4. Fiscal Policy Response to the Crisis: How Much Space for Countercyclical Policy?

International Monetary Fund. Western Hemisphere Dept.
Published Date:
October 2009
  • ShareShare
Show Summary Details

The global crisis put fiscal policymaking at the forefront, highlighting differences in policy frameworks and preparedness within the region. Countries' circumstances prior to the crisis, largely reflecting past fiscal behavior, shaped the varied fiscal policy responses that Latin American and Caribbean (LAC) governments have recently taken. The experience of 2009 confirms that some LAC governments do have “space” to support economic activity during a major downturn. But the experience also draws attention to limits on such space, as well as the need for fiscal policymaking and frameworks to evolve—to be prepared for future shocks.

As the financial crisis in advanced economies quickly triggered a global recession, economic activity in LAC economies cooled down and, as a result, public finances deteriorated. Governments across the region soon saw their own revenues decline significantly, including those related to commodity exports. Moreover, financing conditions suddenly tightened, though in a number of cases this effect was neither severe nor long lasting. Amid these conditions, countries faced a basic choice: Should fiscal policy actively respond to the crisis and, if so, in which direction? In broad terms, three kinds of responses could be considered, each with its own rationale depending on country circumstances:

  • A country might actively tighten fiscal policy, aiming to contain the increase in the fiscal deficit and financing needs triggered by the shock to government revenue. Such a response, known as procyclical, would result in expenditure cuts (or tax increases), reducing domestic demand, at a time when output is in decline. Yet it might be the best available alternative if financing possibilities are limited, perhaps because public debt is already very high and the country is unable to assure creditors of future debt servicing, or the government does not have sizable liquid assets of its own.

  • A country could hold its expenditure and tax policies essentially constant, fully accepting a weaker fiscal balance as an automatic consequence of the revenue shock. This approach can be called acyclical in the sense that it involves no discretionary steps or revisions to policies at a time when output is falling. Even so, the public sector would be helping to stabilize demand and support economic activity by allowing so-called automatic stabilizers to operate, at least by collecting less tax revenue.7 In the case of a temporary drop in output and revenues, such an approach would be a natural reaction for governments with good access to financing (or to their own liquid assets), more so if preexisting automatic stabilizers are judged to be adequate in size, and if monetary policy is free to play an active and effective role in stabilizing domestic demand.

  • A country might decide to actively loosen fiscal policy, taking discretionary steps to raise expenditures (or cut taxes), seeking to provide extra support for demand at a time when activity is weakening. Such a countercyclical response implies an even larger increase in the fiscal deficit than that triggered by the revenue shock alone, so financing possibilities would need to be plentiful. Considerations favoring this approach could include automatic fiscal stabilizers that are relatively small and confidence that discretionary easing can be implemented—and reversed—in a timely manner. More fundamentally, this approach may be called for when monetary policy alone cannot be sufficiently effective in countering an unusually large drop in domestic demand, as may have been the case recently in some countries.8

In fact, fiscal policy has been playing a role to buffer the impact of the crisis in several countries in the region during 2009, but in greatly varying degrees. While fiscal balances are likely to weaken in nearly all LAC countries this year, some countries took steps to tighten their fiscal positions, while others actively raised expenditures—in some cases, substantially. This chapter aims at measuring the fiscal response to the crisis in the region, understanding the basis for the observed patterns, and drawing lessons that may inform fiscal policymaking in the future.

The first part of the chapter briefly looks at LAC countries' fiscal policies prior to the global crisis, noting differences in how countries reacted to strong revenue growth in the last several years of “good times.” Next, the chapter analyzes fiscal developments during the crisis by decomposing the revenue shock, finding that most revenue losses were automatic consequences of the crisis itself. Turning to the expenditure side, the chapter documents wide variation across countries in the size—and direction—of expenditure policy responses to the crisis. Although the focus is on the level of primary expenditure, it is noted how some countries acted to change its composition, seeking to mitigate the effects of the crisis on vulnerable groups.

To evaluate the contribution of the public sector to domestic demand in 2009, the chapter looks at the change in fiscal balance measures that exclude revenues from commodity exports and foreign grants. Then, to identify the part of that contribution that arose from discretionary policy decisions rather than automatic stabilizers, the chapter presents fiscal impulse estimates and assesses their effect in stabilizing output. In turn, these policy responses are linked to a set of factors likely to influence countries' choices of whether to pursue an active countercyclical policy.

Finally, drawing lessons from the recent experiences of LAC countries, the chapter identifies key fiscal policy issues, including the need to eventually unwind recent stimulus and to adapt to a new global environment that will likely be less favorable than in the precrisis expansionary years. The recent experience suggests ways in which fiscal policy frameworks could evolve, not only to support fiscal discipline over time, but also to establish space for maintaining public expenditure and buffering macroeconomic shocks in the future.

The Precrisis Context: Prudent Fiscal Policies and Some Profligacy

Previous studies have emphasized that fiscal policies in the LAC region have often been procyclical, with fiscal policy contributing to output fluctuations. Actively countercyclical policies, and sometimes even the normal functioning of automatic stabilizers, were not possible owing to the absence of financial buffers and large levels of public debt that restricted financial market access during downturns, as prospective creditors were reluctant to lend (Gavin and Perotti, 1997; and Reinhart, Rogoff, and Savastano, 2003, among others).9

The last decade witnessed significant change in the fiscal policies of several countries in the region. Better fiscal discipline over time brought down public debt levels in many cases (Vladkova-Hollar and Zettelmeyer, 2008). Moreover, some countries managed to build enough political consensus for the application of fiscal rules, at times embedded in fiscal responsibility laws (FRLs).10 Successful implementation of fiscal rules supported fiscal discipline and therefore contributed to avoiding crisis episodes in which fiscal policy is forced to overreact in response to a cutoff in financing. The combination of improved policy frameworks and better fundamentals resulted in a number of countries in the region gaining investment-grade credit ratings, which brought better access to international capital markets, even in times of stress.

This shift has resulted in considerable heterogeneity in the conduct of fiscal policy. While some countries have broadly continued with procyclical policies, in particular with expenditures closely tracking revenues upward in favorable times, others have attained considerable countercyclicality.

