Argentina 2001 and the End of Convertibility

Key Takeaways

• The exchange rate was legislation. The Convertibility Law of 1 April 1991 set one peso to one dollar and required the central bank to hold reserves at least equal to the monetary base, so ending the peg took an Act of Congress.
• Argentina attempted the only adjustment a hard peg permits and it ran the wrong way. Consumer prices fell 1.2, 0.9 and 1.1 percent in 1999, 2000 and 2001 on Fund figures, while output shrank in all three years.
• The run came before the freeze. Total deposits fell 20.3 percent in the last six months of 2001 on the Republic's own figures, and reserves fell from $25.2 billion at end-2000 to $14.6 billion at end-2001.
• The freeze was quantified. The Republic states that the December 2001 limits trapped roughly $60 billion of demand deposits; the Congressional Research Service records the cap as $1,000 per month.
• The exit rewrote private contracts. Dollar deposits converted at 1.40 while dollar loans owed to banks converted at 1.00, an asymmetry that blew a hole in bank balance sheets the state then filled with bonds.
• The dollar value of the country collapsed far faster than its output. Real output fell 10.9 percent in 2002; gross domestic product in dollars fell from $300.4 billion to $112.5 billion, about 63 percent.
• Gross public debt went from 48.0 percent of gross domestic product in 2001 to 147.2 percent in 2002, almost entirely because dollar obligations were revalued against a peso economy, not because Argentina issued anything new.
• The economy recovered before the debt did. Real output regained its 1998 level in 2005; the last defaulted bonds were not settled until 2016.

What Was Convertibility, and Why Was It a Law Rather Than a Policy?

Most fixed exchange rates are administrative choices that the central bank defending them can revise on a Sunday evening. Argentina built something harder to reverse. On 1 April 1991 Congress enacted the Convertibility Law, which guaranteed conversion of the peso into dollars at one to one and limited the printing of pesos to what the central bank could buy dollars to back.

The Republic of Argentina describes the machinery in its own filings: from 1991 through 2001 monetary policy "was governed by the Convertibility Law of 1991, which pegged the peso to the U.S. dollar at a one-to-one exchange rate and required the Central Bank to maintain international monetary reserves at least equal to the monetary base." Every peso in public hands had a dollar sitting behind it by statute.

That was the point. Argentina had come through hyperinflation in the late 1980s and the Menem government wanted a commitment the central bank could not undo. A statute can only be repealed by the legislature that passed it, in public, with a recorded vote.

It is also a wall with nothing behind it. Removing the ability to create pesos removed every conventional response to a downturn: no meaningful independent interest rate, no exchange rate to adjust, and no lender of last resort that could print, because printing was what the statute forbade. The Republic lists the consequence among the regime's shortcomings, noting that it "limited the use of monetary policy to stimulate the economy in response to downturns in economic activity" and left the country dependent on foreign capital. For a decade the trade looked worthwhile: output grew 9.1 percent in 1991, 7.9 percent in 1992 and averaged 4.8 percent a year from 1993 to 1998.

Then the neighbourhood moved. The Asian crisis reached Russia in 1998, Argentina entered a four-year recession in the third quarter, and on 15 January 1999 Brazil devalued the real, making Argentine goods abruptly more expensive to the customer that bought most of them. Under a floating rate that is an inconvenience; under convertibility there was no channel for it. The fiscal side compounded it: from 1996 the government funded its deficits increasingly from domestic banks and the new private pension funds, which is how sovereign credit and household savings stopped being separate risks. See the Asian financial crisis case study and Long-Term Capital Management and the 1998 Russian default.

What Did Three Years of Falling Prices Do to a Country That Could Not Devalue?

This is the part most retellings skip. Argentina did not fail to adjust. It adjusted in the only direction the law allowed, and the adjustment made the arithmetic worse.

Fund figures put the annual average change in Argentine consumer prices at 0.9 percent in 1998, then minus 1.2 percent in 1999, minus 0.9 percent in 2000 and minus 1.1 percent in 2001. Three consecutive years of falling prices is what internal devaluation looks like when it is actually attempted rather than merely recommended. Output fell alongside, by 3.4 percent in 1999, 0.8 percent in 2000 and 4.4 percent in 2001.

Deflation does two things to a borrower at once. It reduces the nominal revenue available to service a fixed nominal obligation, whether a mortgage, a corporate bond or a government's tax base. And it raises the real interest rate on every contract already signed. A country trying to grow out of a debt burden needs nominal growth; Argentina by 2001 had negative real growth and negative inflation at the same time, the worst available combination for a debt ratio and the one a hard peg makes most likely.

