The Asian Financial Crisis of 1997

The miracle, and what it was hiding

One has to begin with the period the crisis interrupted, because the interruption is unintelligible without it. Between roughly 1960 and 1996 the economies that would come to be grouped as the “Asian Tigers” — Thailand, Indonesia, South Korea, Malaysia, the Philippines, with Hong Kong and Singapore as more developed cousins — compounded real GDP at somewhere between 7 and 10 per cent a year for three decades. The World Bank called it the East Asian Miracle in a 1993 report that aged, on balance, less well than its authors intended. Hundreds of millions of people left agricultural poverty; container ports rose where paddy fields had been; a manufacturing middle class came into existence in places where it had no precedent.

The recipe was, by the mid-1990s, a cliche: high domestic savings, export-led industry, comparatively cheap labour, aggressive infrastructure build-out, and — this is the point that matters most for what followed — exchange rates pegged, formally or de facto, to the US dollar. The dollar peg was the stabiliser of the whole construction. It gave foreign lenders a currency they understood, allowed domestic borrowers to take on cheap dollar-denominated debt without apparent exchange risk, and kept export prices predictable. For a while, it was an elegant arrangement. Elegant arrangements in macro tend to mature into brittle ones.

The vulnerabilities accumulated rather quietly. Thailand’s current account deficit reached roughly 8 per cent of GDP by 1996, financed almost entirely by short-term foreign capital. Banks and corporates across the region had borrowed in dollars at 5 or 6 per cent and lent at home in local currency at 12 or 15, pocketing the carry and assuming the peg would hold. The mismatch was classic and, one imagines, visible to anyone who went looking: dollar liabilities, baht or rupiah assets, no hedge. Bangkok office vacancy crossed 15 per cent while new towers were still being topped out. Connected corporates received loans on terms no credit committee anywhere else would have approved. The foreign capital financing all of this was what the literature politely calls portfolio flows and what traders, less politely, call hot money. None of this was a secret. It was simply inconvenient to price.

The peg breaks

Thailand: July 1997

By the spring of 1997 the positioning was already well-established. Macro funds — Soros’s among others, but by no means only his — had identified the baht as the weakest peg in the system and were pressing it from the offshore forward market. The Bank of Thailand did what central banks always do in this position, which is to say it defended. It burned through roughly $23 billion of its $30 billion in reserves in a few months of intervention. On July 2 it stopped. The baht moved from 25 per dollar to, within months, above 50. A 50 per cent devaluation is not a repricing; it is a regime change, and in a system where half the corporate balance sheet is denominated in the foreign currency, it is a solvency event disguised as an exchange-rate event.

Contagion: August to November

What followed is the part of the story that looks most familiar to anyone who has sat through a subsequent crisis. Foreign capital did not stop to distinguish between Bangkok and Jakarta and Seoul. It left. The currencies went over the cliff together:

Indonesia was the worst of it. An 84 per cent depreciation in the rupiah converted any reasonable dollar liability into an unpayable one overnight. Bank balance sheets evaporated. Unemployment doubled and then doubled again. The price of rice, of all things, became a political variable. By May 1998 there were riots in Jakarta that killed more than a thousand people, and by the end of the month Suharto, who had governed Indonesia for thirty-two years, had resigned. Pause on this for a moment. A currency crisis toppled a government that had survived the Cold War, a coup attempt, three recessions and the end of Dutch colonial memory. Markets are not, in the end, an abstract sphere.

Hong Kong: October 1997

Hong Kong’s defence is the case study that should be taught in every macro seminar that treats “pegs can be defended” as a serious proposition. The Hong Kong Monetary Authority had the reserves (roughly $90 billion) and the institutional seriousness that Thailand had lacked. It defended the linked exchange rate by taking overnight rates briefly to 280 per cent — a figure that is not a typo — making the carry trade against the currency prohibitively expensive. The peg held. The Hang Seng, meanwhile, fell from 16,800 in August 1997 to 6,660 a year later. The reserves win; the equity market pays the cost. The choice is rarely the one the textbooks advertise.

