Currency and Financial Crises: The Case of Thailand

A Report on the First Country Meeting of the NBER Project on
Exchange Rate Crisis in Emerging Market Economies

John McHale
May, 1999


1. Introduction

On April 26, 1999, a group of Thailand experts, IMF officials, and academic researchers with an interest in currency crises, met at the NBER offices in Cambridge for a free ranging discussion on Thailand's recent crisis. This meeting, organized by Martin Feldstein and Takatoshi Ito, was the third a series organized by the NBER on the financial crises that have swept the world in recent years. The Thailand meeting was the first in the series to focus on a specific country--with Thailand being an appropriate choice given that the recent turbulence started with the Thai devaluation in July 1997. The first and second meetings focused capital controls and theoretical models of currency and financial crises, respectively. Reports on these meetings are available on the NBER web site at www.nber.org/crisis/capital_report.html and www.nber.org/reporter/winter99/eichengreen.html. A fourth meeting held on May 6, 1999, focused on the Mexican crisis of 1994/95 and further country specific meetings are planned for the fall of 1999.

The day's proceedings were anchored around morning presentations by Takatoshi Ito of Hitotsubashi University and the NBER on the events leading up to the crisis, and by Veerathai Santiprobhob of the Thai Ministry of Finance on the dimensions and management of the crisis. This brief report summarizes the two presentations and then provides a condensed account of the discussion. The differing perspectives of those who view the recent crises as inevitable punishments for the sins of crony capitalism, and those who view them as avoidable crises of confidence, led to some lively exchanges. So too did the discussions on how the crisis should have been managed, including issues such as the appropriateness of contractionary policies to stem the depreciation of the exchange rate, and the degree of restructuring forced on the banking and corporate sectors. Given the key--if sometimes criticized--role that the IMF played in managing the crisis, the value of the discussions was considerably enhanced by the perspective of the fund officials that attended.

Since the discussion was not tightly constrained, particular came up at various points through the day. To add some continuity to the account of the day's proceedings, I try, where appropriate, to group related interventions together, rather than giving a strictly chronological account of the proceedings. To encourage a free flowing conversion, participants were told that names would not be used in the report on the discussion.

2. Before the Crisis

Although late to the Asian Miracle, the rapid growth of the Thai economy in the latter half of 1980s and early 1990s was the showcase of the Asian region. Having experienced double-digit growth rates around the turn of the decade, Thailand had steady growth of between seven to nine percent through the early to mid nineties, underpinned by rapidly expanding exports, and a gradual evolution toward more sophisticated and higher value added products. (1) With imports rising quickly as well--fueled, in part, by an investment boom--the current account deficit rose to 8 percent of GDP by mid decade. The large net capital inflow of 9 to 10 percent of GDP meant that this deficit could be financed with international reserves actually rising, while holding to its de facto dollar peg.

In 1996 export growth fell dramatically, forced down by a factors such as the depreciation of the Japanese yen, the earlier (1994) depreciation of the Chinese yuan, and a downturn in the electronics business cycle. By the beginning of 1997 export growth was essentially zero. And with GDP growth dropping below 7 percent, Thailand entered into a "growth recession." Even prior to the growth slowdown, however, the Thai financial sector was experiencing difficulties. In the early part of the decade a financial bubble had developed and then burst with a major fall in the stock market in 1994. Rapid credit expansion, driven in large measure by imperfectly sterilized capital inflows, pushed asset prices--including real estate prices--back up. Much of the evolving financial trouble revolved around the Thai "finance companies." These financial institutions specialized in obtaining foreign currency deposits, and then lending the money domestically (often for real-estate projects). As the economy slowed down in 1997 the financial bubble burst. This, in turn, caused a sizeable increase in volume of non-performing loans in the balance sheets of banks and especially the finance companies. With the finance companies' depositors starting to flee, the government provided liquidity to troubled finance companies in the first half of 1997. These funds were channeled through the Financial Institutions Development Fund (FIDF), an entity of the Bank of Thailand.

Speculators attacked the Thai bhat on May 13, and the days following. A common strategy was to buy the bhat in the spot foreign exchange market while selling the bhat in the forward market. The central bank, as the counter-party to this swap transaction, acquired foreign currency assets, but also acquired an off balance sheet foreign currency liability due to the forward contract. On June 26 the Bank of Thailand announced that had reserves of $33.3 billion, reflecting a relatively small fall of $4 billion during the month of May. The central bank did not, however, reveal the size of its forward contracts, which amounted to US$23.4 billion. The next day, the central bank suspended the operations of 16 of the finance companies, announcing that depositors were to be protected by the exchange of bonds for deposits. (This guarantee turned out to be less than complete, as a less than market interest rate was paid on the bonds.) Overall, this action was viewed positively by the markets, with the stock market rising by about one and a half percent on the day of the announcement.