This difference is illustrated by the behavior of primary fiscal revenues and expenditures in the three years before the crisis (2005–07). These were years of above-trend growth, increasingly positive output gaps, and rising commodity export prices. As in Chapter 2, it is useful to refer to four groups of LAC countries (Figure 4.1).11 In commodity exporting, financially integrated countries (CEFI), growth of primary expenditures was lower than that of total revenues, and was about equal to the contribution of noncommodity revenues to total revenue growth. In contrast, primary expenditure growth slightly exceeded total revenue growth in the other commodity exporting countries (OCE), and far surpassed the contribution of noncommodity revenues to total revenue growth. In the commodity importing, tourism intensive countries (CITI), primary spending grew at higher rates than revenues, while in the other commodity importing countries (OCI), primary expenditures grew somewhat less than revenues.

Figure 4.1.Some countries have continued with procyclical policies, while others attained countercyclicality.

Source: IMF staff calculations.

A similar message comes from comparing growth of primary expenditures to that of trend GDP. In the other commodity exporting countries, primary expenditures grew much faster than trend GDP growth during the commodity boom years preceding the global crisis. As discussed below, this behavior implied a strong fiscal impulse to domestic demand for these countries during this period, compared with only a mild impulse in commodity exporting, financially integrated countries.

The Crisis Hits: Lower Output and Commodity Prices Slash Fiscal Revenues …

LAC governments have experienced significant revenue losses during the crisis, in some cases more than expected. Losses were linked to weak (or negative) output growth and, in many of the region’s larger economies, steep drops in commodity export prices. The declines in fiscal revenues during 2009 seem to be mostly “automatic” consequences rather than arising from discretionary actions on tax policy. The size of automatic stabilizers is, however, relatively small in many LAC countries (compared with those of most OECD countries, for example), given their lower tax ratios.

Indeed, countries where fiscal revenue is strongly linked to commodity exports have had very large revenue drops in 2009 (Figure 4.2).12 In such countries, the decrease in commodity export prices accounts for the majority of the decline in revenues (with Bolivia, Ecuador, Trinidad and Tobago, and Venezuela showing the largest decreases, though drops are also important for Chile and Peru). Only 10–20 percent of the revenue loss seems attributable to the decrease in GDP growth rates (based on IMF staff assumptions of revenue elasticities to GDP). The remaining portion of the revenue loss is not identified and could reflect any tax policy changes, other structural changes (such as falling production of commodities), or that revenues are more sensitive to the output cycle than assumed.

Figure 4.2.Revenues declined significantly in 2009, linked to both lower output and commodity prices.

Source: IMF staff calculations.

For commodity importing countries, the residual unidentified part of the revenue loss is the most important. These losses can neither be attributed to changes in policy, as most of these countries are not implementing discretionary decreases in taxes in 2009, nor to drops in commodity production, since these countries are not large commodity exporters. Instead, the residual must relate mostly to the actual value of the short-term elasticity of tax revenues with respect to GDP turning out to be larger than the standard assumption of a value equal to 1.13

… But Only Some Countries Can Afford to Keep Expenditure Growing …

Expenditure developments in the region can be interpreted as mainly the result of discrete policy actions, since automatic stabilizers on the expenditure side are also relatively small. In fact, expenditure policies differ significantly across countries—in some countries, expenditures seem to mimic the behavior of revenues; in others, this is not the case.

In particular, there is a strong contrast in the behavior of primary expenditure (in real terms) among the commodity exporting countries (Figure 4.3). While primary expenditure growth in financially integrated countries is expected to be somewhat similar in 2008 and 2009 (increasing by about 9 percent and 8 percent, respectively), it is expected to decline significantly between 2008 and 2009 in the other commodity exporting countries (having grown by about 12 percent in 2008, but only increasing by about 3 percent during 2009), as the decrease in commodity-related fiscal revenue turns into a binding constraint. Primary expenditures also show positive growth rates in 2008 and 2009 for other commodity importing countries, though the expected rate for 2009 is lower than that observed in 2008. In commodity importing, tourism intensive countries, real primary spending slightly contracted in 2008, and a further decrease is expected in 2009, reflecting a tightened constraint from the revenue side.

Figure 4.3.Some countries had to curtail expenditure growth in response to the revenue shock.

Source: IMF staff calculations.

A more detailed look at the cross-country pattern for 2008–09 shows that the growth of real primary spending came to a halt for some countries of the region during 2009. In particular, primary expenditure is contracting in real terms in a number of countries of the region suffering a reversal of the terms of trade linked to energy export prices, including in Bolivia and Mexico, and especially in Ecuador, Trinidad and Tobago, and Venezuela. This behavior seems to indicate, for some of these countries, a continuation of the procyclical expenditure patterns observed in the past. In contrast, primary expenditure is increasing at rates similar (or higher) than those observed in 2008 in other countries of the region (including Argentina, Chile, Costa Rica, and Peru, among others). In many countries, the expected growth in primary expenditure (at a rate higher than trend GDP growth) reflects the implementation of stimulus packages (Box 4.1).

… Resulting in Varied Fiscal Support During the Crisis

The discussion so far has concentrated on the individual pieces of the fiscal policy reaction, that is, on the behavior of revenues and expenditures separately. Here, the pieces are brought together to gauge the size of the contribution of the public sector in stabilizing domestic demand and output in 2009. While the net change in domestic fiscal revenues and expenditures affects domestic demand, in turn its impact on output is assumed to depend on the size of the fiscal multiplier: The larger the multiplier, the larger the impact in output of changes in fiscal policy.

Box 4.1Fiscal Stimulus Packages in Latin America and the Caribbean

A number of countries in the LAC region announced specific “anticrisis fiscal packages” during 2009 aimed at stabilizing aggregate demand, providing economic stimulus, and granting relief, on a mostly temporary basis, to vulnerable groups. In the majority of cases, the measures were part of countries' budgets (in their original or reformed versions) and were to be implemented by central governments as, for instance, increases in public infrastructure spending or compensatory transfers to vulnerable groups. However, there were other measures that involved extrabudgetary transfers to state-owned enterprises (SOEs) or the provision of liquidity to state-owned development banks with the purpose of sustaining domestic credit. Most fiscal packages concentrated on increasing expenditures, transfers, or both, but some countries announced temporary tax reductions or administrative measures that allowed taxpayers to defer payment of their tax obligations. On the other hand, some countries were forced to introduce tax measures and improve tax administration to offset the decrease in revenues. The main elements of some of the fiscal stimulus packages announced across the region are described in greater detail in this box.