The Great Inflation of 1965 to 1982 runs the opposite way and makes the same point: the United States had a debt burden inflation quietly reduced and a credibility problem inflation created, while Argentina had a credibility framework that worked and a debt burden deflation quietly enlarged. By mid-2001 the government was reaching for the exchange rate anyway, applying an effective 7 percent devaluation to foreign trade transactions on 19 June 2001. A state that applies a different rate to some transactions has conceded that one peso is not always one dollar, and every holder of a peso contract could do that arithmetic.

How Much Money Left Argentine Banks Before the Government Locked the Doors?

The freeze is the image everyone remembers, so it is worth being precise that the money had largely already gone. The Republic's own account is unambiguous: during the last six months of 2001 the growing perception that a devaluation was imminent triggered a run on deposits and an acceleration of capital flight, and "total deposits in the Argentine banking system fell by 20.3% in the last six months of 2001 and the Central Bank's international reserves fell by 42.1% in the same period."

Independent series show the same shape annually. World Bank data on total reserves including gold put Argentina at $26.4 billion at end-1999, $25.2 billion at end-2000, $14.6 billion at end-2001 and $10.5 billion at end-2002. The 2001 fall works out to 42 percent, and the Republic attributes essentially all of it to the second half. The World Bank's end-2002 figure matches the one the Republic states in its own filing, a useful cross-check on both.

Two features of an Argentine deposit in 2001 explain the speed. A large share were denominated in dollars but held in Argentine banks, so a depositor who feared devaluation already held the harder currency on paper; what they had to fear was the bank's ability to deliver those dollars and the state's willingness to let it. And the backing for the whole system was published, so depositors could watch their own collateral disappear in the statistics.

Under convertibility the central bank could not simply lend banks the pesos to meet withdrawals, because that would have broken the reserve rule the regime rested on. When Silicon Valley Bank failed in 2023 the Federal Reserve opened a facility within seventy-two hours that lent against securities at par, as set out in the Silicon Valley Bank case study. Argentina's central bank was forbidden from anything of the kind. The only lender of last resort a currency board has is its own reserves, and the Republic records those falling 42.1 percent in the six months to December 2001.

What Was the Corralito, and What Did It Actually Prohibit?

At the end of November 2001 the government stopped the run by making withdrawal impossible. The Congressional Research Service chronology places the measure on 30 November 2001, after more than $1.3 billion left the banks in a single day, and records the limit as $1,000 per person per month. The Republic's filing dates the implementing instrument, Decree No. 1,570/2001, to 3 December 2001. The announcement and the implementing decree are three days apart in the record, and this page gives both dates rather than choosing one.

The name it acquired, corralito, is a diminutive of corral, the little pen a toddler is put in. It was not a haircut and not a bank holiday. Balances still existed and could still move between accounts inside the banking system. What they could not do was leave it in cash, and money could not be sent abroad. The Republic quantifies the scope: the limits effectively prevented withdrawal of approximately $60 billion of peso and dollar demand deposits, alongside strict foreign exchange restrictions.

A cap on cash withdrawals in a country where much of ordinary economic life settles in cash is not a technical banking measure. It interferes directly with wages, rent, small business receipts and pensions. Protests began on 1 December. On 13 December unemployment was reported at 18 percent and unions called a national strike. On 14 December supermarket looting began.

Two things follow and should be kept apart. The corralito stopped the outflow, which was its purpose, and it worked. It also converted a financial crisis into a political one, by turning a diffuse loss most people could not see into a daily obstruction every depositor met at a cash machine. The escalation from 1 to 20 December runs through that mechanism, not through any market price. The restrictions outlasted the emergency: the peso limit was not lifted until 25 November 2002, releasing about $6 billion, while the separate freeze on term and dollar accounts, the corralon, ran alongside it.

How Did a Withdrawal Limit Turn Into Five Presidents in Twelve Days?

The collapse of the Argentine government in December 2001 is not background colour. The terms on which convertibility ended were set by an administration that had been in office for days and owed nothing to the decisions of the previous four years.

Dated milestones with the average pesos-per-dollar rate for the month in which each fell. Rates are monthly averages of daily rates published by the Organisation for Economic Co-operation and Development and retrieved from the Federal Reserve Bank of St. Louis, so they are not the rate on the day itself.

The chronology behind those rows: de la Rua resigned on 20 December, Senate President Ramon Puerta was named provisional president for forty-eight hours on 21 December, Congress appointed Rodriguez Saa on 23 December, Rodriguez Saa resigned on 30 December and Puerta refused a second turn the same day, and Congress selected Duhalde on 31 December. The popular count of five includes the Chamber of Deputies president who bridged the two days in between. Rodriguez Saa announced a three-part plan on 26 December: suspend payment on the public debt, start a jobs programme, and introduce a new non-convertible currency called the Argentino. Only the first part outlived his week in office.

