Public Debt Sustainability Analysis
Reflecting a severe recession, difficulties with implementation of the fiscal policy agenda, and the cost of the FX swap program, gross debt of the nonfinancial public sector (NFPS) increased by 10½ percentage points of GDP in 2015, reaching 73.7 percent of GDP. Net debt of the public sector rose by only 3 percentage points of GDP owing to the countervailing impact of valuation gains on gross international reserves. The debt maturity profile and composition have remained broadly unchanged since the last Debt Sustainability Analysis (DSA), and foreign holdings of government securities has remained broadly constant in percent of GDP. However, net interest on NFPS debt picked up significantly in 2015 due to more elevated borrowing costs and because of the large losses on the FX swaps, which are recorded in the interest bill (Box 1). Primary surpluses above 3 percent of GDP (excluding interest revenue) would be needed to keep gross debt from growing as a ratio to GDP beyond the projection horizon. In the baseline scenario, fiscal consolidation based on the government’s proposed reforms on the government expenditure side, if fully implemented, would restore the sustainability of debt, but debt ratios would stabilize after the end of the 5-year projection period, with gross debt reaching 93.5 percent of GDP in 2021. The trajectory of gross debt is highly sensitive to fiscal policy performance, owing to the adverse dynamics determined by modest GDP growth and high borrowing costs. Failure to implement the fiscal reforms proposed by the authorities would jeopardize debt sustainability.
Debt coverage. Brazil’s gross debt statistics cover the NFPS, defined to exclude Petrobras and Eletrobras, and consolidate the Sovereign Wealth Fund. In line with the GFSM 2014 manual, the NFPS debt includes all Treasury securities on the Central Bank’s balance sheet, including those not used under repurchase agreements. At end-2015, the gross debt amounted to 73.7 percent of GDP.2 As reported by the government, net debt statistics correspond to the public sector, defined to include the Central Bank. The consolidated public sector has a large stock of assets, amounting to 42 percent of GDP in 2015, which include 24.1 percent of GDP in international reserves. Non-financial public sector assets amounted to 34.3 percent of GDP. Brazil’s debt is reported at nominal value.3
Debt profile. Federal government (FG) domestic tradable securities accounted for 90 percent of total NFPS gross debt in 2015, of which 2/3 were held by the public.4 Active debt management in recent years has improved the profile of these instruments, which now display longer maturities, of about 4.6 years on average, up from less than 3 years in 2008. Still, a large share of domestic tradable securities, nearly 22 percent of total, will mature in 2016.5 The government has the objectives of raising average maturity to 5.5 years and bringing total short-term debt (on both original and residual maturity basis) down toward 20 percent of all debt over the medium term, which is challenging because of a high term premium. The average maturity of new placements declined somewhat in the course of 2015, but was still well above average residual maturity. Fixed-rate and inflation-linked domestic bonds have gradually replaced foreign-currency linked instruments and floating-rate bonds over the last half a decade. A greater share of newly issued securities was, however, linked to the policy rate (Selic) in 2015. Zero-coupon bonds with original maturities over one year constitute slightly more than half of FG domestic tradable securities held by the public, or 21.7 percent of GDP. Foreign holding of domestic securities was stable at 18 percent of total at end-2015.
Composition of Domestic Tradeable Securities Held by the Public
Source: Tesouro Nacional, and IMF
In addition to tradable securities, other NFPS debt consist of liabilities issued by SOEs and subnational governments, as well as direct bank lending. Foreign currency denominated tradeable securities accounted for only 6.7 percent of the total in 2015, representing slightly below 5 percent of GDP. Gross financing needs have tended to be high, above 15 percent of GDP, and interest charged on new issuances has increased by 250 bps since end-2014. However, this figure overstates rollover risk, as a large fraction of the federal government debt (about 21 percent of GDP, with a maturity profile in line with that of overall debt) is held by the Central Bank, which follows a policy of automatic rollover of its holdings of government securities.
