Brazil 1999 and the Float of the Real

Key Takeaways

• The peg was policy, not law. Brazil's central bank ran a crawling band with an intended nominal devaluation of 7.5 percent a year, a mechanism it could and did abandon by decision rather than by repealing a statute.
• The currency overshot hard and then partly recovered. The Federal Reserve's daily series puts the real at 1.2074 to the dollar on 4 January 1999, a low of 2.20 on 3 March, and 1.809 at year-end, a round trip that took the real from about 83 cents to about 45 cents and back to about 55 cents.
• Rates did the work a banking freeze did in other crises. The Selic target hit 45 percent on 5 March 1999 and was cut in 11 steps to 19 percent by 23 September, according to the central bank's own published series.
• Inflation barely moved. Full-year IPCA inflation for 1999 was 8.94 percent, computed from the central bank's own monthly index, against a currency that lost roughly 45 percent of its dollar value.
• The $41.5 billion international package announced in November 1998 bought about two months. Reserves kept falling through the defense, and the peg broke anyway on 13 and 15 January 1999.
• Brazil built a new monetary framework in six months. Inflation targeting took effect 1 July 1999 under Decree No. 3,088, and the first target, 8 percent plus or minus 2 percentage points, was met.
• Real output grew 0.5 percent in 1999 on International Monetary Fund figures. The dollar value of Brazil's economy fell from $864.3 billion to $599.6 billion in the same year, almost entirely a currency effect rather than an output collapse.
• No deposits were frozen and no bank holiday was declared. The transmission channel was the exchange rate and the interest rate, not the banking system.

What Was Brazil's Crawling Peg, and How Was It Different From a Currency Board?

Brazil's stabilization plan, the Plano Real, began in mid-1994 and brought inflation down from four-digit annual rates to single digits in under three years. The exchange rate did most of the disinflationary work, but Brazil did not fix it the way Argentina fixed the peso in 1991. Brazil's own central bank economists describe the mechanism plainly in a working paper written the following year: the official policy "consisted of an intended nominal devaluation of 7.5% p.a., while annual inflation was near 2%." The real was allowed to slide against the dollar inside a managed band, adjusted in small steps, rather than held at one fixed number by statute.

That distinction matters for everything that follows. A currency board like Argentina's Convertibility Law removes the exchange rate as a policy tool entirely and can only be ended by the legislature that created it. Brazil's crawling band was a central bank decision, defended with reserves and short-term interest rates, and it could be widened or abandoned by the same institution that ran it, on a timetable of days rather than the months a repeal would take. The August 1997 Asian crisis was survived this way: Brazil's central bank, under governor Gustavo Franco, raised interest rates sharply and held the band. What followed in 1998 and 1999 was a second, larger version of the same defense, run against a bigger financing gap.

See the Argentina 2001 case study for the statutory version of a currency peg and what it took to break one.

What Vulnerabilities Had Built Up by the Time Russia Defaulted in August 1998?

Brazil ran twin deficits through the mid-1990s that a crawling peg does not fix by itself. The current account deficit averaged around 3 to 4 percent of gross domestic product across 1996 to 1998 on International Monetary Fund figures, meaning the country needed a steady inflow of foreign capital just to hold its position, before any fiscal shortfall is added. The stabilization plan itself had been gradualist on the fiscal side: Brazil's central bank later wrote that "a much-needed definitive fiscal adjustment was continually postponed because, in part, the government coalition was not sufficiently convinced of its urgency."

That combination, a currency defended at cost and a fiscal position not yet secured, is what turned an external shock into a domestic crisis. Reserves were the pressure valve: World Bank data show Brazil's total reserves including gold at $60.2 billion at the end of 1996 and $52.2 billion at the end of 1997, already declining before Russia's default reached Brazil at all.

Then the external shock landed. Russia defaulted on its domestic debt and devalued the ruble on 17 August 1998, an event covered in detail in the Long-Term Capital Management and Russian default case study. Capital that had been willing to fund Brazil's current account deficit stopped being willing to fund anyone's, and Brazil, as the largest emerging market still running an explicit, defendable exchange-rate target, became a natural next test.