Read the two columns together. Through the entire political collapse the rate stays at one. The peg did not give way under market pressure the way a floating currency does; it was legally intact while the state around it was not, and it moves only when a new president and a new legislature change the law. The peg outlived the administration that had defended it, and it took a new legislature to kill it.

What Exactly Did Argentina Default On in December 2001?

The Republic's own language is that "on December 24, 2001, the Government (under the temporary administration of President Rodriguez Saa) declared a moratorium on a substantial portion of the Republic's public debt," and that Duhalde endorsed it days later. The Congressional Research Service places the announcement of the accompanying economic plan on 26 December. The declaration and the plan were separate acts; the moratorium came first.

Two numbers circulate for the size and they measure different things. The Congressional Research Service describes Argentina as "defaulting on $151 billion of debt," covering the public debt stock affected by the crisis. The Republic's filings state that the 2005 exchange "pertained to approximately U.S.$81.8 billion of defaulted debt," counting eligible securities plus past due interest to 31 December 2001, and note that running interest to the end of 2003 would have raised that subset to at least $102.6 billion. Neither is wrong: one is the total in distress, the other the securities a specific offer was made against.

What was excluded matters as much as what was included. Argentina kept paying multilateral lenders for as long as it could; Decree No. 256 of 6 February 2002 formally suspended payments and authorised a restructuring, and later resolutions refined what stayed in. The prioritisation held only so long. On 14 November 2002 Argentina paid $79 million of interest to the World Bank while missing a $726 million principal payment due on 15 October, and on 15 January 2003 it missed a $680 million payment to the Inter-American Development Bank.

The order of operations is the thing to remember. The default was declared while the peg was still legally in force, before the exchange rate moved. Argentina defaulted first and devalued second, the reverse of the sequence in most currency crises, which meant creditors were repriced against a still-fictional one-to-one rate for two more weeks.

What Did the Public Emergency Law Do on 6 January 2002?

The Public Emergency Law is the hinge of the episode and its operative provision is disarmingly narrow. As the Republic puts it, the law ended the regime "by eliminating the requirement that the Central Bank's gross international reserves be at all times equal to at least 100% of the monetary base." Once that clause was gone the central bank could issue pesos it had no dollars for, and the parity became unenforceable arithmetic rather than law.

What replaced convertibility on day one was not a float but a dual rate: a devaluation of about 29 percent, to 1.40 pesos per dollar, for major foreign commercial transactions, with a floating rate for everything else. A dual rate is a way of choosing who receives a devaluation and who pays for it, and it lasted four and a half weeks. On 3 February 2002 the government closed the foreign exchange market, abolished the dual system for a single floating peso, and ended the dollar's status as an officially circulating currency. The market reopened on 11 February, and the peso fell about 20 percent to 2.5 before recovering into a 1.9 to 2.2 range.

The Duhalde package did far more than move a price. Alongside the float and the conversion of dollar contracts, the central bank's charter was amended so that it could print currency, make short-term advances to the government and act as a lender of last resort.

That charter amendment deserves its own sentence. The institution forbidden by statute from creating money without a dollar behind it was, within a month, statutorily permitted to lend to the government and to the banks. Argentina did not merely change its exchange rate in January 2002. It reconstituted the central bank as a normal central bank, and the reason the crisis has a monetary phase at all is that the institution capable of causing one had just been recreated.

What Was Pesification, and Who Was Made to Pay for It?

A country whose residents hold dollar deposits in domestic banks and owe dollar loans to those same banks cannot devalue without deciding who eats the difference. Leave dollar contracts in dollars and every borrower with peso income is instantly insolvent. Convert them all at the old rate and every depositor is expropriated. Argentina split the difference, unevenly.

Conversion rates applied under the pesification decrees of early 2002, as described by the Republic of Argentina in its filings with the Securities and Exchange Commission. All rates are pesos per United States dollar.

Set the first two rows against each other and the asymmetry is the whole story. A bank had liabilities converted at 1.40 and assets converted at 1.00. That gap was not a rounding convention but an instruction to transfer wealth from savers and banks to borrowers, executed by decree, and it left the banking system with a hole roughly proportionate to the dollar loan book. The state filled the hole with paper: $44 billion of bonds for banks and depositors announced on 5 March 2002, of which $8 billion reached the banks on 25 September. A compensation bond from a government that has just declared a moratorium on its other bonds is a particular kind of asset, and the recipients knew it.