Among commodity exporting financially integrated countries:

  • Brazil announced several on-budget tax and spending measures to stimulate aggregate demand and support specific groups (through cash transfers to the poor, tax breaks for auto purchases, and subsidies for housing construction for the poor). States were partially shielded from transfer declines, with some in negotiations to improve debt terms with federal bodies (for example, social security). The fiscal package also included small-scale loan guarantee programs for small and medium-sized enterprises. The largest measure was off budget: a below-the-line loan of 3.3 percent of GDP to BNDES, a state-owned development bank, over two years for lending, including to the state-owned oil company Petrobrás. Other public banks also increased lending.

  • In Chile, the fiscal package included both on-budget and off-budget measures; the former comprised temporary tax reductions and increases in transfers to the poor, while the latter included transfers to state entities providing credit guarantees, the capitalization of the state-owned bank, and the state-owned copper mining company, CODELCO.

  • Mexico announced measures including increases in social and infrastructure spending (above the line), freezes and reductions of administered prices, increases in lending limits, and the provision of additional guarantees by development banks.

  • In Peru, measures included large increases in infrastructure spending, transfers to foster social protection (education, health, and agriculture), guarantees of funds, transfers to sectors affected by the crisis, and drawbacks, in that order.

Among other commodity exporting countries:

  • Argentina announced large increases in infrastructure and other capital spending, a widening of coverage under antipoverty programs and pensions, measures to support consumer credit, and some reductions in export taxes, among other measures.

  • In Paraguay, the economic reactivation plan includes short-term measures, including a larger (2 percent of GDP more expansionary) budget in 2009–10 than that of 2008 and some refocusing of conditional cash transfers and capital spending (financed partly by international financial institutions), and measures addressing medium-term structural issues, including public financial management improvement, SOE reform, and stronger financial sector supervision.

Among commodity importing, tourism intensive countries, The Bahamas shifted the composition of spending to maintain expenditures on infrastructure and provide transitory unemployment benefits, though revenue measures were also announced to contain the fiscal impact of the shock. In Grenada, measures included the acceleration of capital spending within a tighter budget, temporary tax relief to hotel and guest houses (to support the tourism industry and limit job losses), and targeted social spending (for example, a road maintenance program).

Among other commodity importing countries, Costa Rica announced a number of measures, including increased spending on education and labor-intensive infrastructure projects, an increase in public sector employment (through the hiring of teachers and police officers), and strengthening of cash transfer programs. In El Salvador, authorities announced increases in social spending and in public investment projects, free medicines for hospitals, and urban conditional cash transfers and temporary employment programs. In Guatemala, measures focused on transfers to vulnerable groups and labor-intensive investment projects. In Panama, measures focused on infrastructure works and expanding the social safety net; moreover, authorities granted noncontributory pensions to the elderly and increased wages for the police. The Fiscal Responsibility Law was modified in June 2009 to allow a deficit of up to 2.5 percent of GDP (up from 1 percent) in 2009 and 2 percent in 2010. The Canal expansion project, though it is not part of the stimulus strategy, is expected to support economic activity during 2010, despite its high import content.

The importance of timely implementation of the stimulus measures (and of budgets, more generally) is particularly high given the generalized slowdown in economic activity across the region. Although a detailed report on cross-country budget implementation (including fiscal stimulus measures) is difficult, given some data limitations, available data for 2009 indicate that progress has been mixed. In this connection, although a number of measures requiring only administrative decisions were implemented relatively quickly, spending proceeded at a pace that was slower than planned in a number of countries during the first half of 2009 and is picking up during the second half of the year. The reasons behind the delays include problems with implementation capacity (particularly in regard to investment), delays in approving budgets or budget reforms, and delays in approving external debt contracting, among others.

Changes in “domestic primary fiscal deficits,” that is, excluding revenue related to commodity exports and foreign grants, provide a simple summary view of the contribution of the public sector to domestic demand, one that includes the effects of both discrete policy actions and automatic stabilizers.14 An expansion of the domestic primary deficit, as shown below, signifies a push to domestic demand from the public sector. From this perspective, governments of commodity exporting, financially integrated countries are contributing most strongly to domestic demand in 2009, with a deficit increase of about 3.5 percent of GDP (following several years of approximately neutral contributions). The situation is starkly different for other commodity exporting countries, where domestic primary deficits are expanding by about 0.7 percent of GDP (in contrast to the strong expansion of 1.9 percent of GDP during 2008). In the two other country groups, domestic primary deficits are growing in 2009, albeit not by as much as in the first group.

Movements in Domestic Primary Deficits 1/(Average change in ratios to GDP, from previous year, in percentage points)
Commodity exporting financially integrated countriesOther commodity exporting countriesCommodity importing tourism intensive countriesOther commodity importing countries
Source: IMF staff calculations.

Excluding commodity-related fiscal revenues and foreign grants.

Source: IMF staff calculations.

Excluding commodity-related fiscal revenues and foreign grants.

Next, to distinguish the actively supportive role of countercyclical fiscal policy from the support coming from automatic stabilizers, the cyclically adjusted primary balance (CAPB) is first estimated. A standard framework is used to identify the cyclical reaction of domestic primary revenues and primary expenditures to the output gap. For countries where a significant portion of fiscal revenues are commodity related, an estimate is provided as to what part of the observed commodity-related revenues is “structural,” using estimates of long-term commodity prices.

The sign and size of the CAPB gives an indication of the underlying fiscal position, abstracting from temporary influences. It is also useful for projecting the future path of public debt, when taking into account the interest rate at which the government may borrow and the economy’s trend growth rate, among other factors.

