Mexico's 1994 Peso Crisis and the Tequila Effect

Key Takeaways

• The devaluation came in two doses because the first one failed. The 20 December band-widening of about 15 percent lost the government more than $4 billion in reserves the very next day, and the peso was floated freely on 22 December, less than 48 hours later.
• Tesobono debt, government notes payable in pesos but indexed to the dollar, grew from $3.1 billion outstanding at the end of March 1994 to $29.2 billion by December, nearly all of it maturing within 1995. A currency crisis and a government financing crisis arrived on the same debt.
• Reserves had been falling for eight months before the devaluation, not eight days: $24.4 billion at the end of March 1994, $17.3 billion by the end of April after the Colosio assassination alone cost $7.1 billion in a single month, more than $16 billion at the end of July, and $10.0 billion by 9 December.
• The rescue package totaled $48.8 billion, announced 31 January 1995: $20 billion from the U.S. Treasury, $17.8 billion from the IMF, in a program that was the largest the Fund had ever approved relative to a member's quota, $10 billion from the Bank for International Settlements, and $1 billion from Canada. A further $3 billion from private banks and $1 billion from Argentina and Brazil were pledged and never drawn.
• Mexico's real GDP contracted 5.9 percent in 1995 and inflation reached 35.0 percent on an annual-average basis, while the current account deficit that had reached 5.4 percent of GDP in 1994 collapsed to 0.4 percent in 1995, computed by Swoopr Investment from World Bank data, as the peso's depreciation and the recession together crushed import demand.
• Mexico repaid every dollar of the U.S. loan more than three years ahead of schedule, with the final $3.5 billion transferred on 15 January 1997. The White House stated that interest payments left the U.S. Treasury with a net gain of nearly $580 million.
• The shock did not stay in Mexico. Argentina's real GDP growth swung from 5.8 percent in 1994 to a 2.8 percent contraction in 1995, and its modeled unemployment rate rose from 11.8 percent to 18.8 percent over the same period, even though Argentina's currency board worked by a different mechanism than Mexico's exchange-rate band.

What Happened During Mexico's Peso Crisis of 1994 and 1995?

Mexico entered 1994 running a crawling exchange-rate band that had held, with periodic widenings, since 1991. The peso traded near the top of that band all year, a sign that the market wanted more pesos sold than the government wanted to sell, and the government financed the gap by selling reserves and, increasingly, by issuing tesobonos rather than raising interest rates enough to choke off the outflow on its own. Two political shocks, a guerrilla uprising in January and an assassination in March, tested that arrangement without breaking it. A poorly designed devaluation in December did.

The proximate failure was specific and dated. On 20 December, newly inaugurated president Ernesto Zedillo's finance secretary, Jaime Serra Puche, announced that the peso's trading ceiling would move about 15 percent weaker, with no accompanying fiscal tightening and no interest-rate move to make holding pesos more attractive at the new rate. Reserves fell more than $4 billion the following day as the market tested the new ceiling and found it undefended. On 22 December, with reserves nearly exhausted, the government abandoned the band altogether and let the peso float. Serra Puche resigned a week later, replaced by Guillermo Ortiz Martínez, and international support, first proposed as a congressional loan-guarantee package that stalled, then delivered through the U.S. Treasury's own executive authority over the Exchange Stabilization Fund, arrived through late January and February 1995.

Chronology of the crisis and its resolution

Dated events against the peso's exchange rate from the Federal Reserve's daily series, where a rate is available for that date.

Read that table for the gap between the political-shock rows and the devaluation row. Colosio's assassination in March cost Mexico $7.1 billion of reserves in a month and did not force a devaluation. The 20 December announcement, made after eight months of further reserve decline with no accompanying program, did. The trigger and the vulnerability are different questions, and the table is deliberately built to keep them visually separate.

Why Was Mexico's Exchange-Rate System Fragile Before December 1994?

Mexico ran a managed exchange-rate band, not a hard peg, from November 1991: the peso could trade anywhere within a range whose ceiling depreciated by a small preannounced amount each day. Through 1993 and into 1994 the peso traded persistently near that ceiling, which is a market telling a government it wants to sell more of the currency than the government wants to buy. Meeting that gap without moving the rate meant selling reserves, and the government did so for most of 1994 rather than let the ceiling move or raise interest rates enough to make holding pesos attractive on its own terms.

The current account deficit that the capital inflows were financing was not a new problem in 1994; it had been widening for several years. Computed by Swoopr Investment from World Bank data, Mexico's current account deficit stood at 4.4 percent of GDP in 1993 and widened to 5.4 percent in 1994, or roughly $30 billion at that year's exchange rate and nominal GDP. A deficit of that size is sustainable only for as long as foreign capital keeps arriving to fund it, on terms the country can roll over. Much of the capital that had financed Mexico's deficit through the early 1990s had come in as relatively liquid portfolio investment attracted by an economy that was liberalizing and about to join NAFTA, not as long-term direct investment that does not reverse itself on a bad news day.