The ripple

The crisis did not stay in Asia, and the mechanism by which it travelled deserves more attention than it usually gets. Commodity prices fell as Asian import demand collapsed; oil in particular went from around $25 a barrel in early 1997 to below $11 by late 1998. This was fatal for Russia, whose post-Soviet fiscal arrangement was essentially a commodity-price regime masquerading as a budget. In August 1998 Russia defaulted on domestic GKO debt and devalued the rouble — the first major sovereign default since the war — which in turn caught Long-Term Capital Management wrong-footed on a suite of convergence trades that had assumed, in the style the firm was famous for, that the relationships would mean-revert before the margin calls arrived. They did not. The Fed orchestrated a $3.6 billion private-sector rescue to prevent what the participants at the time believed, with reason, to be a systemic event. The S&P 500 fell 19.3 per cent between July and October 1998. The Fed cut rates three times that autumn. Those cuts, on any honest reading, helped ignite the final phase of the dot-com bubble that would burst eighteen months later. Crises do not end; they metastasize.

The IMF argument, which has not fully ended

The IMF arrived with conditions attached. Fiscal consolidation. High interest rates to stabilise the currency. Bank closures. Structural reform of the relationship between state, banks and favored conglomerates. The debate over whether this package helped or deepened the slump has never fully resolved, and I have views.

The critics were led, loudly, by Joseph Stiglitz — at the time the Bank’s chief economist, a post he used to rather publicly disagree with the Fund next door — and the critique was not silly. Raising policy rates to 30 or 40 per cent to defend a currency does indeed crush domestic firms whose only sin was being headquartered in the wrong country. The medicine was, in places, genuinely worse than the disease. Mohamed El-Erian, writing later, put it carefully: the Fund applied a template designed for fiscal-profligacy crises to what was fundamentally a private-sector balance-sheet crisis, and the template did not fit.

The defenders had their case too. Without IMF liquidity the dollar debts would have defaulted outright, foreign credit would have remained closed for a decade rather than three or four years, and the recoveries — which were genuine — would have been longer and uglier. The honest answer is that both sides are partly right. The Fund was too austere in 1997 and learned something from it; its 2008 interventions in Europe, whatever else one thinks of them, were not designed on the same template.

Malaysia’s decision to refuse the Fund and impose capital controls is the interesting counterfactual. Mahathir was pilloried in the English-language financial press at the time — I remember the tone, which was close to incredulous — and his recovery turned out to be no worse than the region’s, and arguably a little better. Capital controls, which had been a heretical instrument in the 1990s consensus, have since become something the IMF itself admits can be appropriate under specific conditions. The profession updated. Rather quietly, but it did.

The human cost

Before the numbers, a qualification. The macro profession, including me, has a tendency to treat GDP contractions as line items. Indonesia contracted 13.1 per cent in 1998; the poverty rate went from 11 per cent to roughly 40; tens of millions of people who had climbed out of subsistence fell back into it in a single calendar year. Thailand contracted 10.5 per cent; hundreds of thousands of construction workers lost work when projects were halted literally overnight; the half-finished concrete frames of what had been speculative office towers stood unfinished across Bangkok for the better part of a decade. Korea contracted 5.8 per cent in a society whose social contract had been lifetime employment at the chaebols, and the mass layoffs that followed broke that contract in a way that had genuine consequences for suicide rates and for the political centre of gravity of the country.

The Korean gold campaign, in which ordinary citizens queued at banks to donate wedding rings and heirloom jewelry to help repay the sovereign’s foreign debt, is the detail that tends to come back to me. It is the sort of collective act that is hard to imagine in most Western societies, and, one imagines, tells you something about Korea that matters beyond the crisis. The instruments a macro desk trades have, at the far end of the chain, people queuing for rice and people donating gold. The textbooks underweight this.

What changed in the plumbing

The crisis left a set of structural adjustments that shape markets to this day.