Depositors were offered an exit and refused it. A voluntary swap of frozen deposits for bonds ran from 18 June 2002 and was abandoned on 16 July with 25 percent participating: Argentines who had spent seven months unable to reach their own money were being asked to exchange it for a claim on the entity that had frozen it. The transferable lesson is about denomination rather than about Argentina. A dollar deposit in a domestic bank is not a dollar; it is a domestic-law claim whose currency of settlement can be changed by domestic legislation, a distinction that underlies country, sovereign and political risk.

Why Did the Supreme Court Ruling Fail to Reopen the Accounts?

On 1 February 2002 the Argentine Supreme Court ruled that the restrictions on bank deposit withdrawals were unconstitutional. On 4 February the executive eased some restrictions by decree while placing a six-month moratorium on enforcement of the court's order, and Congress moved to impeach all nine members of the Court.

This is the least financial and most important passage in the case. A court had told depositors they were entitled to their money. The state suspended the judgment and threatened the judges. In September 2002 an appellate court reached the same conclusion about the peso freeze, with the same practical effect, which is none: the restriction was not lifted until 25 November 2002, nearly ten months after the first ruling.

Investors habitually model sovereign risk as a probability of non-payment. Argentina in early 2002 shows a harder failure: the legal answer was clear, favourable to the claimant, formally delivered, and simply not executed. The value of a claim is its merits multiplied by the probability a judgment on those merits is enforced against an unwilling party, and Argentine depositors held the first without the second. The same structure reappears thirteen years later in the holdout litigation.

The choice was not gratuitous. Honouring the ruling in February 2002 meant releasing dollar deposits from a banking system that had no dollars. Reserves stood at $14.6 billion at the end of 2001 and were down to $9.6 billion by 30 June 2002, against frozen claims the Republic itself sizes at roughly $60 billion. There was no version of compliance that did not collapse the banks. The state chose the banks and suspended the judgment, and both halves of that belong in an honest account.

How Far Did the Peso Fall, and Why Did Prices Not Follow It Down?

Monthly average exchange rates trace the path. December 2001 averaged 1.01 pesos per dollar. January 2002 averaged 1.62, February 1.99, March 2.39, April 2.93, May 3.29 and June 3.61. The monthly average peaked at 3.66 in October 2002 and had eased to 3.50 by December. The Republic states its own daily low as 3.87 on 26 June 2002 and 3.36 at 31 December 2002. In the direction a saver would have felt it, a peso that bought one hundred United States cents in November 2001 bought about twenty-six cents at the June low.

That is the exchange rate. It is not what happened to the price of bread. Consumer prices, on the Republic's citation of the national statistics institute, rose 40.9 percent during 2002 and wholesale prices rose 118.0 percent. The Republic notes that both figures were "relatively low in comparison to the more than 240.1% depreciation of the peso against the U.S. dollar during that year." Fund figures, measuring the annual average rather than December against December, put 2002 consumer inflation at 25.9 percent. Both are correct and answer different questions.

The gap between a currency losing three quarters of its value and consumer prices rising two fifths is the pass-through question, and in Argentina pass-through was heavily incomplete. Domestic demand had collapsed, so sellers of non-traded goods and services could not raise prices. Utility tariffs were frozen by decree from January and still being litigated in December. Wages did not chase prices, because unemployment was above 20 percent.

The devaluation therefore did what a devaluation is supposed to do and rarely manages: it cut real domestic costs sharply against world prices without an inflationary spiral. The cost fell on households rather than on the price index, and it shows in the trade account. Argentina's current account went from a deficit of 4.4 percent of gross domestic product in 1998 to a surplus of 7.7 percent in 2002, a swing of twelve points achieved mostly by the country becoming unable to afford imports.

Why Did the Debt Ratio Triple in a Year Argentina Barely Borrowed?

General government gross debt, on Fund figures, was 34.1 percent of gross domestic product in 1998, 40.8 percent in 2000 and 48.0 percent in 2001. In 2002 it was 147.2 percent. Argentina had been shut out of international markets, had declared a moratorium and was arguing with multilateral lenders over instalments. It did not borrow its way to 147 percent.

The ratio moved because its two halves were denominated in different currencies. The numerator was substantially dollar debt, whose peso value tripled with the exchange rate. The denominator was Argentine output, measured in pesos. A country whose liabilities are in a currency it cannot print and whose income is in a currency it can does not have a debt ratio; it has a foreign exchange position with a debt ratio attached.

Argentine macroeconomic series through the crisis. Growth, inflation, unemployment and gross public debt are International Monetary Fund World Economic Outlook figures; inflation is the annual average change in consumer prices, which differs from a December-to-December comparison. Reserves are total reserves including gold at year end, from World Bank data, and continue to $19.7 billion in 2004 and $28.1 billion in 2005. World Bank and Fund real growth agree to one decimal place in every year here except 2003, where the World Bank records 8.8 percent against the Fund's 9.0.

