RDP 9805: The Origin of the Asian Financial Turmoil 2. What Happened?

When a country's foreign exchange market is under pressure, that pressure is typically reflected in exchange rates, international reserves, interest rates, and the temporary imposition of controls on credit or foreign exchange transactions.^[3] The combination depends on the circumstances of the particular country affected and on the defensive strategy chosen by the authorities. In some cases, currency crises also spill over into (or are caused by) other asset markets, including banking, equity and bond markets. When the authorities are unable to regain the confidence of markets on their own, they may ask the IMF and the World Bank, along with other countries, for assistance. As background to the subsequent discussion on origins, this part of the paper reviews the main developments.

2.1 Exchange Rates

Although the Thai baht had been subject to episodic pressures during the previous 18 months, it was the intensification of these pressures in May 1997 and the subsequent forced floating of the baht on 2 July that really ushered in the recent turbulence in currency and equity markets. As shown in Figure 1, the second half of 1997 and the first weeks of 1998 witnessed sharp declines in the dollar exchange rates of the ASEAN-4 economies. The worst affected was the Indonesian rupiah, which at one stage was worth only a fifth of its June 1997 value against the US dollar. The baht itself lost half its value while the Malaysian ringgit and Philippines peso were down around 40 per cent at their lows. There were more modest declines in the Singapore and Taiwan dollars over the same period. South Korea's currency initially held up better but depreciated heavily in November and December 1997, also losing half its value by the low point. The Hong Kong dollar maintained its parity despite strong market pressure and the Chinese renminbi was little affected.^[4] With hindsight, January 1998 now appears to have marked the worst of the depreciations. All the affected currencies regained some strength during the first quarter of 1998. The recovery has been most hesitant in Indonesia, the economy about which most uncertainties remain.

On an effective (i.e. trade-weighted) basis, the ASEAN-4 and Korean currencies recorded slightly smaller depreciations, the Singapore and Taiwan dollars were relatively stable while the Hong Kong dollar and Renminbi appreciated.^[5] (Figure 2.)

Real effective exchange rates over the 1990s are shown in Figure 3 for seven Asian emerging economies. Two observations merit mention. First, the combination of the close (explicit or implicit) tracking of the US dollar by many Asian currencies and large swings in the yen-dollar exchange rate have meant that the ASEAN-4 currencies have shown marked medium-term fluctuations both with respect to the yen and to a wider trade-weighted basket of exchange rates. Specifically, after depreciating vis-à-vis the yen in the first half of the 1990s, the ASEAN-4 currencies displayed real appreciation relative to the yen between early 1995 and mid 1997. Their real effective exchange rates showed the same pattern, albeit much damped in magnitude. These real appreciations relative to non-dollar currencies in the 18 month run-up to the current crisis are noteworthy because they presumably contributed to any market concerns about overall exchange rate overvaluation.

Second, the recent depreciations of the ASEAN-4 currencies have come from different levels of their real effective exchange rates (relative to historical averages). Whereas the real effective exchange rates of both the Philippine peso and the Singapore dollar were around 15 per cent above their average level of the previous decade prior to the recent fall, the Thai baht and Indonesian rupiah were less than 10 per cent above and the South Korean won was below its medium-term average. (Figure 3).

This comparison of movements of real effective exchange rates in the 1990s suggests that, uncertainty about equilibrium real exchange rates notwithstanding, in mid 1997 the extent of any exchange rate misalignments was quite modest in comparison with the size of the subsequent depreciations.

The size of the Asian depreciations is placed into perspective in Table 1. This shows the 20 largest real effective depreciations over six-month periods since 1970 (excluding countries with inflation over 50 per cent as such calculations are unreliable during hyperinflations). By comparison, the largest six-monthly change in the G3 currencies over this period was the 1995 real depreciation in the yen of around 20 per cent; some European currencies recorded similar real depreciations in the ERM crisis and the Australian dollar had a similar real depreciation in the mid 1980s. Taking Table 1 and Figure 3 together one gets an impression of just ‘how large’ have been these real depreciations in emerging Asia, both with respect to historical experience and to simple proxies for equilibrium rates.

