II. Real Exchange Rate Appreciation: can Fiscal Policy Help?1
Brazil has experienced a large real exchange rate appreciation in recent years, generating concerns about competitiveness and prompting the authorities to respond with a combination of policies. This paper shows that fiscal policy can play a role in alleviating these pressures. In particular, we find that a permanent fiscal adjustment is associated with a real exchange rate depreciation over the long term. Furthermore, increases in public investment could also reduce the real effective exchange rate. As the magnitude of these two channels is roughly equal in size, the implication for Brazil is that increasing public investment is likely to ease appreciation pressures but, to be an effective tool, the increase cannot be deficit-financed. This highlights the importance of tackling long-standing budget rigidities to generate fiscal space for public investment.
1. Brazil’s real effective exchange (REER) rate is at historical highs. Since 2000, the REER has appreciated over 50 percent, surpassing most emerging markets (Figure 1). Among other factors, this partly reflects large terms of trade gains and capital inflows. Indeed, just in 2011 gross capital inflows (defined as direct investment, portfolio investment and other flows) exceeded $133 billion (5¼ percent of GDP). Strong economic growth prospects in the aftermath of the global crisis and structurally high interest rates have been elements behind this surge. This trend raises important challenges for Brazil because of the potential loss of competitiveness and the increased exposure to volatile capital flows.
Figure 1.Real and Nominal Effective Exchange Rate, Trends
2. The Brazilian authorities have used all aspects of the policy toolkit to manage these pressures. The exchange rate has appreciated, the macro-policy mix has been adjusted, and reserves have been built. Furthermore, macroprudential measures (such as reserve requirements limiting short dollar position of banks) and capital flow management measures (notably the tax on foreign purchases of domestic bonds and equities, “IOF”) have been used in an adaptive manner to stem the large inflow of foreign capital and to slow the pace of nominal appreciation (see Benelli et al, 2011 for further discussion of these issues). Notwithstanding these efforts, the reality is that the real effective exchange rate in Brazil remains somewhat overvalued.
3. In this context, we ask to what extent fiscal policy can help reduce these appreciation pressures. Several strands of the literature have highlighted possible channels through which fiscal policy can indeed affect the REER (see section B). To be sure, fiscal policy is often cited as an important instrument in the policy toolkit available to countries preoccupied with this issue. For example, Ostry et al. (2010) highlight “using available scope to tighten fiscal policy” as a fundamental part of the macroeconomic response to capital inflows when there are concerns about excessive exchange rate appreciation. At the same time, theoretical analysis and the empirical evidence to date are somewhat inconclusive about the effect of fiscal policy on the real exchange rate.
4. The purpose of this paper is to assess empirically the relationship between fiscal policy and the REER in emerging markets and draw policy implications for Brazil. Specifically, the paper analyzes whether (1) fiscal adjustment can have a permanent effect on the REER; and (2) to what extent the composition of public spending can play a role. Overall, the findings in this paper suggest that both fiscal adjustment and an increase in public investment are associated with a reduction in the real exchange rate. The strength of these two channels is approximately the same for Brazil, which suggests that increases in public investment are likely to reduce appreciation pressures only to the extent that they are financed through a compositional shift within the budget (i.e. reducing government consumption to increase public investment) rather than financed via additional public debt. The rest of the paper is organized as follows: section B briefly reviews the literature. Section C describes the data and model specification. Section D presents the results. Section E draws policy implications for Brazil and section F concludes.