This defense of the currency worked during June, but continuing speculative pressures and the deteriorating net reserve position (gross reserves less forward liabilities) led the government to abandon the 13 year old dollar peg on July 2. The exchange rate immediately depreciated by 17 percent, but then remained relatively stable (unlike Mexico two and a half years earlier).

Later in July, the Thai government approached Japan and the United States for liquidity support, but was told to consult with the IMF. The weak financial condition of the finance companies was quickly identified as the key policy issue. It was believed that investor confidence--and thus currency stabilization--would depend on an effective response to the non-performing loan problem. In early August, the government announced that 42 of these companies would be closed in addition to the 16 companies already suspended.

On August 11, a $17.2 billion IMF-coordinated package was announced. Direct support from the IMF was limited by the usual 5-times quota rule, but this was evidently insufficient given Thailand's $23 billion forward liability. The package included support from a number of Asian countries in addition to the IMF, World Bank and Asian Development Bank. The United States did not pledge funds. The IMF board approved a three-year stand-by agreement of just under $4 billion on August 20, allowing for an immediate disbursement of $1.6 billion. As usual, IMF conditionality applied to these funds. The main requirements were: identifying and effectively closing the 58 financial institutions; (2) targeting fiscal surpluses of 1 percent of GDP; maintaining the exchange rate system as a managed float (with interventions limited to smoothing fluctuations); and a broad money growth target of 7 percent for 1997. The program also targeted a growth rate of 2.5 percent for 1997 and 3.5 percent in 1998, and an inflation rate of 4 to 5 percent.

2. The crisis and its Aftermath

2.1 The Magnitude of the Crisis

Over the six months following the floating of the currency, the Thai bhat lost around 55 percent of its value in the midst of a deepening regional crisis. The currency also fluctuated wildly, with daily swings sometimes in the order of about 7 to 8 percent. Given openness of the economy these swings were highly disruptive to economic activity, making exchange rate stabilization a policy priority. With private sector external debts of US$69 billion--about US$30 billion of which was short term--the financial and corporate sectors were placed in severe financial difficulty by the depreciation. Many international loans were called back, as foreign creditors became worried about foreign exchange availability and the possibility that the bhat would continue to fall.

With the deteriorating economic conditions and high interest rates, the extent of non-performing loans grew rapidly in the Thai financial system. 56 of the 58 institutions signaled out for intervention by the Bank of Thailand were closed in 1997. Since then 12 more finance companies and 6 commercial banks have been "intervened," with one of the banks subsequently closed. Commercial banks recorded large losses as they made provisions against the increasing value of non-performing loans. The large loss of capital caused by the deterioration of their assets implied a major loss of lending capacity given the need to meet capital adequacy requirements. To put it another way, a major recapitalization would be necessary to maintain outstanding credit given the prudential standards.

Overall, the combined currency and financial crisis, together with deteriorating conditions in the region, led to a decline in GDP of 0.4 percent in 1997. This was about 3 percentage points worse than the IMF target set in the middle of the year. In 1998, GDP fell by 8 percent--almost 12 percentage points worse than the IMF target. The Thai economy was now experiencing a full-blown crisis in both its financial and real sectors.

2.2 The Policy Response

Macroeconomic Stabilization

With the speed at which the bhat sank, the stabilization of the currency was considered the most urgent policy priority. Interest rates were raised to stem the capital flight. This added to the strain on the financial institutions and corporate sector, and ran the risk of leading to a further loss of confidence and further withdrawals of capital. It was believed, however, that injecting liquidity into the system would lead to larger capital outflows and reserve losses.

Santiprobhob offered five reasons why exchange rate stabilization is essential for economic recovery:

1. Exchange rate instability is highly disruptive to trade and investment activity, given the openness of the economy and the absence of exchange rate hedging instruments.

2. Further depreciation adversely effects the willingness of foreign lenders to roll over loans. The roll over rate dropped significantly in January 1998 when the bhat reached its lowest point.

3. The rate of depreciation also affects domestic confidence in the local currency, and thus affects the degree of capital flight.

4. The rate of depreciation affects domestic inflation. One consequence is the lowering of living standards, as the prices of imported goods rise. This exacerbates social tensions and affects the recovery of domestic consumption and investment.

5. Bhat depreciation exacerbates contagious effects within the region, leading to a new round of financial instability.

With priority given to achieving exchange rate stability, the Thai government pursued a policy package to improve confidence in the currency. This package included: restrictive monetary and fiscal policy; securing funds from multilateral and bilateral agencies to build reserves; measures to stabilize the financial system; and efforts to expand exports by seeking dedicated credit lines and seeking new markets.