How Did Brazil Defend the Real Between August 1998 and January 1999?

The first line of defense was the interest rate. Brazil's central bank published rate data show the Selic averaging 19.23 percent annualized in August 1998, the month of Russia's default, then jumping to 34.29 percent in September and peaking at 41.60 percent in October as capital flight accelerated. That is a defense measured in weeks, not months: a policy rate more than doubled to hold a currency band.

Selic figures are the monthly average of the annualized daily rate, Banco Central do Brasil series 4189. The rate eased through late 1998 once the November support package was announced, then rose again after the January float, a second, separate defense captured in the table further below.

Reserves absorbed the rest of the strain. World Bank data put Brazil's total reserves including gold at $44.5 billion at the end of 1998, down from $52.2 billion a year earlier, a fall of roughly 15 percent even with the rate defense in place and an international support package already announced. The exchange rate itself barely moved during this period: the Federal Reserve's daily series shows the real at 1.1730 to the dollar on 17 August 1998 and still only 1.2020 by 2 December, the day the International Monetary Fund's board approved its share of the rescue package. The defense was working, at a cost, right up until it wasn't.

What Did the November 1998 International Support Package Actually Buy?

On 13 November 1998 the United States government announced an international support package for Brazil totaling approximately $41.5 billion, assembled from the International Monetary Fund, the World Bank, the Inter-American Development Bank and twenty bilateral creditors. The White House's own fact sheet from that day itemizes it: $18.0 billion from the IMF, $4.5 billion from the World Bank, $4.5 billion from the Inter-American Development Bank, and $14.5 billion in bilateral financing, of which the United States committed $5 billion.

The IMF board approved its share on 2 December 1998: a three-year stand-by arrangement equivalent to SDR 13,025 million, about $18.1 billion, with 70 percent available through the Supplemental Reserve Facility, a higher-cost, faster-disbursing IMF instrument built for exactly this kind of confidence crisis. Roughly $15.7 billion of the total was scheduled for release through the end of 1999, with about $5.3 billion available immediately.

It bought Brazil roughly six weeks of a stable exchange rate and a further five weeks of active, ultimately unsuccessful, defense. Brazil's own central bank account of the period is direct about why: "the government was initially successful in implementing the fiscal package, but market confidence continued to erode up to January 1999, also reflecting concerns over the newly elected governors' commitments to adjusting their states public finances." A package sized for a financing gap does not repair a credibility gap, and the credibility gap was about to widen.

What Happened in Minas Gerais on 6 January 1999?

Itamar Franco, a former president of Brazil, took office as governor of the state of Minas Gerais on 1 January 1999. Within days he declared a moratorium on his state's debt payments to the federal government, a move his own English-language Wikipedia biography summarizes as enacted "as soon as he took office," reported in contemporaneous coverage as a 90-day suspension on a debt stock above 16 billion reais.

A single state's debt moratorium is not, by itself, a national balance-of-payments event. What it did was confirm the exact fear the November package had been meant to defuse: that state governments elected in October 1998, several from parties outside President Fernando Henrique Cardoso's coalition, were not committed to the fiscal discipline the federal government had promised the IMF. Brazil's central bank had already flagged "concerns over the newly elected governors' commitments to adjusting their states public finances" as a live risk before Minas Gerais moved. The moratorium turned that risk from a possibility analysts discussed into a headline foreign investors could act on immediately, and it landed one week before the band was widened.

What Happened on 13 and 15 January 1999?

Reais per U.S. dollar, daily rate, Federal Reserve Bank of St. Louis series DEXBZUS. Where a listed date fell on a weekend or holiday, the nearest published trading day is shown.

Two separate decisions sit inside those two dates. On 13 January the central bank widened the band it had defended since 1994, an adjustment rather than a surrender, and the real moved from 1.2108 to 1.3210, about 9 percent. It did not hold. On 15 January the central bank abandoned the band altogether and let the real float, and most of its board left office the same day. Brazil's own account of the sequence, published the following year, is unambiguous about the order: "after a brief attempt to conduct a controlled devaluation, the real was forced to float on January 15."