5. While exchange rates are one of the most studied topics in international economics, most papers analyzing their determinants do not focus on fiscal variables. The empirical literature on the long-term behavior of exchange rates is dominated by attempts to test the PPP theory. In the international finance literature, the focus is more on short-term dynamics, with an emphasis on tests of the uncovered interest parity theory. These papers focus mostly on the interaction between monetary policy, interest rates and the nominal exchange rate.1
6. Moreover, there is no consensus in the existing theoretical literature about the relationship between fiscal policy and the real exchange rate:2
In Keynesian models, an expansionary fiscal shock raises the demand for home goods and money, thereby inducing a real appreciation either through higher interest rates and arbitrage capital inflows or a rise in domestic prices (see, Mundell, 1963; and Flemming, 1962).3 However, Sachs and Wyploz (1984), argue that the Mundell-Fleming framework ignores a number of critical factors that may be associated with a different result.4
The composition of government spending could also matter. In particular, increases in government spending—whether tax or debt financed—will result in a real appreciation if skewed toward nontradable goods. The effect of public investment, on the other hand, is ambiguous. An increase in public investment may lead to a real appreciation if it raises productivity in the tradable sector through the Balassa-Samuelson mechanism (see, Balassa, 1964; and Samuelson, 1964). But the opposite effect may result if public investment disproportionately increases productivity in the nontradables sector. Moreover, if productivity increases symmetrically in both sectors, there will be no impact on the real exchange rate (Galstyan and Lane, 2009).
In real business cycle models, increases in government spending trigger a decline in domestic private consumption and an increase in labor supply leading to a real appreciation (see, Backus et al, 1994). In contrast, more recent models find that, under incomplete financial markets this is not necessarily the case (Kollmann, 2010). In particular, when faced with an increase in government spending, domestic households experience a negative wealth effect and, thus, they work harder and increase domestic output. Limited risk sharing exacerbates the negative wealth effect, with the resulting supply-side response leading to a deterioration of the country’s terms of trade and real exchange rate depreciation.
7. The empirical evidence is also relatively inconclusive.5 Results vary depending on the methodology, specification, and sample used in the estimation. For example, Cardarelli, Elekdag, and Kose (2007) estimate a model based on a cross-section of countries (including advanced and emerging economies) and show that real appreciation and demand growth is more contained in countries that respond to capital inflows by pursuing a tighter fiscal policy in the form of slower growth of government expenditure. In the same direction, IMF (2008) and Ricci et al (2008) estimate panel cointegration models and find that an increase in government consumption appreciates the REER. Guajardo et al (2011) use a historical approach to identify changes in fiscal policy in advanced economies and find that the real exchange rate depreciates in response to fiscal consolidation. On the other extreme, several studies based on dynamic VARs have found that fiscal expansions in advanced economies are associated with real depreciations. For example, Kim and Roubini (2008) find that an increase in the government primary deficit induces a real exchange rate depreciation for the United States. Similarly, Monacelli and Perotti (2007) look at the United States, United Kingdom, Canada, and Australia and show a negative relation between government spending and the real exchange rate.
C. Data and Econometric Methodology
8. Given data constraints, we focus on a parsimonious set of economic fundamentals to explain the REER. Our sample covers an unbalanced panel comprising 28 emerging economies for the period 1983-2011.6 In the baseline model, we relate the real effective exchange rate to five underlying determinants drawn from the literature:7
Relative GDP per capita (GDPPC) in constant 2005 U.S. dollars is measured relative to a weighted average of trading partners. Since it works as a proxy for the level of productivity, we expected to find a positive correlation between GDPPC and the REER in line with the Balassa-Samuelson conjecture. Also, richer countries tend to spend more on services that have higher income elasticity of demand (see, Bergstrand 1991) which would result in a higher real exchange rate.
Balance of goods and services (TB) is measured as ratio of GDP and is used as a proxy for the international investment income. In steady-state, the trade balance surplus should equal the international investment income deficit and, thus, we expect to find a negative relation between the TB and the REER.8
Structural balance (SB) is defined as the non-financial public sector cyclically-adjusted balance excluding one-off adjustments. Since we want to focus on discretionary fiscal policy we prefer to use the SB as opposed to the headline fiscal balance. Also, by using the SB we limit endogeneity problems as the effect of automatic stabilizers is excluded in this measure (although the possible endogeneity effect from countercyclical fiscal policy is not corrected).9 We hypothesize that a higher SB will be associated with a depreciation of the real exchange rate, in line with the conventional Keynesian model. Preliminary evidence seems to suggest that this is the case, i.e. there is a negative relation between the changes in the structural balance and the changes in the real exchange rate (Figure 2).