This strategy appears to have been successful from January 1998 onwards, with the exception of the export promotion, hindered as it was by the deep regional slump. Over the first quarter of the year, the bhat appreciated from 56 bhat per dollar to around 41 bhat per dollar--and has since remained relatively stable. Net international reserves also increased as around US$21 billion of forward contracts were unwound. This turnaround in the net reserve position was driven by a large current account surplus--12 percent in 1998 compared to a deficit of 8 percent in 1996--and a slowdown of private capital outflows. Despite the 40 percent depreciation of the bhat, inflation has stayed low as the economy contracted. Inflation peaked at 10.7 percent in June 1998, averaging about 8 percent for the year. The expected average inflation rate for 1999 is just 2.5 percent. The low inflation rate is one reason that interest rates have come down. All major banks have now lowered their deposit rates to below pre-crisis levels.

Economic Stimulation

The crisis management strategy was dictated by the belief that exchange rate stabilization should be the primary objective. It was also believed that in an economy as open as the Thai economy, stabilization and the stimulation of real activity were not compatible in the midst of a crisis of confidence. This meant that a liquidity squeeze and a contraction of real activity were unavoidable until the exchange rate stabilized. As the exchange rate stabilized in early 1998, a more expansionary policy stance began to be pursued. The move to economic stimulation was also aided by the good inflation performance and rapid turnaround in the current account, combined with the (not unrelated) greater than expected drop in real GDP growth.

In February 1998 the fiscal target for Fiscal Year 1997/98 was adjusted from a surplus of 1 percent of GDP to a deficit of 3 percent of GDP--a 4 percentage point swing in just six months. Greater stress was placed on fiscal expansion than liquidity expansion because it was believed that fiscal measures could be targeted to assist those most adversely affected by the crisis. Moreover, the weak state of the financial system reduced its capability to perform its normal intermediation function. On the composition of the fiscal measures, the government initially favored expenditure over tax measures, believing that it could better target the former. The slow pace of disbursement, however, has led to an increased emphasis on tax measures in its more recent fiscal stimulus efforts. Despite the emphasis on fiscal stimulus, interest rates came down in line with inflation, form a peak of 25 percent in January 1998 to 10 percent in June 1998, and to just 1 percent in April 1999. Importantly, the move to a more expansionary stance did not cause renewed currency instability or increased inflation.

Restructuring the Financial Sector

The delay in restructuring the financial sector in the early stages of the crisis was a major contributor to the loss of confidence among depositors; notwithstanding the blanket guarantee for depositors and the shutting down of the 56 finance companies. The support from the FIDF to ailing institutions was costly, and required large borrowing in the short-term repurchase market to finance its operations. This pushed up short-term interest rates, discouraging sound banks from lending to the real sector. Moreover, as the balance sheets of financial institutions deteriorated, they became less willing to lend funds, causing a severe credit crunch in the economy. As the new government came to office in November 1997 it placed priority on limiting the operations of the FIDF, and minimizing the distortions created by its operations.

A number of steps were taken by the new government were taken to address the crisis in the financial system:

Despite these efforts, the intensifying crisis in the region led the government to introduce further measures in a plan announced in August 1998.

Given that the terms offered by the government for capital support, the take up of this support by the banks and finance companies has been low. However, the government believes that by "providing a wall that financial institutions can lean against," this has increased the confidence of domestic and foreign investors, leading to a better chance of private sector driven recapitalizations.

Santipraphob closed by drawing together what he saw as the lessons of Thailand's crisis. First, the proper sequencing of policies is crucial. Economic stimulus can only proceed after the exchange rate has been stabilized. Second, financial sector problems must be addressed early on. Delay raises the fiscal cost of the ultimate government bailout, and sound institutions are hampered by the continued presence of ailing institutions and the bail out operations. Third, fiscal expansion is more effective than monetary expansion amid a crisis in which financial intermediaries do not play their normal roles. Fourth, domestic funds are unlikely to be sufficient for restructuring the real and financial sectors. To compete for the limited pool of funds available for emerging markets, the government must commit to the free movement of capital, transparency, and market friendly policies. Lastly, given the possibility that external conditions might deteriorate further, the best way to insulate the economy is by putting in place "good internal systems" and stimulating domestic demand (primarily through fiscal expansion).

3. General Discussion

3.1 The Ito Questions: Considering Some Counterfactuals

Could the Thai crisis have been avoided if different policy decisions had been taken during the key period from 1996 to early 1997? Takatoshi Ito asked a number of questions to stimulate discussion about potential turning points. As reference to issues raised by Ito arose at various points during the day, I try to group the relevant discussion together.

