Chapter 4. Brazil’s Experience in Managing Capital Inflows: The Role of Capital Controls

R. Gelos
Published Date:
March 2014
  • ShareShare
Show Summary Details
Roberto Benelli, Alex Segura-Ubiergo and W. Christopher Walker 

This chapter presents Brazil’s experience in managing capital flows. It discusses the Brazilian authorities’ use of a broad range of instruments to deal with capital inflows, including foreign exchange (FX) intervention, macroprudential tools, fiscal adjustment, and capital controls. Although it is difficult to separate the individual effects of each policy measure, they seem to have collectively reduced the speed of exchange rate appreciation. Capital controls have affected mostly the composition of flows (higher foreign direct investment [FDI] and less fixed-income portfolio investment), rather than the overall quantity of flows. The IOF (Imposto sobre Operações Financeiras) also appears to have curtailed incentives for carry trades.

Brazil’s recent experience managing capital inflows has generated a substantial amount of interest, reflecting both the nation’s prominence as a major emerging market and the fact that it has been one of the front-runners in adopting capital flow management measures (CFMs). This chapter provides a comprehensive account of Brazil’s recent experience by reviewing the authorities’ policy response to the surge in capital inflows in the recovery from the global financial crisis. Given the difficulty in identifying empirically the effect of CFMs—and hence a proper counterfactual of what might have happened to capital inflows under different policy responses—this chapter refrains from strong causal inferences, but rather relies on a variety of empirical facts to disentangle the possible impact of CFMs. 1 The chapter is organized in three sections, beginning with some stylized facts about capital inflows into Brazil in the period 2009–11, followed by a description of authorities’ policy response, and concluding with an assessment of the role and effectiveness of capital controls.

Stylized Facts on Capital Inflows

Strong economic growth prospects for Brazil in the aftermath of the global crisis reinforced structural factors (including high nominal interest rates), resulting in large capital inflows and strong appreciation pressures (Figures 4.1 and 4.2). In 2010, gross capital inflows2 exceeded US$150 billion (6.8 percent of 2010 GDP), compared with US$92 billion in 2009. Gross capital inflows remained strong in 2011, exceeding US$15 billion per month.

Figure 4.1Gross Capital Inflows

(Billions of U.S. dollars; 3-month moving average)

Source: IMF staff calculations.

Note: IOF = Imposto sobre Operações Financeiras.

Figure 4.2Foreign Investment

(Billions of U.S. dollars; monthly average)1

Source: IMF staff calculations.

1 “2006 turbulence” indicates the period from April 2006 through June 2006; “Pre-Lehman” indicates the period from July 2006 through August 2008; “Post-Lehman crisis” indicates the period from September 2008 through March 2009; “Post-Lehamn recovery” indicates the period from April 2009 to date.

Deep capital markets and high interest rates make Brazil one of the preferred destinations for capital flows to emerging market economies. Brazil’s equity market is one of the world’s ten largest by market capitalization, containing several globally known firms, such as Petrobras, Vale, Embraer, and Odebrecht. The sovereign bond market is also deep, featuring a variety of maturities and types of bonds, including fixed-rate, floating-rate, inflation-linked, and foreign-currency-linked bonds. The total stock of domestic bonds outstanding exceeds US$900 billion. As a result, Brazil is a bellwether, attracting a large share of total inflows to emerging and developing economies and reflecting global views toward such economies as an asset class (Figure 4.3).

Figure 4.3Flows to Brazil: Share to and Correlation with Emerging Markets

Source: IMF staff calculations.

Foreign investors’ interest in Brazil is also influenced by the large derivatives market. Volume trading in the market for derivatives tripled from 2003 to 2011, transforming the domestic BM&F futures market into one the largest clearinghouses for derivatives in the world. The market for most classes of derivatives is liquid out to maturities of more than two years, with most of the trading concentrated in interest-rate and currency-based derivatives. There is also an active offshore market in nondeliverable currency forwards (NDFs) in the Brazilian real, centered on banks located in New York. There are offshore markets as well for several other “access” products that allow foreign investors to trade as if they were in Brazil. These include credit-linked notes and derivatives based on Brazilian interest rates, all settled in U.S. dollars.

As a result, Brazil has dominated capital inflows to Latin America, and accounted for a large share of global equity issuance in 2010. This was due in part to the record Petrobras issuance in the third quarter that year.3 In addition to FDI and equity flows, fixed-income inflows were steady during most of 2010, reflecting both “real money” flows as well as retail flows (especially from Japan), while external corporate bond issuance rose to near-record highs.