Box 1.Impact of Swaps and Valuation Changes on Interest and Debt in 2015
The Central Bank of Brazil (BCB) uses a variety of instruments to intervene in foreign exchange markets but the most frequently used instrument is the Brazilian FX swap (the swap), a non-deliverable future settled in local currency. The swap program has grown over time and, at end-December 2015, the notional value of the outstanding stock of swaps was about US$110 billion, equivalent to 30 percent of gross international reserves or 7.3 percent of GDP. The FX swap is meant to provide hedging to agents with exposures in dollars. If the Brazilian real depreciates too fast, the central bank makes cash losses on the FX swaps; if it depreciates too slowly or appreciates, it makes cash profits. Gains and losses on FX swaps operations are reported by the BCB under the item “Equalização Cambial” (EC) in public debt statistics.
The BCB considers EC in the compilation of the General Government (GG) net debt (a positive result decreases net debt; a negative result increases the net debt), but not in compiling gross debt. EC is calculated daily and, aside from results of the swap operations and derivatives, includes changes in the valuation of reserves (net external debt) and the carry cost of reserves. The treatment of EC applies to realized, as well as accrued-but-unrealized gains or losses incurred by the BCB. A positive result gives rise to a transfer from the BCB to the Treasury Single Account (TSA) while to cover a negative result the government issues debt to the BCB.
EC closely tracks exchange rate movements and is influenced by the size of reserves. Because of the large size of foreign exchange reserves and the dominance of the U.S. dollar in their composition (over 2/3 of reserves are in U.S. dollars), movements in EC are strongly affected by changes in the real/U.S. dollar exchange rate. For example, when the dollar value of reserves increased from 17.8 percent of GDP in 2014 to 24.1 percent of GDP at end-December 2015, on account of a 49 percent depreciation of the Brazilian real during the year, EC was 1.9 percent of GDP.
Interest on EC, which includes the interest on swaps, is accounted together with interest on public debt. During the period elapsing between the calculation of the balance sheet of the BCB and the effective payment, the values of EC accrue interest at the same rates applied to the FG deposits in the TSA. Gains/losses on the swaps are also included in the net interest bill and thus in the overall government balance. These occur when swap contracts mature, and the BCB pays its counterparts the observed exchange rate change plus the pre-set coupom cambial and receives the ex ante Selic rate in return (Selected Issues Paper for the 2015 Staff Report and Law No. 11.803, of November 5, 2008.). This part of the interest bill reached 1.5 percent of GDP in 2015, pushing up the implied interest on net debt.
Net NFPS Debt and Exchange Equalization Implicit Cost of Net Debt Gains and Losses on FX Swaps Exchange Equalization Decomposed
B. Baseline Scenario
Macroeconomic assumptions. The projections assume a real GDP contraction of -3.3 percent in 2016, ½ percent output growth in 2017, and a gradual return to potential growth of 2 percent by 2019. The assumed fiscal adjustment brings the primary balance to 0.8 percent of GDP by 2021, which implies an improvement of about 3 percentage points of GDP during 2017-21, significant but short of what would be needed to stabilize gross debt by 2021. The nominal interest rates on new borrowing are between 12 and 15 percent over 2016-21, bringing the effective interest rate to about 12 percent on average.6 The baseline scenario assumes limited structural reforms and a modest world output recovery. In this baseline, gross debt remains on an upward path, reaching 93.5 percent of GDP by 2021 and likely to keep growing for some time still. Gross financing needs are projected to remain above the high-risk threshold of 15 percent of GDP on average over the medium term. By the end of the projection horizon, the primary balance required to stabilize debt in the baseline scenario is 3.3 percent of GDP (excluding interest revenue).7 The baseline assumptions do not include the early repayment of BNDES credit to the government, which is currently under analysis at the TCU (the Federal Court of Accounts, whose concurrence is being sought).8
Heat map. The heat map in Figure 1 suggests that the stock of debt and the gross financing needs are high. The premium asked by the market on government borrowing and the reliance on external financing are elevated, pointing to a moderate risk to the debt profile. As noted earlier, however, the gross financing need indicator overstates actual rollover risk as a result of the policy of continuous rollover maintained by the Central Bank, which holds a significant amount of FG bonds. Gross financing needs would amount to about 15 percent of GDP in 2016 excluding rollover of Central Bank held bonds, still at the threshold level for high risk.
Figure 1.Brazil: Public Sector Debt Sustainability Analysis (DSA)—Risk Assessment
Source: IMF staff.
1/ The cell is highlighted in green if debt burden benchmark of 70% is not exceeded under the specific shock or baseline, yellow if exceeded under specific shock but not baseline, red if benchmark is exceeded under baseline, white if stress test is not relevant.