1. With hindsight, what advice should have been given about the appropriate timing of the devaluation?

2. Given the problems in the banking system, what policies could have been recommended for dealing with the financial sector weakness? And when should these recommendations have been given? He also asked, more generally, what was wrong with the central bank's supervision of the banking system.

3. Would a bigger reserve-building package from the international community have impressed the markets and avoided the crisis?

Exchange Rate Policy

Participants discussed the sufficiency of the July 1997 devaluation. Attention focused on the extent of the overvaluation. Pointing to another source of fundamental weakness, one participant argued that a severe overvaluation could be traced back to the early 1980s. If 1980 is taken as a year in which the real exchange rate was properly valued, the cumulative overvaluation was greater by 1997 than Ito was allowing. As a counterpoint, it was noted that the current account deficit was significantly less in the early part of the 1990s, so that the overvaluation did become more acute as the crisis neared. For this reason, it was not unreasonable to believe that the initial 17 percent devaluation could have restored a reasonable external balance.

A participant pointed that the recent crises have tended to occur in countries with fixed-but-adjustable exchange rate regimes. This, he argued, should have implications for international support packages. To receive exceptional international support (large multiples of their IMF quota), he suggested, countries should be required to move to either floating rates or a currency board type arrangement. Countries can have whatever regime they want, but they shouldn't be allowed to expect big bailouts if they get into trouble with their exchange rate pegs.

Takatoshi Ito found it hypocritical for the US government to call for the two extreme exchange rate system solutions, after it had twisted the arm of the IMF to provide funds to support exchange rate pegs in Russia and Brazil. He also pointed out that Indonesia's floating regime did not allow it to stabilize. In response the first participant said the important thing was that the international community should not give ex ante signals that it will support a country with a peg exchange rate that has got into difficulty. It is not that floating your exchange rate is an automatic way to get out of a crisis, but having a floating rate might give you a better chance of avoiding a crisis to begin with.

One of the IMF officials present questioned that idea that international help is wasted on pegged rate regimes. Having a reasonable stock of reserves often does calm down the markets. Someone else questioned whether any government pursues a truly clean float, arguing that the idea that only the extreme exchange rate regime types is feasible over simplifies the issue. Another point raised was that it is not just countries that peg their currencies that develop large current account deficits. The example of Australia was raised to point to a case of a floating rate regime, significant net capital inflows, and sometimes rather large current account deficits.

Various other objections to a floating exchange rate were made. The example of Mexico was given as case where a country has had a "terrible time" managing its floating exchange rate. "Mexican business complains loudly about the difficulties of dealing with exchange rate uncertainty," this participant said. Another participant pointed out that the choice of regime should not be separated form broader monetary policy concerns, such as the need for a nominal anchor. If the main reason for pegged exchange rates is to constrain monetary policy, someone responded, then the failure of these regimes to do just that leads one to question the wisdom of their use.

Financial Sector Weakness

To stimulate discussion on the role of banks and the corporate sector, Simon Johnson of MIT was asked to make a brief presentation on the relationship between corporate governance structures and macroeconomic performance in Thailand. He began by referring to recent (joint) cross country comparative work he had done that looks how corporate governance affects how emerging market economies have weathered the crisis. His presentation to the NBER meeting focused solely on Thai listed companies, however.

These listed companies were divided into three categories: large groups (somewhere between 8 and 14 depending on where your draw the line); small groups (around 100); and government owned firms that are publicly held (around 10). In terms of outcomes, the large groups tended to outperform the others up to June 1997. From June to December of 1997 there was a severe deterioration in the performance of these large groups, however. This was in large part due to a change in the perception of these groups. Before the crisis they were seen as relative safe havens for investors, but as the crisis intensified the lack of transparency of the connections between the companies in the large groups was a troubling factor for investors. Veerathai Santiprobhob pointed out that the large groups had large foreign debt, so it was not too surprising that they would have been hardest hit. Johnson pointed out, however, that even controlling for debt levels, the large groups tended to perform poorly.

Johnson ended by listing some questions he hoped would provide the basis for further discussion.

In the discussion, the issue of whether the banks (and by extension their corporate debtors) would have been okay without the deep recession was raised. One response was that non-performing loans representing between 7 to 10 percent of total loans by the banking system was not a small number. Another participant asked the Thai experts about what was likely to happen to the companies with loans that were non-performing. He asked if it is expected that these companies will recover as the economy recovers, and also whether they will be given the time and opportunity to recover. It was noted that banks have delayed in restructuring loan portfolios in the hope that the quality of their assets would recover. They have not been forcing their corporate borrowers into bankruptcies. An IMF official asked what people thought should be done to get corporate restructuring going. One suggestion was that the banks should be forced to accept injections of capital in return for writing down these loans, rather than making such injections voluntary. The Thai official present did not believe that such effective nationalization of the banks was an attractive option. This led to some discussion of the allowing foreign banks to acquire ownership. It was noted that having world class banks was very attractive, but foreign banks were wary coming given the high degree of uncertainty about the state of the bank balance sheets.

