Chapter

2. How Countercyclical and Pro-Poor Has Fiscal Policy Been during the Downturn?

Author(s):
International Monetary Fund. African Dept.
Published Date:
April 2010
Share
  • ShareShare
Show Summary Details

Introduction and Summary

Since the economic downturn began, policymakers throughout sub-Saharan Africa, as elsewhere, have to varying degrees sought to use fiscal policy to counter the slowdown. Using information from a survey of IMF country desk officers,1 this chapter addresses two policy questions:2

  • How countercyclical and pro-poor has the fiscal policy response to the crisis been?

  • What explains differences between policy intent and outcomes?

This study’s main findings are that planned and implemented fiscal policies in most sub-Saharan countries have indeed been countercyclical, and that social spending has been protected. Specifically, on the basis of preliminary budgetary outturn data for 2009, the study’s findings include:

  • In formulating their 2009 budgets, about half of the countries expected economic growth to fall below average growth rates posted during 2003–07. To counter the slowdown in growth, a large number of them planned to respond countercyclically, mostly through spending increases. Where spending plans did not increase, it was mainly due to concerns about macroeconomic stability and financing constraints (including aid disbursements).

  • Preliminary data for 2009 indicate that fiscal policy has indeed been mostly countercyclical. This is in stark contrast to the past when fiscal policy in sub-Saharan African countries was overwhelmingly procyclical. In 2009, fiscal deficits increased in two-thirds of the countries in the region experiencing a slowdown in growth largely because of discretionary spending increases beyond medium-term trends. To a large extent, this reflects the stronger fiscal positions in most countries heading into the crisis, and the availability of additional external financing. In a number of countries, still prevailing macroeconomic imbalances made it difficult to implement countercyclical policies even as the anticipated slowdown materialized.

  • The trend of rising health and education expenditures established before the crisis in all sub-Saharan Africa country groups does not seem to have been interrupted, with real growth rates in outlays remaining robust. Capital expenditures generally seem to have held up, although there were significant disparities between countries. A growing number of countries have put in place cash transfers, which have good targeting mechanisms and typically offer high impact at low cost. And an increasing number of countries are taking a more developmental approach to social protection, focusing on public works, and food security, especially through agricultural input subsidies.

  • Countercyclical responses may have been undermined by errors in growth and other macroeconomic budget assumptions and implementation constraints. These problems are a symptom of the more general difficulty of implementing fiscal policy that is particularly challenging in sub-Saharan Africa. There is also some empirical evidence that overly ambitious fiscal plans, inaccurate growth forecasts, and inadequate budget institutions tend to exacerbate implementation errors throughout the region.

The main policy messages are fourfold:

  • Where growth is expected to rebound to precrisis levels, spending plans need to be cast consistent with medium-term fiscal objectives, unwinding any short-term stimulus that might have been provided. For many countries, this implies starting to withdraw any stimulus that has been put in place beginning with 2011 budgets. For others, such as those of the West African Economic and Monetary Union (WAEMU)—which seem to have escaped the crisis with a relatively small impact on growth—tightening fiscal policy as early as 2010 would be appropriate to reduce debt vulnerabilities that have increased during the crisis. Among the fragile states, which have more pressing social and infrastructure needs, further increases in expenditure—if financed by increased revenue mobilization and concessional aid—would be appropriate.

  • With the resumption in growth, countries should first work to increase their revenue-to-GDP ratios, which are still quite low in many countries. Second, reforms to improve the way public finances are managed and to increase the quality and amounts spent on human and physical capital should be accelerated. Finally, countries should avoid reverting to unsustainable financing sources—such as accumulation of payment arrears.

  • Solid budget institutions and forecasting ability are important to improving fiscal responses to the economic cycle in sub-Saharan Africa. Such strong budget institutions ensure adequate financing in bad times and help contain overspending in good times. They help to understand the scale and scope of fiscal challenges to design the appropriate response. Enhancing forecasting capacity reduces the risk of policy formulation errors.

  • While measures taken to protect capital and social spending are a step in the right direction, more needs to be done. Sustained increases in both the quantity and quality of human capital and infrastructure spending are needed to raise long-term growth. Social safety nets can also play an important role, and should be reinforced through systematic implementation and expansion of targeted programs.

Has Fiscal Policy Been Countercyclical?