Headline (that is, unadjusted) primary balances improved during the expansion years (2003–07) for the commodity exporting countries. In turn, among commodity importers, the primary balance improved initially and then deteriorated, especially in tourism intensive countries. For 2009, the headline primary balance is expected to deteriorate for all groups, although the other commodity exporting countries are expected to tighten their CAPB, while commodity exporting, financially integrated countries are relaxing theirs. This pattern appears to reflect differing policies during the expansion years, in particular the use of past revenue windfalls to build financial buffers, as opposed to finance expenditure increases (Figure 4.4).

Figure 4.4.Not all countries have been able to ease cyclically adjusted fiscal balances in the crisis.

Source: IMF staff calculations.

Estimations of CAPBs are subject to uncertainty, particularly related to commodity revenue.15 Countries with larger shares of commodity-related revenues in total revenues (as in Ecuador and Venezuela) will be subject to greater uncertainty as to the actual size of the CAPB. Bearing in mind this uncertainty, the largest countries in the region appeared to have had cyclically adjusted surpluses during the first part of the decade, in part reflecting improvements in long-term commodity prices. However, during the three years prior to the crisis, CAPBs improved in some cases (as in Chile and Peru) and worsened in others (as in Argentina, Ecuador, Mexico, and Venezuela). For 2009, a number of countries are letting their CAPBs deteriorate in varying degrees in response to the downturn (including Argentina, Brazil, Chile, and Peru), while others are consolidating their cyclically adjusted positions (including Ecuador and Venezuela).

While the level of the CAPB is the relevant indicator for questions of solvency and debt sustainability, its year-to-year changes are useful for questions of the short-term impact of fiscal policy decisions on domestic demand and output. Relatedly, the fiscal impulse (FI) is defined as the difference between the domestic CAPB (that is, excluding commodity-related fiscal revenues and foreign grants) in the current year and that of the previous year (Box 4.2). This FI is not influenced by commodity price issues, though it is still subject to measurement uncertainty associated with the size of the output gap.

The FI during the years preceding the crisis as well as that expected for 2009 shows some interesting differences across the region. Commodity exporting, financially integrated countries displayed a neutral to mildly procyclical fiscal policy stance during the expansion years (2003–07), while current plans imply the implementation of a clearly countercyclical fiscal policy during 2009, withdrawing some stimulus in 2010 as economies recover. FI changes have been less pronounced and fiscal policy has been less volatile, when compared with other country groups. In contrast, other commodity exporting countries implemented a clearly procyclical fiscal policy during the expansion years, and they will continue to do so during the downturn. For these countries, fiscal policy changes are at times abrupt and fiscal policy in general has been more volatile than for commodity exporting financially integrated countries.

Interestingly, fiscal policy in the other commodity importing countries appears on average to have been mildly countercyclical in past years—a trend that continues in 2009. In the case of Costa Rica, the fiscal impulse in 2009 is especially substantial. Finally, commodity importing, tourism intensive countries have implemented procyclical fiscal policies through most of the years analyzed, with such trends also continuing during 2009–10 (Figure 4.5).

Figure 4.5.Fiscal impulses varied greatly. Commodity exporting, financially integrated countries achieved the greatest countercyclicality in 2009.

Source: IMF staff calculations.

As the table shows, the entire role of the public sector in stabilizing domestic demand, as captured by the change in the headline domestic primary deficit, can be decomposed into two parts. The FI measures the part of such a change explained by discretionary policy responses, whereas the rest of the change is explained by the operation of automatic stabilizers.

Box 4.2.Measuring the Fiscal Stance: The Cyclically Adjusted Primary Balance and the Fiscal Impulse

The methodology for estimating the cyclically adjusted primary balance (CAPB) used in this chapter allows the portion of the changes in fiscal variables that result from policy decisions to be identified and differentiated from the portion resulting from the economic cycle (orautomatic stabilizers).1 In turn, the estimation for the CAPB allows calculation of how much impulse fiscal policy is providing to domestic demand in a given year. Such fiscal impulse is measured as the difference between the cyclically adjusted primary deficit in two consecutive years.

Estimating the CAPB requires country-specific information, including the elasticity of fiscal revenues and expenditures with respect to changes in output, as well as an estimation of the output gap. The “typical” case for the region in 2009 is one in which (1) GDP growth is negative, and the output gap is deteriorating with respect to 2008, which is also negative; (2) the income elasticity of domestic fiscal revenues (that is, excluding commodity-related revenues and foreign grants) is equal (or larger) than one, meaning that changes in output (in percentage terms) are reflected in similar (or larger in absolute terms) changes in fiscal revenues; and (3) the income elasticity of primary expenditures is equal to zero, meaning that the level of spending does not change as a result of the economic cycle, but rather that changes in its level are mostly the result of discrete policy measures.

In the typical case for 2009, the primary expenditure–to–GDP ratio is increasing and the domestic primary revenue–to–GDP ratio is decreasing (or staying about constant). Absent discretionary policy measures, domestic revenues will decline, but the ratio of domestic revenues to GDP will remain about constant or decrease, given estimated elasticities. In turn, the levels of primary expenditures will remain about constant, but the ratio of expenditures to GDP will increase.

Automatic stabilizers are estimated by comparing fiscal variables (as percentages of GDP) in the downturn with those same variables when the output gap is zero:

  • On the expenditure side, the level of expenditures remains about constant, but increases in terms of GDP and provides some stabilization, which is measured by subtracting the ratio of expenditures to GDP in bad times to that ratio when the output gap is zero. Such a difference constitutes the automatic stabilization provided by expenditures or, in other words, the expenditure (in percentage of GDP) that occurs for cyclical reasons.

  • On the revenue side, the level of domestic revenues falls and the domestic fiscal revenues–to–GDP ratio decreases (or remains about constant). The portion of the decline in fiscal revenues providing automatic stabilization would again be measured by the difference between the domestic revenue–to–GDP ratio in the downturn and that ratio when the output gap is zero.

The domestic CAPB results from excluding from the observed domestic primary revenues–and expenditures–to–GDP ratios, the portion that results from automatic stabilization. Any changes in the domestic cyclically adjusted primary balance between two consecutive years (that is, the fiscal impulse) would add one to one to (or subtract one to one from) domestic demand. A deterioration in the CAPB (or analogously, an increase in the primary deficit) implies a positive contribution of fiscal policy to domestic demand.