Policy Response

The Brazilian authorities used a broad range of measures to manage the surge in capital inflows, including exchange rate appreciation, reserve accumulation, and fiscal adjustment, as well as macroprudential and CFM measures.

Rationale for Managing Capital Flows

Large inflows in the context of a flexible exchange rate system and open capital account can lead to strong swings in the exchange rate. The Brazilian real appreciated by more than 40 percent against the U.S. dollar between December 2008 and July 2011, while the real effective exchange rate reached a level stronger than before the financial crisis (when it had already reached its highest level in 30 years). As a result of this rapid appreciation, exchange rate valuation assessments showed that the Brazilian real was substantially overvalued (Figure 4.4).

Figure 4.4Gross and Net Capital Inflows

Source: IMF staff calculations.

Note: NEER = nominal effective exchange rate.

Large capital inflows also raise the risk of disruptive sudden stops (Ostry and others, 2011). Such stops tend to be highly synchronized across countries, particularly among countries with open capital accounts. Brazil, in particular, has in the past experienced firsthand the consequence of volatile capital flows. This “excessive” volatility is well captured by an exercise that compares actual capital flows to Brazil with their predicted value based on a panel regression of capital flows to a range of pull and push factors (Figure 4.5).4

Figure 4.5Predicted and Actual Capital Inflows

Source: IMF staff calculations.

Reserve Accumulation

The Brazilian real would likely have appreciated even further without strong FX intervention (Figure 4.6). In this regard, the central bank (Banco Central do Brasil, or BCB) stepped up its intervention in both the spot and forward markets during the period of heavy inflows, often holding daily (or even twice-daily) auctions. From December 2008 to July 2011, Brazil’s stock of foreign reserves rose by more than $139 billion, to $346 billion. The BCB also intervened in the onshore forward market, buying U.S. dollars forward against the real.

Figure 4.6Nominal Exchange Rate, Central Bank Interventions, and IOF Imposition

Source: IMF staff calculations.

Note: FX = foreign exchange; IOF = Imposto sobre Aperaçóes Financeiras.

Fiscal Adjustment

Although fiscal adjustments are generally understood to be aimed at maintaining debt sustainability and containing inflation, their positive effect as a tool to deal with large capital inflows should not be underestimated. In 2011, the combination of budgetary adjustment and reductions in quasi-fiscal activities (including transfers from the Treasury to the national development bank) was equivalent to a structural fiscal adjustment of about 2 percent of GDP. As in many other emerging markets, the difficulty of estimating fiscal multipliers and the interest rate channel of monetary policy makes it difficult to assess trade-offs between monetary and fiscal policy. This uncertainty is reflected in the broad range of responses in a poll of domestic analysts conducted by the BCB (Figure 4.7). However, this survey also suggested that, according to the mode of the response distribution, a 1.2 percent of GDP fiscal adjustment announced would be equivalent to about 120 basis points of otherwise necessary monetary tightening that would have attracted even more inflows.

Figure 4.7Change in Policy Rate Equivalent to Fiscal Measures

Source: Central Bank of Brazil, based on a survey of market participants.

Macroprudential Instruments

In January 2011, the BCB introduced reserve requirements that sought to limit the short U.S. dollar positions of banks. The BCB noted that this measure was designed to reduce leverage in the system, particularly for banks with restricted Tier I capital but easy access to international liquidity. The measure required banks to deposit the equivalent of 60 percent of their short U.S. dollar positions—a form of short-term U.S. dollar borrowing—in cash at the central bank, at no interest. By accessing offshore U.S. dollar liquidity, domestic banks could facilitate the carry trade via the domestic currency futures market5 (see Appendix 4.1 for a description of the mechanisms of this trade). Besides reducing foreign currency leverage, the measure increased the cost to banks of becoming counterparties to other investors in the derivatives market, thus affecting carry trade liquidity. Banks’ short dollar positions decreased in the wake of this action.

Figure 4.8Change in Policy Rate Equivalent to Macroprudential Measures1

Source: Central Bank of Brazil, based on a survey of market participants.

1 Refers to a single change in policy rate (effective for four quarters).

The authorities also introduced other macroprudential measures aimed at restraining credit growth and further reducing incentives for external borrowing by banks. In December 2010, the BCB raised bank capital requirements for most consumer credit operations with maturities over 204 months, which apply primarily to car loans. The BCB also raised unremunerated reserve requirements on time deposits from 15 percent to 20 percent. The additional (unremunerated) reserve requirements for banks’ sight deposits were increased from 8 percent to 12 percent.