The last Regional Economic Outlook (IMF, 2009b) provided initial evidence that most countries in the region were planning to use fiscal policy as a stabilization tool. This represents a break from the past—previously, fiscal policy in the region tended to be strongly procyclical. For instance, Thornton (2008) found that real government consumption in African countries was overwhelmingly procyclical in 32 of 37 countries for the period 1960 to 2004. More recent studies suggest that this tendency has declined over the years, partly owing to increased availability of financing and better fiscal discipline as measured by reductions in public external debt (Lledó, Yackovlev, and Gadenne, 2009). Building on this body of work, this section:

  • Reviews countries’ policy intentions as reflected in their 2009 budgets;

  • Assesses actual policy implementation in terms of 2009 budget outturns and estimates; and

  • Explores factors that may explain the differences between planned and actual fiscal responses.

Countercyclical fiscal policy (expansionary when growth is below trend and contractionary in good times) is generally desirable because it helps to smooth output volatility. The extent to which policies are countercyclical is typically measured by correlations between cyclically adjusted measures of government activity and the output gap.3 However, given the lack of reliable estimates for cyclically adjusted fiscal positions and for potential output for sub-Saharan Africa (see IMF, 2009b), this method cannot be applied reliably. In line with previous studies (IMF, 2009b; Kaminsky, Reinhart, and Végh, 2004), this section therefore uses a simplified approach to assess fiscal stances and measure cyclical patterns (Box 2.1). The same approach is used to assess both policy intentions and implemented policies. Briefly, we characterize fiscal policy as countercyclical if a country was able to increase real primary spending in the face of an economic downturn (defined as growth in 2009 falling below its 2003–07 trend). As a robustness check, we also look at alternative measures, including the evolution of real non-oil primary and overall balances.

Were Policy Intentions Countercyclical?

When most sub-Saharan African countries formulated their 2009 budgets,4 about half of them anticipated that economic growth would fall below average rates posted from 2003–07. The projected negative growth gap for sub-Saharan Africa as a whole amounted to about 2½ percentage points (Figure 2.1),5 but on average it was more pronounced and more prevalent among middle-income countries. For instance, South Africa’s 2009 budget projected real output to decline by 1½ percent, which represents a growth slowdown of 6½ percentage points relative to the 2003–07 growth benchmark. In contrast, growth gaps were positive in more than three-fifths of the 11 fragile countries in our sample. The Democratic Republic of Congo, for instance, assumed growth at more than 10 percent in its 2009 budget, an increase of more than 4 percentage points relative to the medium-term growth benchmark.

Among countries anticipating a slowdown in growth, about three-fourths planned to respond countercyclically (Figure 2.2). Real primary spending was planned to increase in 15 out of the 21 countries for which 2009 budgets assumed below trend output growth rates. It accelerated beyond recent (2003–07) medium-term trend spending growth in 12 of these countries. Half of the oil exporters and most middle-income and low-income countries planned to respond countercyclically.

Figure 2.1.Sub-Saharan Africa:1 Projected Growth Gaps2

Source: IMF, African Department database.

1 Excludes Eritrea, Guinea, and Zimbabwe.

2 Growth gap is defined as the difference between 2009 projected real GDP growth and 2003–07 average real GDP growth.

Three out of the four fragile countries also intended to be countercyclical. Real changes in non-oil primary balances and overall balances provided broadly similar results.6 The size of these countercyclical spending plans was typically quite large. Plans for spending as a percent of GDP were on average about 5 percentage points above 2003– 07 averages and accounted for the observed increase in deficits relative to the same period in all country groupings (Table 2.1). The increase in spending was remarkable even relative to medium-term spending plans in all groups except oil exporters. In real terms, primary spending plans grew by 5 percentage points above the 2003–07 averages (Table 2.2). Planned increases were particularly dramatic among some fragile countries such as Togo, where primary spending was set to increase by more than 30 percentage points above recent (2003–07) medium-term trend increases. Few countries planned to introduce discretionary cuts in revenues.7

Figure 2.2.Sub-Saharan Africa:1 Fiscal Cyclicality, 2009 Budget Plans

Source: IMF, African Department database.

1 Excludes Eritrea, Guinea, and Zimbabwe.

Table 2.1.Sub-Saharan Africa: 2009 Budget Plans vs. 2003–07 Average
Overall BalanceTotal SpendingTotal Revenue and Grants
(Difference in percent of GDP)
Sub-Saharan Africa-5.75.1-0.6
Oil Exporters-7.45.4-2.1
Middle-income Countries-7.87.90.2
Low-income Countries-2.32.2-0.1
Fragile States1.72.44.1
Source: IMF, African Department database.

