Brazil: Selected Issues Paper
Author: International Monetary Fund. Western Hemisphere Dept.
Publication Date:
12 May 2015
eISBN:
9781475521320
Language:
English
Keywords:
ISCR; CR; minimum wage; Brazil; wage; bond yield; U.S. dollar; FX repo; intervention program; channeling FX hedge; FX derivative; Exchange rates; Currency markets; Currencies; party FX futures contracts
Download PDF (2.7 MB)

This Selected Issues paper examines infrastructure investment in Brazil. Brazil has inferior overall infrastructure quality relative to almost all its export competitors. Brazil’s infrastructure endowment ranks low by international standards, and its low quality affects productivity, market efficiency, and competitiveness. Areas in which Brazil’s competitiveness has lagged include, but are not limited to, education, innovation, governance, and justice. Brazil’s infrastructure gap has become a major obstacle to growth and filling this gap will entail increasing investment and also stepping up other reforms.

Abstract

This Selected Issues paper examines infrastructure investment in Brazil. Brazil has inferior overall infrastructure quality relative to almost all its export competitors. Brazil’s infrastructure endowment ranks low by international standards, and its low quality affects productivity, market efficiency, and competitiveness. Areas in which Brazil’s competitiveness has lagged include, but are not limited to, education, innovation, governance, and justice. Brazil’s infrastructure gap has become a major obstacle to growth and filling this gap will entail increasing investment and also stepping up other reforms.

Assessment of Foreign Exchange Intervention1

This chapter assesses the Central Bank of Brazil’s (BCB) foreign exchange intervention since June 2013, with a particular focus on the program of pre-announced interventions launched in August 2013 and in effect through end-March 2015. We find that the BCB used adequate tools to contain excessive exchange rate volatility following the Bernanke tapering speech; preannouncing the program succeeded in lowering volatility by reassuring investors with open FX positions; the size of the program and the maturity of its instruments likely filled a market gap beyond increasing the supply of hedge at the margin. However, the program may have slowed exchange rate convergence to more competitive levels. Its continued renewal—including during periods of low global volatility—raised questions about the BCB’s objectives and prevented the real’s depreciation to its equilibrium level. Our recommendation is not to continue the program beyond March. Further intervention would be limited to episodes of excessive volatility.

A. Background

1. Intervention in foreign exchange markets is a well established policy tool in many countries. Central banks in a large number of countries have intervened in foreign exchange markets since June 2013, mostly through purchases or sales of foreign currencies in spot markets (e.g. China, Malaysia, Russia, Singapore, Thailand and Turkey). Such transactions are typically sterilized through the issuance/sale or purchase of domestic debt securities. Sterilized intervention can be successful at influencing exchange rate movements to the extent that private market participants do not fully take advantage of the resulting arbitrage opportunities. Reasons why this may not happen are imperfect capital mobility as well as portfolio rebalancing or signaling effects.2

2. The BCB has intervened frequently in foreign exchange markets since the adoption of the floating exchange rate regime in January 1999.

  • In 2001–2002, pressures on the Brazilian real intensified amid three major shocks—spillovers from the Argentine debt crisis, the ‘9.11’ market jitters in the U.S., and public debt solvency concerns following the presidential election. The real depreciated by up to 44 and 71 percent in 2001 and 2002, respectively, as international reserves fell below US$40 billion. The BCB responded by using U.S. dollar linked debt instruments (Treasury securities (NTN-D) and BCB notes (NBC-E)) in support of the currency, at a high costs to public accounts. The more U.S. dollar linked debt instruments were used, the less they were seen as good substitutes for actual foreign currency by the public. The increased perception of credit risk carried by these instruments caused their rollover rates to drop.

  • From April 2002, the BCB replaced U.S. dollar linked instruments with Brazilian FX swaps, which were perceived as carrying a lower credit risk because they were traded and settled at the BM&F (the mercantile exchange) and offered daily margin adjustments (Figures 1 and 2);3 However, market participants continued to fear sovereign defaults which could endanger the payments of U.S. dollar linked debts and the convertibility of outstanding FX swaps. As onshore futures prices dropped below spot rates and offshore forwards, perceived convertibility risk pushed up the cupom cambial which reached a peak at end-2002 (Figures 3 and 4).4, 5

  • Since 2003, as the newly elected government strengthened the macroeconomic framework and resilience to external shocks, the BCB gradually unwound the stock of FX swaps; the outstanding amount of U.S. dollar linked debt instruments also decreased substantially by end-2005, reducing the sensitivity of public debt to currency movements;

  • Between 2006 and 2012, besides a short period following the Lehman Brother’s bankruptcy, the Brazilian real came under appreciation pressure amid buoyant capital inflows. The BCB began to engage in reverse FX swaps and spot dollar purchases, building international reserves to about 17 percent of GDP. During September 2008–May 2009, the BCB intervened via a range of tools simultaneously, including spot dollar sales, auctions of FX swaps and FX repos, and even indirect dollar loans to Brazilian firms through the banking system (Stone and others, 2009).