There was also some discussion about whether there was a secondary market for the non-performing loans. Simon Johnson's reading of the evidence is that such a market does exist. But others expressed skepticism, noting that the market was considered a failure. Johnson responded that there was disappointment with the prices, but transactions were taking place. At this point one frustrated participant asked why not just shut down the troubled banks. One response was that the banks are meeting prudential regulations. Shutting them down would mean applying a stricter standard to these banks than regulations and international norms dictated. This led to some sharp exchanges about whether the banks are in fact insolvent. One participant described the banks as black holes that were obstacles to growth. They will eventually be able to extract fiscal transfers and will not disappear of there own accord. One of the IMF officials countered by asking this participant if he had ever closed a bank, and stressed the logistical difficulties of doing so. A different participant drew attention to the value of the relationship between the bank and the borrower. The value of the information that the bank has about the borrower is lost if the bank is lost if the bank is shut down and the loan sold.

The discussion came back to a question that had been asked in various through the day--were non-performing loans bad loans per se, or were they loans that had gone bad because of the crisis. Without an answer to this question it is difficult to know what should be done with the banks. Johnson also thought that it was important to distinguish between the finance companies and the banks. It was the finance companies that had made the bad real estate loans that resulted in the famous white elephant projects. At this point, the IMF official who had earlier been exasperated at the suggestion that the banks should be shut pointed out that before the crisis the larger Thai banks had the reputation for being among the best managed financial institutions in the region. Another official added that the banks had been given A ratings. Further support was given for the banks based on the fact that they were true private institutions rather than agents of the government, and that their officials were highly qualified, often with overseas training and experience. As the debate continued, the historical example of bank closings in the US during the 1930s was raised to support the view that a radical approach to shutting down banks can end a banking crisis.

The Size of the IMF Package and the Importance of Reserves

The thread running though the discussion of the reserves issue was a disagreement among participants about whether the crisis stemmed from the fundamental weakness of the Thai economy, or some sort of a self fulfilling loss of confidence that might have been avoidable with different policies.

The discussion opened with one participant expressing surprise at the suggestion that a larger package would have helped, noting that the bulk of Ito's presentation pointed to a very sick economy. Given the weakness of the financial system, he asked if the right response have been to bail everybody out. An IMF representative asked what would have been done with greater reserves. Noting that the main problem seemed to lie with the financial sector, he asked what good the reserves would have done if they were not used to bail out this sector. Another participant posed the general question of what exactly the level of reserves stand for--why is it that they are so important? They are important, a Cambridge-based participant offered, because their level is indicates how much pain the country would be able to take before being forced to devalue, and thus can deter speculators. But it was countered that this just delays the adjustment with the result that the economy ends up with even larger foreign liabilities. As another reason for a larger package, Takatoshi Ito added that such a package might also have bought more time to fix the financial sector. He also stressed that when talking about bailouts it is important to distinguish the domestic and foreign investors. He argued that the foreign creditors got bailed out anyway. A larger package of reserves might have allowed more time to fix--but not necessarily bail out--the domestic institutions.

It was also suggested that stabilizing the exchange rate makes the fundamentals actually better, and not just seem better. Echoing this theme, another participant added that the banks in particular would not have suffered the balance sheet problems that they did if the exchange rate had not fallen by as much as it did. This led to renewed arguments that the economy had developed severe imbalances by the mid-1990s that were independent of the level of the currency. One participant pointed to the fast rate of domestic credit expansion (which fueled the real estate boom) even as the rate of export growth slowed as the real exchange rate became overvalued. Every effort was being made, she argued, to keep the expansion going with easy money, which led to accumulating weakness in the financial sector.

An IMF official sought to directly answer the question of whether a $30 billion package would have helped achieve more stabilization than the $17 billion package. His answer was probably. But he noted that the political situation at the outset of the crisis led the government to have little credibility. For this reason, it is conceivable that a larger package would not have helped.