2/ The cell is highlighted in green if gross financing needs benchmark of 15% is not exceeded under the specific shock or baseline, yellow if exceeded under specific shock but not baseline, red if benchmark is exceeded under baseline, white if stress test is not relevant.
3/ The cell is highlighted in green if country value is less than the lower risk-assessment benchmark, red if country value exceeds the upper risk-assessment benchmark, yellow if country value is between the lower and upper risk-assessment benchmarks. If data are unavailable or indicator is not relevant, cell is white. Lower and upper risk-assessment benchmarks are: 200 and 600 basis points for bond spreads; 5 and 15 percent of GDP for external financing requirement; 0.5 and 1 percent for change in the share of short-term debt; 15 and 45 percent for the public debt held by non-residents; and 20 and 60 percent for the share of foreign-currency denominated debt.
4/ Long-term bond spread over U.S. bonds, an average over the last 3 months, 03-Jun-16 through 01-Sep-16.
5/ External financing requirement is defined as the sum of current account deficit, amortization of medium and long-term total external debt, and short-term total external debt at the end of previous period.
Past forecast errors. There is no evidence of a systematic projection bias in the baseline assumptions that would undermine the assessment of sustainability and the projected fiscal adjustment is in line with other countries’ experiences. The median forecast error for GDP growth, the primary balance and inflation are below 2 percent and in line with other countries’ forecast errors.
Realism of projections. Brazil’s projected fiscal adjustment (an improvement of about 2½ percentage points in the cyclically-adjusted primary balance/GDP over the medium term) is in line with other countries’ experiences, with a 3-year adjustment in cyclically-adjusted primary balance at the 23 percentile among all surveillance countries.9
C. Shocks and Stress Tests
Growth shock. The impact on the debt-to-GDP ratio of the 1 standard deviation shock (3.4 percent) to real GDP growth is large. Over 2017-18, real GDP growth rates become negative while inflation declines by 200 bps in each year. The primary balance worsens before recovering in 2019; interest rates increase by 55 bps for each percentage point decline in output growth. In this scenario, gross debt increases to 112 percent of GDP by 2021 while gross financing needs reach 39 percent of GDP.
Primary balance and real interest rate shocks. In the primary balance shock, consolidation is delayed by one year. In the interest rate shock, borrowing costs increase by 200 bps in the first projection year and remain the same over the medium term. The impact of these shocks on debt and gross financing needs is modest, pushing up debt-to-GDP by 3 percentage points above the baseline in 2021.
Combined macro-fiscal shock and contingent liabilities shocks. The macro-fiscal shock combines the growth and interest rates shock and a primary balance shock as in the standard examples above, together with the real exchange rate shock consistent with a maximum movement of the real exchange rate over the past 10 years and a pass-through coefficient of 0.25. The contingent liability shock simulates a loss of 10 percent of banking system assets, in line with the experience with past banking crises in other countries, and assumption of contingent liabilities by the public sector. The latter results in a primary deficit of -5.5 percent of GDP in 2017 and a 1 standard deviation shock to GDP growth in 2017 and 2018. The effect of these shocks on debt ratios and gross financing needs is large and reflects in particular the sensitivity of debt to growth shocks, which make the largest contribution to the change in the debt path. Debt-to-GDP reaches about 109-119 percent by 2021 in the two shocks.
No reform. A custom-made scenario assumes that the government fails to implement its proposed reforms on the fiscal front, so that primary deficits are registered throughout the projection period. The currency depreciates by 5 percent in 2016 and by another 20 percent in 2017. Spreads on domestic securities pick up by 500 bps in 2016 and 2017, and remain elevated thereafter. Growth recovery is weaker and the depreciation is steeper. Debt reaches 106 percent of GDP by 2021.
Debt Reduction Begins by 2021 (“IMF recommended scenario”). An alternative custom-made scenario projects debt and financing needs under an assumption of more aggressive fiscal consolidation than under the baseline. The scenario is calibrated so that the primary balance reaches the debt-stabilizing level within the 5-year projection period. Faster disinflation and higher confidence supported by fiscal consolidation allows interest rates to drop lower than in the baseline scenario, providing growth support in outer years while fiscal reforms send a positive confidence signal. Gross financing needs are lower and gross debt is put on a declining trajectory by 2021.