Although these measures were not designed specifically to deal with capital inflows, to the extent that they reined in credit growth and aggregate demand, they could help manage flows. This was of particular importance in light of the fact that, given Brazil’s already high interest rates by international standards, further increases in the policy rate would have attracted more capital inflows. Thus, while prudential measures are directed at maintaining stability, they also complement monetary policy by having a moderating influence on demand. In a survey, the BCB found that market participants believed, on average, these macroprudential measures were equivalent to an increase of about 90 basis points in the policy rate.

Capital Flow Management Measures

In response to heavy postcrisis inflows in 2009, Brazil imposed an entry tax on foreign purchases of domestic bonds and equities. The tax, known as the IOF (Imposto sobre Operações Financeiras), was imposed on October 19, 2009, at a rate of 2 percent on purchases by foreigners of domestic bonds and equities. Other types of capital inflows, including direct investment and foreign currency borrowing abroad by Brazilian banks and firms, were not initially affected. The finance ministry announced that the measure was intended to combat speculation in capital markets and to counteract the appreciation of the real, which it viewed as damaging export industries and employment. Rapid credit growth or monetary policy considerations were not cited as issues at the time of the introduction of the IOF.

Strong capital inflows and appreciation pressures prompted the authorities to adopt a battery of new measures, tightening potential loopholes in the IOF, over the course of 18 months (Table 4.1).

Table 4.1Regulatory Changes Affecting the IOF
October–November 2010The initial IOF was raised twice in October 2010, but only for flows into debt securities. It was first raised to 4 percent on October 5 and increased further to 6 percent on October 19. The authorities subsequently announced its extension to margin calls on derivative positions. New regulations were shortly introduced to register as new operations flows that initially came into equities (taxed at 2 percent) and subsequently migrated to debt securities (thereby avoiding the higher 6 percent tax), as well as the conversion of IOF-exempt purchases of American Depository Receipts traded at the New York Stock Exchange into stocks traded at the Sâo Paulo Bovespa stock market.
March 2011On March 28, the IOF was extended to foreign borrowing by corporations and banks for loans with a maturity of less than 360 days. One week later, the maturity was extended to 720 days as it became clear that the maturity of loans was shifting to avoid the higher IOF tax. For domestic banks, this measure complemented the reserve requirement on short foreign currency positions adopted in early January 2011, effectively restricting their ability to raise foreign currency borrowing and participate in the carry trade via the derivatives market.
April 2011On April 4, the National Monetary Council mandated that any changes in foreign exchange loans/bonds (that is, renovation, rollover, restructuring, or transfer) had to be registered as a new foreign exchange operation. This was intended to close another potential loophole as corporations and banks could modify initial longer-term agreements (exempt from IOF) into short-term loans that should have paid the IOF.
Sources: Central Bank of Brazil; and press reports.Note: IOF = Imposto sobre Operações Financeiras.
Sources: Central Bank of Brazil; and press reports.Note: IOF = Imposto sobre Operações Financeiras.

This was not the first time Brazil had employed CFMs on portfolio inflows—up until October 2008, it had levied a 1½ percent tax on bond (but not equity) inflows, but the authorities had eliminated the tax in response to the financial crisis. But the experience with capital controls subsequent to October 2009 stands out for the frequency and scope of the interventions as the authorities used an adaptive strategy to adjust to financial market creativity in bypassing the controls.

CFMs: Role and Effectiveness

Many researchers have examined the impact of capital controls on capital flows. Interest in such research rises during periods of heavy flows into developing economies. Brazil, which has often made use of CFMs during the postwar era, has been the focus of many of these studies. Most, such as Cardoso and Goldfajn (1998), conclude that CFMs can be effective in the short term, but not over the longer term. Carvalho and Garcia (2006), who also focused on the case of Brazil, came to a similar conclusion in finding a limited and temporary impact of CFMs. Work in other Latin American countries has yielded similar conclusions. Clements and Kamil (2009), looking at the case of Colombia, find the total volume of inflows unaffected by CFMs, but exchange rate volatility reduced.

The effectiveness of CFMs during 2009–11 in Brazil needs to be assessed against the initial objectives. The Brazilian authorities stressed the negative macroeconomic effects of excessive currency appreciation. Thus, one criterion for assessing the effectiveness of the controls would be whether total capital inflows, and hence exchange rate appreciation pressures, abated. However, there may be other reasons for introducing CFMs in the face of heavy inflows. These include the desire to limit hot money inflows and the need to prevent credit bubbles that could endanger financial stability. Accordingly, a second possible criterion would be whether the composition of capital flows changed. Finally, capital controls can provide some maneuvering room for monetary policy, which can be particularly useful in a country such as Brazil with high interest rates.6 This is a consequence of the “impossible trinity”—that a country cannot simultaneously have an independent monetary policy, a fixed exchange rate, and an open capital account.7 Assessment of the controls’ effectiveness might therefore focus on whether capital controls provided some additional room to maneuver for monetary policy. One way of testing this hypothesis is to analyze whether capital controls reduced the incentive for carry trade, as these are the types of capital inflows that more directly respond to interest rate differentials.