What changed through 1994 was the composition of the financing, not the deficit itself. As confidence in the peso weakened after the political shocks described below, foreign investors became reluctant to hold peso-denominated government paper, called Cetes, without being compensated for currency risk. Rather than raise Cetes rates enough to hold that money, or defend the band with a larger reserve commitment and a credible fiscal program, the government increasingly issued tesobonos instead, a choice that solved the immediate rollover problem while building a much larger one for December.

What Was a Tesobono, and Why Did It Turn a Currency Problem Into a Debt Problem?

A tesobono, formally a Certificado de la Tesorería de la Federación denominated in dollars, was a short-term Mexican federal government note. It paid in pesos, but its principal value was indexed to the peso-dollar exchange rate, so a dollar-based buyer took on essentially none of Mexico's currency risk. If the peso fell, the tesobono's peso payout rose by exactly enough to keep its dollar value constant. That structure made tesobonos an easy sell to nervous investors throughout 1994 without the government having to raise peso interest rates, which would have signaled distress and slowed the domestic economy in an election year.

The scale of the shift is the single most important balance-sheet fact in this case study. Tesobonos outstanding grew from $3.1 billion at the end of March 1994 to $29.2 billion by December, an increase of more than nine times in nine months, and virtually all of it fell due within 1995. Every peso of that growth was currency risk the government had moved off investors' books and onto its own, at the exact moment its capacity to absorb that risk, measured in reserves, was falling.

The mechanism explains why the devaluation could not simply "fix" Mexico's competitiveness problem the way a currency adjustment sometimes does. When a government owes debt in its own currency, inflating or devaluing that currency reduces the real burden of the debt. A tesobono did the opposite: because its peso value rose in lockstep with the devaluation, the December slide made every tesobono more expensive to redeem in real terms, precisely when the government's reserves to redeem them with were most depleted. The debt did not get cheaper because the currency fell. It got more expensive, in the currency the government could actually create.

How Did the Zapatista Uprising and the Colosio Assassination Change the Political Backdrop in 1994?

Two shocks in 1994 tested Mexico's political and financial stability before the December devaluation, and neither one directly caused it, though both left a mark on reserves and on investor confidence.

On 1 January 1994, the day NAFTA took effect, the Zapatista National Liberation Army occupied several towns in the highlands of Chiapas, including San Cristóbal de las Casas, Ocosingo, Las Margaritas and Altamirano, in explicit opposition to the trade agreement's effect on the rural poor. The government and the Zapatistas agreed to a ceasefire on 12 January, and the uprising did not escalate into a broader national conflict, but it introduced a political risk premium into a year that had begun with Mexico presenting itself internationally as a newly liberalized, NAFTA-ready economy.

The larger shock came on 23 March 1994, when Luis Donaldo Colosio, the ruling PRI party's presidential candidate and heavy favorite to win the August election, was assassinated at a campaign event in Tijuana. Reserves fell from $24.4 billion at the end of March to $17.3 billion by the end of April, a decline of $7.1 billion in a single month, as investors reassessed political risk in the world's most direct terms. In the aftermath, the Clinton administration set up a $6 billion currency-swap line with Mexico, split evenly between the Federal Reserve and the Treasury under the North American Framework Agreement, to help it brace for a possible run on the peso. That line was still in place, separate from the $48.8 billion package, when the December crisis hit. Ernesto Zedillo, Colosio's replacement as the PRI candidate, won the presidential election on 21 August 1994 and was inaugurated on 1 December, nineteen days before the devaluation his own government would announce.

The throughline across both shocks is that reserves fell and did not fully recover between them. The Colosio assassination cost $7.1 billion in a month; the pre-devaluation decline from July to December cost roughly $6 billion more, spread across five months in a slower bleed the market noticed less because it lacked a single dramatic headline. By the time Serra Puche announced the band widening on 20 December, reserves had already absorbed most of a year's worth of political shocks and capital flight before the trigger event was even a rumor.

What Happened on December 20 to 22, 1994?

The three days from 20 to 22 December are the hinge of this case study, because they show how a policy choice, not just an underlying imbalance, decided the shape of the crisis.

20 December: the band widens, alone. Finance secretary Jaime Serra Puche announced that the peso's trading ceiling would move roughly 15 percent weaker against the dollar. Devaluations of a fixed or managed rate are not unusual and are not automatically destabilizing when they arrive with a credible program: a fiscal tightening, an interest-rate increase, or both, signaling that the new rate will actually hold. Mexico's announcement carried neither. The rate itself moved from 3.4662 to 3.9500 pesos per dollar that day.