The largest is the reserve accumulation. Asian central banks, having discovered what it feels like to run out of ammunition in the middle of a speculative attack, spent the following decade stockpiling dollars with a determination that was not, at first, fully understood in Washington or Frankfurt. China’s FX reserves crossed $3 trillion by 2014. This had a second-order consequence Ben Bernanke would later call the “global savings glut,” and it was a non-trivial contributor to the low dollar yields that enabled the US housing bubble. The 1997 crisis, in other words, has a direct lineage to 2008. One does not usually see it drawn on the same chart.

Most of the affected countries moved from hard pegs to managed floats. This is the right lesson, and it was learned. Banking supervision was tightened, unevenly, across the region. Corporate governance in Korea improved visibly; Thai financial supervision, less visibly; Indonesia’s reforms were slower and had periodic reversals. The emerging market asset class, which had been priced for most of the 1990s as a single risk factor, began to be analysed country by country on reserve adequacy, external debt composition, institutional quality. It is obvious now. It was not then.

What a macro reader should take away

I will resist the temptation to itemise this as a numbered list of timeless principles, because the pattern of numbered lists of timeless principles is itself a warning sign. A few observations, in no particular order.

Pegs that diverge from fundamentals do not survive. The Thai baht in 1997, the pound in 1992, the Argentine peso in 2001, the Swiss franc floor in 2015, the Egyptian pound at several points: the mechanism is always the same. The central bank holds the line until it cannot, and the line breaks between the close and the open of a single trading day. A macro trader reading positioning in a pegged currency should assume that the binary outcome — peg holds, peg snaps — does not have a 50/50 distribution over time; it has a 99/1 distribution until it does not, which is why the carry looks free for years and then, rather quickly, is not.

Contagion moves faster than fundamental reassessment. In the weeks after the baht devaluation, foreign capital pulled out of Indonesia and Korea without a serious reappraisal of either economy, because in a liquidation the unit of sale is not the individual credit but the category. This is the Bismarckian observation, if I can borrow it carefully: systems of alliance hold because the participants choose to treat them as binding, until one defects and the whole apparatus reprices. Capital flows have the same property. When one peg breaks, the market stops treating the others as credible regardless of their individual merit.

Leverage is the common denominator. Thai banks borrowing dollars to fund baht assets, Korean chaebols running debt-to-equity ratios above 400 per cent, LTCM at thirty times, the Russian fiscal position in 1998 — these are the same crisis in different currencies. Leverage converts a volatility event into a solvency event. That is its defining feature. The 1997 crisis is, at its core, a leverage crisis; so is every other crisis in this guide.

Hot money creates fragility at every scale. What was true of Thailand in 1996 — that the capital funding it could leave faster than the economy could adjust — is true of a small-cap equity whose bid is momentum-driven, or of a sovereign whose financing rolls quarterly. If the marginal buyer can become the marginal seller without friction, the asset is more fragile than its price suggests. This applies up and down the capital structure. Bond investors, in particular, tend to discover it only after the fact.

Diversification across similar-looking exposures is not diversification. A portfolio of Thai, Indonesian and Korean equities in July 1997 was not three positions; it was one position dressed in three languages. Sector rotation and geographic rotation matter, but the correlations that protect you are the ones across fundamentally different regimes, not across flavours of the same regime. True diversification is harder than the brochure suggests.

And a closing observation. Every economy in the crisis recovered. Korea emerged with Samsung and Hyundai on a global footing; Thailand rebuilt its financial system; Indonesia eventually stabilised, democratised, and grew. The crisis was not the end of the Asian story. It was a violent reset that forced adjustments that had been politically impossible the year before. Short-sellers got paid in 1997 for identifying the mispricing; long-term holders who bought near the bottom in 1998 were paid, eventually, too. Both are true. The difficulty, as always, is the middle distance between the two, which was where the real cost was absorbed.

See also: The 2008 Financial Crisis · The Dot-Com Bubble · George Soros · What Is a Bear Market? · The Federal Reserve and Interest Rates