However, any such assessments are complicated by the difficulty of comparing the realized strategy to a counterfactual scenario without capital controls. In principle, econometric techniques would be able to control for other factors that could affect the dependent variables of interest (that is, exchange rate, total capital flows, and so on). However, the IOF is—to a large extent—endogenous to capital flows and to the exchange rate (that is, it is imposed when capital flows are high and the exchange rate is appreciating fast) and hence it would be difficult to assess the direction of causality. Finally, the IOF cannot be modeled as a dichotomous variable (defining an existing or absent characteristic) given frequent changes in rates and regulations affecting it, which are likely to generate problems of measurement error.

Total Capital Flows and the Exchange Rate

The pace of capital inflows remained high following the introduction of CFMs. Total net capital inflows averaged US$5.9 billion per month after their recovery from the financial crisis and before the imposition of the IOF in October 2009; they increased to a monthly average of US$8.5 billion from 2009:Q4 through 2010:Q3, when the IOF was reintroduced at a rate of 2 percent, and accelerated further to a monthly average of US$11.5 billion between 2010:Q4 and 2011:Q1, when the IOF was raised to 6 percent for flows into debt securities and several loopholes were being closed. Similarly, the average monthly pace of reserve accumulation remained on an upward trend, accelerating from US$4.2 billion during the first phase of the IOF to US$7 billion during the most recent phase of the IOF, thereby exceeding the previous record pace set before the financial crisis (Figure 4.9).

Figure 4.9Average Monthly Increase in Reserves

(in millions of U.S. dollars)

Source: IMF staff calculations.

Note: IOF = Imposto sobre Operações Financeiras.

However, the average rate of currency appreciation declined. The rate of average monthly exchange rate appreciation during October 2010–March 2011 (about 0.6 percent) was less than half of that in the period before the financial crisis, when appreciation pressures were also strong (Figure 4.10). Excluding the period of sharp depreciation during the financial crisis, and the rebound during the recovery period, it seems that the degree of exchange rate flexibility declined as the Brazilian real became more overvalued (Table 4.2). Given that the volume of inflows did not materially decline during this period, it may be that the slower pace of appreciation was a consequence mostly of larger intervention by the BCB.

Figure 4.10Average Monthly Rate of Real Appreciation against the US Dollar


Source: IMF staff calculations.

Note: IOF = Imposto sobre Operações Financeiras.

Table 4.2Average Monthly Flows, Reserves Accumulation, and Exchange Rate Dynamics
Net Capital Flows (U.S. dollars)Monthly Reserve Accumulation (U.S. dollars)Average Monthly US$/Real Exchange Rate Appreciation (percent)
Precrisis$5.2 billion$1.9 billion1.5
Crisis$−3.5 billion$ −0.6 billion−4.6
Recovery$ 5.9 billion$ 3.6 billion3.9
IOF 1$ 8.7 billion$ 4.6 billion0.4
IOF 2$11.5 billion$ 7.0 billion0.6
Sources: Central Bank of Brazil; and IMF staff calculations.Note: The precrisis period covers 2007:Q1–2008:Q3, when capital inflows were reaching peak levels; the crisis period covers 2008:Q4 and 2009:Q1, when capital flows turned negative and GDP collapsed; the recovery period covers 2009:Q2/Q3; “IOF 1” covers 2009:Q4 through 2010:Q3, when the Imposto sobre Operaçóes Financeiras (IOF) was imposed at 2 percent; and “IOF 2” covers 2010:Q4 and 2011:Q1, when the IOF was increased to 6 percent for debt securities and several loopholes were closed.
Sources: Central Bank of Brazil; and IMF staff calculations.Note: The precrisis period covers 2007:Q1–2008:Q3, when capital inflows were reaching peak levels; the crisis period covers 2008:Q4 and 2009:Q1, when capital flows turned negative and GDP collapsed; the recovery period covers 2009:Q2/Q3; “IOF 1” covers 2009:Q4 through 2010:Q3, when the Imposto sobre Operaçóes Financeiras (IOF) was imposed at 2 percent; and “IOF 2” covers 2010:Q4 and 2011:Q1, when the IOF was increased to 6 percent for debt securities and several loopholes were closed.