Figure 2.Brazil: Public DSA—Realism of Baseline Assumptions
Source: IMF Staff.
1/ Plotted distribution includes surveillance countries, percentile rank refers to all countries.
2/ Projections made in the spring WEO vintage of the preceding year.
3/ Data cover annual obervations from 1990 to 2011 for advanced and emerging economies with debt greater than 60 percent of GDP. Percent of sample on vertical axis.
Figure 3.Brazil: Public DSA—Baseline Scenario
Source: IMF staff.
1/ Public sector is defined as non-financial public sector.
2/ Based on available data.
3/ Long-term bond spread over U.S. bonds.
4/ Defined as interest payments divided by debt stock (excluding guarantees) at the end of previous year.
5/ Derived as [(r - π(1+g) - g + ae(1+r)]/(1+g+π+gπ)) times previous period debt ratio, with r = interest rate; π = growth rate of GDP deflator; g = real GDP growth rate; a = share of foreign-currency denominated debt; and e = nominal exchange rate depreciation (measured by increase in local currency value of U.S. dollar).
6/ The real interest rate contribution is derived from the numerator in footnote 5 as r - π (1+g) and the real growth contribution as -g.
7/ The exchange rate contribution is derived from the numerator in footnote 5 as ae(1+r).
8/ In 2015 includes the issuance of bonds to the BCB of 1,5 percent of GDP.
9/ Includes asset changes and interest revenues (if any). For projections, includes exchange rate changes during the projection period.
10/ “Interest revenue” is a reconciliation series calculated as a difference between the gross interest of the NFPS and the net interest of the PS. This concept is used to maintain consistency between the fiscal accounts, in which the net interest used to compute the overall balance includes also the net interst bill of the BCB, and gross interest of the NFPS in the DSA.
11/ Assumes that key variables (real GDP growth, real interest rate, and other identified debt-creating flows) remain at the level of the last projection year. Up to 2021, the primary balance shown is a non-interest balance, with interest income showing in the residual, but from 2022 onwards interest income counts toward the required primary balance.
Figure 4.Brazil: Public DSA—Composition of Public Debt and Alternative Scenarios
Source: IMF staff.
Figure 5.Brazil: Public DSA—Stress Tests
Source: IMF staff.
The analysis of public debt sustainability is based on the framework developed for market access countries. See Staff Guidance Note for Public Debt Sustainability Analysis in Market Access Countries, IMF, May 2013.
In contrast, the national definition of gross debt includes the stock of Treasury securities used for monetary policy purposes by the Central Bank (those pledged as security in reverse repo operations), but excludes the rest of the government securities held by the Central Bank. Thus, according to the national definition, gross debt of the general government amounted to 66.5 percent of GDP at end-2015. Note also that after completion of the 2015 DSA, published revisions to the historical nominal GDP series have caused the debt ratios up to 2014 to be revised down.
The nominal value is calculated as the PDV of future interest and principal payments at the security’s contractual interest rate(s), and generally differs from face value.
The rest is held by the Central Bank on its balance sheet. The Central Bank uses about ¾ of its holdings as security in liquidity-draining operations with the banking system.
This is broadly a recurring situation. Short-term debt at original maturity is very small, amounting some 2 percent of total NFPS debt.
Interest rates on new borrowing are projected to behave in line with the forecast for the Selic rate augmented by a spread that varies depending on the maturity of the bond. Note also that effective interest rates show persistence because of the high proportion of zero coupon bonds with multi-year maturities. In particular, zero-coupon bonds issued with very high rates in 2015 will have their cumulated interest falling due in coming years.
The debt stabilizing primary balance shown on the right-most column of the table on “Contributions to changes in public debt” in the Baseline Scenario (4.6 percent of GDP) corresponds to a concept of the primary balance that includes interest revenue. The definition of the primary balance as a non-interest concept is customary in Brazil and corresponds to 3.3 percent of GDP quoted in the main text.
The proposed schedule is for a repayment on R$100 billion in three installments starting in 2016. The estimated savings on interest from this transaction amount to about R$7 billion per year.
The fiscal multipliers and persistence were estimated in the 2015 Article IV Selected Issues Paper at 0.3-0.5 percent.