The same effect shows in the size of the economy. Nominal gross domestic product in dollars was $300.4 billion in 2001 and $112.5 billion in 2002, a fall of about 63 percent, while real output fell 10.9 percent. The gap between those numbers, roughly fifty percentage points, is not economics. It is the unit of account changing, and it is the clearest illustration of why a cross-border investor's return and a resident's experience of the same year can diverge beyond recognition. The mechanics for an ordinary portfolio are in international ETFs, currency risk and hedging.

Note what the ratio did next, with no restructuring completed: 125.2 percent in 2003 and 117.9 percent in 2004 on nominal growth and a recovering currency alone, before the 2005 exchange took it to 80.3 percent. The fall from 147.2 to 117.9 is 29.3 points of the 66.9 the ratio eventually gave back, so nearly half of the improvement happened before a bond was written down.

What Happened to Argentine Households?

The World Bank's poverty assessment of July 2003 measured what the macroeconomic series cannot. Poverty as traditionally measured in Argentina rose from 37 percent in 2001 to about 58 percent in urban areas by October 2002, and the number of people below the indigence line, meaning without enough income for a basic food basket, doubled.

The composition of that shift is unusually well documented. Between April 2001 and April 2002 the Buenos Aires poverty line rose 26 percent and the indigence line, which is purely food, rose 29 percent, while the total consumer price index rose only 18 percent. Devaluation raises the price of tradable goods first, and food is tradable, so a household spending most of its income on food faced a 29 percent rise in the cost of its basket against a published 18 percent.

Employment moved less than expected and income moved more. Urban unemployment rose from 13 percent in 1998 to 22 percent in May 2002, then fell to 18 percent by October. Between May 2001 and May 2002, 825,000 jobs were lost, 90 percent in the formal sector. The real income of the poorest tenth fell 47 percent between October 2001 and May 2002, then rose 79 percent by October. That reversal is not a recovery in living standards but the arithmetic of a very low base plus a large emergency transfer: Plan Jefes paid 150 pesos a month to about two million household heads.

Two observations carry past Argentina. The fastest-moving variable in a currency crisis is the real value of income, not employment, because prices move in weeks while payrolls move in quarters. And a devaluation is a regressive tax before it is a competitiveness gain, in proportion to how much of a household's basket is tradable.

Which Clock Says Argentina Recovered, and Why Do They Disagree?

Argentine real output peaked in 1998 and bottomed in 2002. Chaining World Bank annual growth rates, and cross-checking against Penn World Table constant-price levels, an index of real gross domestic product with 1998 set to 100 runs 96.6 in 1999, 95.9 in 2000, 91.6 in 2001, 81.6 in 2002, 88.9 in 2003, 96.9 in 2004 and 105.5 in 2005. The two independent series agree to two decimal places at every point.

So the peak-to-trough contraction was 18.4 percent over four years, and real output regained its 1998 level during 2005, seven years after the peak. Growth ran at 9.0, 8.9 and 8.9 percent in 2003, 2004 and 2005, which is what a recovery looks like from a very deep hole in an economy with idle capacity and a newly competitive currency.

The currency, the banking restrictions and the labour market each kept a separate calendar. The exchange rate stabilised first, in the second half of 2002, as a trade surplus and tightened controls supplied foreign currency. The peso withdrawal restrictions ended in November 2002. Urban unemployment, 22 percent in May 2002 on World Bank figures, had fallen to 18 percent by that October, and the Fund's national rate did not reach 11.6 percent until 2005. The Fund relationship closed later still: Argentina owed $14.0 billion at the end of 2001 and repaid the remaining $9.5 billion on 3 January 2006, borrowing from the central bank to do it and cutting reserves 33.3 percent from $28.1 billion to $18.6 billion.

A reader looking for one number to describe the recovery will not find one, and the absence is the point. Output took seven years, employment longer, the banking restrictions eleven months, the currency about nine, and the debt fifteen years. Any statement of the form "Argentina recovered in X years" answers a question nobody asked without saying which clock it read.

Why Did the Default Take Fifteen Years to Finish?

Argentina made its first restructuring offer more than three years after the moratorium. The Republic launched the 2005 exchange on 14 January 2005, inviting holders of 152 defaulted series. It applied to approximately $81.8 billion of defaulted debt counted as principal plus past due interest to 31 December 2001. Holders tendered about $62.3 billion of eligible value, which the Republic puts at 76.2 percent, leaving roughly a quarter of the stock outside the deal and unpaid.

A second offer followed in 2010. On 30 April the Republic invited holders of 149 defaulted series to exchange, reopened the offer domestically in December, and reported an acceptance rate of 81 percent. Measured against the eligible amount rather than by participation, the securities tendered came to about $12.4 billion, or 66.2 percent. Cumulatively about 92 percent of the defaulted debt had then been restructured, with roughly $6.1 billion untendered.