21 December: the market tests the new ceiling and finds it undefended. Investors and Mexican residents alike, having just watched the government devalue once without a program, had every reason to ask whether it would do so again, and to sell pesos before it did rather than after. The government tried to defend the new, wider band and lost more than $4 billion in reserves in a single day doing it, a scale of loss that made clear the new ceiling would not hold either.

22 December: the float. With reserves nearly exhausted and the band strategy having failed within 48 hours of its announcement, the government let the peso float freely rather than continue defending any fixed rate. The exchange rate closed at 4.85 pesos per dollar that day, up 40 percent from its level three days earlier.

Serra Puche resigned on 29 December, six days after the float, and was replaced by Guillermo Ortiz Martínez. The sequencing matters for how the episode should be read: the underlying imbalance, current account deficit, reserve decline, tesobono buildup, had been building since at least early 1994, but the specific shape of the crisis, a controlled decline that became an uncontrolled one within three days, was a direct consequence of trying a partial, unaccompanied devaluation first. A larger, credible move on 20 December, or no move at all until a program was ready, might have produced a different December. The one Mexico chose invited exactly the test it failed.

How Far and How Fast Did the Peso Fall?

The Federal Reserve's daily exchange-rate series lets the decline be measured precisely rather than described impressionistically. Every figure below is computed by Swoopr Investment from that series.

Two shapes are worth noticing in that table beyond the headline number. First, the collapse was front-loaded: nearly all of the eventual damage was done in the eleven days between 19 and 30 December, when the peso lost 30.7 percent of its dollar value; the further slide to the November 1995 low added another 26 percentage points across the following ten months, a slower bleed rather than a second crash. Second, there was no meaningful recovery within the window this page covers. The peso closed 1995 at 7.74, essentially at its post-crisis level rather than partway back toward its pre-devaluation range, which is the opposite of how equity indices in most of the crashes in this library behaved within a year.

The mechanical arithmetic behind the two right-hand columns matters for reading any currency-crisis chart correctly. A rate that rises from 3.4662 to 8.05 pesos per dollar is a 132 percent increase in the price of a dollar, but it is not a 132 percent decline in the peso's value; the peso's dollar value is the reciprocal of the rate, and it fell by 56.9 percent, not 132 percent. Reports and charts that quote the larger number as "the size of the devaluation" are describing the exchange rate's movement, not the currency's loss of value, and the two are easy to conflate.

What Was in the $48.8 Billion International Support Package?

A larger, congressionally approved loan-guarantee package had been proposed in mid-January 1995 and stalled amid opposition in Congress. Rather than continue seeking a vote, President Clinton used his own executive authority over the Treasury's Exchange Stabilization Fund on 31 January 1995 to commit up to $20 billion directly, structured as short-term currency swaps of up to 90 days, medium-term swaps of up to five years, and securities guarantees of up to ten years. The Department of Justice's own contemporaneous legal opinion on the authority describes the assistance as designed around "an assured source of repayment," language that reflects how carefully the loans were structured to be repaid rather than granted.

The IMF's own component is worth reading on its own terms, because its size relative to Mexico's IMF quota, the amount a member country is entitled to borrow under normal rules, was unprecedented. The Fund approved SDR 12,070.2 million, about $17.8 billion, equal to 688.4 percent of Mexico's quota, the largest financing package the IMF had approved relative to a member's quota up to that point. SDR 5,259 million, about $7.8 billion, or 300 percent of quota, was available immediately rather than phased in over the program's life, a design meant to answer the acute liquidity problem directly rather than dole out support in increments a fast-moving run could outrun.

The headline figure most often quoted for this rescue, close to $50 billion, is the sum of everything announced rather than everything actually put to use. The U.S. General Accounting Office lists $20 billion from the United States, $17.8 billion from the IMF, $10 billion from the BIS and $1 billion from Canada as the four components of the $48.8 billion total, but its own account of what was activated is narrower than that headline figure suggests: the BIS facility, alongside $3 billion pledged by private banks and $1 billion pledged by Argentina and Brazil, was never drawn upon. Of the U.S. Treasury's own $20 billion commitment, the amount actually disbursed to Mexico reached $13.5 billion. The distinction is not pedantic: a headline package size built from every facility that was announced, rather than the money that actually changed hands, overstates how much external support Mexico drew on to stop the run.

What Did Mexico Have to Do in Return for the Loan?

The support came with three distinct layers of obligation, not one. The IMF's 18-month stand-by arrangement carried the Fund's standard conditionality: a program of fiscal and monetary targets Mexico had to meet to keep drawing on the credit, reviewed periodically by IMF staff, the mechanism through which the Fund's own history describes the loan as supporting "the Government's 1995-96 economic and financial program" rather than an unconditional transfer.