How Far Did the Real Fall, and Did It Overshoot?

In the currency's own value rather than the quoted rate, a real that bought about 83 U.S. cents on 4 January 1999 bought about 45 cents at the 3 March low, a fall of roughly 45 percent. Reais per dollar rose from 1.2074 to 2.20 over the same window, an 82 percent move in the quoted rate, the same fact stated the other way around.

Brazil's own central bank data, published in the working paper that describes the transition, states the exchange rate averaged R$1.52 per dollar in January and R$1.91 in February, against R$1.21 prior to the change in regime. The Federal Reserve's independent daily series produces almost identical monthly averages when computed directly: 1.5120 for January and 1.9261 for February 1999. Two different institutions counting the same event land within half a percent of each other.

The rate then partly reversed. By 6 April it was back to 1.7300, and it spent most of the rest of 1999 oscillating between roughly 1.70 and 1.95, closing the year at 1.8090, or about 55 cents to the real. That round trip, overshoot followed by partial recovery, is the pattern the exchange-rate models in Brazil's own inflation-targeting framework were built to describe: the working paper's reference list cites Rudiger Dornbusch's 1976 paper on exchange-rate overshooting directly, because the Brazilian data matched the textbook shape closely enough to be worth naming.

Who Ran Brazil's Central Bank During the Worst of the Crisis?

Gustavo Franco, central bank governor since August 1997 and the architect of the defense described above, left the post in the same window the real was floated. Francisco Lopes, a career central bank economist, ran monetary policy through the following weeks without ever being confirmed by the Senate to the post in his own right; contemporaneous reporting from early February 1999 describes him as removed "less than one month after he was nominated to the post." Both Lopes and Finance Minister Pedro Malan offered to resign during the worst of the pressure; President Fernando Henrique Cardoso accepted neither.

Arminio Fraga, previously a managing director at Soros Fund Management, was named to replace Lopes around 1 February 1999. Because Brazil's Federal Senate must formally approve central bank board nominees through a two-step hearing and vote process, Fraga and the rest of the reconstituted board did not formally take office until 4 March 1999, seven weeks after the float. Brazil's own account of the gap is candid about the cost: with no confirmed board in place, the exchange rate averaged R$1.52 per dollar in January and R$1.91 in February "in the absence of a well-defined guidance for monetary policy," precisely the period the real fell hardest.

How High Did Interest Rates Go, and How Fast Did They Come Back Down?

Selic target rate, Banco Central do Brasil series 432 (from 5 March 1999, when a formal daily target was first published) preceded by the monthly average annualized rate, series 4189, for the months before that series begins.

The Monetary Policy Committee, Copom, took office on 4 March 1999 and immediately voted to raise the Selic target from a prevailing 39 percent to 45 percent, effective 5 March and close to where futures contracts were already trading. The committee's own first public statement, quoted in the central bank's working paper, laid out the logic in five points ending with: "the basic interest rate should be sufficiently high to offset exchange-based inflationary pressures... but with a downward bias, for if the exchange rate returns to more realistic levels, keeping the nominal interest rate that high would be unjustified." That bias was used twice before the committee's next scheduled meeting, cutting the rate to 42 percent and then 39.5 percent as the currency began to recover.

From there the descent was steady rather than abrupt: 29.5 percent by mid-May, 21 percent by late June, and 19 percent by late September, where it held for the rest of 1999. A policy rate that had briefly matched some of the highest sovereign-crisis rates on record came down over roughly six months without a second currency shock, which is itself part of the story: the rate cuts tracked a currency that was recovering and an inflation number that never spiked, not a policy retreat under renewed pressure.

Why Didn't the Devaluation Turn Into Hyperinflation?