1 A more formal discussion of this topic is included in the Technical Appendix.

A look at the largest countries of the region suggests that a number of them ran fiscal policies that were procyclical, in varying degrees, during the precrisis period. For 2009, the response is expected to be clearly countercyclical in the cases of Brazil, Chile, Mexico, and Peru, and less so in the cases of Argentina and Colombia; responses are clearly procyclical in the cases of Ecuador and Venezuela.16

Finally, the contribution of fiscal impulse to output stabilization is assessed, depending on both the estimated size of the impulse and that of the assumed fiscal multiplier. In this regard, a number of studies estimate that, for developing countries, the fiscal multiplier is about 0.5, although the range of estimates is wide.17 Using that estimate and assuming a one-quarter lagged effect for fiscal policy, the commodity exporting, financially integrated countries show the largest fiscal contribution to output stabilization.

Why Did Active Fiscal Policy Responses Differ? A Look into Fiscal Space

Fiscal impulses in 2009 are expected to differ across countries. While in some countries fiscal policy will provide stimulus, in others it will withdraw stimulus, potentially contributing to the downturn. Why such different responses?

Decomposition of the Changes in Domestic Primary Deficits into Fiscal Impulses and Automatic Stabilizers(Percent of GDP)
Commodity Exporting, Financially Integrated CountriesOther Commodity Exporting Countries
YearPrimary deficit change

(1 + 2)
Automatic stabilizers

Fiscal impulse

Primary deficit change

(1 + 2)
Automatic stabilizers

Fiscal impulse

Commodity Exporting, Financially Integrated CountriesOther Commodity Exporting Countries
YearPrimary deficit change (1 + 2)Automatic stabilizers

Fiscal impulse (2)Primary deficit change (1 + 2)Automatic stabilizers

Fiscal impulse (2)
Source: IMF staff calculations.
Source: IMF staff calculations.
Commodity Importing, Tourism Intensive(Percent of GDP, simple averages)
Commodity Exporting, Financially Integrated CountriesOther Commodity Exporting CountriesCommodity Importing, Tourism Intensive CountriesOther Commodity Importing Countries
Source: IMF staff calculations.
Source: IMF staff calculations.

Some analysis suggests that the answer to this question is related to differences in the conditions prevalent in countries of the region before they were hit by the crisis. Indeed, the smaller fiscal impulses of some countries may have reflected constraints on their ability to secure more financing. In other words, countries implementing countercyclical policies in 2009 are those that have the ability to do so. This is the notion of fiscal space, which is related to the availability of financial buffers, the ability to access international capital markets, and having stronger precrisis primary balances. Countries with larger debt ratios, lower debt ratings, and no financial buffers would have less space to conduct fiscal policies than those with lower debt ratios, investment grade debt ratings, and available financial buffers.

This conclusion is supported by a simple crosscountry regression (with each country weighted by its relative size) that relates fiscal impulse in 2009 with the level of the cyclically adjusted fiscal balance in 2008 and the credit rating on external debt, as indicators of each country’s “initial” conditions.

The results indicate that FIs in 2009 are higher in countries with stronger initial CAPBs and with stronger credit ratings (Figure 4.6). Relating these results with the analysis above suggests that countries running procyclical policies tend to take advantage of abundant liquidity during moments of favorable international conditions, but when market access turns more difficult, they are bound to adjust their fiscal positions, resulting in negative fiscal impulses. In contrast, governments running countercyclical policies enjoy continued access to international capital markets. A few countries stand out as having an FI different from the average reaction. Among others, Chile and Costa Rica (the latter in the context of an IMF-supported program) appear to have larger FIs, while Ecuador and Venezuela have smaller ones.

Figure 4.6.Countries with stronger fiscal positions and credit ratings were able to ease fiscal policy.

Source: IMF staff calculations.

1/ Bubble sizes reflect country sizes in 2008 (valued at PPP exchange rates).

There are other possible explanations for the differing fiscal policy responses. One possibility is that responses were constrained by fiscal rules that limited overall deficits. What was observed, however, is that countries with functioning fiscal rules relaxed or modified rules in ways that accommodated larger fiscal deficits, owing to the notion that circumstances in 2009 were extraordinary. On the other hand, the fiscal response of some Caribbean countries appears to have been bound by their debt-reduction target.

Initial Conditions and Cross-Country Variation in Fiscal Impulses, 2009
CoefficientsRobust Standard ErrorP-value
Debt rating 1/-0.310.010.00
Cyclically adjusted primary balance 20080.600.010.00
Adjusted R-squared0.68
Observations (weighted by PPP GDPs)31
Source: IMF staff calculations.

Lower values imply stronger ratings.

Source: IMF staff calculations.

Lower values imply stronger ratings.

Fiscal policy responses could in principle be constrained by problems of coordination between monetary and fiscal authorities. However, the largest fiscal response occurred in countries with inflation-targeting regimes and flexible exchange rates, and, in those countries, monetary policy conditions were relaxed along with fiscal policy. Moreover, the fact that many countries with fixed exchange rates are actually running procyclical fiscal policies might indicate the presence of limited access to external capital markets, and thus that authorities are reluctant to accumulate further debt to avoid jeopardizing their exchange rate system.

Some countries might have refrained from actively implementing countercyclical fiscal policy because they were uncertain whether it would be effective in stabilizing output, or they expected that letting existing automatic stabilizers operate would be enough. This might explain the policy response in economies with very open current accounts, as openness is thought to reduce the size of the multiplier. However, such a pattern is not clear: some countries with very open current accounts are implementing procyclical fiscal policies (as in most of the Caribbean), while others are applying countercyclical policies (as in Chile and Peru). Moreover, there is no clear pattern between the size of FIs and that of automatic stabilizers.

Looking Ahead: Balancing Fiscal Sustainability with Sustaining the Recovery

What path should the fiscal balance take once the economic recovery takes hold and medium- and long-term fiscal goals regain importance? There are a number of aspects to this question.