Composition of Capital Flows

Although total flows remained high despite CFMs, the IOF seems to have affected the composition of capital inflows (Figure 4.11). Average monthly FDI following the second round of IOF measures (“IOF 2”) tripled, to almost US$8 billion compared with previous periods. Portfolio inflows into equities (which are taxed at the lower IOF of 2 percent) also continued on an upward trend.8 Conversely, flows into domestic debt securities appeared to be strongly affected by the IOF, with a monthly average flow during the period only one-tenth of the previous rate. The IOF clearly seemed to have reduced the purchase of domestic real-denominated bonds by foreigners.

Figure 4.11Composition of Capital Flows

(Monthly Averages in millions of U.S. dollars)

Source: IMF staff calculations.

Note: IOF = Imposto sobre Operações Financeiras.

However, the fall in portfolio flows into domestic securities led to a large increase in external borrowing by residents. A major driver of these flows was likely the behavior of domestic banks, which substituted away from short U.S. dollar positions (subject to the new reserve requirements) and toward other forms of funding not subject to the IOF (Figure 4.12). This substitution kept FX onshore liquidity abundant and allowed the carry trade through the domestic derivatives market to continue.

Figure 4.12Private External Debt

(Billions of U.S. dollars)

Source: IMF staff calculations.

The increase in FDI inflows was due in large measure to an increase in intercompany loans, whose monthly average more than doubled, reaching close to US$3 billion (Figure 4.13). This suggests that there may have been some circumvention of the IOF via FDI, which was virtually exempt from the IOF (taxed at 0.38 percent). Garcia (2007) documented that in the past it had been common practice in Brazil during previous periods of capital controls to create domestic shell companies that received an intercompany loan from a foreign party (classified in the balance of payments accounts as FDI) and use the proceeds to invest in the local equity or bond market, thereby bypassing the IOF (Table 4.3).

Figure 4.13Average Monthly Intercompany Loans

(Millions of U.S. Dollars)

Source: IMF staff calculations.

Note: IOF = Imposto sobre Operações Financeiras.

Table 4.3Composition of Net Capital Inflows(U.S. dollars)
Foreign Direct InvestmentPortfolio into EquitiesPortfolio into Debt Securities
IOF 12.483.652.28
IOF 27.934.780.22
Sources: Central Bank of Brazil; and IMF staff calculations.Note: The precrisis period covers 2007:Q1–2008:Q3 when capital inflows were reaching peak levels; the crisis period covers 2008:Q4 and 2009:Q1 when capital flows turned negative and GDP collapsed; the recovery period covers 2009:Q2/Q3; “IOF 1” covers 2009:Q4 through 2010:Q3 when the Imposto sobre Operações Financeiras (IOF) was imposed at 2 percent; and “IOF 2” covers the period 2010:Q4 and 2011:Q1 when the IOF was increased to 6 percent for debt securities and several loopholes were closed.
Sources: Central Bank of Brazil; and IMF staff calculations.Note: The precrisis period covers 2007:Q1–2008:Q3 when capital inflows were reaching peak levels; the crisis period covers 2008:Q4 and 2009:Q1 when capital flows turned negative and GDP collapsed; the recovery period covers 2009:Q2/Q3; “IOF 1” covers 2009:Q4 through 2010:Q3 when the Imposto sobre Operações Financeiras (IOF) was imposed at 2 percent; and “IOF 2” covers the period 2010:Q4 and 2011:Q1 when the IOF was increased to 6 percent for debt securities and several loopholes were closed.

Despite the pickup in FDI inflows, the extension of the IOF (at 6 percent) to external borrowing appeared to reduce onshore FX liquidity substantially. Especially notable was the sudden widening of the spread between onshore U.S. dollar interest rates (the cupom cambial) and offshore rates, suggesting that the ability to arbitrage between onshore and offshore markets was severely curtailed (see Appendixes 4.1 and 4.2). This spread subsequently abated somewhat, though changes in other external factors (for example, global risk, commodity prices) complicate the impact assessment. However, it still appears that the expansion in the field of coverage of the IOF reduced the previous incentives for recomposition of inflows, in favor of a reduction in total flows.

Monetary Policy Independence and Incentives for Carry Trade

One useful approach to analyzing the impact of the IOF is to consider it from the perspective of a carry trade investor into Brazil.9 Foreign investors who aim to profit from high Brazilian interest rates have several choices:

  • The most straightforward approach is to invest funds directly in the domestic Brazilian bond market. Starting with a zero net position, this would be done by borrowing U.S. dollars, converting them to Brazilian currency at the spot exchange rate, and then using the proceeds to buy a Brazilian bond of a certain maturity.

