisis stems from the fact that the institutions and professional observers unwittingly permitted the situation to go on so long and degenerate so far.
This week's Learning Curve was written by Michel-Henry Bouchet, head of CERAM's global finance chair and principal adviser at ING Barings, Bertrand Groslambert is finance professor at CERAM and Ephraim Clark is finance professor at Middlesex University.
CERAM is a graduate school of management and technology in Sophia Antipolis, France.