Figure 1.
Figure 1.

Foreign Currency Intervention

(In billions of U.S. dollar)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil and Fund staff calculations.
Figure 2.
Figure 2.

USD Linked Debt and Foreign Reserves

(In billions of U.S. dollar)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil; CEIC; and Fund staff calculations.
Figure 3.
Figure 3.

Cupom Cambial, FX Rate, and FX Swap

(In percent and billions of U.S. dollar)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil; Bloomberg; and Fund staff calculations.
Figure 4.
Figure 4.

Convertibility Risk

(Difference between offshore NDF and onshore NDF (DNDF), in percent of DNDF, annualized)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Bloomberg and Fund staff calculations.

B. Intervention Tool Kit

3. Sterilized intervention in spot markets is one instrument at the disposal of the BCB but by no means the only one. The BCB uses a variety of instruments to intervene in foreign exchange markets, including outright USD sales, FX repos and FX swaps. However, it has not intervened in spot markets since May 2012, in part due to the spot market’s limited size and access. While access to domestic spot markets is restricted to chartered banks, laws preclude trading or delivering the real offshore, and domestic bank accounts denominated in foreign currency are forbidden by law.

4. Intervention in Brazil’s large and vibrant derivatives markets has become more frequent than spot intervention in recent years. Brazil’s large derivatives markets developed amid investor interest in trading interest and exchange rate risk onshore (including due to Brazil’s history of high inflation, high nominal interest rates and devaluations). Foreign investors, in particular, have used derivatives to hedge exposures, including from carry trades. Today, Brazil’s derivatives markets are among the largest in the world.6 In particular, Brazil has active FX futures markets (BM&F BOVESPA) and dollar forward OTC markets (CETIP) whose trading volumes are four times larger than those in domestic spot markets (Figure 5). Importantly, the Brazilian exchange regime prohibits financial instruments traded in Brazilian markets from settling in foreign currency with a few exceptions.7 As a result, policymakers can make use of a highly liquid market for FX derivatives that settle in local currency.

Figure 5.
Figure 5.

Ratio of Daily Turnover in Derivatives Markets to Spot Market

(As of April 2013, net-net basis)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: BIS Triennial Central Bank Survey and Fund staff calculations.

5. Two important instruments in the BCB’s toolkit are the so-called Brazilian FX repos and Brazilian FX swaps:

  • The most frequently used instrument is the Brazilian FX swap, a non-deliverable future settled in local currency (unlike conventional cross-currency swaps). While other countries (e.g. Colombia, Korea, South Africa, and Mexico) have intervened in foreign exchange markets via derivative products, the FX swap is unique in that it settles in local currency. Formally, the BCB announces details on each FX swap auction one business day in advance of receiving market participants’ bids through the SELIC system (“Special System for Settlement and Custody”).8 As contracts mature, it pays its counterparts the observed exchange rate variation plus the pre-set cupom cambial and receives the ex ante SELIC rate in return.9 At the auction, participants bid the cupom cambial, since SELIC is known. While the FX swap has no direct impact on international reserves, the central bank implicitly takes a short dollar position as part of the transaction.

  • The Brazilian FX repo is akin to a conventional FX swap, resembling a dollar credit line. It has traditionally been used to provide FX liquidity to the market during periods of seasonal shortages. Formally, the BCB conducts an auction to sell spot dollars to its counterparts and receives Brazilian reais as collateral with an agreement to repurchase these dollars at maturity according to the auctioned forward rate.10 In other words, the BCB pays the SELIC rate and receives the cupom cambial. While FX repos temporarily reduce international reserves, these recover fully upon maturity of the FX repo; both spot and forward rates are set at the time of the auction.11

6. The FX swap can be used as a substitute to spot market intervention to the extent that convertibility risk is negligible. FX swaps provide hedge to investors with open real positions, thus directly bidding down the forward exchange rate. The crucial difference to non-deliverable forward contracts (NDF) is that the Brazilian FX swap settles in local currency without an exchange of notional principal. In other words, at time of settlement, the BCB pays its counterparty the additional amount of reais necessary to keep the dollar value of her initial (notional) real position unchanged (that is, the excess depreciation of the real relative to the forward rate implicit in the auction parameters). From the investor’s perspective, the FX swap is thus a good substitute for spot dollars (or an NDF) to the extent that the real will be fully convertible to dollars at the date of settlement. This rests on the expectation that the BCB will provide spot dollars at maturity if necessary and that convertibility will otherwise not be restricted (e.g. by capital outflow controls, a decision to alter the FX regime or rapid depreciations around the settlement date).12

7. FX swaps are well suited to meet demand for hedge while spot dollar supply may increase depending on market participants’ willingness to “transform” the instrument. Given that FX swap contracts are settled in real and do not involve the exchange of principal, any given swap contract can be settled without recourse to foreign exchange reserves.13 The instrument is thus well suited to meet demand for hedge or speculative FX exposure amid excessive volatility in foreign exchange markets. However, a different situation arises if there is a need to service either current obligations or capital outflows. In this case, the BCB can only substitute spot intervention with FX swap intervention to the extent that market participants are willing to bring dollar liquidity onshore as a counterpart to their long dollar derivative position. In normal times, commercial banks can fill this role (Garcia and Volpon, 2014).