  • A second approach would be to borrow a smaller amount of dollars, convert the dollars to reais, and use the proceeds to pay the margin on a forward currency contract (buying the real forward against the U.S. dollar) on the domestic Brazilian futures exchange. This transaction offers the same return profile as the outright bond purchase. The IOF would apply only to the margin posted to the exchange, becoming very small.

  • A third possibility, favored by investors who wish to avoid transactions onshore in Brazil, is to buy a nondeliverable forward contract (denominated in reais but settled in dollars) offshore (for example, in New York or London), normally from a global bank that makes a market in such contracts.

The first mechanism (direct purchase of a Brazilian bond) was immediately affected by the introduction of the IOF (as demonstrated by the reduction of flows into debt securities), but the other two were not. This left two important loopholes in the IOF. Although the purchase of domestic bonds lost attractiveness, investors flocked to the derivatives market, where the incidence of the IOF tax was minimal. A 6 percent tax on the margin (typically 10 percent or less) would reduce the IOF effective tax rate to 0.6 percent, insufficient to have any impact. The IOF’s favorable treatment of positions in the futures market contributed to the buildup of large long Brazilian real and short U.S. dollar positions by nonresident investors. Furthermore, the rate of return from NDFs offshore was also not affected, which continued to attract carry trade flows.

The authorities moved to close the loopholes by first imposing reserve requirements on banks’ short U.S. dollar positions and later by extending the IOF to foreign borrowing. The introduction of reserve requirements on short U.S. dollar positions limited the ability of banks to operate as counterparties for nonresident investors engaged in the carry trade. However, banks were able to offset their short U.S. dollar positions by borrowing abroad, and the loophole remained in place during the first quarter of 2011. But when the IOF was extended to foreign borrowing by corporations and banks in April 2011, the onshore U.S. dollar rate (cupom cambial) spiked to unprecedented levels and the corresponding return in Brazilian reais from onshore and offshore NDFs collapsed (see Figure 4.14), as banks were no longer able to conduct low-cost arbitrage between offshore U.S. dollar markets and the onshore cupom cambial.

Figure 4.14Brazil—Domestic and Futures-Implied Interest Rates

Source: Bloomberg, L.P.; and IMF staff calculations.

Note: NDF = nondeliverable currency forwards.

Collateral Damage: Downsides of CFMs

The usefulness of CFMs needs to be assessed against their costs.

  • Reducing desirable flows. Controls may end up creating barriers to the entry of all foreign savings—including for productive investments—that are needed in the context of Brazil’s low domestic savings and current account deficit (for example, long-term debt or trade financing). Indeed, shortly after the IOF was raised in October 2010, domestic government bond yields started to trend upward and bid-cover ratios fell in some public debt auctions.

  • Hampering financial sector development. For growing middle-income countries, capital controls may inhibit or reverse development of the domestic financial sector.

  • Distortions. Controls may create distortions that shut off legitimate hedging activities, with a potentially negative impact on the real economy. This is most apparent in the spike in the onshore U.S. dollar rate. For example, the effective interest rate charged to an importer wanting to buy U.S. dollars forward as part of normal hedging operations rises as a result of these distortions. That higher cost could induce the would-be hedger to incur greater exchange rate risk, which could amplify the magnitude of unexpected shocks.

  • Stability risks. Although capital controls are intended to stabilize or reduce inflows, making them less risky, they might in fact force capital to go through channels that are actually more destabilizing. Some examples include forcing inflows through derivatives markets rather than spot markets, diverting inflows through nonfinancial firms, and (possibly) through trade finance flows.


The usual policy prerequisites for the application of CFMs in Brazil were broadly met. The exchange rate had become overvalued on a multilateral basis. Fiscal policy was being tightened. Sterilized intervention had been undertaken to slow the pace of appreciation. Moreover, the authorities supported these steps with prudential measures to contain credit and funding risks.

Nonetheless, the Brazilian experience with capital controls provides a mixed picture of their usefulness. Effectiveness has to be defined relative to a specific policy goal, and the study considers three—restricting capital inflows, containing exchange rate appreciation, and carving out space for the autonomous working of monetary policy. The overall level of flows remained high despite the introduction of CFMs, but their composition changed, suggesting a number of channels for circumvention (derivatives market, intercompany loans, and so on). However, the last stage of tightening of the IOF seems to have been effective at curtailing onshore U.S. dollar liquidity, indicating that the tax may become more effective at insulating onshore from offshore markets. This could allow more room for monetary policy to react without attracting additional flows.