Relative public consumption (PC) is defined as government consumption as a share of GDP relative to a weighted average across trading partners.10 We measure this variable in relative terms in order to capture the forces driving the structure of relative prices captured by the real exchange rate. We expect an increase in public consumption to raise the relative demand for nontradables, thereby leading to a real appreciation (Figure 3).
Relative public investment (PI) is defined as government investment as a share of GDP relative to trading partners. As discussed above, the effect of PI on the real exchange rate is ambiguous. Public investment will lead to a real appreciation (depreciation) if it improves disproportionally productivity in the tradable (nontradable) sector. If, on the other hand, productivity improves symmetrically in the tradable and nontradable sectors, there will be no impact on the real exchange rate. Figure 4 indicates there could be a negative relationship between public investment and the REER in our sample.
Figure 2.Change in REER and Structural Balance, 2000-2011
Sources: IMF; and staffs calculations.
Figure 3.REER and Public Consumption
Figure 4.REER and Public Investment
where x is a vector including the explanatory variables described above and t is a time variable. In this model β is the vector of long-run cointegrating coefficients, A denotes the first-difference operator and y is the vector of coefficients of leads and lags of changes in the determinants11, and ϵit is the residual term. Fixed effects are necessary because the real effective exchange rate measures are index numbers, making their levels not comparable across countries. They also account for time-invariant country-specific factors, reducing the omitted variable bias. We favor the use of a panel DOLS because: (1) given the limited length of the sample, estimating separate real exchange rate equations for each country would result in imprecise estimates; and (2) data series are non-stationary.12
10. Our results show that fiscal policy has a non-negligible permanent effect on the REER:
Permanent fiscal adjustment is generally associated with a depreciation of the real exchange rate (Table 1, columns 1 and 3). The estimated coefficient of the structural balance is about -0.017. Since the dependent variable is estimated in logs, this means that an improvement in the structural balance of 1 percent of GDP would imply a depreciation of the real exchange rate of 1.7 percent over the long term. This is line with the results of Guajardo et al (2011) for advanced economies who find for a sample of advanced countries that a 1 percent of GDP consolidation is associated with a 1.57 percent real depreciation.
The composition of spending also matters. An increase in relative government investment depreciates the real exchange rate in the long run while government consumption does not have a statistically significant impact (Table 1, columns 2 and 3).13 As an illustration of the effect of these relativities, a 1 percentage point increase in relative public investment in Brazil would mean increasing public investment by 7½ percentage points of GDP; such a sizable increase would be associated with 12.6 percent depreciation in the real exchange rate. These results are in contrast with findings for advanced economies where government consumption appreciates the real exchange rate while public investment does not have an effect (Galstyan and Lane 2009). A possible explanation for this difference is that public investment is more likely to increase productivity in the nontradable sector among emerging markets given likely lower levels of infrastructure development. An additional argument could be associated with the different composition of government spending: emerging markets have relatively higher public investment but lower public consumption compared to advanced economies (Figure 5).14
|Relative government consumption||-0.027||0.164||0.001||0.119||0.184|
|Relative government investment||-0.044||-0.126||-0.114||-0.144||-0.129|
|Relative GDP per capita||0.127||-0.013||0.139||0.121||0.161||0.154|
|Balance of goods and services||-0.303||-0.632||-0.198||-0.278||-0.076|
|Structural primary balance||-0.013|
Figure 5.Composition of Government Spending
Sources: IMF, World Economic Outlook.