8. Intervention through FX swaps directly targets the market where price discovery takes place. Garcia et al (2014) show that the Brazilian FX futures market dominates the spot market in terms of price discovery; it adjusts more quickly to the arrival of new information and is faster to recover to equilibrium in case of a shock to fundamentals. FX swap auctions take advantage of the powerful distribution channel to spot and derivatives markets. Institutions that purchase FX swaps, such as financial institutions and institutional investors, can take offsetting exposures in the CETIP OTC market or BM&FBOVESPA exchange and thereby provide FX cover to other market participants such as corporates with open FX positions.

9. FX swaps may fill a market gap for hedging instruments with long maturities when extended in large volumes. Intervention based on FX swaps may fill a market gap for long-maturity hedging contracts. Private party FX futures contracts traded on BM&F BOVESPA tend to have shorter maturities, and OTC market derivative products with longer maturities are not sufficiently liquid to meet steep increases in demand.14 In particular, the net long dollar derivative position of corporates and households alone increased seven-fold from US$7 billion at end-May 2013 to 48 billion by end-October 2014, while non-residents shifted their US$6 billion short position into a US$36 billion long position over the same period.

C. Intervention Program Following the ‘Tapering Talk’

10. Volatility in FX markets following the Fed’s tapering talk led the BCB to intervene actively since June 2013. The Brazilian currency was hit hard following Governor Bernanke’s “tapering” speech on May 22, 2013. Following a period of discretionary intervention between June and August 2013, the BCB announced a program of pre-announced interventions on August 22. The instruments of choice have been Brazilian FX swaps and, to a much lesser degree, FX repos (Figure 6).

  • The pre-announced intervention program initially consisted of daily auctions of foreign exchange swaps and repos equivalent to US$3 billion a week—US$500 million in FX swaps from Monday to Thursday and US$1 billion in repos on Friday—through end-2013;

  • In December, the program was extended on a smaller scale—US$200 million in swaps from Monday to Friday and FX repos only “on demand”—until end-June 2014. In late June, the program was extended on unchanged terms until end-2014;

  • By the end of 2014, the notional balance of outstanding FX swaps amounted to US$110 billion, which were primarily bought by commercial banks (52 percent) and domestic institutional investors (45 percent), with foreign investors directly holding only a small portion (Figure 7).

  • The FX repos with commercial banks amounted to US$17 billion by end-2013, but were unwound by end-August 2014. To provide FX liquidity to the market during a period that is generally market by capital outflows every year, the BCB carried out auctions for US$10.3 billion in FX repos in December 2014; and

  • At the end of December 2014, the BCB announced a further extension of its FX swap program, but with a reduced scale and period. It decided to offer US$100 million per day until March 2015, while previous extensions lasted for six months with twofold volume.

Figure 6.
Figure 6.

Outstand Balance of FX Swap and FX Repo since 2013

(In billions of U.S. dollar)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil and Fund staff calculations.
Figure 7.
Figure 7.

FX Swap Exposure by Counterparties

(In billions of U.S. dollar)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Source: Itau (2015).

11. Recourse to FX repos and FX swaps paired with the policy preannouncement was adequate as demand for hedge dominated. Convertibility risk remained negligible throughout the current intervention episode as shown in Figure 4. Exchange rate pressures did not appear to be driven by outflow pressures as FDI and portfolio inflows remained stable amid the BCB’s monetary policy tightening cycle and measures to ease the IOF tax (Figure 11). While FX repos addressed temporary liquidity needs, FX swaps bid down the price of hedge for open FX positions and may have filled a market gap as discussed above.

Figure 8.
Figure 8.

FX Swaps at Maturity

(In billions of U.S. dollar, as of January 20, 2015)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil and Fund staff calculations.
Figure 9.
Figure 9.

Net Increase of FX Swap

(In billions of U.S. dollar)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil and Fund staff calculations.
Figure 10.
Figure 10.

Rollover Rate of FX Swap

(In percent of maturing FX swaps)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil and Fund staff calculations.
Figure 11.
Figure 11.

Capital Flows

(In millions of U.S. dollar)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil and Fund staff calculations.

12. The large outstanding stock of FX swaps runs the risk of losses on FX swaps. Given the depreciation pressure in the FX market, the BCB could make considerable realized losses from the ongoing intervention program. The loss would, however, be more than offset by unrealized valuation gains on international reserves.