The U.S. Treasury's own loans, running separately from the IMF program, were structured, in the Department of Justice's contemporaneous language, around "an assured source of repayment," a condition the Treasury required before committing taxpayer-backed funds through the Exchange Stabilization Fund. This page does not describe the specific mechanics of that repayment assurance beyond the Justice Department's own description, because a primary source detailing the exact routing was not confirmed during this session's research, and an unverified mechanism is worse than an acknowledged gap.

The third layer was procedural. Because a congressionally approved package had stalled, the Treasury used executive authority that did not require a direct vote, which meant the assistance came with ongoing oversight obligations to Congress instead. The GAO's own 1996 report exists because of that oversight requirement, and its central finding, that neither the Federal Reserve nor the Treasury foresaw the crisis's magnitude despite documented concern over Mexico's exchange-rate policy, is a product of that accountability structure working after the fact.

How Did Mexico's Banking System Nearly Fail, and What Was PROCAPTE?

The devaluation and the interest-rate response it forced hit Mexican banks from two directions at once: borrowers who owed dollar-linked debt or who simply could not service loans at much higher peso rates began defaulting, while the banks' own funding costs rose. The government's response, the Programa de Capitalización Temporal, or PROCAPTE, was announced on 24 February 1995 and became operational on 31 March.

PROCAPTE let banks issue subordinated debentures to Mexico's deposit insurance fund, mandatorily convertible into equity after five years if the bank had not otherwise recapitalized itself, in exchange for capital that lifted their regulatory ratios immediately. On 31 March 1995 six banks were recapitalized under the program, including Serfin, Inverlat and Bital, then the third, fourth and fifth largest banks in Mexico, for a combined $950 million. The program itself was funded principally through a $3.25 billion loan from the Inter-American Development Bank and the World Bank, of which $2.25 billion went directly into the banking system, and it required participating banks to reach a capital ratio of at least 9 percent of net capital to risk-weighted assets. Across the six banks, the average capital ratio rose from 5.8 percent to 9.6 percent as a direct result.

PROCAPTE was a bridge, not the full resolution. It addressed capital ratios at a moment banks needed a public backstop to keep operating, but it was one program among a larger and longer sequence of Mexican bank-support measures through the rest of the 1990s that this page does not attempt to size, because a verified total cost figure for that broader effort was not confirmed during this session's research. What PROCAPTE demonstrates on its own, cleanly, is the same lesson the currency side of the crisis teaches: a solvency problem that starts on one side of a balance sheet, here, a currency mismatch and a rate shock hitting borrowers, becomes a capital problem on the other side, here, the banks that lent to them, on a timeline measured in weeks rather than years.

How Did the Crisis Spread to Argentina and Other Emerging Markets?

The pattern investors and journalists later named the "Tequila effect" was capital leaving emerging markets broadly, not just Mexico, once Mexico had demonstrated that a seemingly stable currency arrangement could fail within three days. Argentina is the clearest documented case. Its currency board, a different and in principle more rigid mechanism than Mexico's managed band, fixed the peso to the dollar by law and was not itself devalued during this period, yet the Argentine economy absorbed a severe shock anyway as depositors and investors, unable to easily distinguish which emerging market might be the next Mexico, pulled money out broadly.

Computed by Swoopr Investment from World Bank data, Argentina's real GDP growth swung from 5.8 percent in 1994 to a 2.8 percent contraction in 1995, and its unemployment rate, by the World Bank's modeled estimate, rose from 11.8 percent in 1994 to 18.8 percent in 1995. That swing happened without an Argentine currency devaluation, which is the clearest evidence that the transmission mechanism was investor behavior rather than a shared underlying fundamental: capital that had been indiscriminately available to emerging markets before December 1994 became indiscriminately unavailable for a period afterward, only later differentiating again between borrowers on the basis of their own individual conditions.

The episode is often described as a dress rehearsal for the much larger Asian financial crisis less than three years later, and the comparison is informative on scale. The Mexican package totaled $48.8 billion; when Thailand, Indonesia and South Korea needed support beginning in July 1997, the international community mobilized $118 billion, more than double, for a crisis that shared Mexico's core mechanism, short-term foreign-currency exposure whose risk had been masked by an exchange-rate arrangement investors trusted more than the fundamentals underneath it warranted. Details of that later, larger episode, including how differently Thailand's and Korea's pegs broke compared with Hong Kong's, which held, are covered in the Asian financial crisis.

What Happened to Mexico's Economy and Inflation in 1995?