Another official questioned the conventional wisdom that Thailand did not get much support. He pointed out that the support amounted to 12 percent of GDP, compared to 13 percent for Korea and 17 percent for Indonesia; levels of support that were not all that different, he agued. In the case of Korea, part of its support came what has become known as the second line of defense, which was never actually drawn upon. In response it was also pointed out that these numbers were large compared with past bailouts. Another issue raised related to the appropriateness of measuring the amount of support as a fraction of GDP. "What is relevant," this participant claimed "is how much money is on the other side of the table." Someone else added that it is the "same house you are playing against" no matter whether you are a big country or a little country. Others thought that the size of the economy does matter, since the ability to speculate against a currency typically requires a domestic seller. Thus the ability to speculate against a country tends to be limited by the size of the market.

3.2 Managing the Crisis: Comments on the Santiprobhob Presentation

The first issue raised in the general discussion stimulated by Veerathai Santiprobhob's presentation was the appropriateness of the high interest rate in the early stages of the crisis. One participant expressed doubts that high bhat interest rates attracted foreign investors. In response, Santiprobhob noted that the authorities were also concerned with domestic capital flight. This led to a discussion of who was that was taking the money out. In particular, it was asked if the problem was foreign banks not rolling over lines of credit, or a more general capital flight. An IMF official informed the group that the non-roll over of credit lines did become a major problem until December of 1997. He also stressed that the increase in the interest rate was never all that great--it reached a maximum of about 24 percent--and was less the result of a proactive policy to raise interest rates than a foreign investors selling Thai assets. In his view, speculators would not be deterred by interest rates of 24 percent given the wild swings of the exchange rate.

The participant who initially raised the question about the source of the capital outflows argued that the answer to his question is important for how manageable the crisis is. If the problem is foreign banks not rolling over credit lines, then this problem is easier to deal with policies such as high levels of prudential foreign exchange reserves or controls on short time capital movements. On the other hand, if the problem is a more generalized capital flight--involving speculators and domestic residents--then these policy responses are less likely to be effective. In response to a question about the behavior of Thai residents during the crisis, Santiprobhob said it is important to distinguish between their taking money out of the country and taking money out of the banking system and holding it as cash. Both were important. The banks that were losing money had to turn to the FIDF for liquidity support. As he noted in his presentation, the need to fund this support led the FIDF to borrow in the short-term money market, which created upward pressure on interest rates. Another IMF official noted that there was not an overall run on the banking system. There was a run on weak institutions, but the FIDF recycled funds (through their money market borrowings) from the strong to the weak institutions.

At this point some participants expressed surprise that domestic residents were not taking their money out, noting that this situation was different to the typical Latin American capital flight scenario. Santiprobhob pointed out that this did take place to a certain extent, as corporations with export earnings tended to keep this money abroad. This then led to questions about the value of the government guarantees to depositors. One participant added that there was considerable uncertainty about what the guarantee was really worth.

The next topic that came up in discussion was the appropriateness of the government's fiscal stance. It was asked whether the abandonment of the 1 percent surplus target was a bow to reality--a belated recognition that this was unattainable. The answer given was that it was recognized that it would be counterproductive. There were also question about how tendentious was the debate between the government and the fund on the easing of the target. One of the IMF officials shared that there was general agreement on the appropriateness of fiscal easing. He observed that the Thai officials were at least as conservative as the fund on this issue. The officials were pressed on whether the fund changed its view on the appropriateness of the fiscal contraction. They admitted that they did.

The issue of financial sector restructuring also initiated some debate. An official from the US government asked if the conditions imposed by the government--notably the dilution of ownership--in return for injections of new capital were two severe. An IMF official observed that government recapitalization was meant to be a safety net, and not to drive the process. Someone else asked about the politics of bailouts. He wondered if fear of political opposition prevented bailouts that might be warranted on economic grounds. The answer offered was that the key political constraint is that shareholders are not to be bailed out. Another participant asked what was the incentive for putting equity into a bank with a large share of non-performing loans. The answer given was that the new money comes in on a preferred basis, so that the original owners are forced to take some pain.

Questions were also raised about the timing of reforms. Santiprobhob's account painted a picture of a rational policy progression from stabilization to stimulus. One participant asked if, with hindsight, policy mistakes were made. As an example, it was asked if the initial fiscal policy was too tight. Santiprobhob agreed that it was. But that in the early stages of the crisis, it was not known how severe the economic contraction would be. Another participant noted that it is very hard for governments or international institutions to offer forecasts that are more pessimistic than the consensus. In part, to avoid making the confidence crisis worse. He speculated that the size of the economic slump was apparent to policy makers even as the fiscal stance was kept tight.