11. Several sensitivity analyses confirm the robustness of these results. The first question is whether these findings are driven by some groups of countries. In particular, Asian emerging economies have particularly large investment rates that could explain these results. Thus, we begin by estimating the model adjusting for possible outliers and find a similar message as in our baseline specification with the size of the coefficient on investment being only slightly smaller (Table 1, column 4). Also, estimating the model with a dummy for Asia yields the same results. Second, we look into a different measure of fiscal adjustment. In particular we use the structural primary balance instead of the overall structural balance. This variable may be more accurate to capture the true policy stance as interest rates (which are outside the control of the government) may fluctuate distorting the size of fiscal adjustment. Consistent with our previous results, we find that an increase in the structural primary balance depreciates the REER, although the impact is smaller (Table 1, column 5). Finally, we introduce capital inflows as an additional control and results remain unchanged (Table 1, column 6). Interestingly, though, capital inflows do not seem to have an effect on the REER over the long term irrespective of whether we use portfolio inflows or other inflows as our preferred measure.15 This is a question we leave for further investigation in future research given our focus on fiscal policy variables.
E. Implications for Brazil
12. What role can fiscal policy play in efforts to contain real exchange appreciation pressures in Brazil? In order to make an assessment it is important to look at fiscal performance in Brazil and place it in an international perspective.
Fiscal policy. Since the introduction of the Fiscal Responsibility Law in 2000 Brazil has maintained primary surpluses of around 3¼ percent of GDP, one of the highest among emerging markets (Figure 6). However, the overall deficit is still relatively high—because of large interest payments. In terms of the fiscal policy stance, there was a large adjustment during the period 2002-2008. This allowed the creation of buffers that were used in part during the crisis in the form a discretionary stimulus.16 Following a large fiscal withdrawal in 2011, the structural deficit has declined to 3¼ percent of GDP, still larger than pre-crisis levels. Further improvements will likely require addressing budgetary rigidities going forward.
Composition of spending. Relative to other emerging markets, Brazil is an outlier. In particular, public consumption, at 21¼ percent of GDP in 2011, is one of the highest among emerging markets and almost double the level of peers in Latin America (Figure 7). Public consumption in percent of GDP has increased by 2 percentage points in Brazil since 2000, in contrast to most other emerging markets where it has declined. This is striking taking into account that public consumption does not include transfers (where increases have been large). On the other hand, public investment in Brazil has increased somewhat since 2000 but, at about 2¼ percent of GDP, is less than half the average of other emerging markets. Moreover, the level of public investment is now 70 percent below that of trading partners (a marked deterioration since 2000). This evidence suggests that, by reallocating spending, Brazil could make some space for public investment and reap additional benefits.
Figure 6.Emerging Markets: Fiscal Performance
Sources: IMF, World Economic Outlook; and staff’s calculations.
Figure 7.Emerging Markets: Composition of Government Spending
Sources: IMF, World Economic Outlook; and staff’s calculations.
1/ Excluding Brazil.
2/ Relative public consumption (investment) is calculated as the ratio of Brazil’s public consumption (investment) in percent of GDP to a weigthed average of its trading partners’ public consumption(investment) in percent of GDP. A number above 1 means Brazil has higher public consumption (investment) in percent of GDP than is trading partners.
13. Simulation analysis suggests that fiscal policy in Brazil could help reduce real appreciation pressures over the long term. In particular, a 1 percent of GDP increase in public investment in Brazil would lead to a 1.7 percent real depreciation. However, this is roughly the same effect but with an opposite sign as a corresponding 1 percent of GDP deterioration of the structural balance. Thus, if both investment and the structural deficit were to increase by similar amounts, the REER would not change. In other words, increasing public investment could only help if accompanied by offsetting measures to generate savings (for example, by reducing public consumption). To put this into context, we consider two scenarios. Scenario 1 assumes Brazil improves the structural balance by 1 percent of GDP. In addition, we assume public investment in Brazil converges to the level of its Latin American peers. This would require finding additional fiscal space of 2¼ percent of GDP. Scenario 2 assumes the same improvement in the structural balance but public investment converging to the average in emerging markets (requiring fiscal space of 3¼ percent of GDP). These scenarios imply that an appropriate combination of fiscal policy actions could, ceteris paribus, support a real depreciation in the range of 6¼ to 7¼ percent in the long term (Figure 8).
Figure 8.Brazil: Simulation Analysis
Sources: IMF, Information Notification System; World Economic Outlook; and staff’s calculations.