That residual 8 percent produced fifteen years of litigation, because creditors who refuse an exchange keep their original contractual rights in full. The Republic published a settlement proposal on 5 February 2016, conditional on approval by the Argentine Congress and the lifting of injunctions issued by United States courts under the equal treatment clause in the old bonds. Congress passed Law No. 27,249 on 31 March 2016, repealing the statutes that had prohibited any payment on untendered debt, and on 22 April Argentina issued $16.5 billion of new securities, of which $9.3 billion funded settlement payments.

The arithmetic of holding out is worth stating rather than moralising about. A creditor who accepted in 2005 got restructured bonds immediately and gave up most of the face value. One who refused waited eleven more years, litigated at their own expense, and was eventually paid on terms negotiated against the leverage of a court injunction. None of that was knowable in 2005, and a strategy needing fifteen years of legal budget, an unenforceable judgment throughout and a change of government at the far end is not one most investors can execute. What the episode changed is the paperwork: this litigation is why sovereign bonds now routinely carry collective action clauses letting a supermajority bind a dissenting minority. That is a market convention rather than a law, so read the terms of the instrument in front of you.

Which Warning Signs Were Readable in 2001, and Which Only Afterwards?

Visible before the event

• The spread, continuously. The Congressional Research Service records Argentine bonds at 610 basis points over United States Treasuries on 10 December 1999 and 1,300 basis points by 10 July 2001. A market demanding thirteen extra points of yield is not expressing an opinion about liquidity.
• Three years of deflation alongside three years of contraction. Both series were published, and a debt ratio rising while prices fall has no benign interpretation.
• Two debt exchanges in five months. The exchange of 16 and 17 June 2001 covered $29.5 billion of short-term debt for longer maturities at higher rates. The 6 November swap covered about $60 billion and cut the average rate from 11 or 12 percent to 7. Rating agencies called the second an effective default at the time, not afterwards.
• The reserve position. Under a currency board the reserve stock is the whole collateral for the currency and it is published. Watching it fall from $25.2 billion to $14.6 billion during 2001 required arithmetic, not insight.
• The deposit outflow. A fifth of the deposit base left in six months before any restriction was imposed. Argentine residents were the earliest and best-informed sellers of Argentine risk.

Only clear afterwards

• That the exit would be asymmetric. A depositor could reason correctly about devaluation risk and still not anticipate deposits converting at 1.40 while loans converted at 1.00, a split decided in weeks by an administration that did not exist in November 2001.
• That the courts would be overridden. The Supreme Court ruled for depositors and the ruling was suspended. No prior Argentine experience made that the base case.
• That inflation would be contained. A currency losing three quarters of its value would normally produce far more than 40.9 percent consumer inflation. It did not, because domestic demand had been destroyed first.
• That growth would return this fast. Three consecutive years near 9 percent from 2003 was not the consensus in mid-2002, and it followed from the same devaluation that impoverished half the population.

The honest summary is that the direction was legible for two years, the timing was not, and the terms of the resolution, which decided most of the actual loss allocation, were unknowable because nobody had chosen them yet. An investor who read the spread correctly in July 2001 still could not have known in which of several very different ways they would be repaid.

Why Is Argentina a Poor Template for the Next Currency Peg Failure?

Very few pegs are statutory. Convertibility was an Act of Congress constraining the central bank's balance sheet directly. Most fixed and managed rates today are policy commitments adjustable administratively, which makes them easier to abandon early and cheaply and correspondingly less credible while they last. Argentina bought a decade of credibility by making the exit expensive, then paid for the decade in one year.

Domestic dollarisation of contracts was extreme. The exit required rewriting private loan and deposit agreements by decree because Argentine households and firms had signed dollar contracts under Argentine law in enormous volume. A country with a floating currency and mostly local-currency contracts has no equivalent problem, because a devaluation changes relative prices without invalidating balance sheets.

The political vacuum was not incidental. The terms of the exit were set by a president Congress selected on 31 December, the day after his predecessor resigned, and who was sworn in on 1 January with the currency board still legally standing. Most crises are resolved by an administration with a mandate and a memory of the decisions that caused the problem. Argentina's was not.

The comparison with a modern banking stress does not hold. Central banks now answer deposit flight with facilities measured in days, as in the Silicon Valley Bank case study and the response to the 2020 COVID crash. Argentina's central bank was prohibited by the constitution of the regime itself, and by the time the prohibition was lifted the run was over. What travels is narrower: a promise about an exchange rate is an asset with a credit quality of its own, and when a government's liabilities are denominated in a currency it does not issue, its solvency is a foreign exchange question rather than a fiscal one.