The first aspect is the appropriate timing for withdrawing the fiscal stimulus implemented in 2009. As discussed in Chapter 2, the answer depends on a number of factors, including the timing and strength of the recovery and also on the degree of remaining space or buffers available. While there is always concern with removing stimulus too early, countries that have few buffers and unclear access to financing may need to err on the side of caution, keeping their remaining ammunition in reserve. More generally, it will usually make sense to remove fiscal stimulus before monetary stimulus.

The second aspect is related to the impact of current policies on medium-term debt ratios. The functioning of automatic stabilizers implies that fiscal balances deteriorate in recessions and improve in expansions, and thus, there is a public debt cycle associated with the fiscal balance cycle. However, the long-term path for public debt will depend on the CAPB. In the current context, the discretionary response of fiscal authorities in many countries has weakened the CAPB, and thus, if the extra impulse is not withdrawn, debt ratios may end up increasing. Thus, it is important to assess the magnitude of fiscal consolidation needed to ensure that public debt ratios at least remain stable at end-2008 levels, or decline in cases of overly high levels. This analysis is particularly important for countries with already high public debt ratios and those with difficulties accessing capital markets, as well as for those with potentially high contingent liabilities (for instance, the capitalization of state-owned enterprises, development banks, or the central bank, which can cause discrete jumps in the public debt level).

A basic debt sustainability analysis shows that LAC countries will need eventually to strengthen their fiscal balances significantly (with respect to that expected for 2009), if public debt ratios are to remain constant at the levels observed at end-2008. However, the adjustment needed is of a different nature for different countries. For instance, the adjustment needed in commodity exporting, financially integrated countries will be produced mostly automatically as the economy recovers, output gaps are closed, and commodity prices return to long-term values. In contrast, the adjustment needed in the remaining groups is mostly structural, that is, it will need policy action by fiscal authorities. Finally, if the global crisis persistently affects trend growth and interest rates (that is, more so than already assumed for the baseline scenario—see Chapter 2), the fiscal adjustment needed to keep debt ratios constant will increase, with the increase being larger in countries with a legacy of larger debt ratios (Figure 4.7).18

Figure 4.7.Fiscal positions will need to strengthen significantly to keep debt from rising.

Source: IMF staff calculations.

1/ Values in parentheses refer to public debt in percent of GDP at end-2008.

2/ Structural adjustment to headline primary balance in 2009 that leaves debt ratios unchanged.

3/ Including for commodity prices.

4/ If trend growth is lower by 0.35% and interest rates are higher by 90 basis points.

The third aspect is related to fiscal policy frameworks, targets, and rules. The current crisis has shown that countries that had secured enough political consensus for the application of prudent fiscal frameworks entered the crisis better prepared, in particular owing to the accumulation of large fiscal buffers during good times (as in Chile and Peru). This is particularly important for countries with a large proportion of commodity-related revenues in total revenues, as fiscal revenues in such countries tend to be more unpredictable, with commodity gaps not necessarily coinciding with domestic output gaps. The presence of such fiscal rules has not constrained the implementation of discretionary policy measures, as credibility in their resumption allows some flexibility in their application. Even if not formally adopting a rule, targeting a cyclically adjusted fiscal balance can be useful in policy analysis and discussion, as well as to broadly guide fiscal policy. The estimation of CAPBs would, at least, provide a reference point and help increase accountability, promoting saving during good times, raising credibility of fiscal policy, and allowing countries to increase net debt during bad times by accessing capital markets. Carefully designed “escape clauses,” when fiscal rules are well entrenched, would also be desirable.

The fourth aspect is concerned with the functioning of automatic stabilizers. These are generally small in the region on both from the revenue and the expenditure sides—and arguably too small.19 Regarding those from the expenditure side, stabilization is only provided by the size of primary spending with respect to output. Only a few countries have some form of automatic transfers linked to the cycle, and these are often minor or negligible. It would be worthwhile to consider and carefully design further automatic stabilizers, including larger, temporary, automatically triggered transfers to the most vulnerable sectors of the population (perhaps in the form of unemployment insurance), and even consider mechanisms that provide relief to taxpayers (both individuals and corporations) during downturns. Increasing the size of automatic stabilizers would reduce the need for discretionary fiscal policies during bad times, and enhance credibility by avoiding the political cost of withdrawing discretionary stimulus when the economy recovers.

The fifth aspect is related to the role of discretionary policies at current and future junctures. Discretionary fiscal policies were justified given the extraordinary magnitude of the current crisis, as well as the need for a coordinated international response. For less extraordinary, more normal times, a good strategy would be based on the following: a careful design of automatic stabilizers, ensuring that they are not too small; the estimation of a CAPB to inform, and perhaps to guide, fiscal policy, to accumulate financial buffers during good times (reducing debt and improving liquidity access); and continuous coordination between monetary and fiscal policy, subject to the constraints imposed by the monetary framework. While fiscal policy has regained prominence with the crisis, and will maintain a central role in less flexible exchange rate regimes, monetary policy should normally be at the forefront in cyclical management in countries with more flexible exchange rate arrangements.

This chapter concludes with two observations about the impact of fiscal policy choices implemented in 2009. First, regarding the overall level of national income, it is extremely difficult for any one country to quantify what would have happened if fiscal policies had responded differently to the crisis. Extensive literature addresses this question, and suggests a wide range of possible fiscal “multipliers,” depending on many factors that vary across countries and over time. The presumption here is that, first, those LAC countries that had the fiscal space to raise expenditure in 2009 were in general wise to do so, even without being certain of its degree of success in supporting output; in the unusual circumstances of large and growing output gaps, a positive impact was reasonably likely, and worth the attempt. And second, beyond the effort to support total demand and output, there was also a special challenge in 2009 for fiscal policy to play an important social role, to cushion the effect of the downturn on vulnerable groups, irrespective of demand management objectives. Ideally, such social responses should be designed and planned far in advance of downturns, with their implementation being stepped up (and subsequently phased out) in an essentially automatic manner. However, where such systems were not already adequately established, the size of the shock in 2009 called for a discretionary policy response—and vulnerable groups benefited to the extent that their governments were ready and able to act, having previously accumulated the necessary fiscal space.

Technical Appendix

The Fiscal Impulse

The fiscal impulse measures the impact on domestic demand of discretionary changes in fiscal policy. A positive fiscal impulse is defined as a discretionary change in policy that results in an increase in domestic demand.