2.2 Exchange Market Intervention

It is not straightforward to describe the role played by exchange market intervention as available data capture only part of these operations. Although the concept most relevant for vulnerability to a speculative attack would be net international reserves relative to some measure of potential liquid liabilities of the authorities (Section 3.1), published data generally refer only to gross reserves and thus miss intervention by the authorities in the forward market. As widely reported, such forward market operations were an important element of the Thai defence, involving a position which peaked at US$24 billion. The individual-country reserve figures may also miss co-ordinated intervention undertaken in support of currencies by (third party) monetary authorities or intervention that is funded through a country's fiscal accounts rather than its international reserves. Furthermore, it has been suggested that some of the Korean international reserves were not useable as they had been lent to banks.

With these caveats in mind, Figure 4 shows the behaviour of gross international reserves over the 1996–1997 period for eight Asian economies. It shows that Thailand and South Korea initially relied much more on their reserves in defending the exchange rate than did Indonesia, Malaysia or the Philippines. One interpretation is that the Malaysian and Indonesian authorities were mindful of the large reserve losses suffered not only by Thailand but also by some other countries in earlier vigorous but ultimately unsuccessful currency defences (e.g., the huge interventions during the ERM crisis) and decided not to go far down that road; that is, they intervened heavily for a short period and then turned to other measures. In the case of the Philippines, low initial reserve levels dictated limited reliance on intervention as part of the defensive strategy. Figure 4 also illustrates the large reserve holdings of Hong Kong, Singapore and Taiwan relative to the ASEAN-4 countries. Despite its large stock of reserves, Taiwan opted on 20 October to allow its currency to depreciate.

2.3 Interest Rates and Credit Controls

An important element in most currency defences is played by short-term interest rates. These are shown in Figure 5, with corresponding real rates shown in Figure 6.^[6] Four developments should be noted. Firstly, the ASEAN-4 countries and South Korea had increased interest rates during 1996, either to rein back excessive credit growth or support the exchange rate. Consequently, they found themselves in the first quarter of 1997 with already high short-term real interest rates. Some might argue these high interest rates increased the vulnerability to attack because efforts by speculators to push up the cost of ‘holding on’ to exchange rate commitments begun from a relatively high base.

Secondly, as exchange market pressures intensified in 1997, each of the ASEAN-4 countries further increased short-term interest rates to combat those pressures. As shown in Figure 5, short-term (nominal) interest rates in Thailand and Indonesia displayed high volatility in 1997, especially from May onwards; in some episodes, government policies and market pressures drove overnight rates to very high levels (near 100 per cent in Indonesia, and over 350 per cent in the offshore Thai market), albeit for relatively short time periods. Malaysia's use of high interest rates was the most limited among the ASEAN-4 countries. There was a large spike in early July but rates then returned to pre-crisis levels.

Thirdly, reminiscent of the tactics employed by several European countries during the 1992–93 ERM crisis (e.g., France, Ireland, Portugal and Spain), each of the ASEAN-4 countries resorted to a mixture of administrative controls, taxes, and moral suasion, aimed at discouraging capital outflows and short sales of both the domestic currency and domestic equities, and at reducing the need for even larger increases in domestic interest rates.^[7] In May 1997, the Thai authorities introduced restrictions on the ability of local banks to extend baht credit to offshore banks, to conduct foreign exchange swaps (baht for dollars), and to sell baht for dollars to speculators in the spot offshore market. In July, the Philippines authorities prohibited local banks for three months from engaging in offshore, nondeliverable forward peso contracts with offshore banks, and there are reports that the central bank also used moral suasion to discourage local banks from making peso credit available to those who might be expected to use it to speculate against the peso. The Indonesian authorities resisted for longer any temptation to impose capital controls but in August introduced limits on foreigners' access to swap markets. In Malaysia, a ban on short selling of equities was introduced but was soon lifted. A $20 billion plan was also announced to support stock prices, under which shares sold by Malaysians to state pension and investment funds would fetch a premium whereas shares sold by foreigners would occur at market prices. In addition, Prime Minister Mahathir has threatened on several occasions to introduce controls on foreign exchange trading but these have not been implemented. When the Hong Kong dollar was under strong attack in late October, Hongkong and Shanghai Banking Corporation (a private bank, but one that undertook some central bank functions before the establishment of the Hong Kong Monetary Authority) temporarily restricted withdrawals on time deposits to aid in the currency's defence.