The real economy absorbed the currency and banking shock with a lag of roughly a quarter, then contracted sharply for the rest of the year. World Bank data puts Mexico's real GDP growth at 4.4 percent in 1994, followed by a 5.9 percent contraction in 1995, one of the sharpest single-year reversals in the modern data for any economy this library covers. Annual average consumer price inflation rose from 7.0 percent in 1994 to 35.0 percent in 1995, as the devaluation passed through directly into import prices and domestic costs.

The current account column tells the adjustment story the other two cannot. A deficit of 5.4 percent of GDP in 1994 nearly vanished by 1995, not because Mexico became a more attractive place to lend to, but because a currency that had lost roughly a third of its value and an economy in recession together crushed the country's ability to buy imports, which is the textbook mechanism by which a floating currency eventually rebalances a current account, at the cost of a severe recession along the way rather than a gradual adjustment. Measured in dollars rather than pesos, Mexico's GDP fell from $553.6 billion in 1994 to $380.2 billion in 1995, a decline of 31.3 percent, computed by Swoopr Investment from World Bank data. Almost none of that dollar-denominated collapse reflects a smaller Mexican economy in real, peso terms; it reflects the same output being worth roughly a third less in dollars, a distinction worth holding onto whenever a crisis-era GDP figure is quoted in dollars without the real, local-currency figure alongside it.

The recovery that followed was faster than the contraction. Real GDP growth returned to 6.2 percent in 1996 and 7.2 percent in 1997, helped by a currency that had made Mexican exports substantially cheaper and by proximity to a growing U.S. economy under a newly implemented NAFTA. Inflation took longer to normalize, still running at 34.4 percent in 1996 before falling to 20.6 percent in 1997, a reminder that a currency crisis's price-level damage typically outlasts its growth damage by a year or more.

Why Did U.S. Financial Markets Barely React?

Unlike some of the episodes in this library where the same week that broke an emerging-market currency also moved U.S. equity indices sharply, Mexico's devaluation and float did not produce a comparable shock in U.S. markets. Several structural reasons explain the gap rather than any one dramatic data point.

First, the direct financial exposure sat in a narrower place than a broad equity index: tesobono holders, disproportionately U.S. institutional investors, and banks with direct Mexican lending exposure bore the immediate mark-to-market losses, rather than the corporate earnings that set the level of the S&P 500. Second, the U.S. macro backdrop going into the crisis was one of a Federal Reserve already well into a tightening cycle for reasons unrelated to Mexico: the federal funds effective rate rose from about 2.85 percent in January 1994 to about 5.59 percent by mid-November, a sequence of increases through the year that had already repriced U.S. rate expectations before the peso ever moved. Third, the U.S. response itself, an assurance of an "assured source of repayment" through Treasury-structured loans rather than a grant, was designed and communicated in a way that limited it from reading, to markets, as an open-ended U.S. fiscal commitment.

The broader lesson generalizes: a sovereign-financing crisis in an emerging market, even one newly integrated with the United States through NAFTA, does not automatically transmit into a domestic equity shock the way a crisis inside the U.S. financial system does. The channel matters more than the headline size of the loss, a point macro and market regimes covers in more depth.

When Did Mexico Repay the U.S. Loan, and Did the Rescue Turn a Profit?

The formal emergency support agreement was signed on 21 February 1995, and Mexico began repaying ahead of its own schedule almost immediately once its access to international capital markets recovered. The White House's own fact sheet on the repayment, issued 15 January 1997, lays out the schedule: Mexico repaid $3 billion by January 1996, prepaid $7 billion of the $10.5 billion still outstanding in August 1996, nearly four years ahead of that portion's original schedule, and transferred the final $3.5 billion on 15 January 1997, completing the full repayment more than three years earlier than originally required.

The financial outcome for the United States was positive, not merely break-even. The same White House fact sheet states that Mexican interest payments on the loans resulted in a net gain of nearly $580 million for the American taxpayer, meaning the ESF loans, structured at market-related rates rather than concessional ones, paid the U.S. Treasury more than an equivalent Treasury investment would have earned over the same period. Combined with the fact that $3 billion in pledged private-bank support and $1 billion pledged by Argentina and Brazil were never even drawn, the actual fiscal exposure the United States carried through this episode was smaller, shorter, and more profitable than the $50 billion headline figure suggests on its own.

That outcome is worth separating clearly from the question of whether the intervention was good policy. A loan can be repaid in full, with profit, and still represent a consequential use of executive authority that bypassed a stalled congressional vote, and it can still have left Mexican households absorbing a 5.9 percent GDP contraction and 35 percent inflation regardless of how the U.S. side of the ledger closed. The repayment figures answer a narrow question, whether the United States government made money on the loans, cleanly and in the affirmative. They do not answer the much larger question of what the crisis cost Mexico.

What Warning Signs Were Visible Before December 1994, and What Wasn't?

This distinction matters because a case study that only says "the warning signs were there" without specifying which ones, and which weren't, teaches false confidence rather than a usable skill.