1/ The countries included in Latin America are: Argentina, Colombia, Peru, and Mexico.
14. In order to reap these benefits, it would be important for Brazil to create fiscal room. Our results show that strengthening the structural fiscal position could play a role in alleviating appreciation pressures. As an added benefit, this could help reducing real interest rates, thus creating additional fiscal space (see, Segura-Ubiergo, 2012). A particularly promising avenue to facilitate a real depreciation would be to increase public investment, which is already an important priority for the authorities as demonstrated in their strategy under the Growth Acceleration Program (Programa de aceleragao do crescimento, PAC). Nevertheless, to be an effective tool for the exchange rate, the increase in public investment would need to be financed by savings, and not by an increase in the deficit. A similar logic applies to financing investment through quasi-fiscal operations (such as policy lending to BNDES). Beneficial effects on the exchange rate would likely be maximized if these operations were matched by higher public savings. Else the external current account could deteriorate, pressing up the real exchange rate. Moreover, the interest subsidy on BNDES lending directly lowers net public saving, while an increase in contingent liabilities here could gradually push up risk premia. Similarly, public investment projects undertaken via concessions or PPPs could also result in higher current account deficits (if not accompanied by an increase in public savings) and crowding-out of private investment.
15. The most promising route to create that space would be lowering government consumption. Achieving this end would require reducing fiscal earmarking/mandates that lock current spending at very high levels and create a bias against public investment. While some of these earmarks/mandates, like those for health and education spending floors have positive social objectives in their design, improvements at the margin in their design could be explored. The priorities could include (i) reducing revenue-earmarking and mandatory spending in combination with more effective medium-term planning and rolling multi-year budget plans; and (ii) strengthening the costing, monitoring, and evaluation of public spending with a view to increasing its efficiency.
16. Fiscal policy in emerging markets does have an effect on the real exchange rate. This works through two channels. First, increases in public savings (i.e. a stronger structural fiscal position) could reduce appreciation pressures over the long term and hence might be an important instrument to ensure higher competitiveness. Second, the structure of government spending matters, with increases in public investment leading to a reduction in appreciation pressures. This has important implications for Brazil since current spending accounts for almost 90 percent of total spending and, thus, there is scope to increase public investment. One caveat, however, is that both channels have roughly the same impact on the REER. What this means in practice is that increases in public investment that are not accompanied by offsetting measures to reduce current spending would likely have little effect on the REER. Therefore, creating room for investment by a reallocation of public spending would have multiple beneficial effects, both for improving public service delivery but also for helping address real appreciation pressures. Just as an example, Brazil would need to increase public investment by 2½ to 3¼ percent of GDP to converge to levels in emerging market peers. Given already high primary surpluses, achieving this solely through fiscal adjustment is likely to be challenging, which highlights the importance of addressing budgetary rigidities to reallocate resources from public consumption to investment. Equally important to increase public investment would be to improve project’s delivery and spending execution. This is an area where lack of capacity in planning and management, difficulties in obtaining necessary licenses and procedural problems have resulted in long delays in the past (for further discussion, see OECD, 2011).
The sample includes 28 emerging countries for the period 1983 to 2011: Argentina, Brazil, Bulgaria, Chile, China, Colombia, Hungary, India, Indonesia, Jordan, Kazakhstan, Kenya, Lithuania, Malaysia, Mexico, Morocco, Nigeria, Pakistan, Peru, Philippines, Poland, Romania, Russia, Saudi Arabia, South Africa, Thailand, Turkey, and Ukraine. Time span varies depending on the countries with shorter data available for the fiscal aggregates.
Variables are defined as follows:
Real effective exchange rate is based on consumer price index and taken from the IMF, Information Notification System.
Balance of goods and services is defined as the difference between exports and imports of goods and services. The data are taken from the IMF, World Economic Outlook.
Real GDP per capita (in constant 2005 prices) is taken from the IMF, World Economic Outlook.