13. The maturity structure of the outstanding stock of FX swaps is distributed relatively evenly going forward (Figure 8). About US$10 billion in outstanding FX swaps is set to mature during the first business day of each month until November 2015. The FX swaps sold in January 2015 will mature in September and December 2015. In addition, the central bank has continued to roll over contracts worth US$10.4 billion maturing in February 2015 at a pace of 10,000 contracts per day.

14. Although auctions of new swaps follow preannounced schedules and amounts under the program, the BCB has the discretion to determine the rollover rate of maturing swaps. Initially, maturing swaps were fully rolled over, except those due on November 1, 2013; but once exchange rate volatility started to abate around April 2014, the rollover rate was reduced to the point that the outstanding notional value of FX swaps decreased on a net basis in June 2014 (Figure 9). As the Brazilian real came under renewed pressure in the run-up to the first-round presidential elections in October 2014, the BCB increased rollover rates back to almost 100 percent (Figure 10).

D. Distribution Channels to Spot and Derivative Markets

15. Commercial banks have been key players in channeling FX hedge to market participants due to their exclusive ability to trade in both FX spot and derivatives markets:

  • Commercial banks have sold DDI&USD futures in the BM&F BOVESPA and U.S. dollar forwards in the OTC market (registered at the CETIP exchange) against their long swap positions to market participants (mainly corporates and foreign investors), allowing them to hedge their BRL exposures. At end-2014, banks’ short positions in the two FX derivative products totaled about US$20 billion and US$30 billion, respectively. Summing these positions up against their FX swap position, commercial banks’ net long position in FX derivatives amounts to US$8 billion;

  • In the spot market, commercial banks have sold dollars, either obtained through FX repos from the BCB or borrowed from abroad at Libor plus a spread (Figure 12). At end-2014, commercial banks held short spot positions of US$28 billion, compared with their long spot positions of US$5 billion in May 2013.

  • Overall, commercial banks’ net open position currently reaches US$20 billion, or 0.7 percent of banking system assets.

Figure 12.
Figure 12.

FX Flows Financed by FX Repo and Commercial Banks’ Short FX Position

(In millions of U.S. dollar)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil and Fund staff calculations.

16. Corporates, foreign investors, and domestic institutional investors are the main “end-users” of the BCB’s swap program (Figure 13). As the main beneficiary of a lower cost of hedge, corporates’ long position in FX derivatives totals US$48 billion, or 44 percent of the stock of outstanding FX swaps. Foreign investors, in turn, hold long positions in FX derivatives of about US$34 billion; this amounts to about 19 percent of foreign holdings of domestic debt securities. Domestic institutional investors purchase a large share of FX swaps from the BCB but also sell futures and forwards to corporates and foreign investors such as commercial banks, partially dollarizing their portfolios but keeping long positions relatively small at US$20 billion.

Figure 13.
Figure 13.

“End Users” of FX Swaps

(In percent of the BCB’s FX swap position)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Central Bank of Brazil and Fund staff calculations.

E. Effectiveness

17. The intervention program aims at taming volatility in the foreign exchange market. A key factor in the design of the policy has been the policy’s pre-announcement. Both foreigners’ and private domestic market participants’ demand for hedge increased following Bernanke’s tapering speech amid mismatches between local currency assets and foreign currency liabilities on corporate balance sheets (see SIP on corporates). Preannouncing the program aimed at lowering excess volatility by reassuring investors with open FX positions; in addition, the size of the program and the maturity of its instruments likely filled a market gap beyond increasing the supply of hedge at the margin. Figure 14 illustrates that the volatility in the Brazilian real indeed dropped substantially on impact of the program announcement while later program announcements had less discernible effect. In order to substantiate this finding, we move to a simple regression approach.

Figure 14.
Figure 14.

FX Implied Volatility Relative to Peers

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Bloomberg and Fund staff calculations.

18. A simple regression approach is taken to assess whether the BCB achieved its policy objective (Table 1). We define the dependent variable as the implied volatility in the Brazilian real relative to the implied volatility in the currencies of a number of peers.15 The data is daily, ranging from January 2010 to March 2014 in the baseline sample and to June 2014 in a robustness check.16 The main variables of interest are dummies that indicate the time of announcement of the initial program as well as of its extensions. The dummies take the value one starting on the day following the relevant announcement and the value zero before that. The estimated coefficients can thus be interpreted in a cumulative fashion and indicate the contribution of each announcement to the volatility in the Brazilian real relative to peer currencies throughout the remainder of the sample period.17 Control variables include the Bernanke “tapering” announcement on May 22, 2013, the Federal Reserve’s “delay taper” announcement on September 18, 2013, announcements to scale back the IOF tax as well as macro (policy) controls. The latter include the SELIC rate (to proxy for monetary policy) and a consumer confidence indicator (to proxy for volatility resulting from domestic confidence).18 Finally, alternative regressions 2 and 3 additionally include roll-over announcements by the central bank as controls, and regression 3 extends the sample period to June 2014.

Table 1.