IPCA is Brazil's official consumer price index, Instituto Brasileiro de Geografia e Estatística, published by Banco Central do Brasil as series 433. Selic is the monthly average annualized rate, series 4189.

Two mechanisms did most of the work. The first was the interest rate itself: a Selic averaging above 43 percent in March made holding reais expensive to bet against and gave importers and retailers a reason to defer price increases rather than pass through a currency move that real rates that high suggested might reverse. The second was demand. Brazil's central bank working paper is direct that the economy was not running an inflationary process in the technical sense: "as there were no indications of the presence of an inflationary process in Brazil, a gradualist disinflation strategy was not recommendable," because the price shock was a one-time relative-price realignment rather than a self-sustaining spiral, and a weak domestic economy left little room for sellers to make a devaluation stick as a permanent repricing.

Compounded across the twelve IPCA readings in the table above and the three remaining months of the year, full-year 1999 inflation came to 8.94 percent, calculated directly from the central bank's own published index. Against a currency that lost about 45 percent of its dollar value, that is a pass-through ratio nowhere near one-for-one. It is also, not coincidentally, almost exactly the 8 percent target the government would set for that same year in June, a target set after eleven of the twelve months' inflation was already known.

What Did Brazil Adopt on 1 July 1999, and How Was the Target Set?

President Cardoso issued Decree No. 3,088 on 21 June 1999, establishing inflation targeting as Brazil's monetary policy framework effective 1 July. The decree set the mechanics rather than the numbers: targets and their tolerance bands would be set by the National Monetary Council on the finance minister's proposal, the price index would be chosen the same way, and a breach would require the central bank governor to publish an open letter to the finance minister explaining the cause, the remedy, and the expected time to return to target.

The National Monetary Council filled in the numbers nine days later, in Resolution No. 2,615 of 30 June 1999. It chose the IPCA, the broad consumer price index compiled by Brazil's national statistics institute across nine metropolitan areas plus Goiânia and the Federal District, as the target index, and set targets of 8 percent for 1999, 6 percent for 2000 and 4 percent for 2001, each with a tolerance band of plus or minus 2 percentage points. Brazil's own working paper on the design explains the declining path: the January devaluation was treated as a one-time relative-price shock rather than the start of an inflationary process, so the target could fall each year without assuming a slow, gradualist disinflation was needed.

The choice to target headline inflation rather than a core measure that strips out volatile items was deliberate and, by the authors' own account, costly in flexibility: "adopting a headline index was essential for credibility reasons... Brazilian society has experienced several price index manipulations in a not so distant past, and so would be suspicious about any change related to suppressing items from the target index." The framework also carried no escape clause for a missed target beyond the open letter, which the same paper argues is why the tolerance band needed to be as wide as 2 percentage points in each direction.

Did Brazil Hit Its First Inflation Target?

Yes. The 1999 target was 8 percent with a band of 6 to 10 percent, and IPCA inflation for the full year came to 8.94 percent, inside the band and close to the central number, on an index compiled independently of the target-setting decision that named 8 percent nine days before the second half of the year began. Brazil's central bank, writing about a year later, was careful not to overclaim the result: "even though the target for 1999 has been met, it is too early to discuss its success," noting that a single year is not enough to judge whether a new framework has earned lasting credibility.

That caution reads as reasonable rather than false modesty. A regime adopted in the same year as a currency crisis, hitting a target that was set with most of that year's inflation data already known, is a real result and also a favorable first test. The harder test, whether the framework could hold a target through a year without a currency shock behind it, was still ahead in mid-2000 when the paper was written.

What Happened to Output, Jobs and the Dollar Size of the Economy?

Real output grew 0.3 percent in 1998 and 0.5 percent in 1999 on International Monetary Fund figures, effectively flat across both years rather than contracting the way a headline currency crisis of this scale might suggest. Unemployment, the same source shows, rose from 10.1 percent in 1998 to 11.1 percent in 1999, a real but modest deterioration. Growth then accelerated to 4.4 percent in 2000 as the new policy framework bedded in and the currency stabilized.