While it is difficult to know how exchange and interest rates would have behaved in the second half of 1997 in the absence of these administrative measures, one thing is clear: as in the ERM crisis, such administrative measures did not prevent ASEAN-4 exchange rates from declining sharply, nor did they obviate the need for domestic interest rates to increase significantly.

2.4 Equity Prices

Figures 7 and 8 trace the evolution of equity prices in eight Asian emerging economies, in the short term (since January 1997) and medium-term (since 1990), respectively. The ASEAN-4 countries suffered steep falls in their equity markets in the second half of 1997. As regards the four other Asian emerging economies shown, Hong Kong and South Korea registered falls comparable with those in three of the ASEAN-4 countries, whereas the decline has been more moderate (around 20 per cent) in Singapore and Taiwan. As with their exchange rates, the equity markets turned around in January 1998. By April 1998, most were higher than at the start of the year, but still well down on mid 1997.

Turning to equity prices over the medium term, there are again significant differences across countries. Most dramatically, even before the crisis, Thai stock prices had fallen by over 60 per cent from their peak in early 1994 and by end 1997 were about half their level at the beginning of the decade. In contrast, stock prices had tripled in the Philippines over this decade, doubled in Malaysia and risen by 60 per cent in Indonesia.

The other east Asian emerging economies whose stock market performance in the 1990s have been as disappointing as Thailand's are South Korea and Taiwan. In the former, stock prices have fallen over 60 per cent since the peak (in late 1994) and by end 1997 were about half their 1990 level. In contrast, even after the sharp fall in late 1997, Hong Kong's Hang Seng index had gone up over threefold since the beginning of the decade.

The behaviour of equity markets is important because they serve as an alternative to bank financing for some companies, because sharp changes in stock prices can generate non-trivial wealth effects, because banks either use equities as collateral for bank loans and/or invest themselves in equities, and because they are an important channel for foreign capital inflows. Moreover, they are often regarded as a barometer of confidence in the domestic economy and as either an early warning signal or precipitating factor in many past financial crises in emerging economies.^[8]

2.5 Real Impact of the Financial Problems

One immediate effect of the financial crisis has been a contraction in real activity. In Thailand, this is already apparent in the industrial production data which was 17 per cent lower in January 1998 than in January 1997. The other economies were later to suffer their financial problems and there are longer lags in data releases. However, an indication of the magnitude of the problem can be gleaned from a comparison of the ‘consensus’ growth forecasts for 1998 made in June 1997 with the latest forecasts made in April 1998. (The revision should capture the impact of the financial crisis rather than any other factors thought likely to slow output before the crisis.) Indonesia, Thailand and South Korea are the economies whose growth has been most marked down, and the downgrades are the most severe recorded in the 1990s. Real GDP is now expected to contract this year, despite the stimulus to net exports from their depreciations, reflecting, inter alia, high interest rates, vastly increased foreign currency debt obligations, more widespread corporate and banking failures, higher layoffs, and the generalised uncertainty and loss of confidence associated with such crises.

There should be large movements in the current accounts of the affected economies. Mexico changed its current account position by 6 per cent of GDP in the first year after its mid 1990s crisis. Already South Korea and Thailand have recorded some monthly current account surpluses. The April consensus forecasts imply the combined merchandise trade balance of Indonesia, Thailand and South Korea would rise by almost US$50 billion between 1997 and 1999.