Explaining Volatility

Dependent variable: Implied volatilitty of Brazilian real as a share of implie volatility in peer currencies

article image
* p<0.1, ** p<0.05, *** p<0.01** Peers are defined as Chile, Colombia, India, Indonesia, Mexico, Peru, South Africa and Turkey.Source: IMF.

19. Our findings suggest that the intervention program successfully tamed volatility in foreign exchange markets in the aftermath of Bernanke’s tapering speech. The Bernanke speech on May 22, 2013 increased volatility by between 6 and 9 percent relative to Brazil’s peers. In other words, the announcement to taper the Fed’s asset purchases had a relatively stronger impact on the real than on the currencies of Brazil’s peers. Interestingly, even the Fed “delay taper” announcement increased the real’s relative volatility by between 8 and 12 percent.19 At the same time, the regressions suggest that the initial announcement of the intervention program reduced volatility by 13 to 15 percent relative to peers; combined with a further reduction through the December 18 announcement, the program managed to almost entirely offset the excess-volatility effects of the Fed announcements. Interestingly, the announcement of IOF tax reduction on derivatives transactions reduced volatilily as well; as did tighter monetary policy, while worsening consumer confidence tended to increase volatility relative to peers. In other words, the effectiveness of intervention can, to a degree, be linked to the presence of other reinforcing policies.

20. However, as volatility subsided in international markets, the extension of the intervention program in June 2014 sent conflicting signals about the program’s objectives. Volatility in global financial markets dropped to lows by mid-2014.20 Indeed, the regression results in Table 1 suggest that the two announcements regarding the program’s extension in June 2014 reduced volatility only at the margin.21 One reason may have been the fact that the outstanding stock of FX swaps already accounted for some 80 percent of short term external debt as of end-June 2014. The decision to extend the program therefore raised questions among market participants as to the BCB’s policy objectives and the program’s nature as an extraordinary and temporary measure.

21. While not an explicit objective of the central bank, the intervention program likely supported the real in a very similar way as sterilized spot intervention would have done. As discussed previously, sterilized spot intervention is effective to the extent that private market participants do not offset the arbitrage opportunity resulting from increasing deviations from UIP following the intervention. If indeed successful, the change in the spot rate feeds through to the forward rate. Assuming that the intervention through NDFs is effective, it would operate as follows: the NDF would directly bid down the forward exchange rate. This provides an arbitrage opportunity as the cupom cambial increases, inducing market participants to bring dollar liquidity onshore, thus supporting the spot exchange rate. Ex ante, there is no reason to believe that NDFs would be less effective in moving the exchange rate level, especially when derivatives markets are significantly more liquid than spot markets. The FX swap, in turn, is no different from the NDF aside from it being settled in domestic currency. In other words, it is unlikely to affect the exchange rate in a different way as long as convertibility concerns are negligible. Indeed, Kohlscheen and C. Andrade (2013) find that discretionary intervention using FX swaps during 2012 had sizable effects on the exchange rate. A glance at the real’s evolution relative to peer currencies suggests that the current pre-announced intervention program also strengthened the real: it appreciated significantly relative to its peer currencies in the immediate aftermath of the program announcement (Figure 15).

Figure 15.
Figure 15.

Foreign Exchange Rate

(August 22, 2013 = 100)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Sources: Bloomberg and Fund staff calculations.

22. Given that the intervention program was largely pre-announced, however, assessing its impact on exchange rate levels is difficult. Starting on August 22, market participants would have priced in any pre-announced action by the central bank in advance of it being taken. Even before August 22, market participants may have at least expected a continuation of the discretionary interventions before the pre-announced intervention program was announced. The only elements of discretion for policymakers were the extension announcements in December 2013 and June 2014, as well as the monthly roll-over rates for maturing swaps.

23. Some elements of surprise are found in the initial program announcement, the announcements to extend the program and in some roll-over announcements. We use intraday data to search for surprises both on the day of each announcement and on the day the auction was actually implemented (Table 2). These event studies confirm the regression results that the intervention announcements had a significant impact on volatility. The August 22 announcement reduced volatility by about 3 percent within 3 hours and more than 5 percent within 24 hours, and the real appreciated by 2–2.5 percent in both forward and spot FX markets within 3 hours and by around 3.5 percent within 24 hours after the announcement. It appreciated a further 1.4 percent once the program was announced to continue on December 6, 2013. The June 6th 2014 announcement to continue the program further appreciated the real by about 1 percent. In addition, announcement effects were observed before and/or after other events, although these are mostly found to be small.

Table 2.

Impacts of Important Announcements of FX Swap Auctions 1/

(In percent, the appreciation of the real before and after the announcement)

article image
Sources: Bloomberg and Fund staff calculation.

Yellow, pink, and green color show the appreciation of more than 0, 1, and 3 percent, respectively.

F. Limits to Intervention

24. An interesting question is where the limit of the intervention program lies. The notional value of the outstanding stock of FX swaps stood at some US$110 billion at end-2014. This is equal to about 29 percent of international reserves. However, because FX swaps are settled in local currency, the stock of reserves is unaffected by FX swap based intervention.

