Volume II · After Gold · Money Research

05 — Globalization and the Emerging-Market Crises, 1990–2001

The trade treaties

Floating currencies and freer capital movement developed alongside efforts to reduce trade barriers. The North American Free Trade Agreement took effect on 1 January 1994 and reduced barriers among the US, Canada and Mexico; it was a free-trade agreement, not the EU-style single market with free movement of people, services, goods and capital. The European single market dates from 1993. Council of the EU, single-market overview. The Uruguay Round of GATT negotiations, launched in 1986, concluded in 1994 and established the World Trade Organization on 1 January 1995. China joined the WTO on 11 December 2001. The scale and distribution of the subsequent trade, employment and Treasury-investment effects require separate period-matched evidence; chronology alone does not assign them to accession.

This is the mechanism economists later called "Bretton Woods II" (Dooley, Folkerts-Landau and Garber, 2003): Asian exporters kept their currencies cheap against the dollar, accumulated dollar reserves, and lent them back to the United States, exactly as Europe and Japan had done under Bretton Woods I, except that the dollar was now backed by nothing but the expectation that the arrangement would continue. China's foreign-exchange reserves grew from about $150 billion in 1999 to nearly $4 trillion in 2014.

The Tequila crisis (1994–95)

Mexico had pegged the peso within a crawling band, run current-account deficits of 7% of GDP financed by short-term dollar-linked bonds (tesobonos), and in 1994 suffered a peasant uprising in Chiapas and two political assassinations. On 20 December 1994 the new Zedillo government devalued by 15%; the market took the peso down 50% within weeks and Mexico could not roll over $30 billion of tesobonos. The Clinton administration, bypassing Congress, put together a $50 billion package from the Treasury's Exchange Stabilization Fund and the IMF — at the time the largest rescue ever. Mexico repaid early, but the crisis established two precedents: that the United States and IMF would bail out a country whose failure threatened the system ("too big to fail" for sovereigns), and that pegged exchange rates with open capital accounts were a trap. It also produced the "tequila effect," contagion to Argentina and Brazil, which introduced the vocabulary of every crisis since.

The Asian financial crisis (1997–98)

The "Asian tigers" had pegged their currencies to the dollar, liberalised their capital accounts under IMF and US Treasury encouragement, and borrowed heavily in dollars short-term to fund long-term local investment — the same currency and maturity mismatch as Herstatt and the S&Ls. When the dollar rose against the yen in 1995–97, their exports lost competitiveness. On 2 July 1997 Thailand, having exhausted its reserves defending the baht, let it float; it fell 50%. Contagion hit Indonesia (rupiah down 80%, Suharto's thirty-one-year rule ended in May 1998), South Korea (the world's eleventh-largest economy needed a $58 billion IMF package in December 1997, the largest ever), Malaysia, the Philippines and, in stock markets, Hong Kong, which spent HK$118 billion buying shares in August 1998 to defend its currency board against speculators and won.

The IMF's programmes — high interest rates, fiscal austerity, bank closures — were widely blamed for deepening the recessions, and the crisis discredited the Washington Consensus in Asia. Malaysia's Mahathir imposed capital controls in September 1998, was denounced, and recovered as fast as anyone. The lasting consequence was that Asian governments resolved never again to depend on the IMF: they built enormous foreign-exchange reserves as self-insurance (from about $250 billion for emerging Asia in 1997 to over $5 trillion by 2014), ran current-account surpluses, and lent the proceeds to the United States. The "global savings glut" that Ben Bernanke identified in 2005 as the source of low US interest rates and the housing bubble was born in the Asian crisis; Chiang Mai Initiative swap lines (2000) and the later push for yuan settlement were the same instinct.

Russia's default and LTCM (1998)

Russia, running deficits it financed with high-yield short-term GKO bonds, was hit by the fall in oil prices to $10 after the Asian crisis. On 17 August 1998 it devalued the ruble, defaulted on its domestic debt and declared a moratorium on foreign bank debt — the first default by a nuclear power and a G8 member. The shock reached Wall Street through Long-Term Capital Management, a hedge fund run by Nobel laureates Myron Scholes and Robert Merton, which had $4.7 billion of capital, $125 billion of assets and over $1 trillion of derivatives notional, on bets that spreads would converge. When every spread widened at once, LTCM lost $4.6 billion in four months. On 23 September 1998 the Federal Reserve Bank of New York summoned fourteen banks and organised a $3.6 billion private rescue, and the Fed cut rates three times. The lesson was that a single leveraged fund could threaten the global system and that the Fed would act to prevent it — the Greenspan put extended to hedge funds.