Box 2.1.Characterizing Fiscal Policy Responses—A Simplified Approach

Fiscal policy is considered countercyclical if expansionary in “bad times” and contractionary in “good times”; for procyclical fiscal policies, the relationship is opposite:

Measuring “good” and “bad” times. Bad times are characterized as periods where real GDP growth is below trend (negative growth gap) and good times as periods where growth is above trend (positive growth gap). An alternative and more traditional approach would be to measure good and bad times by looking at deviations between output levels from their long-run trends (output gaps) by using the Hodrick–Prescott filter. Bad times or recessions would be defined as periods where output gaps are negative and good times or booms as periods where output gaps are positive. However, owing to a shortage of high-frequency data and the presence of structural breaks in most output series, this method cannot be applied reliably in the region. Our approach here has the appeal that it is nonparametric and free from these estimation problems.

Estimating trend growth. Most African countries experienced sustained growth accelerations between 1995 and 2007 (IMF 2008b; Arbache, Go, and Page, 2008) with growth arguably converging to its medium-term potential. This convergence has likely been interrupted since 2008 following the food and fuel shocks and the recent 2009 global economic slowdown. Therefore, growth has likely accelerated above trend among oil exporters (and decelerated below trend among oil importers) in 2008 given historically high oil prices. Growth also likely decelerated below trend in most African countries as a result of the global economic slowdown. Taking that into account, we estimate the trend growth rate for each country using average growth rates during 2003–07 as a proxy for latest growth acceleration stage in the post-1995 growth takeoff.

Assessing fiscal policy. We estimate fiscal expansions and contractions mainly by looking at how much real primary spending grew on an annual basis. Real primary spending excludes interest payments (largely outside the control of the policymaker). It is the best available fiscal policy measure for our sample given the lack of systematic data on tax rates (see Iltzetki and Végh, 2008). A positive value would indicate a fiscal expansion and a negative value a fiscal contraction. Table 1 summarizes the resulting cyclical fiscal patterns.

Table 1.Fiscal Policy: Cyclical Patterns
Real Primary Spending Growth
Positive

(Fiscal Expansion)
Negative

(Fiscal Contraction)
NegativeCountercyclicalProcyclical
Growth Gap
PositiveProcyclicalCountercyclical

The following methodological points are also worth noting:

  • Real primary spending growth is computed as real percent changes in primary spending in the 2009 budget relative to 2008 budget outturns. Budget outturns reflect the latest estimates available and are reported on a fiscal year basis, and are thus comparable to corresponding budget plan numbers. In principle, we should have used 2008 budget outturn estimates available when 2009 budget plans were formulated, but such data were not available for most countries.

  • We estimated real changes using real GDP deflators in all countries except for oil exporters. We used consumer price inflation indexes to deflate oil exporters’ fiscal numbers given that the recent decline in oil prices significantly lowered the GDP deflator, tending to overestimate expenditure growth.

  • We also looked at overall and non-oil primary balances to help identify cases where fiscal expansions were planned and implemented by letting automatic stabilizers work or by accommodating declines in commodity-related revenues. As with primary spending, we also measured fiscal balances as real percent changes relative to 2008 budget outturns rather than as percent of GDP to filter out any impact caused by output movements related to the cycle (see Kaminsky, Reinhart, and Végh, 2004).

  • To help isolate cyclical from medium-term structural patterns in fiscal policy, we also looked at how much real primary spending grew beyond spending growth rates assumed in medium-term spending plans. We used the 2003–07 average annual real primary spending growth to proxy for primary spending growth rates assumed in medium-term spending plans.

Caveats. Our findings should be taken as indicative and interpreted with caution because:

  • Our approach may provide different results relative to the traditional output gap approach in some cases. For example, a fiscal expansion right after the trough of a recession when output growth is high but the output gap is still negative would be classified as countercyclical under the traditional approach and procyclical under our approach if the growth gap is positive. Similarly, a fiscal contraction at the end of the boom when growth is low or negative but the output gap remains positive would be classified as countercyclical under the traditional approach but procyclical under ours if the growth gap is negative.

  • Our approach may also underestimate trend growth for countries where growth accelerations have been more recent. This is particularly the case among fragile countries where the 2003–07 average growth rate is likely to underestimate their medium-term trend growth. As a result, positive growth gaps are likely to be overestimated and negative growth gaps underestimated. Our focus on cases of negative growth gaps minimizes this problem by ensuring that negative growth gaps among fragile countries and other instances of late growth takeoff are indeed associated with economic downturns in those countries. In any case, de-trending techniques (parametric or not) should always be used with caution, especially in developing countries (see Aguiar and Gopinath, 2007).