25. The stock of reserves can be seen as a natural limit for the intervention program. While FX swaps allow the central bank to intervene in the market without committing foreign reserves, these implicitly serve as collateral for the transaction. 22 Any potential buyer would presumably purchase FX swaps under the premise that both principal and proceeds can be converted into dollars at the time of maturity. If the notional outstanding stock of FX swaps becomes larger than available reserves, the central bank can no longer guarantee that dollar liquidity will be available once the contract matures. Moreover, FX swaps may become unattractive instruments before the notional stock outstanding actually reaches 100 percent of reserves. Historical experience, nevertheless, shows that the central bank has approached this natural limit in the past. In 2003, the ratio of the FX swaps to international reserves reached a high of about 90 percent in March and April, leaving little room for the BCB to intervene further.

26. Defending the public sector’s net creditor position in foreign exchange would likely limit the program at an earlier stage.23 A net creditor status can work as an insurance against capital account crises that involve exchange rate depreciations. The higher the net creditor position, the stronger the fiscal position in a crisis scenario, everything else equal. In Brazil, the public sector has held a net creditor position in foreign exchange since 2007. Table 3 illustrates that its positive territory is still comfortable at a surplus of about US$141.3 billion. Everything else equal, the net creditor position would thus be eroded if the notional value of the outstanding stock of FX swaps reaches around US$250 billion.

Table 3.

Public Sector Net Creditor Position

(In billions of U.S. dollar)

article image
Source: Authorities’ Data.

27. Reserve adequacy would be an argument to limit the outstanding stock of swaps even before the net creditor position is compromised. To the extent that market participants view reserves as collateral for FX swaps, one may want to examine reserve adequacy after subtracting FX swaps from the stock of reserves. Figure 16 illustrates that reserves are currently more than adequate from the perspective of the Fund’s reserve adequacy metric. Even when subtracting the stock of FX swaps, reserves would be well within the adequacy range of 100–150 percent. However, reserves would fall below this range if the stock of FX swaps reaches about US$180 billion.

Figure 16.
Figure 16.

International Reserve Adequacy

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Source: Fund staff calculations.

28. Prudential limits on commercial bank balance sheets may constitute another limit to FX swaps based intervention. As discussed previously, their growing long dollar FX swap position against the BCB has allowed commercial banks to increase their short spot dollar positions to service onshore dollar liquidity needs (Figure 14). While the short spot position is covered by the long derivative position, commercial banks likely face self-imposed prudential limits as to the size of the short spot position they are able to take (Volpon and Lei, 2013 and 2014). If such a limit were to be reached at some point in the future, FX swaps would no longer be an instrument suited to alleviate actual dollar liquidity needs in the market.24

29. The stock of reserves forms an upper limit for the outstanding stock of FX swaps.

G. Discussion

30. The intervention program was well calibrated and achieved its objectives. The BCB used adequate tools to contain excessive exchange rate volatility following the Bernanke tapering speech; preannouncing the program succeeded in lowering volatility by reassuring investors with open FX positions. What is more, the size of the program and the maturity of its instruments likely filled a market gap beyond increasing the supply of hedge at the margin.

31. However, the program’s continued extension raised questions about the BCB’s objectives. The June 2014 announcement to extend the program came in an environment of low volatility and does not appear to have contributed to lowering volatility further. At the same time, the notional value of the outstanding stock of bonds already covered more than 80 percent of short-term debt at remaining maturity. While volatility increased subsequently in an environment of uncertainty in the run-up to the elections, observers increasingly questioned the objectives behind the program extension. Similarly, at the time of the most recent extension at end-2014, the program amounted to US$110 billion and thus more than 100 percent of short-term debt at remaining maturity, some 60 percent of foreign holdings of domestic debt, 35 percent of non-financial corporate external debt, and about 20 percent of total external debt. The coverage of all four measures of external debt is now higher than during previous FX swap programs (Figure 17).25

Figure 17.
Figure 17.

Size of FX Swap to Measures of External Debt

(In percent)

Citation: IMF Staff Country Reports 2015, 122; 10.5089/9781475521320.002.A005

Source: Fund staff calculations.

32. The program may have slowed exchange rate convergence. According to staff’s assessment, the real continues to be moderately over-valued (see external sector assessment). The BCB’s intervention program likely contributed to preventing its depreciation to a level in line with fundamentals.

33. The exchange rate should remain the main external shock absorber, with intervention limited to episodes of excessive volatility. Volatility in global markets has fallen since the tapering speech in May 2013. Our recommendation is not to continue the program beyond March. Further intervention should be limited to episodes of excessive volatility.

References

  • Bayoumi, Tamim, Joseph Gagnon, and Christian Saborowski, 2014, “Official Financial Flows, Capital Mobility and Global Imbalances,” Journal of International Money and Finance, forthcoming.