The IMF officials present were asked if they wanted to add anything to the Ito and Santiprobhob presentations. By and large, they agreed with the accounts given, but sought to underline a few points. They stressed that the crisis was a mix of a traditional balance of payments crisis and a financial sector crisis. The former required a standard response, including fiscal contraction and a step devaluation. It was also pointed out that the IMF did not have access to key data in the early stages of the crisis. In particular, they did not know the magnitude of the forward dollar contracts, or the amount of liquidity support that had been provided by the FIDF (close to 100 percent of base money by July 1997). This information conditioned the policy response sought by the fund. The IMF's key objectives at the outset were to reduce the current account deficit and to stop the money flowing from the central bank through the FIDF.

At this point, discussion turned again to the issue of non-performing loans. One participant wanted to know what fraction of the non-performing loans was due to the crisis, and what fraction were already non-performing in the first half of 1997. Although no one was able to give a firm estimate, a participant quoted a JP Morgan study that put the fraction of non-performing loans at 25 percent of total loans at the beginning of the crisis. An IMF participant referred to the pre-crisis fund report from June 1997, which, he claimed, laid out the problems that were present. Another IMF official reported, however, that when he visited Thailand in March of 1997, the official figures on non-performing loans in the banking system put them at just 7 percent. This official continued by saying that the Thai's should could a lot of credit for how they have managed the crisis.

Meeting chair Martin Feldstein attempted to switch discussion to the issue of whether the IMF conditions too tight, leading to a damaging credit crunch. Takatoshi Ito responded first, saying that he believed the 1 percent initial fiscal surplus target was overly strict (especially given the tightening from the monetary side). He noted that the IMF's excuse that they did not foresee the drop in GDP was not convincing given that this drop partly a function of the fiscal stance. In defense, one of the IMF officials distinguished between doing the right thing ex ante (with the information that they had in July 1997) and the right thing ex post (what would have been the best policy with hindsight). Given that the situation looked like a conventional balance of payments crisis at the outset, he felt that their policy prescription was appropriate. One participant asked if there were not enough contractionary forces anyway--including the wealth effects of the exchange rate depreciation. Why was a further contraction through fiscal tightening necessary? Another IMF official admitted that their program was predicated on the assumption that capital would not flow out as much as it did and that the exchange rate would not fall as much as it did. They knew the signs of these effects, but not the magnitudes.

Following up on the IMF's approach to fiscal targeting, a participant asked why the Fund does not use a structural deficit target, or, assuming that this is too difficult to define, at least a target that is conditional on the GDP performance. The IMF is doing that now, the official responded, noting that this is one of the changes induced by the crisis. Another important constraint noted by the IMF officials is that it proved to be difficult to bring expenditure programs on line quickly in Thailand. So that it was hard to reverse the contractionary fiscal stance. Given everything else that was going on, another participant wondered if a more expansionary fiscal stance of a couple of percentage points of GDP would have made all that much difference. Following on from this, the IMF officials were asked what model they used to predict the impact on output of various levels of government spending. Their model, they admitted, is basically an accounting framework, and predictions of policy effects involve a large amount of judgement. Based on his observations of how the IMF operates, one participant suggested that the fund bases its predictions to a large extent on its experience with other, what it believes to be similar, countries. Thailand, he surmised, turned out to be a very different country from what the fund initially thought. An official responded that some people were comparing Thailand to Mexico, which he agrees was not appropriate.

Returning more directly to the issue of the credit crunch, a participant asked how the fund takes into account the importance of the lending channel for monetary policy. One of the officials argued that this channel is hard to operationalize, given that credit from the private sector to the private sector is not an instrument of the government. Another official who had worked more directly on Thailand disagreed, arguing that the lending channel was considered to be very important in the Thai case. They understood that the credit crunch had little to do with the high interest rates, so that alleviating the credit crunch was not simply a matter of getting interest rates down.

Following up on the question of what policies were required to get credit flowing again, it was asked if it was now believed to be a mistake to initially suspend the finance companies, rather than going quickly for a more complete resolution. He argued that such suspensions only add to the uncertainty. Veerathai Santiprobhob of the Thai Ministry of Finance agreed, but pointed out that weaknesses in the law and political obstacles made more radical action different.

Another participant was interested to know if there was any discussion of allowing the banks to default on foreign liabilities. Was there any support for allowing the foreign creditors to take some pain, he wondered? The answer is that it was considered, but fears of bank runs led to the giving of the blanket guarantee of depositors, both foreign and domestic. A meltdown was to be avoided at all costs. The power that the government had to shut down the banks was also queried. One of the IMF officials said the power was present, but a political consensus was not.