Observable at the time

• Reserves published monthly and visibly declining. $24.4 billion at the end of March 1994, $17.3 billion by the end of April, more than $16 billion at the end of July, $10.0 billion by 9 December. None of this was hidden data; it was the kind of series any analyst covering Mexican sovereign risk could and did track.
• A current account deficit near 5.4 percent of GDP in 1994, a level widely recognized at the time as large for an emerging economy, financed increasingly by short-term rather than direct investment.
• Tesobono issuance rising nearly tenfold in nine months. The shift from peso-denominated Cetes toward dollar-indexed tesobonos was itself a visible signal that the government could not sell peso risk to the market at rates it was willing to pay.
• A peso trading persistently at the top of its band through 1993 and 1994, the market's own running vote that the currency was overvalued at the official rate.

Clear only in hindsight

• That the December 20 band-widening would fail within a single trading day. Partial devaluations of managed currencies had, in other countries and other years, sometimes held. This one did not, and the specific reason, no accompanying fiscal or monetary program, was legible only once the market's reaction to its absence played out.
• That the government would have no program ready. A devaluation paired with credible tightening is a materially different event than one paired with nothing, and outside observers had no way to know in advance which version Mexico would attempt.
• That the shock would transmit to Argentina without an Argentine devaluation. The "Tequila effect" as an indiscriminate emerging-market capital flight was a pattern that became legible only as it happened, not one the reserve and current-account data available in November 1994 predicted for countries other than Mexico itself.

The U.S. General Accounting Office's own 1996 investigation is useful precisely because it was designed to answer this question institutionally rather than rhetorically: it found that the Federal Reserve and the Treasury had expressed documented concern over Mexico's exchange-rate policy before December, and separately found that neither foresaw the crisis's eventual magnitude. Awareness of a vulnerability and anticipation of its timing and scale are not the same finding, and the report's own structure keeps them apart.

Why Is the Mexican Peso Crisis a Poor Template for the Next Currency Crisis?

Three features of this episode were specific enough that expecting them to recur in the same shape is likely to be the wrong preparation for the next one.

The instrument was unusual, and the next crisis is unlikely to use it. A sovereign government issuing its own short-term notes indexed to a foreign currency, rather than borrowing directly in that currency or borrowing from foreign banks, is a specific financing choice. The Asian financial crisis less than three years later ran through a different channel entirely, private corporate and bank borrowing in dollars against domestic-currency revenue, and Argentina's own later default in 2001, covered in Argentina's 2001 default, unwound a currency board rather than a managed band. The common thread across all three, a currency mismatch that a fixed or managed exchange rate had made to look smaller than it was, generalizes; the specific tesobono mechanism does not.

The scale and speed of the U.S. response were unusual, and largely unrepeated since. A sitting U.S. president committing $20 billion through his own executive authority, within six weeks of the trigger event, to a neighboring country bound to the United States by a trade agreement four days into effect, reflects a specific set of political circumstances, NAFTA's fresh implementation and Mexico's direct border relationship with the United States, that most emerging-market crises since have not had available to them. The scale of support for later crises rose rather than fell, $118 billion for the Asian crisis alone, but the mechanism shifted almost entirely toward IMF-led multilateral programs rather than a unilateral U.S. Treasury commitment of this size and speed.

Mexico's own earlier 1982 default shows the same country can produce two structurally different crises. On 12 August 1982 Mexico told the Federal Reserve, the Treasury and the IMF it could not service roughly $80 billion in debt, owed mostly to syndicated international bank loans, an episode covered in full in Swoopr's case study on Mexico's 1982 debt crisis, with the broader regional chain reaction it set off, and the global interest-rate shock that triggered it, covered separately in the Volcker disinflation. Numerous Latin American countries eventually rescheduled similar debt over a period of years. By 1994 that bank-loan channel had been replaced by tradable government securities that investors could sell in an afternoon, which is a large part of why the 1994 crisis moved in days rather than years. A country's financing structure, not just its underlying fundamentals, decides how fast a crisis it is capable of having.

Common Myths About the Mexican Peso Crisis

"NAFTA caused the crisis." The Zapatista uprising began the same day NAFTA took effect, and the coincidence has fused the two in popular memory ever since. The vulnerabilities that produced the December devaluation, the current account deficit, the reserve decline, the tesobono buildup, predate NAFTA's implementation and would have existed in some form without it. NAFTA is a genuine part of the story of Mexico's post-crisis recovery, through cheaper exports and U.S. demand, but it is not a credible cause of the crisis itself.

"The devaluation caused the crisis." The devaluation was the trigger and, in its poorly designed first form, the immediate mechanism of the acute panic, but the underlying imbalance had been visible in reserve and current-account data for at least a year beforehand. A currency adjustment was arguably necessary given the size of the deficit; what turned it into a crisis was the absence of a credible program alongside the first, partial move on 20 December.