    • Search Google Scholar
    • Export Citation

  • Cime de Toledo, Marcelo, 2014, “The Central Bank’s Intervention Tools in the Foreign Exchange Market, its Costs and Benefits: at this Momento, Swaps are Preferable to International Reserves,” Depec-Bradesco Economic Highlights, No. 75 (São Paulo: Bradesco).

    • Search Google Scholar
    • Export Citation

  • Garcia, Márcio, 2013, “Should Brazil’s Central Bank be Selling Foreign Reserves?”, CEPR’s Policy Portal (VOX).

  • Garcia, Márcio, Marcelo Medeiros, and Francisco Santos, 2014, “Price Discovery in Brazilian FX Markets,” Working Paper, Department of Economics (Rio de Janeiro: Pontifical Catholic University of Rio de Janeiro).

    • Search Google Scholar
    • Export Citation

  • Garcia, Márcio, and Tony Volpon, 2014, “DNDFs: a More Efficient Way to Intervene in FX markets?” Discussion Note, No. 621 (Rio de Janeiro: Pontifical Catholic University of Rio de Janeiro).

    • Search Google Scholar
    • Export Citation

  • Kohlscheen, Emanuel, and Sandro C. Andrade, 2013, “Official Interventions Through Derivatives: Affecting the Demand for Foreign Exchange,” Working Papers, No. 317 (Brasilia: Central Bank of Brazil).

    • Search Google Scholar
    • Export Citation

  • Itaú, 2015, “Brazil: FX and Capital Markets,” January 5.

  • Sarno, Lucio, and Mark P. Taylor, 2001, “Is Official Exchange Rate Intervention Effective?” Economica, No. 26, 1, pp. 24–36.

  • Stone, Mark R., and W. Christopher Walker, and Yosuke Yasui, 2009, “From Lombard Street to Avenida Paulista: Foreign Exchange Liquidity Easing in Brazil in Response to the Global Shock of 2008-09,” IMF Working Paper 09/259 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Volpon, Tony, and George Lei, 2013, “Brazil: Who is Buying What From the BCB,” Nomura Strategy Insights, November 21.

  • Volpon, Tony, and George Lei, 2014, “Who’s Buying What from the BCB-January Update,” Nomura Strategy Insights, January 17.

1

Prepared by Heedon Kang and Christian Saborowski.

3

Not all FX swap contracts were adjusted daily in 2002.

4

Convertibility risk is defined in this paper as the percentage difference between the offshore non-deliverable forward rate (NDF) and the onshore non-deliverable forward rate (DNDF) in percent of DNDF.

5

The cupom cambial reflects the onshore dollar interest rate implied by the domestic local currency interest rate, the spot exchange rate, and currency futures prices traded in BM&F BOVESPA. As an example, a one-year local currency interest rate of 12 percent, a spot rate of 2.5 reais per dollar, and a one-year forward exchange rate of 2.75 reais per dollar, would imply an onshore dollar interest rate of 2 percent. Assuming that onshore and offshore dollar instruments are close substitutes, the cupom cambial would be equal to the prevailing one-year domestic interest rate in the U.S. (Stone and others, 2009). In practice, however, instances of perceived convertibility risk have in the past driven a significant wedge between the two.

6

See Dodd and Griffith-Jones (2007) and Garcia and Volpon (2014).

7

Brazilian law (Decree-Law No. 857) states that every contract, security, document or obligation, in order to be fulfilled in Brazil, can’t stipulate payment in gold or foreign currency, or, in any form, restrict or refuse fulfillment in the Brazilian currency. The exceptions to that law are: currency exchange operations, import/export contracts, export financing (when a Brazilian bank buys, paying in reais, in advance, the amount of foreign currency to be received by an exporter in an export operation) or loans or any obligations in which the creditor or debtor is domiciled outside Brazil.

8

The BCB announces detailed information prior to each auction, such as the exact time of the auction, the maximum quantity of contracts that the BCB offers, and the maturity. Bidders are allowed to place up to five bids, specifying the quantity and price quotation for the bids. But, every bid-winner pays the same SELIC rate and receives the same cupom cambial and exchange rate variation. The BCB has its discretion to accept any volume of contracts up to the maximum that is on offer. The cut-off procedure is decided based on market analysis, such as auction consensus and the cupom cambial curve. The FX swap contracts are registered at the BM&F BOVESPA.

9

In the past, the instrument was defined such that BCB received the CDI rate rather than the SELIC rate.

10

The counterpart includes the dealer banks that are authorized by the BCB.

11

For accounting purposes, the forward points (forward rate–spot rate) are accrued daily by the maturity date (the maturity varies from 30 to 365 days).

12

Note that, by law, the Brazilian real can only be bought and sold in Brazil. Hence, a dollar shortage in Brazil represents a convertibility risk.

13

However, if the BCB accumulates a large short position, it can incur significant losses if the real depreciates.

14

About 90 percent of private party U.S. dollar futures trades are concentrated in maturities of less than 30 days, while the maturity of about 90 percent of the BCB’s FX swaps has been over 200 days.