Structural balance is defined as the overall balance adjusted for the cycle and excluding one-offs. Due to data availability, we take the cyclically adjusted balance for Mexico and Philippines. Cyclically adjusted balance is defined as the overall balance minus cyclical balance whereby the cyclical revenues and expenditures are computed using country-specific elasticities with respect to the output gap. Data are from the IMF, World Economic Outlook.
Public consumption is defined as current primary spending excluding transfers. The data are based on national accounts and come from IMF, World Economic Outlook.
Public investment is defined as public gross fixed capital formation. Data come from IMF, World Economic Outlook.
Trade weights are calculated using Direction of Trade Statistics data. For each country we focus on the top trading geographic destinations of its exports that account for at least 80 percent of exports during the period 1980—2010. Because of data limitations, coverage is below 80 percent at the beginning of the sample.
Capital inflows are defined as gross flows including direct investment, portfolio investment and other flows. Data are from the IMF, World Economic Outlook.
2011, “Fiscal Policy and the Current Account,”IMF Economic Review, Vol. 59, No.4, pp. 603–629.
2005, “Exchange Rates and Fundamentals: Evidence on the Economic Value of Predictability,”Journal of International Economics, vol. 66(2), pp. 325–348, July.
1994, “Dynamics of the Trade Balance and the Terms of Trade: The J-Curve?,”American Economic Review, Vol. 84, No.1, pp. 84–103.
1964, “The Purchasing Power Parity Doctrine: A Reappraisal,”Journal of Political Economy, No. 72,pp. 584–96.
2011, “Brazil’s Experience Managing Capital Inflows: The Role of Capital Flow Capital Flow Management Measures,”Selected Issues Paper, (unpublished: Washington: International Monetary Fund).
1991, “Structural Determinants of Real Exchange Rates and National Price Levels: Some Empirical Evidence,”American Economic Review, NO. 81, pp. 325–34.
2007, “Managing Large Capital Inflows,”Chapter 3, World Economic Outlook, October (Washington: International Monetary Fund).
2005, “Exchange Rates and Fundamentals,”Journal of Political Economy, Vol. 113, No. 3, pp. 485–517.
1962, “Domestic Financial Policies Under Fixed and Under Floating Exchange Rates,”International Monetary Fund Staff Papers, Vol. 9, pp. 369–79.
2009, “The Composition of Government Spending and the Real Exchange Rate,”Journal of Money, Credit and Banking, Vol. 41, No.6 pp. 1233–1249.
2011, “Expansionary Austerity: New International Evidence,”IMF Working Paper 11/158 (Washington: International Monetary Fund).
International Monetary Fund,2008, Exchange Rate Assessments: CGER Methodologies, IMF Occasional Paper No. 261 (Washington: International Monetary Fund).
2008, “Twin Deficit or Twin divergence? Fiscal policy, Current Account and the Real Exchange Rate in the U.S.,”Journal of International Economics 74, pp. 362–383.
2010, “Government Purchases and the Real Exchange Rate,”Open Economies Review, Springer, vol. 21(1), pages 49–64, February.
1995, “Exchange Rates and Fundamentals: Evidence on Long-Horizon Predictability,”American Economic Review, Vol. 85, No. 1, pp. 201–18.
2008, “Fiscal Policy, the Trade Balance, and the Real Exchange Rate: Implications for International Risk Sharing”. IGIER, Working Paper.
1963, “Capital Mobility and Stabilizing Policy Under Fixed and Flexible Exchange Rates,”Canadian Journal of Economics and Political Science, Vol. 29, pp. 475–85.
OECD,2011, Economic Survey of Brazil, (Paris: Organization for Economic Cooperation and Development).
2010, “Capital Inflows: The Role of Controls,”IMF Staff Position Note10/04 (Washington: International Monetary Fund).
Exchange Rate Forecasting, Order Flow and Macroeconomic Information (March2009). CEPR Discussion Paper No. DP7225.
2008, “Real Exchange Rates and Fundamentals: A Cross-Country Perspective,”IMF Working Paper 08/13 (Washington: International Monetary Fund).
2010, “The Macroeconomic Effects of Tax Changes: Estimates Based on a New Measure of Fiscal Shocks,”American Economic Review, Vol. 100, No. 3, pp. 763–801.