The dollar value of the economy tells a different story, because it is mechanically a currency story rather than an output story. Brazil's nominal gross domestic product measured in dollars fell from $864.3 billion in 1998 to $599.6 billion in 1999, a drop of about 31 percent, on the same International Monetary Fund series, purely because each real bought fewer dollars, not because Brazil produced 31 percent less. Anyone whose exposure to Brazil was measured in dollars, a foreign bondholder, an equity investor without a hedge, an importer with dollar costs, took a currency loss the domestic economy itself did not generate.

Reserves, the resource spent defending the old band, continued falling even after the float: World Bank data show total reserves including gold at $36.3 billion at the end of 1999, down from $44.5 billion a year earlier, before recovering only partially to $33.0 billion in 2000 and $35.9 billion in 2001. The defense cost more than the numbers defending the band alone suggest, because the outflow did not stop the moment the exchange-rate policy changed.

Which Warning Signs Were Visible Before January 1999, and Which Only Made Sense Afterward?

Three signals were public well before the float. The current account deficit, running near 4 percent of gross domestic product on International Monetary Fund figures, was published quarterly and showed Brazil's dependence on continuous foreign inflows. Reserves were falling in the official data every month from 1996 onward, a decline visible before Russia ever defaulted. And the fiscal side was explicitly conditional: the IMF's own package was structured around states adjusting their finances, meaning the market already knew that piece was unresolved before Minas Gerais moved.

What was not readable in advance was the timing and the sequencing: that a single state's declared moratorium, rather than a larger federal event, would be the specific trigger, or that the central bank's own board would turn over in the same week as the float, leaving monetary policy without a confirmed decision-maker for seven weeks. Most read-throughs written in 1998 about a "twin deficits" vulnerability were directionally right; almost none named 6 January as the date or Minas Gerais as the mechanism. That is the standard shape of a currency-crisis warning: the structural vulnerability was legible for years, the specific trigger was not.

What genuinely surprised most contemporaneous forecasters, on the evidence of how sharply the central bank's own account discusses the passthrough question, was how little of the devaluation reached prices. A currency losing 45 percent of its value with no meaningful inflation acceleration was not the base case going into February 1999.

Why Is Brazil 1999 a Useful Contrast to Argentina 2001?

The two cases are often mentioned in the same breath because both were emerging-market currency crises within three years of each other in South America's two largest economies, and both followed the same 1998 Russian shock. The mechanics diverged sharply from there, and the divergence is more instructive than the similarity.

Brazil's peg was administrative; Argentina's was statutory. Brazil's central bank widened its band on 13 January and floated on 15 January, two decisions taken two days apart by an institution empowered to make them. Ending Argentina's Convertibility Law required Congress to repeal the reserve-backing requirement, which did not happen until 6 January 2002, more than a decade after the peso was fixed and more than a month after Argentina had already defaulted on its debt. See the Argentina 2001 case study for the full mechanism.

Brazil floated its currency without touching its banks. No withdrawal limit, no corralito, no deposit freeze; the crisis moved through the exchange rate and the interest rate, leaving the payments system running throughout. Argentina capped cash withdrawals at $1,000 a month at the end of November 2001 and did not lift the restriction on peso accounts until 25 November 2002, eleven months later.

Output diverged the most. Brazil's real economy grew slightly through 1999 while its dollar-denominated size fell by about a third on the exchange rate alone. Argentina's real output fell 10.9 percent in 2002, on top of the exchange-rate collapse, because the same peso that had been a unit of account for wages and prices for a decade abruptly stopped being one. A floating currency that had already been sliding by design absorbed a shock that a currency fixed by law could not.

Common Myths About Brazil's 1999 Currency Crisis

"Brazil defaulted on its debt like Argentina did." It did not. Brazil devalued and floated its currency; it did not declare a moratorium on its own federal debt, and its central government continued servicing its obligations, including the new IMF program, through the crisis. The Minas Gerais moratorium was a single state's action against its debt to the federal government, not a national default.