3.4 The Easing of the Fiscal Policy Stance

After the lengthy discussion about banks and credit, Martin Feldstein used his power as chairman to switch the focus back to fiscal policy. Again the issue was the major shift in the policy stance on fiscal policy. He asked if the IMF was fully on board. The official agreed that they were. This led to questions probing exactly when it is that the IMF supports--or at least is willing to allow--large budget deficits. A number of reasons were given for why the fund looks more favorably on Thai deficits than they did at the outset of the crisis, including an improved current account balance and the elimination of the offshore swaps that had initially so alarmed the fund. Overall, the balance of payments is in a very stable position, leaving much more scope for fiscal stim.ulus. Another participant pointed out that estimate of the cost of bailing out the financial system had fallen, which also allowed for a larger deficit. One participant stressed that the fiscal remedies being advocated now by the fund are appropriate for the present state of the crisis. In the early stage of the crisis--when major balance of payments problems existed--the contractionary policy stance was appropriate. The fact that the IMF is supporting a fiscal deficit now is unlikely to imply any change in their standard approach to a balance of payments crisis.

Other questions were raised about the evolution of the fund's thinking on fiscal policy. First, the officials were asked if there had been any change in thinking on how investor confidence is affected by the fiscal policy stance. Especially after Mexico, the conventional wisdom was that fiscal austerity was a way to prove to investors that macroeconomic management was in good hands. One official responded by taking issue with what he called the Ito view that the initial fiscal austerity had been partly behind the contagion in the region, doing more harm than good to investor confidence.

The next issue raised was the sufficiency of IMF review to adjust targets. One participant wanted to know if target adjustments did not come quickly enough because of the elapsed time between reviews. The fund officials noted that in more recent cases, reviews are being conducted at much shorter intervals in crisis situations--every two weeks for Korea and monthly for Indonesia. Takatoshi Ito reminded the gathering that there was no sense of crisis in the summer of 1997. It was believed to be adequate to come back and reassess the situation in October. One IMF official cautioned that the severity of the crisis should not be exaggerated. Comparing Thailand with Indonesia--where things did go badly wrong--he noted that the situation in Thailand did stabilize reasonably quickly. Even though there was a major recession, manufacturing output did begin to turn around fairly early in 1998, and catastrophes such as a total loss of nominal control were avoided. Indeed, inflation never rose too far and interest rates have been brought down to below (the low) pre crisis levels.

Questioning then turned to the ultimate cost to the Thai government of bailing out the financial sector. The IMF estimate of the cost--which even the IMF officials offered tentatively--was $33 billion (or around 25 percent of GDP). Veerathai Santiprobhob thought that this estimate, which was based on the accumulated bad debt of the financial sector to date, was too high since it did not take into account the secondary market value of that debt. However, it was asked if anyone knew of a case where the ultimate cost of a bailout was overestimated in the midst of a crisis.

3.5 Some Final Comments

Asked if all the important his questions had been addressed, Takatoshi Ito mentioned that the role of hedge funds and the related issue of capital controls had not been discussed. Given the activities of speculators, he asked if it was a mistake for Thailand to have open capital markets. One response was that there were many balance of payments crises in the 1950s and 60s with very restricted capital accounts. The countries experiencing these crises ultimately experienced very high costs. This participant was not convinced that the ability to "pull the trigger faster" in today's capital markets was a bad thing for countries receiving capital inflows. When the macro fundamentals are wrong, problems will follow--free capital account or not. Another participant stressed the importance of distinguishing between controls on inflows and controls on outflows. "These are very different and should be discussed separately," he said.

Some discussion developed on the subject of implicit guarantees to foreign investors. One participant pointed to the bail out of foreign creditors to private sector banks in Chile in the 1980s as the beginning of a period where the foreign investors always ended up getting bailed. "Did the failure to bail out the private investors in Russia signify that something had changed," it was asked. Adding that this failure to bailout investors gave a lot of bankers quite a shock. Some participants were doubtful that much had changed, so that investors would continue to lend on the belief that there was little chance they would not be paid back.

The meeting ended with Martin Feldstein thanking Takatoshi Ito and Veerathai Santiprabhop for their efforts, and applauding the assembled group for the quality of the discussion. He gave a special thanks to the IMF officials who attended, noting that the usefulness of the debate was greatly enhanced by their presence. He closed by looking forward to the up coming meetings on other crisis affected countries.


1 The Asian economies are said to follow a "flying geese" pattern, whereby the countries are lined up in order of the sophistication of the products they produce. Despite its recent difficulties, Japan leads in advanced technology, and has (partially) shed industries such as consumer electronics and shipbuilding to countries such as South Korea and Taiwan. These latter countries in turn yield industries such as textiles to the ASEAN countries. Thailand has been shifting from textiles to consumer electronics and auto assembly. One problem that Thailand faced is that it was not able to move along the sophistication line fast enough given growing competition from China (Ito background paper).

2 Temporary guarantees were provided to remaining financial institutions.