The measurement of the fiscal impulse is complicated by the fact that the observed fiscal position is influenced by cyclical factors. This chapter uses a standard framework that identifies what portion of each country’s fiscal position can be attributed to cyclical factors (that is, the workings of automatic fiscal stabilizers) and what portion to discretionary fiscal responses.

The size of automatic stabilizers depends on the size of the public sector, on revenue and expenditure elasticities with respect to output, and on fiscal-related legislation (for example, the existence or absence of unemployment-related transfers, the composition of tax revenues between indirect and direct taxes, and the progressivity of the tax system, among others). Concretely, the framework is defined by the following expressions:

In the expressions above, gapt=(ytyt*)/yt* where y is real output and y* refers to real trend output, T denotes domestic primary revenues (that is, excluding commodity-related fiscal revenues and foreign grants), G denotes primary expenditures, Y refers to output, and an asterisk denotes cyclically adjusted concepts. In addition, ηT, Y is the income elasticity of domestic revenues, and ηG, Y is the income elasticity of primary expenditures. Finally, pdeftD,* refers to the cyclically adjusted domestic ftt denotes the fiscal impulse: An increase in the cyclically adjusted domestic primary deficit implies a positive fiscal impulse.20

The measurement of fiscal impulse excludes commodity-related fiscal revenues and foreign grants, as they mostly originate from abroad. For instance, an increase in fiscal revenues in the form of grants will not imply a decrease in the disposable income of a resident, but rather, in that of a nonresident. In contrast, the impact of the use of such revenue will be fully captured in primary expenditures. The same applies for commodityrelated revenues, which mostly originate from exports.

The Cyclically Adjusted Primary Balance

In countries where the share of commodityrelated revenues and foreign grants in total fiscal revenues is not significant, the cyclically adjusted primary balance (CAPB) is defined by the negative of Expression (3).

From a policy perspective, the sign of the CAPB can have different interpretations depending on whether the output gap is positive or negative, and also depending on the cost of servicing public debt. For instance, at times of positive output gaps, a positive CAPB could be interpreted as a relatively strict fiscal policy, which could be the result either of prudence or of the need to service an onerous public debt. Conversely, a negative CAPB in times of positive output gaps could be associated with a lax fiscal policy, more so in the presence of a large interest bill. More important, the sign of the CAPB will determine the speed of public debt accumulation through the economic cycle.

In countries with significant commodity-related fiscal revenues (or foreign grants), the CAPB will need to include an estimate of the long-term value of such fiscal revenues.

Commodity-related fiscal revenues can take a number of country-specific forms (including royalties, resource rent taxes, regular income taxes, production-sharing schemes, indirect taxation, or other nontax payments) or could result from the direct involvement of the state in the production of natural resources (Davis, Ossowski, and Fedelino, 2003).

Accounting for the commodity-related impact on fiscal accounts will be more challenging in countries that are resource rich in a number of commodities as opposed to just one. Moreover, it is necessary to differentiate commodity-related revenues that originated from exports from others that originated domestically. The finiteness (or not) of the natural resource under analysis introduces additional complications, as it blurs the distinction between temporary and permanent commodity-related revenues.

Concretely, cyclically adjusted commodity-related fiscal revenues are estimated as follows:

In Expression (5), crt denotes observed commodity-fiscal revenues (as percent of output), and crt* denotes cyclically adjusted commodity revenues (as percent of trend output) expressed at long-term commodity prices, CPt* (as opposed to observed prices, CPt).

The analysis assumes that commodity revenues depend on variables that are mostly beyond the control of fiscal authorities; it also assumes that commodity revenues do not always move together with output.

In countries where foreign grants constitute a large proportion of fiscal revenues, the calculation of the CAPB should proceed using revenues, excluding grants, for reasons analogous to those for which commodity-related revenues are excluded. Interestingly, if grants are large and volatile and the recipient government is cash constrained, the fiscal impulse will reflect such volatility. As grants can have different purposes (some negatively correlated with the cycle, as, for instance, humanitarian relief, while others are not), the analysis in this chapter assumes that such correlation is zero for the aggregate of grants.

To sum up, the CAPB is defined as follows:

In Expression (6), pbt* denotes the CAPB and grt* the ratio of grants to trend output.

Estimation Uncertainty

Estimations of CAPBs and associated measures of fiscal impulse are subject to uncertainty, as their calculation implies the estimation of unobserved variables, including trend output and long-term commodity prices, as well as that of unobserved parameters, including relevant income elasticities.

Such uncertainty is compounded by a number of methodological assumptions regarding the relation between the observed and long-term values of certain price ratios. In particular, Expressions (1), (2), (5), and (6) assume that the observed and longterm price levels are identical; moreover, Expressions (1) and (2) further assume that the long-term ratio of relevant price deflators to the output deflator are identical to those observed. Finally, Expressions (5) and (6) assume that the ratio of the observed exchange rate to the price level is identical to that in the long term.

The approach in this chapter is to accept the methodological assumptions for the sake of simplicity. Regarding estimation uncertainty, the approach is to attempt to measure it. In this vein, different estimates of trend output (and associated output gaps) were obtained by setting the smoothing parameter of a Hodrick-Prescott (HP) filter at different values (6, 10, 100, and 250). To avoid the end-of-period bias, the filter was applied to series that include forecasts through 2010.

In turn, alternative estimates for long-term commodity prices were estimated by applying a onesided HP filter, with different values for the smoothing parameter (6, 10, 100, and 250), as well as by calculating moving averages of different lengths (3, 5, and 10 years) of observed commodity prices. A one-sided HP (instead of two-sided) was chosen to avoid biasing the long-term commodity price estimations with too much hindsight.

Regarding income elasticities of fiscal revenues and primary expenditures, this chapter uses the estimations in IMF (2007) and working assumptions in Horton, Kumar, and Mauro (2009). Accordingly, the income elasticity of tax revenues was set at a value equal to 1.00 for all countries, except for Brazil (1.01), Chile (1.05), Colombia (1.10), Costa Rica (1.11), El Salvador (1.36), Nicaragua (1.20), Panama (0.90), and Peru (1.09). Income elasticities of primary expenditures were assumed to be zero in all cases.