Common Myths About Argentina's 2001 Default

"The peso was devalued and that caused the default." The order was the reverse. The moratorium was declared on 24 December 2001 while the Convertibility Law was still in force and the official rate was still one to one. The peg was not repealed until 6 January 2002 and the market did not float until 11 February. Argentina defaulted with its exchange rate intact, which is why loss allocation between creditors, depositors and borrowers had to be legislated rather than left to the currency.

"The corralito confiscated people's savings." Not directly. It was a withdrawal restriction: balances stayed on the books and could move within the banking system but could not be taken out in cash or sent abroad, and roughly $60 billion of demand deposits were caught by it. The confiscatory event was the separate decision, weeks later, to convert dollar deposits into pesos at 1.40 while the market rate ran past 3. Conflating the two hides that the freeze and the haircut were different decisions with different authors.

"Argentina was punished for defaulting and lost market access for a generation." The economy grew 9.0, 8.9 and 8.9 percent in 2003, 2004 and 2005, and regained its 1998 output level in 2005 while still in default on most of its bonds. The debt ratio fell from 147.2 percent of gross domestic product in 2002 to 117.9 percent in 2004 before the exchange offer completed. What persisted was the litigation, not the recession.

"The devaluation caused hyperinflation." The peso lost roughly three quarters of its dollar value and consumer prices rose 40.9 percent in 2002 on national statistics cited by the Republic. That is a serious inflation and nothing like the hyperinflation of the late 1980s that convertibility was built to end. Prices did not chase the currency because domestic demand had collapsed, tariffs were frozen and unemployment above 20 percent left nobody able to reclaim real wages.

"The International Monetary Fund refused to help." The Fund lent repeatedly and at scale: a three-year $7.2 billion stand-by on 10 March 2000, a $40 billion multilateral package in December 2000 with a $7.0 billion Fund augmentation in January 2001, and a further $7.2 billion on 7 September 2001. The withheld $1.24 billion instalment of 5 December 2001 came after twenty-one months of programmes built on growth forecasts wrong in the same direction every time: the 2000 arrangement assumed 3.5 percent growth in a year output fell 0.8 percent, and the January 2001 augmentation assumed 2.5 percent in a year output fell 4.4 percent.

What a Reader Can Actually Carry Forward

Most of what happened in Argentina between 1998 and 2002 belongs to Argentina. A statutory currency board, a fortnight of five presidents and a decree rewriting every dollar contract in the country are not a recurring pattern. Four things do generalise, and none requires forecasting a crisis.

• Currency of denomination is a risk, not a label. A dollar deposit in an Argentine bank in 2001 was a peso deposit with a dollar sign on it, and the difference was invisible until the day it mattered. Ask of any holding which legal system decides what currency it settles in, and whether that system's interest matches yours.
• A country's investable size and its economic size are different numbers. Argentine output fell 10.9 percent in 2002 and its dollar value fell 63 percent. A foreign investor's loss was overwhelmingly the exchange rate, not the economy.
• A judgment is not a payment. Argentine depositors won in the Supreme Court in February 2002 and waited until November for their pesos. Holdout creditors won in United States courts and waited until 2016. Legal merit and enforceability are separate variables, and the second is the one that pays.
• Watch what residents do. The most reliable early signal was not a rating or a spread but the deposit series: a fifth of the country's bank deposits left in six months, before any restriction, because the people closest to the risk moved first. Domestic capital flight is published in most countries and is rarely wrong about direction.

The question worth asking now

Not "will another currency board fail", which almost nobody needs to answer, but a narrower one: for each holding, what happens to it if the government whose law governs it decides someone has to take a loss and that it will not be the government? For most portfolios the honest answer is that the question never becomes live. Where it does, the answer is usually visible in the documentation long before it is visible in the price, and reading the documentation costs nothing.