  • Our approach does not estimate the impact of automatic stabilizers and their contribution to countercyclical responses. Attempts to do so are questionable given the lack of reliable estimates of cyclically adjusted fiscal positions and potential output for sub-Saharan Africa. Automatic stabilizers in sub-Saharan African countries have been shown to be small, given the low revenue-to-GDP ratios and general lack of spending programs sensitive to the economic cycle (see Berg and others, 2009). We, therefore, expect our approach to capture the bulk of countercyclical responses in the region.

This box was prepared by Victor Lledó.
Table 2.2.Sub-Saharan Africa: Median Real Primary Spending Growth, Budget Plans, 2008–09 vs. 2003–07 Average
2003–07

Average
Budget PlansCompared with 2003–07 Average
2008200920082009
(Annual percent change)(Percent)
Sub-Saharan Africa8.39.712.21.43.9
Oil Exporters16.49.011.2-7.4-5.2
Middle-income Countries3.40.610.8-2.87.4
Low-income Countries9.219.212.110.02.8
Fragile States4.7-5.830.5-10.425.9
Source: IMF, African Department database.

A small number of countries expected growth to slow but nevertheless intended to tighten their fiscal policies due to financing and macroeconomic constraints. According to our survey of IMF country teams, where spending plans did not increase, regardless of whether growth was expected to slow or not, it was due mainly to concerns about macroeconomic stability and financing constraints, including aid disbursements (Figure 2.3). This was especially true of fragile states, such as Comoros, which have not reached the completion point for heavily indebted poor countries (HIPC). But it is also true of some middle-income countries, such as Seychelles, which struggled with macroeconomic imbalances even before the crisis.

Figure 2.3.Sub-Saharan Africa: Budget Spending Plans, 2009

Source: IMF, African Department Survey.

Both countries intended to tighten their fiscal policies despite projecting a growth slowdown. The presence of IMF-supported programs typically did not negatively impact spending, with spending plans maintained or increased in about two-thirds of program countries. The occurrence or imminence of elections increased the probability that spending would be increased, providing some indication of an electoral cycle effect.

The remaining sub-Saharan African countries formulated their budgets on the assumption that their economies would continue to grow at a rapid clip. Three-fourths of countries nevertheless planned to ramp up spending leading to procyclical policy intentions. Most of them were fragile states, which in part may be the result of post-conflict reconstruction efforts (Figure 2.4).8 However, this group also included some middle-income countries, such as Botswana, Lesotho, and Mauritius, where budgets provisioned for very large increases relative to medium-term spending growth plans.

Figure 2.4.Sub-Saharan Africa:1 Median Real Primary Spending Growth, Budget Plans, 2008–09

Source: IMF, African Department database.

1 Excludes Eritrea, Guinea, and Zimbabwe.

2 Growth gap is defined as the difference between 2009 projected real GDP growth and 2003–07 average real GDP growth.

Did Implemented Budgets End up Countercyclical?

Fiscal policy turned out to be countercyclical in two-thirds of the countries experiencing a slowdown. Preliminary GDP data suggest that growth in 2009 was below trend in more than three-fourths of countries in sub-Saharan Africa. Among countries experiencing negative growth gaps, fiscal policy responded countercyclically in two-thirds of countries in the sense that they had positive increases in real primary spending (Figure 2.5).9 This contrasts sharply with how countries in the region reacted following the last major worldwide recession in 1991 when almost three-fifths of countries with negative gaps implemented procyclical fiscal policies.

  • Countercyclical fiscal policies were the predominant response to the slowdown in most country groupings, including fragile states. Oil exporters were the exception, with less than one-third implementing countercyclical policies, likely the result of unanticipated revenue shortfalls, as discussed below.

  • By and large, countries that aimed to respond countercyclically have been able to do so (Figure 2.6). However, in some cases, fiscal outcomes deviated from plans. On the other hand, fiscal policy remained countercyclical in 14 out of 15 countries originally planning to do so, including all low-income countries and fragile states.10 On the other hand, countercyclical plans turned procyclical at the implementation stage in Namibia.

  • In Botswana, Democratic Republic of Congo, Gambia, Lesotho, Mali, and Mauritius, fiscal policies ended up being countercyclical because large spending increases were implemented as planned but growth turned out worse than anticipated.

  • Macroeconomic constraints continued to prevent countries such as Comoros, Ghana, Malawi, and Seychelles from implementing countercyclical policies even as the anticipated slowdown materialized.