However, the available data thus far, illustrated in Figure 9, as well as the forecasts for 1998, indicate that in the short term the change in trade balances will mainly come from a sharp reduction in imports. At least thus far, the competitiveness benefits for exporters seem to be outweighed by other factors such as difficulties in exporters attracting finance or collapse of neighbouring markets.

These data refer to the US dollar value of merchandise exports and imports. Stability in US dollar value of exports would be consistent with increased export volumes (and so greater real GDP) if prices are being cut in US dollar terms, but still mean that exports are not helping to rebuild reserves or increase national income.

2.6 International Rescue Packages

IMF-led rescue packages have been assembled for Thailand, Indonesia and South Korea. Table 3 compares them with the previous largest rescue package, for Mexico in 1995. The IMF contributions are large, especially relative to the size of their quotas (usually loans do not exceed two or three times a quota) and the contributions from the rest of the world are also substantial. Japan has been the largest contributor, followed by the US, although the latter was not involved with the Thai package. It is noteworthy that Japan also has the largest share of bank lending to the crisis countries, followed by Europe (principally Germany, France and the United Kingdom), Hong Kong and the United States (Table 4).

The packages have been subject to substantial revision, notably making the budgetary demands less onerous, as the economic outlook for the economies has deteriorated. No international package has been sought by Malaysia, although it has voluntarily undertaken some reforms. The Philippines agreed in July 1997 to strengthen an ongoing IMF adjustment programme and received access to over US$1 billion in additional financing.

The IMF's assistance is subject to policy conditionality and in many cases the supplementary facilities offered by other countries are also contingent on compliance with these conditions. The performance criteria on monetary and fiscal policies have presumably been set with an eye toward improving the current account, providing some government contribution to the cost of bank restructuring, maintaining control of inflation, and braking the downward slide in exchange rates. The original growth targets in these programmes have had to be revised downward, as evidence accumulated that the breadth and depth of the crisis would be greater than anticipated. The international rescue targets, in effect, put constraints not only on exchange market intervention but also on the extent to which IMF resources can be used for other purposes.

A distinguishing feature of these rescue packages is that they go much further into reform of the financial sector and corporate governance than most IMF programmes; see Tables 5 and 6. Common elements of the required conditions include improved prudential supervision of the financial system, privatisation and removal of anti-competitive practices.

The rationale for including corporate governance and transparency elements in these programmes reflects the view that weaknesses in this structural area (non-consolidated accounting statements, overleveraged conglomerates, cross-company shareholding arrangements and guarantees, inefficient government monopolies and cartels, absence of official subsidies from published government budget accounts, etc) were one of the underlying elements of vulnerability.

Footnotes

Indeed, these variables typically serve as the components of indices of exchange rate crises in the growing literature on early warning indicators of currency crises; see, for example, Kaminsky and Reinhart (1996), Goldstein and Reinhart (1998) and IMF (1998). [3]

Changes in exchange rates are calculated in terms of units of foreign currency (e.g. US dollar) per unit of domestic currency, that is, the ‘world’ price of the domestic currency. This results in smaller changes than using units of national currency per unit of foreign currency. For example, a movement of the baht from 25 to 40 to the US dollar is regarded as a 37.5 per cent depreciation, whereas the latter approach would calculate the depreciation as 60 per cent. A disadvantage of the latter definition of the exchange rate is that changes of 100 per cent might be interpreted as implying that the currency has lost all its value when that is surely not the case. [4]

The series used in Figure 2 are calculated by the RBA. The series used in Figure 3 are calculated by JP Morgan and are based on baskets of 45 currencies. [5]

Real rates are calculated by subtracting the most recent 12-month-ended percentage change in consumer prices. They therefore are likely to be overstated in late 1997, by which time expected inflation would have incorporated the inflationary impetus from the depreciations. [6]

These measures are described more fully in IMF (1997). [7]

See Mishkin (1996), Kaminsky and Reinhart (1996), and Goldstein and Reinhart (1998). [8]