References

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

• Republic of Argentina: Description of the Republic, Exhibit D to Annual Report on Form 18-K, filed September 2016: the Convertibility Law and its reserve-backing requirement; growth of 9.1 percent in 1991, 7.9 percent in 1992 and 4.8 percent a year from 1993 to 1998; the 20.3 percent fall in deposits and 42.1 percent fall in reserves in the second half of 2001; Decree No. 1,570/2001 and the $60 billion of demand deposits it caught; the 24 December 2001 moratorium; the Public Emergency Law and its elimination of the 100 percent reserve requirement; the pesification rates of 1.40 and 1.00; the 240.1 percent rise in the dollar-peso rate in 2002, the 3.87 low on 26 June and 3.36 at 31 December; consumer inflation of 40.9 percent and wholesale inflation of 118.0 percent in 2002; reserves of $10.5 billion at 31 December 2002; the 2005 and 2010 exchanges and their participation rates; the 2016 settlement and the April 2016 issuance; and the Fund debt of $14.0 billion repaid on 3 January 2006.
• Congressional Research Service: The Argentine Financial Crisis, A Chronology of Events, Report RL31582, 5 June 2003: every dated milestone on this page not attributed above, including the Convertibility Law of 1 April 1991, Brazil's devaluation of 15 January 1999, the spreads of 610 and 1,300 basis points, the Fund arrangements and augmentations of 2000 and 2001 with their growth assumptions, the two debt exchanges of June and November 2001, the $1,000 monthly withdrawal limit, the withheld $1.24 billion instalment, the December rioting and 28 deaths, the presidential succession, the dual rate, the Supreme Court ruling with its enforcement moratorium and impeachment move, the float of 3 February and reopening of 11 February, the $44 billion and $8 billion bank compensation measures, and the end of the peso restriction on 25 November 2002, which is also the entry naming the corralon as the separate surviving freeze on term and dollar accounts. The $151 billion default figure is this report's own summary figure.
• World Bank: Argentina, Crisis and Poverty 2003, A Poverty Assessment, Report No. 26127-AR, 24 July 2003: every household figure here, including poverty rising from 37 to about 58 percent, the doubling of the population below the indigence line, urban unemployment of 13, 22 and 18 percent, the divergence between the poverty, indigence and consumer price indices, the 825,000 jobs lost, the income swing for the poorest tenth, and Plan Jefes.
• International Monetary Fund: World Economic Outlook database, Argentina real GDP growth series NGDP_RPCH: the real growth column. The linked address is one series endpoint of the WEO data API; the other five figures on this page come from the same API with the series code substituted, namely PCPIPCH for annual average consumer price inflation, LUR for unemployment, GGXWDG_NGDP for gross public debt, NGDPD for gross domestic product in dollars, giving $300.421 billion for 2001 and $112.458 billion for 2002, and BCA_NGDPD for the current account balance. Note that imf.org itself returns HTTP 403 to an automated request while this data API answers normally.
• World Bank: World Development Indicators, Argentina annual GDP growth series NY.GDP.MKTP.KD.ZG: the unrounded annual real growth rates chained into the output index here.
• World Bank: World Development Indicators, Argentina total reserves including gold series FI.RES.TOTL.CD: every year-end reserve figure quoted on this page, from $24.9 billion in 1998 to $28.1 billion in 2005.
• Federal Reserve Bank of St. Louis: US Dollar Exchange Rate, Average of Daily Rates, National Currency for Argentina, Series ARGCCUSMA02STM: every monthly average peso rate quoted here, from 1.01 in December 2001 to the 3.66 peak in October 2002.
• Federal Reserve Bank of St. Louis: Real GDP at Constant National Prices for Argentina, Series RGDPNAARA666NRUG: the constant-price series used to cross-check the output index, giving the same 1998 peak, 2002 trough and 2005 recovery year.

Figures deliberately not stated. This page gives no peso rate for any specific day other than the two the Republic states itself, no Buenos Aires equity index level, no sovereign spread beyond the two the Congressional Research Service records, no bank failure count, no recovery value for any individual defaulted bond and no legal cost for the holdout litigation, because no source verified for this page supplied them. The Fund's Independent Evaluation Office report on Argentina, the standard reference on the Fund's own conduct, could not be retrieved: imf.org returned HTTP 403 to every request. Fund programme dates and amounts here therefore rest on the Congressional Research Service chronology and the Republic's filings, and no claim is made about the Evaluation Office's findings.

Method note: the real output index chains World Bank annual growth rates from a base of 100 in 1998 and was cross-checked against Penn World Table constant-price levels; the two agree to two decimal places at every year shown. The 63 percent fall in dollar output is calculated from the Fund's nominal dollar series. Two measures of 2002 inflation appear here and are not reconciled: 40.9 percent is a December-to-December comparison quoted by the Republic, the Fund's 25.9 percent an annual average. Where two sources date the same measure differently, both dates are given rather than silently chosen between. One growth outturn is deliberately not the chronology's own: the Congressional Research Service, writing in 2003, recorded 2001 output as falling 3.7 percent, while the Fund's current series says 4.4 percent, and this page uses 4.4 throughout so the myths section and the table cannot disagree.

Everything above describes Argentina between 1991 and 2016 and nothing above describes Argentina now. The exchange rate regime, the capital controls, the debt position and the banking rules have each been rewritten more than once since, so none of this is investment advice, a forecast, or a guide to any live Argentine instrument. Sovereign bond documentation standards, including the collective action clauses discussed above, are market conventions that continue to evolve.

Frequently Asked Questions