"The $41.5 billion package failed, so the money was wasted." The package did not prevent the float, but it was not wasted: roughly $15.7 billion of the IMF's share alone was scheduled for disbursement through 1999, and that financing supported Brazil's reserves and its new inflation-targeting program through a period when market access on ordinary terms would have been far more expensive, if available at all.

"A 45 percent currency collapse should have produced runaway inflation." The intuition assumes a fixed pass-through ratio that Brazil's own data does not support. Full-year IPCA inflation came to 8.94 percent, and the central bank's own modeling work found the passthrough coefficient itself falls as the currency's prior overvaluation shrinks, which is consistent with a large one-time move producing a smaller-than-expected price effect once demand and interest rates are both working against it.

"Brazil's economy collapsed in 1999." Real output grew 0.5 percent that year on International Monetary Fund figures. What collapsed was the dollar value of the economy, which fell by roughly 31 percent purely on the exchange rate, a distinction that matters enormously to a foreign investor and very little to a Brazilian worker paid and spending in reais.

"Gustavo Franco was fired for the devaluation." The public record does not support a single clean narrative here, and this page does not attempt to supply one. What is documented is that the board turned over in the same week as the float and that his eventual successor did not take formal office until seven weeks later, a gap Brazil's own central bank account attributes to the constitutional requirement that the Senate confirm central bank nominees, not to any stated cause for Franco's own departure.

What a Reader Can Actually Carry Forward

Most of the specific machinery here, a crawling band, a Copom bias mechanism, a Senate confirmation process that left a central bank without a confirmed board for seven weeks, belongs to Brazil in 1999. Four things generalize without requiring a forecast of the next crisis.

• How a peg can be exited changes what a crisis looks like. A currency that can be floated by administrative decision breaks differently, and often less destructively, than one that can only be abandoned by repealing a law. Ask what kind of decision, and by whom, would be required to end any fixed or managed rate a holding depends on.
• The exchange rate and the banking system are separable channels. Brazil's crisis ran entirely through the currency and interest rates; the banks stayed open and deposits stayed available throughout. A currency shock does not automatically become a banking shock, and checking whether deposit or withdrawal restrictions are even legally possible under the relevant framework is a distinct question from checking the currency's flexibility.
• Dollar losses and local losses are different numbers, and both are real. Brazil's economy shrank by about a third in dollar terms in 1999 while growing slightly in real local terms. Know which currency your own exposure is actually measured in before reading a headline GDP or currency number as your own result.
• A credible new framework can be built quickly, but "credible" and "tested" are not the same word. Brazil designed and launched inflation targeting in about six months and met its first target within eighteen months of the crisis that forced the framework's creation. Its own architects said as much at the time: meeting one target under favorable circumstances is a result, not yet a track record.

The question worth asking now

Not whether the next currency crisis will produce a Brazil-style outcome rather than an Argentina-style one, which is not a choice any single investor gets to make, but a narrower one: for a given fixed or managed exchange rate, what is the actual legal and institutional mechanism by which it could end, and does ending it require touching the banking system at all? Brazil and Argentina answered that question two different ways within three years of each other, and the difference in outcome tracked the difference in mechanism far more closely than it tracked the size of either currency's fall.