Appendix 4.1. Mechanisms of the Carry Trade in the Futures Market

This appendix describes the steps involved in a nonresident investor’s carry trade into the domestic futures market, and the related hedging operations by a resident counterparty. To the extent that this counterparty, typically a resident bank, hedges its currency risk in the underlying cash market, this trade results in the same balance of payments pressure that would arise if the carry trade were conducted directly in the cash market (for example, by purchasing domestic bonds). This mechanism relies critically on the resident counterparty’s ability to take the nonresident’s opposite position in the domestic futures market and hedge the resulting currency risk via FX borrowing (not subject to the IOF). The resident counterparty thereby provides liquidity to the nonresident investor’s trade and earns a (risk-free) arbitrage profit proportional to the spread between the domestic U.S. dollar rate implied by the domestic futures market (the cupom cambial) and the offshore U.S. dollar rate paid on external borrowing (typically, the London interbank offered rate plus a spread). The detailed steps are described below and depicted in Figure A4.1.

Figure A4.1Carry Trade via the U.S. Dollar Futures Market1

Source: IMF staff calculations.

Note: BRL = Brazilian real; FX = foreign exchange; OTC = over the counter.

1 The horizontal line denotes the time line. Rectangles above (below) the line denote positive (negative) cash flows; rectangles shaded dark gray (light gray) denote payoffs in U.S. dollars and local currency, respectively.

  • The chain of trades is initiated by a nonresident investor who buys a real futures contract in the domestic futures market (step 1 in Figure A4.1). That is, at maturity, the nonresident investor receives the agreed price in reais per U.S. dollar (settlement in the futures market takes place in local currency).

  • The nonresident investor’s counterparty is a resident investor, that is, a local bank that agrees to receive the value of the U.S. dollar at maturity against a payment in Brazilian reais (step 2).

  • The local bank could choose to maintain its short Brazilian real position, or could choose to hedge its currency exposure. This could be done by borrowing U.S. dollars offshore (step 3); for example, by drawing on its available credit lines. Because the resulting U.S. dollar liability at maturity matches the claim to receive U.S. dollars on the U.S. dollar futures contract, the bank is hedged against currency risk (that is, a zero net position).

  • External borrowing by the bank is recorded in the balance of payments as a capital inflow (step 4). If the central bank chooses to intervene, then the bank sells the proceeds from external borrowing into the spot FX market.

  • By investing the Brazilian reais proceeds from the sale of U.S. dollars on the spot market at the domestic interest rate, the bank is able to earn an arbitrage profit whenever the cupom cambial is higher than the interest rate paid on its external borrowing (step 5). The only risk that is potentially left for the bank is counterparty risk on the futures contract.

Appendix 4.2. Using Forward Prices and Spreads to Test the Effectiveness of Capital Flow Management Measures

This appendix describes a method for using forward prices in offshore markets to test the effectiveness of capital control measures. This method is especially useful for a currency such as the Brazilian real, for which forward contracts are heavily traded in both onshore and offshore markets.

The cleanest test for the impact of the IOF on foreign access to the Brazilian market makes use of offshore nondeliverable currency forward (NDF) prices. This is not the usual measure for testing or gauging the effectiveness of capital controls, but has some important advantages over alternative approaches. The usual tests for the effectiveness of capital controls are based on flows and movements in the exchange rate and reserves. These are useful but they suffer from the fact that the degree of the response cannot, in general, be used to infer the relative effectiveness of the measure. Forward or “basis” spreads, at least in principle, can be used to gauge the implied arbitrage costs that a particular measure inserts into the market.

Futures-implied interest rates can be constructed, and compared with actual interest rates. An offshore forward-implied interest rate can be calculated as:

where interest rates are annualized three-month rates, i$ is the U.S. dollar interest rate (90-day Libor), e is the spot exchange rate, and f90, on is the 90-day NDF, quoted in reais per U.S. dollar. An alternative measure is the onshore future-implied interest rate in Brazilian reais, based on the onshore futures contract, f90,on. This implied rate is calculated as:

Either of these can be compared with the onshore Brazilian three-month interest rate:

These measures can be used to determine “basis spreads,” as either BSoff = (iBRL,offiBRL), or BSon = (iBRL,oniBRL). Full covered interest parity would entail that both of these measures be zero. For most emerging market economies, however, basis spreads are not zero, even under normal market conditions. The departure from covered interest parity, and the existence of a negative basis spread, typically reflects the effects of counterparty credit concerns, global interbank stresses, and transaction costs, possibly including costs induced by capital controls.10 For this discussion, the issue is whether the change in regulations induces a change in basis spreads and, if so, how large a change and in what direction.