"It was purely a bailout of Wall Street." Tesobono holders were disproportionately U.S. institutional investors, and the rescue did protect their principal from a default that a Mexican inability to roll over $29.2 billion in maturing notes would otherwise have produced. But the structure was a set of loans requiring "an assured source of repayment," not a grant, and Mexico repaid every dollar more than three years early with the U.S. Treasury recording a net gain of nearly $580 million. Calling that a bailout in the sense of an unrecovered public loss to protect private creditors does not match how the numbers actually closed.

"Fifty billion dollars of public money went to Mexico." The headline figure combines four formally announced components, $20 billion from the U.S., $17.8 billion from the IMF, $10 billion from the BIS and $1 billion from Canada, at $48.8 billion, with $3 billion from private banks and $1 billion from Argentina and Brazil that were also pledged but never activated. The GAO's own account goes a step further: the BIS facility itself was never drawn upon either, so the money Mexico actually used came from a narrower set of sources than the headline implies. Of the portion that was drawn, it was structured and repaid as loans, with interest, not disbursed as aid.

"Nobody could have seen it coming." Reserves were published monthly and had been declining for eight months before the devaluation; the current account deficit and the shift toward tesobono financing were both visible in public data through 1994. What was not foreseeable was the specific timing, the failure of the first, partial devaluation within a single trading day, and the scale of the contagion to Argentina. The U.S. General Accounting Office's own review draws exactly this line, finding documented awareness of the underlying vulnerability alongside a genuine failure to anticipate the crisis's eventual magnitude.

What a Reader Can Actually Carry Forward

The temptation is to reduce this case to "watch for falling foreign-exchange reserves," which is true and also too generic to be useful on its own. The more specific lessons sit in how the tesobono mechanism worked and in how the December devaluation was executed.

What generalizes

• Debt indexed to a currency the borrower cannot print carries the borrower's currency risk, whatever the debt is called. A tesobono looked like a peso instrument on its face; economically, it functioned as dollar debt, and it behaved like dollar debt does everywhere else, becoming more expensive in real terms exactly when the currency it was denominated against fell. The same logic applies to a household with a foreign-currency mortgage or a company with unhedged dollar-denominated bonds against local-currency revenue.
• A short-term debt stock maturing all at once is a specific, measurable vulnerability distinct from the total debt level. $29.2 billion of tesobonos, nearly all maturing within a single calendar year, mattered more than the total figure alone would suggest, because it converted a solvency question into a rollover question with a hard deadline.
• A partial, unaccompanied policy response can invite a bigger test than either a full response or no response at all. The 20 December band-widening, without fiscal or monetary measures attached, told the market the government recognized a problem it was not yet prepared to solve, and the market responded by testing whether the government could defend even the smaller move it had just made.
• Reserve and short-term external debt data are available well before a crisis and are worth tracking for exactly this reason. None of the vulnerability in this case study was secret; it was published monthly and required an analyst willing to read it, not privileged information.

What does not generalize

• The specific tesobono instrument. Later crises transmitted currency risk through different channels, private corporate dollar debt in Asia, a currency board in Argentina, and a reader looking for the next tesobono by name will likely miss the next crisis's actual mechanism.
• The scale and speed of the U.S. rescue. A $20 billion unilateral U.S. Treasury commitment inside six weeks reflected NAFTA's fresh implementation and Mexico's direct relationship with the United States. Most emerging-market crises since have received support through slower, more multilateral IMF-led channels instead.
• The speed of Mexico's recovery. Growth returned to 6.2 percent in 1996 and 7.2 percent in 1997, aided by NAFTA-driven export growth and a strong U.S. economy next door. A currency crisis in an economy without a similarly positioned trading partner has no guarantee of a comparably fast rebound.

The one question worth asking now

Not "which country will be the next Mexico," which is a forecasting question almost nobody can answer reliably, but a narrower and fully answerable one: for any government or corporate borrower a reader is exposed to, how much of its debt is both short-term and effectively denominated in a currency it does not control, whether by direct issuance or by an indexation clause that achieves the same thing, and does its stock of readily available reserves or cash cover that debt if rollover access disappears for a quarter? That question requires no view on politics, timing, or which currency comes under pressure next. It only requires reading a balance sheet the way this case study reads Mexico's tesobono stock against its reserves.