2012, “The Puzzle of Brazil’s High Interest Rate,”IMF Working Paper 12/62 (Washington: International Monetary Fund).
1964, “Theoretical Notes on Trade Problems,”Review of Economics and Statistics, No. 46, pp. 145–54.
1984, “Real Exchange Rate Effects of Fiscal Policy,”NBER Working Paper No. 1255 (Cambridge, Massachusetts, National Bureau of Economic Research).
Prepared by Marialuz Moreno Badia and Alex Segura-Ubiergo.
For a review of the literature, see Abbas et al, 2011.
Goods market clearing will result in a nominal appreciation assuming prices are sticky.
These include (i) the growth of public debt that may follow a fiscal expansion; (ii) the fiscal measures that may have to be taken to service growing debt; (iii) the wealth and portfolio implications of current account deficits induced by the fiscal expansion; and (iv) forward looking expectations in the asset markets.
The relationship between fiscal policy and the real exchange rate has been much less studied than the relationship between fiscal policy and the current account balance. For example, in a comprehensive review of the literature, Abbas et al, 2011 look at 20 papers studying the impact of fiscal policy on the current account balance, and only 5 analyzed the impact on the real exchange rate as well. Most studies find a positive relationship between budget balances and the current account.
The time dimension varies depending on countries/variables. For a description of the variables and a list of countries, see Appendix I.
Empirical analyses differ in their choices of the underlying real exchange rate fundamentals, sometimes because of data constraints. Alternative specifications were also estimated and some of these results are reported in the robustness checks. We did not include a measure of systemic risks in our estimation, such as VIX, since it is unlikely it would affect the REER long-term dynamics and the effect would in any case be wiped out once a time variable was included.
Standard intertemporal macroeconomic models predict debtor countries will need a more depreciated real exchange rate to generate trade surpluses necessary to service their external liabilities.
An alternative to deal with the endogeneity problem would be to use historical documents to identify changes in fiscal policy as has been done in the literature looking at the impact of fiscal policy on growth (see, for example Romer and Romer, 2010). One limitation of this approach, however, is that retrospective estimates of measures are rarely available and using contemporaneous assessments could be misleading since the size of the fiscal adjustment ex-post may differ from what policymakers believed ex-ante. In any case, Granger-causality tests seem to indicate that the REER does not cause movements in the structural balance.
For each country we focus on the top trading geographic destinations of its exports that account for at least 80 percent of exports during the period 1980—2010.
The choice of one lead and lag is dictated by the sample length.
Standard panel unit root tests do not reject the null hypothesis of a unit root for the real exchange rate. In addition, the tests indicate nonstationary for several of the explanatory variables (trade balance, structural balance). The DOLS methodology adds leads and lags of first differences of right-hand side variables to the set of regressors in order to wipe out the correlation of the residuals with the stationary component of the unit root process of the explanatory variables. Since this introduces serial correlation of the residuals, we use the Newey-West correction method to correct the standard errors. The DOLS residuals were found to be stationary using panel unit root tests, which is consistent with panel cointegration.
An alternative specification with time dummies shows relative public consumption to have a positive significant effect but this result is not robust and thus we do not report it in here.
Brazil has public investment ratios closer to the average of advanced economies; nevertheless, there are sizable infrastructure gaps, suggesting potential productivity gains from public investment could be large.
Nevertheless, in an alternative specification (not reported here) we find that capital inflows have a significant impact on the REER for Brazil although the effect is relatively small.
Public gross debt fell from 79.8 percent of GDP in 2002 to 63.5 percent of GDP in 2008 reflecting this effort. Moreover the composition of debt improved dramatically with substantial reductions in external and short-term indexed debt. Nonetheless, for some perspective, it is useful to recall that debt levels today are the same as in 2000. This reflects partly the spike in debt associated with the economic shock Brazil experienced in 2002-03, as well as the impact on debt of the stimulus extended during 2009-10 to offset the effects of the global crisis.