Brazil and Argentina (1999–2002)

Brazil, whose Real Plan of 1994 had ended decades of hyperinflation with a crawling peg, was the next domino: in January 1999 it floated and the real fell 40%; it recovered quickly under inflation targeting and became a model for emerging-market central banking. Argentina did not. Its Convertibility Law, enacted on 27 March 1991, made the austral convertible from 1 April at 10,000 australes per US dollar for central-bank dollar sales. The peso replaced the austral only on 1 January 1992 at one peso per 10,000 australes, continuing the one-peso-per-dollar parity. Argentina, original Law 23,928, Articles 1–2 and 12; BCRA, “Billetes y monedas | Emisiones anteriores,” Peso section. Inflation fell during the regime; a four-year recession began in 1998 amid an overvalued peso, a rigid labour market and provincial deficits. In December 2001, after the government froze bank deposits (the "corralito"), riots killed 39 people, five presidents served in two weeks, and Argentina defaulted on about $100 billion of debt — then the largest sovereign default in history — and abandoned the peg in January 2002, with the peso losing 70%. Depositors' dollar savings were forcibly converted to pesos. Argentina has since defaulted twice more (2014, 2020), and the memory of "pesification" is why Argentines hold more US dollar cash per head than anyone outside the United States and why President Milei, elected in 2023, campaigned on dollarisation.

The 1990s crises produced a rough consensus, the "impossible trinity" or trilemma: a country can have any two of a fixed exchange rate, free capital movement and independent monetary policy, but not all three. Under the gold standard countries had chosen fixed rates and free capital and given up monetary independence; after 1971 the large economies chose independence and free capital and gave up fixed rates; the crises of the 1990s were what happened to countries that tried to keep all three.

Deregulation and the dot-com bubble (1995–2002)

In the United States, the Gramm–Leach–Bliley Act of November 1999 repealed the Glass–Steagall separation of commercial and investment banking (Citicorp and Travelers had already merged in 1998 in anticipation), and the Commodity Futures Modernization Act of December 2000 exempted over-the-counter derivatives, including credit default swaps, from regulation. Both were justified by the "Great Moderation" — the belief, backed by fifteen years of low inflation and mild recessions, that independent central banks had tamed the cycle. Alan Greenspan's Fed, credited with the 1990s boom, was celebrated as "the maestro."

The internet, commercialised from 1993–95, created a stock-market bubble on the scale of 1929: the Nasdaq rose from 1,000 in 1995 to a peak of 5,048 on 10 March 2000 and fell 78% by October 2002; $5 trillion of market value vanished; the Fed cut rates from 6.5% to 1% by June 2003, the lowest since 1958, and held them there for a year, which fed the housing boom. Enron (2001) and WorldCom (2002) collapsed in accounting frauds and produced the Sarbanes–Oxley Act of 2002. The bubble also seeded the payment innovations of the next decade: Amazon (1994), eBay (1995), PayPal (1998), online brokerage, and the infrastructure that would carry mobile money and crypto (“Innovation”).

Gold's low point

Gold prices were low in 1999, while several central banks planned or conducted sales. HM Treasury sold approximately 395 tonnes of UK gold reserves at 17 auctions from July 1999 to March 2002, at a weighted average auction price of $274.92 per ounce. The Bank of England's account says UK government holdings were around 715 tonnes when the programme was announced; “half its reserves” is only a rough description of that initial stock, not a separately measured share of all UK reserves. Sale proceeds were reinvested in interest-bearing dollar, euro and yen assets that remained in the reserves. HM Treasury, *The sale of part of the UK gold reserves 1999–2002*, Details; HM Treasury, *Gold reserves*, FOI response, 4 March 2019; Bank of England, *An analysis of the UK gold auctions 1999–2002*, introduction and n.3. The Central Bank Gold Agreement of September 1999 limited and coordinated specified sales. These observations do not establish a “lowest real price” without a named deflator and series, or that gold's monetary role had finished or would rise in a direct cycle to 2026.

Key takeaways

Trade liberalisation and freer capital movement developed alongside several 1990s currency and credit crises. NAFTA was a trade agreement, unlike the EU single market; pegged currencies with short-term foreign borrowing faced distinct vulnerabilities in Mexico, Asia, Russia, Brazil and Argentina. IMF programmes, national responses and subsequent reserve accumulation differed, and their size and effects need matched sources. A sequence of crises and rescues does not itself show that reserve purchases funded a later US housing bubble or that each rescue taught the same lesson. Central-bank gold sales coincided with gold's 1999 low, but one price point does not establish the system's confidence or a direct cycle to 2026.