  • In line with fiscal plans, countercyclical measures implemented mainly took the form of discretionary spending increases, though the size of countercyclical responses was not as large as originally planned.

  • Increases in fiscal deficits remained driven by spending increases in all categories except for oil exporters and some low-income countries where revenue declines have prevailed (Table 2.3). As discussed in detail below, this result seems to be a combination of weaknesses in budget execution and unrealistic output growth assumptions and revenue targets.

Figure 2.5.Sub-Saharan Africa:1 Estimated Growth Gaps2 and Fiscal Cyclicality, 2009 Budget Outturns

Source: IMF, African Department database.

1 Excludes Eritrea, Guinea, and Zimbabwe.

2 Growth gap is defined as the difference between 2009 estimated real GDP growth and 2003–07 average real GDP growth.

Figure 2.6.Sub-Saharan Africa:1 Deviations in Fiscal Cyclicality, 2009 Budget Plans vs. Outturns

Source: IMF, African Department database.

1 Excludes Eritrea, Guinea, and Zimbabwe.

Table 2.3.Sub-Saharan Africa: 2009 Budget Outturns vs. 2003–07 Average
Overall BalanceTotal SpendingTotal Revenue and Grants
(Difference in percent of GDP)
Sub-Saharan Africa-5.83.7-2.0
Oil Exporters-8.33.4-4.9
Middle-income Countries-7.57.0-0.4
Low-income Countries-2.20.6-1.5
Fragile States2.53.66.1
Source: IMF, African Department database.

What Undermines Countercyclical Responses?

The ability of policymakers to deliver a cyclically appropriate fiscal policy response may be undermined by unrealistic budget assumptions about growth and other macroeconomic variables, and implementation constraints. They are a symptom of the more general problem of implementing fiscal policy that is particularly challenging in sub-Saharan Africa (Box 2.2). That said, making macroeconomic projections in an unprecedented crisis such as this one is very demanding in any region.

  • Forecast errors in economic growth may undermine countercyclical responses independently of how well fiscal plans are implemented. Failure to correctly identify changes in the economic cycle at the budget planning stage may turn appropriate countercyclical policy intentions procyclical even if fiscal outturns are as planned (for example, a fiscal contraction in good times may cease to be countercyclical if it proceeds to be implemented as planned and simultaneously “times turn bad”). Forecast errors may also delay the implementation of appropriate responses. Macroeconomic assumptions on growth and other variables, such as inflation, the exchange rate, and external financing, may indirectly turn a countercyclical plan procyclical by constraining spending because the projected funding is insufficient.11 In sub-Saharan Africa, forecast accuracy is particularly compromised by the lack of good quality real-time data, larger and more frequent macroeconomic shocks, and weaker forecasting capacity than in other regions. Unrealistic fiscal targets reflecting strategic or political considerations exacerbate this problem.

  • Differences between fiscal plans and outturns may also constrain countercyclical responses independent of forecast errors, reflecting implementation constraints. This is the case when planned fiscal expansions in bad times are underexecuted or reallocated due to unanticipated weaknesses in project execution or revenue collection capacity.12 This may also occur if budget institutions are not capable of shielding fiscal adjustments approved in good times from political pressures to overspend or undertax.

  • Even with accurate fiscal forecasts and proper budget execution, countercyclical fiscal plans may still fail to be implemented if the government is not capable of meeting unanticipated financing shortfalls.

In 2009, most countries could neither fully anticipate the economic slowdown nor adjust accordingly once such forecast errors materialized (Figure 2.7). Preliminary estimates suggest that three-fourths of the countries experiencing a growth slowdown in 2009 underestimated it and about half of them did not relax their fiscal stance once forecast errors were revealed. Failure to correctly identify changes in the economic cycle at the budget planning stage was particularly noticeable among oil exporters such as Angola and low-income countries such as Madagascar. In these countries, fiscal stances remained tight even as real GDP growth was more than 12 percentage points lower than originally anticipated. In most cases, however, fiscal policies remained or turned countercyclical at the implementation stage despite forecast errors.

Figure 2.7.Sub-Saharan Africa: Real GDP Growth Gap1 Forecast Errors,2 and Fiscal Responses,3 2009

Source: IMF, African Department database.