15

In principle, this should allow us to rule out global factors that affect all EMs more or less equally as driving the results.

16

Following June, the volatility in the real reached record levels as domestic factors including the election dominated the picture.

17

This definition appears appropriate since the outstanding stock of swaps should be what matters for volatility. The reason is that its magnitude determine the share of open FX positions that can be hedged.

18

Monthly data is transformed to daily data by assuming it to be constant within each given month.

19

Since the announcement likely reduced volatility in EM currencies, the coefficient can be interpreted as having reduced volatility in the real by less than for peer currencies.

20

Volatility increased markedly in the second half of 2014, largely as a result of domestic factors including the election.

21

It is important to highlight that the effect of the various program announcements is cumulative in the present model. In other words, any effect of the June 2014 would have come on top of the effect of the existing announcements.

22

Depending on the ability and willingness to roll over hard currency liabilities by Brazilian corporations, and the desire of foreigners to hold on to BRL-denominated debt, further demand for spot dollars could arise as these two agents choose to “un-hedge” their position. Thus we would treat these positions as “contingent” demand of BCB hard currency reserves (Volpon and Lei, 2013 and 2014).

23

See Bradesco (2014).

24

There are no specific statutory limits to open short USD positions a commercial bank may carry on June 25, 2013 the BCB revoked regulation that stipulated mandatory reserves be collected against short USD spot positions.

25

The stock of FX swaps in 2003 and 2009 is defined as the maximum stock of outstanding notional principal during that the respective year.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Brazil: Selected Issues Paper
Author: International Monetary Fund. Western Hemisphere Dept.
Volume/Issue:
Volume 2015: Issue 122
Publisher:
International Monetary Fund
ISBN:
9781475521320
ISSN:
1934-7685
Pages:
102
DOI:
https://doi.org/10.5089/9781475521320.002
  • View in gallery

    Foreign Currency Intervention

    (In billions of U.S. dollar)

  • View in gallery

    USD Linked Debt and Foreign Reserves

    (In billions of U.S. dollar)

  • View in gallery

    Cupom Cambial, FX Rate, and FX Swap

    (In percent and billions of U.S. dollar)

  • View in gallery

    Convertibility Risk

    (Difference between offshore NDF and onshore NDF (DNDF), in percent of DNDF, annualized)

  • View in gallery

    Ratio of Daily Turnover in Derivatives Markets to Spot Market

    (As of April 2013, net-net basis)

  • View in gallery

    Outstand Balance of FX Swap and FX Repo since 2013

    (In billions of U.S. dollar)

  • View in gallery

    FX Swap Exposure by Counterparties

    (In billions of U.S. dollar)

  • View in gallery

    FX Swaps at Maturity

    (In billions of U.S. dollar, as of January 20, 2015)

  • View in gallery

    Net Increase of FX Swap

    (In billions of U.S. dollar)

  • View in gallery

    Rollover Rate of FX Swap

    (In percent of maturing FX swaps)

  • View in gallery

    Capital Flows

    (In millions of U.S. dollar)

  • View in gallery

    FX Flows Financed by FX Repo and Commercial Banks’ Short FX Position

    (In millions of U.S. dollar)

  • View in gallery

    “End Users” of FX Swaps

    (In percent of the BCB’s FX swap position)

  • View in gallery

    FX Implied Volatility Relative to Peers

  • View in gallery

    Foreign Exchange Rate

    (August 22, 2013 = 100)

  • View in gallery

    International Reserve Adequacy

  • View in gallery

    Size of FX Swap to Measures of External Debt

    (In percent)

Foreign Currency Intervention

(In billions of U.S. dollar)

USD Linked Debt and Foreign Reserves

(In billions of U.S. dollar)

Cupom Cambial, FX Rate, and FX Swap

(In percent and billions of U.S. dollar)

Convertibility Risk

(Difference between offshore NDF and onshore NDF (DNDF), in percent of DNDF, annualized)

Ratio of Daily Turnover in Derivatives Markets to Spot Market

(As of April 2013, net-net basis)

Outstand Balance of FX Swap and FX Repo since 2013

(In billions of U.S. dollar)

FX Swap Exposure by Counterparties

(In billions of U.S. dollar)

FX Swaps at Maturity

(In billions of U.S. dollar, as of January 20, 2015)

Net Increase of FX Swap

(In billions of U.S. dollar)

Rollover Rate of FX Swap

(In percent of maturing FX swaps)

Capital Flows

(In millions of U.S. dollar)

FX Flows Financed by FX Repo and Commercial Banks’ Short FX Position

(In millions of U.S. dollar)

“End Users” of FX Swaps

(In percent of the BCB’s FX swap position)

FX Implied Volatility Relative to Peers

Foreign Exchange Rate

(August 22, 2013 = 100)

International Reserve Adequacy

Size of FX Swap to Measures of External Debt

(In percent)