This approach results in 4 alternative estimates of fiscal impulse and 28 alternative estimates of CAPB for each country. The country-specific results reported in the chapter correspond to the simple averages of such estimations. Upper and lower bounds correspond to the maximum and minimum estimation values for the concept reported.

From a policy perspective, the existence of upper- and lower-bound CAPBs creates some challenges, as it is not possible to establish with certainty the “true” size (and at times not even the sign) of the CAPB. This is particularly important for countries that target public debt ratios or need to reduce their debt levels. Such countries would need to have larger CAPBs to achieve a target debt ratio with a given probability, if upper- and lower-bound CAPB estimations differ too much (Di Bella, 2008).

There are also policy challenges regarding the appropriate size of the fiscal impulse. These arise, for instance, from the absence of synchronicity between commodity price cycles and output gaps. In this connection, a country with significant commodity-related fiscal revenues may benefit from an increase in the long-term price of a relevant commodity, at a moment in which the economy is overheating. In such a case, the government could increase its long-term primary expenditures without compromising debt sustainability; however, such an increase would overheat the economy further. Analogously, the long-term price of a given commodity may decrease at a time of negative output gaps: while this will require a reduction in long-term primary expenditures to keep public debt ratios constant, an immediate decrease in expenditures may worsen the recession.

Debt Sustainability

The assessment of fiscal adjustment needed to keep debt ratios constant presented in this chapter makes use of the following expression:

where pb¯ is the primary balance (as a percentage of output) that stabilizes the debt-to-output ratio at a given target d¯ (see, for instance, IMF, 2005). In addition, y^=γ+π+γ.π denotes the output growth rate (in nominal terms), γ is the output growth rate (in real terms), and π is the inflation rate; moreover, r^=α+rf+(1α)rd is the average interest rate paid on public debt, α is the proportion of foreign currency–denominated debt in total debt, rf is the interest rate on foreign currency–denominated debt, rd is the interest rate on domestic currency–denominated debt, and ε is the change in the real exchange rate. The analysis in this chapter sets d¯ at its end-2008 value and compares pb¯ with the expected headline primary balance for 2009, pb09, as well as with the estimated CAPB at end-2009, pb09*. Thus, the expression AT=pb¯pb09 will denote the total adjustment in the primary balance needed to keep debt ratios constant at their end-2008 values. In turn, the expression A*=pb¯pb09* will denote the adjustment in the CAPB required to keep the debt ratio constant at its end-2008 value. Then, AC = ATA* can be defined as the “automatic” fiscal adjustment that would occur as output gaps close and commodity prices return to their longterm values.

Note: This chapter was prepared by Gabriel Di Bella.

Although an “acyclical” response implies no discretionary change in fiscal policy, letting automatic stabilizers work can substantially reduce domestic demand fluctuations.

There are limits to the effectiveness of monetary policy, including in cases of financial stress that impair its transmission, or when policy interest rates are close to the “zero bound.”

The procyclicality of fiscal policy in Latin America has been explained by a number of causes (Aizenman, Gavin, and Hausmann, 2000; and Talvi and Végh, 2005, among others).

See Corbacho and Schwartz (2007) and Dabán and others (2003). Countries in the region with FRLs include Argentina, Brazil, Chile, Colombia (only applicable for regional governments), Ecuador, Panama, Peru, and Venezuela; ECCU countries have committed to achieving a debt target of 60 percent of GDP by 2020.

CEFI includes Brazil, Chile, Colombia, Mexico, and Peru; OCE includes Argentina, Bolivia, Ecuador, Paraguay, Suriname, Trinidad and Tobago, and Venezuela; CITI includes Antigua, Bahamas, Barbados, Belize, Dominica, Grenada, Jamaica, St. Kitts and Nevis, St. Lucia, and St. Vincent and the Grenadines; and OCI includes Costa Rica, the Dominican Republic, El Salvador, Guatemala, Guyana, Haiti, Honduras, Nicaragua, Panama, and Uruguay. Figures reported for different groups refer to simple country averages.

Figures for 2009–10 reflect the latest information available on outcomes thus far in 2009, policy announcements (including of “stimulus packages,” Box 4.1), and IMF staff forecasts.

This greater sensitivity of revenue to the cycle may in part be explained by relatively large decreases in imports, following drops in tourism, remittances, and foreign direct investment inflows. This is the case with many countries in Central America and the Caribbean, as the collection of taxes on consumption mainly relies on the withholding of taxes at customs.

Commodity revenues and foreign grants, where sizable, are excluded because such revenues are, to a large extent, received from nonresidents and thus do not directly affect domestic demand. Countries with significant commodity-related fiscal revenues are defined as those in which such revenues exceed 2 percent of GDP; these include Argentina, Bolivia, Chile, Colombia, Ecuador, Guyana, Mexico, Paraguay, Peru, Suriname, Trinidad and Tobago, and Venezuela. The same cutoff is applied for foreign grants, requiring adjustments for Bolivia, Dominica, Grenada, Guyana, Haiti, Honduras, Nicaragua, and Suriname.

Reported cyclically adjusted concepts refer to simple averages of alternative estimates (Box 4.2 and Technical Appendix).

FIs were calculated on the basis of above-the-line figures, that is, without considering below-the-line operations as for instance, those to inject fiscal resources into state-owned banks with the purpose of supporting domestic credit, as in Brazil, with BNDES; and Chile, with Banco de Estado (see Box 4.1).

Ilzetzki and Végh (2008) estimate three-year cumulative multipliers for developing countries at about 0.5; IMF (2008) estimates the three-year cumulative multipliers for taxes at 0.2 and for spending, at −0.2. Estimates also vary widely for advanced economies. For a survey, see Spilimbergo, Symansky, and Schindler (2008).

Owing to data availability, the analysis uses gross debt ratios. The fiscal adjustment needed to keep net debt ratios constant would be larger. This is particularly important for countries with large precrisis fiscal financial buffers.

Notwithstanding the large tax revenue falls observed in a few countries.

Expressions (1) and (2) result from taking a first-order Taylor approximation of the underlying expressions, around an output gap equal to zero.

    Other Resources Citing This Publication