References

Every figure on this page was verified against the following sources, each retrieved on 26 August 2026:

• U.S. General Accounting Office: Mexico's Financial Crisis, Origins, Awareness, Assistance, and Initial Efforts to Recover, GAO/GGD-96-56: the December 1994 chronology, reserve levels through the year, the Colosio-linked reserve decline, tesobono totals, the $48.8 billion package composition, and the $13.5 billion actually disbursed by the U.S. Treasury.
• Peterson Institute for International Economics: "The Mexican Peso Crisis of 1995 and Its Aftermath": which parts of the $48.8 billion package, the BIS facility, the private-bank pledge and the Argentina/Brazil pledge, were never activated, citing GAO 1996 at pages 109 to 110.
• International Monetary Fund: Press Release 95/10, IMF Approves US$17.8 Billion Stand-By Credit for Mexico: the $17.8 billion stand-by arrangement, its 688.4 percent of quota, and the $7.8 billion available immediately.
• The White House: Fact Sheet on Mexico Debt Repayment, January 15, 1997: the repayment schedule and dates, and the nearly $580 million net gain to the U.S. Treasury.
• Yale Program on Financial Stability, New Bagehot: PROCAPTE: the program's dates, the six recapitalized banks, the $950 million total, its IDB and World Bank funding, and the capital-ratio improvement.
• U.S. Department of Justice, Office of Legal Counsel: Use of the Exchange Stabilization Fund to Provide Loans and Credits to Mexico: the swap-authority structure and the "assured source of repayment" language.
• Federal Reserve Bank of St. Louis FRED: Mexican Pesos to U.S. Dollar Spot Exchange Rate, Series DEXMXUS: every exchange-rate figure and percentage change, computed by Swoopr Investment from the daily series.
• Federal Reserve Bank of St. Louis FRED: Federal Funds Effective Rate, Series DFF: the January and mid-November 1994 readings.
• World Bank World Development Indicators: GDP growth (annual %), Mexico: real GDP growth, 1993 through 1997.
• World Bank World Development Indicators: Inflation, consumer prices (annual %), Mexico: annual average inflation, 1993 through 1997.
• World Bank World Development Indicators: Current account balance (% of GDP), Mexico: the deficit for 1993 through 1996, including the dollar figure computed by Swoopr Investment.
• World Bank World Development Indicators: GDP (current US$), Mexico: the 1994 and 1995 dollar GDP figures and the decline computed by Swoopr Investment between them.
• World Bank World Development Indicators: GDP growth (annual %), Argentina: the 1994 growth and 1995 contraction.
• World Bank World Development Indicators: Unemployment, total (% of total labor force) (modeled ILO estimate), Argentina: the 1994 and 1995 unemployment figures.

Figures deliberately not stated. This page gives no specific figure for the peak level of Mexican short-term (Cetes or tesobono) interest rates during 1995, no percentage decline for the Mexican stock market during the crisis, no total cost figure for Mexico's broader banking-system rescue beyond the $950 million PROCAPTE recapitalization, no specific routing or dollar figure for the collateral arrangement behind the U.S. Treasury's loans beyond the Justice Department's own "assured source of repayment" language, and no unemployment or poverty figures for Mexico itself during 1995. No primary or institutional source verified during this session's research supplied a number for any of these that this page could stand behind. Where a mechanism is well documented but a specific magnitude is not, as with the interest-rate spike, the mechanism is described and no number is supplied.

Method note: figures described as computed were derived by Swoopr Investment from the series named above. Exchange-rate percentage changes use the Federal Reserve's daily noon buying rate and are not intraday. World Bank current account and GDP figures are annual, not point-in-time, so the $30 billion 1994 current account deficit figure is a full-year computation rather than a rate observed on any single date. Inflation figures are annual averages, which differ from a December-to-December figure and should not be substituted for one.

This is educational content about a historical episode. It is not investment advice, it is not a forecast, and nothing here should be read as a claim about the condition of any currency, government or financial institution today.

Related Reading

• Market History Case Studies: where a three-day currency collapse sits against the multi-year episodes elsewhere in this library.
• The Asian Financial Crisis: a larger, later episode sharing Mexico's core mechanism, at more than double the rescue-package scale.
• Argentina's 2001 Default: the same country hit by the 1995 Tequila effect, later unwinding a currency board rather than a managed band.
• Mexico's 1982 Debt Crisis: the same country's earlier default on roughly $80 billion of syndicated bank debt, a structurally different crisis than 1994's.
• The Volcker Disinflation: the global interest-rate shock that helped trigger Mexico's 1982 default.
• The 2008 Financial Crisis: for contrast, a crisis transmitted directly through the U.S. financial system rather than around it.
• Fixed Income and Bonds: the mechanics of a security indexed to something other than its own face currency, the structure behind a tesobono.
• Macro and Market Regimes: for reading which channel a given macro shock is actually moving through.
• International Investing: how currency risk shows up in a portfolio holding foreign assets.
• Risk Management: sizing exposure to a single-country risk before a rollover deadline arrives.

Frequently Asked Questions