References

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

• Banco Central do Brasil: Bogdanski, Tombini and Werlang, Implementing Inflation Targeting in Brazil, Working Paper Series No. 1, July 2000: the crawling-band mechanics and its 7.5 percent annual devaluation and 2 percent inflation description; the January 15 float and the board replacement; the January and February monthly average exchange rates of R$1.52 and R$1.91; the wholesale and consumer price moves in February; the March 4 board transition and Senate confirmation process; the Selic path from 39 to 45 percent and back to 42 and 39.5 percent; Decree No. 3,088 of 21 June 1999 and Resolution No. 2,615 of 30 June 1999, including the 8, 6 and 4 percent targets, the tolerance bands, the IPCA index choice and its geographic coverage, the 1998 average CPI inflation of 1.7 percent, the May 1999 Rio de Janeiro inflation-targeting seminar, and the statement that the 1999 target was met.
• The White House: Fact Sheet on International Support for Brazil, 13 November 1998: the $41.5 billion package total and its IMF, World Bank, Inter-American Development Bank and bilateral components, including the $5 billion United States contribution.
• Federal Reserve Bank of St. Louis: Brazil / U.S. Foreign Exchange Rate, Series DEXBZUS: every daily and monthly-average exchange rate figure on this page, from 1.1730 on 17 August 1998 through 1.8090 at year-end 1999, and the independently computed January and February 1999 monthly averages of 1.5120 and 1.9261.
• International Monetary Fund: World Economic Outlook database, Brazil real GDP growth series NGDP_RPCH: the real growth figures of 0.3 percent in 1998, 0.5 percent in 1999 and 4.4 percent in 2000. The same World Economic Outlook data API, with the series code substituted, supplies the other Fund figures on this page: LUR for unemployment (10.1 percent in 1998, 11.1 percent in 1999), BCA_NGDPD for the current account balance (around negative 2.8 to 4.5 percent of gross domestic product across 1996 to 1999), and NGDPD for nominal gross domestic product in dollars ($864.3 billion in 1998, $599.6 billion in 1999). Note that imf.org itself returned HTTP 403 to every direct request made while this page was prepared; this data API answered normally.
• World Bank: World Development Indicators, Brazil total reserves including gold, series FI.RES.TOTL.CD: every reserve figure on this page, from $60.2 billion at the end of 1996 to $35.9 billion at the end of 2001.
• Banco Central do Brasil: Sistema Gerenciador de Séries Temporais, Selic target rate, series 432: the daily Selic target from its first published value of 45 percent on 5 March 1999 through the step-down to 19 percent by 23 September 1999.
• Banco Central do Brasil: Sistema Gerenciador de Séries Temporais, Selic rate, monthly average annualized, series 4189: every monthly Selic figure on this page for periods before the daily target series begins, including the 19.23 percent August 1998 reading and the 41.60 percent October 1998 peak.
• Banco Central do Brasil: Sistema Gerenciador de Séries Temporais, Índice Nacional de Preços ao Consumidor Amplo (IPCA), series 433: every monthly inflation reading on this page, and the compounded full-year figures of 1.66 percent for 1998, 8.94 percent for 1999 and 5.97 percent for 2000, calculated directly from the twelve published monthly readings for each year.

Figures deliberately not stated. This page gives no dollar figure for the cost of any individual bank rescue during the crisis, no precise date for Gustavo Franco's resignation beyond the window his central bank's own account establishes, no Bovespa index level, no sovereign bond spread, and no specific size for Brazil's dollar-indexed domestic public debt, because no source verified for this page supplied a figure meeting this page's sourcing bar for that specific claim. Reports from the period describe a controversy involving the central bank's foreign-exchange support for two brokerage-linked banks, Marka and FonteCindam, around the time of the float; this page does not detail that episode because a reliable primary figure for its cost could not be verified in the time available, and a description built only on secondary summaries would not meet the standard the rest of this page holds to.

Method note: full-year IPCA figures are computed by compounding the twelve published monthly percentage changes for each calendar year from Banco Central do Brasil series 433, not taken from a pre-aggregated annual figure. The 1998 result of 1.66 percent computed this way is consistent with the 1.7 percent average CPI inflation Brazil's own central bank working paper cites for the same year, a useful cross-check between an independently computed figure and the primary source's own restatement of it. Where this page states a currency value in cents on the dollar rather than reais per dollar, that is this page's own arithmetic transformation of the Federal Reserve's published rate, shown for readability, not a separately sourced figure.

Everything above describes Brazil between 1994 and 2001 and nothing above describes Brazil now. The exchange-rate regime, the inflation-targeting framework's specific targets, and Brazil's fiscal and reserve position have all been revised repeatedly since 2001, so none of this is investment advice, a forecast, or a guide to any live Brazilian instrument.

Frequently Asked Questions