If the IOF is effective in breaking the link between domestic and foreign fixed-income markets, or in inserting a wedge between the two, this should be evident in interest rate spreads. If the new regulations eliminate arbitrage, or impose a cost of arbitrage between domestic and offshore markets, then there should be a difference between the implied interest rate in reais available offshore through the NDF market and the interest rate in reais available onshore in Brazil. The implied interest rate in reais should be lower offshore, where the IOF cannot be collected. The basis spread derived from NDF trading should become negative, entailing a lower-than-market interest rate in reais. If the 6 percent IOF is fully binding on all domestic instruments in Brazil, and if there had been full arbitrage before it was imposed, then the basis spread should widen by 2 percent on instruments with a one-year maturity.

In April 2011, both measures of the futures-implied Brazilian interest rate diverged significantly from the onshore rates available in Brazil (Figure A4.2). This divergence, which was widely noted by market participants, indicated that the returns to the carry trade from the U.S. dollar into Brazilian fixed-income markets had been significantly restricted. Based on the timing of the divergence, and the major role played by banks in arbitraging between the spot and futures markets, it appeared that the restrictions placed on bank short U.S. dollar positions in March 2011 had been at least partly responsible for the shift.

Figure A4.2Expected Change in Relative Interest Rates if IOF Is Fully Binding

Source: IMF staff.

Note: BRL = Brazilian real; IOF = Imposto sobre Operações Financeiras; NDF = nondeliverable currency forward.


    BabaNaohiko and FrankPacker2008Interpreting Deviations from Covered Interest Parity during the Financial Market Turmoil of 2007–08,BIS Working Paper 267 (Basel: Bank for International Settlements).

    CardosoEliana and IlanGoldfajn1998Capital Flows to Brazil: The Endogeneity of Capital Controls,IMF Staff Papers Vol. 45 No. 1.

    CarvalhoBernardo and MarcioGarcia2006Ineffective Controls on Capital Inflows Under Sophisticated Financial Markets: Brazil in the Nineties,NBER Working Paper 12283 (Cambridge, Massachusetts: National Bureau of Economic Research).

    ClementsBenedict and MagdaKamil2009Are Capital Controls Effective in the 21st Century? The Recent Experience of Colombia,IMF Working Paper 09/30 (Washington: International Monetary Fund).

    HerreraLuis Oscar and RodrigoValdes2001The Effect of Capital Controls on Interest Rate Differentials,Journal of International Economics Vol. 53 pp. 38598.

    OstryJonathan D.Atish R.GhoshKarlHabermeierLucLaevenMarcosChamonMahvash S.Qureshi and AnnamariaKokenyne2011Managing Capital Inflows: What Tools to Use?IMF Staff Discussion Note No. 11/06 (Washington: International Monetary Fund).

This is particularly the case in assessing the impact of capital control measures such as the IOF, which is highly endogenous. In addition, the absence of a sufficiently long time series precludes obtaining robust results with the use of econometric techniques.

Gross capital inflows are defined as foreigners’ net direct investment plus portfolio investment and other flows.

Foreigners’ subscribed participation in the recapitalization of Petrobras amounted to US$14 billion.

The details underpinning this panel regression analysis are provided in IMF (2011).

Because the IOF applied only to actual payments made in derivatives transactions, rather than to the notional amount of the transactions, carry trading via futures became more attractive vis-à-vis outright purchase of domestic fixed-income securities.

See, for example, Herrera and Valdes (2001).

Although Brazil has a flexible exchange rate, the desire to prevent excessive appreciation suggests that full FX flexibility ceases to be an option for the authorities to manage inflows in the short term. When this point is reached, the dilemmas of the impossible trinity become applicable.

Flows into equities were bolstered by large initial public offerings (Visa and Santander in 2009) and the recapitalization of Petrobras in 2010. Eliminating these one-off transactions, the impact of the IOF on the volume of equity inflows does not appear to have been large.

These may also be described as “carry trade” investors, to the extent that they acquire Brazilian real fixed-income assets to take advantage of the yield differential between the Brazilian currency and their home currency. This type of investment necessarily entails exposure to currency risk—effectively the investor is betting that any depreciation of the destination currency will be less than the relative interest rate differential. To be sure, the domestic currency may also appreciate against the carry trade funding currency, as is often the case during episodes of strong risk appetite in global markets.

See, for example, Baba and Packer (2008).

    Other Resources Citing This Publication