1 Growth gap is defined as the difference between 2009 estimated real GDP growth and 2003–07 average real GDP growth.

2 Forecast error is defined as the difference between projected and actual real GDP growth gaps in 2009.

3 Looser (tighter) fiscal stance is defined as actual real primary spending growth above (below) plan.

Revenue and spending shortfalls relative to plans were large and above precrisis levels in most cases. A large majority of countries reported revenue and expenditure outturns below their 2009 fiscal targets. On the other hand, revenue shortfalls were particularly pervasive among oil exporters, reflecting unanticipated declines in oil revenues. Though large, such shortfalls were broadly in line with 2004–08 precrisis levels (Figure 2.8). On the other hand, revenue shortfalls among low- and middle-income countries contrast with the windfalls generally observed before the crisis in these countries. Spending shortfalls relative to precrisis levels were somewhat larger in the case of oil exporters, middle-and low-income countries. In line with precrisis levels, they were negative (that is, actual spending above plan) in the case of fragile states. Generally speaking, spending plans were typically ambitious enough to withstand the additional unforeseen output and revenue losses.

Figure 2.8.Sub-Saharan Africa: Budget Plans and Outturn Differences, 2008–09

Source: IMF, African Department database.

According to the survey, forecast errors and implementation constraints were the main factors accounting for deviations between fiscal plans and outturns (Figure 2.9). On the spending side, inaccurate forecasts of macrobudgetary assumptions, together with unrealistic expenditure targets, accounted for roughly half of country responses. Implementation constraints related to project execution capacity and political pressures accounted for the other half. Inaccurate and unrealistic forecasts and fiscal targets were more common than implementation constraints among fragile states. Project execution was the most common problem for the middle-income group, as was political pressure among oil exporters.13 On the revenue side, forecast errors were more detrimental to proper implementation than constraints on implementation.

Box 2.2Fiscal Policy Implementation in Africa

Fiscal policy implementation in any country, regardless of its level of development, is subject to a number of constraints. They arise from difficulties in forecasting downturns and recoveries in real time, strategic considerations leading to overambitious fiscal targets (for example, overoptimistic predictions of economic growth and tax revenues to ensure compliance with ex ante fiscal rules), lengthy budget procedures, and political pressures to overspend or undertax.

In sub-Saharan Africa, the problem is particularly challenging. Additional constraints there include poor data quality, weaknesses in forecasting capacity, large and frequent macroeconomic shocks, inadequate budget institutions, slow project execution, and less stable political systems. Such factors have often been identified as reasons why fiscal policies in the region have tended to be more procyclical than elsewhere (Balassone and Kumar, 2007; IMF, 2008b).

But how challenging is fiscal policy implementation in the region? Does it vary by country? What are the main constraints on it? To answer these questions, we first compute fiscal policy implementation errors for a large number of countries, some in other regions, to benchmark fiscal policy implementation in sub-Saharan Africa. We then use an econometric model for sub-Saharan African countries to investigate how relevant for the region are some of the most common constraints: the accuracy of key budget parameters (growth and inflation), the quality of budget institutions, the role of elections, and other characteristics of the political environment.1

Fiscal policy implementation errors are defined as differences between planned and implemented changes in fiscal policy outcomes (for example, fiscal balances, spending, and revenues). Both planned and implemented changes are calculated as annual changes in variables measured as a percent of GDP. Following Beetsma, Giuliodori, and Wierts (2009), planned changes in fiscal outcomes are computed using real-time one-year-ahead fiscal projections, that is, the fiscal forecasts available to policymakers when they are preparing budget plans. Implemented changes in fiscal outcomes are measured on the basis of the latest available fiscal data. World Economic Outlook (WEO) fall projections and historical series are used to ensure comparability across countries.2

In sub-Saharan Africa, fiscal policy implementation errors tend to be comparable to but more dispersed than in other regions because intraregional patterns are quite distinct. As in other regions, planned fiscal consolidation in sub-Saharan Africa sometimes end in fiscal expansions. Average implementation errors at the level of the overall fiscal balance have been lower than in other regions (Table 1).3 On the other hand, implementation errors at the level of spending and revenue tended to be larger than in other regions, with large underestimations of both variables on average, due to quite different intraregional implementation patterns. Revenue shortfalls among oil exporters, overspending by middle-income countries, and a combination of both in the low-income subgroup account for errors toward high deficits or lower surpluses. Fragile states instead tend to underestimate planned surpluses as a result of largely unanticipated revenue windfalls.

Preliminary econometric evidence suggests that planned fiscal adjustments or fiscal expansions in the region are less likely to be implemented the larger they are, the more inaccurate the growth forecasts they are based upon, and the weaker the budget institutions regulating their design, approval, and execution:

Table 1.Fiscal Outcomes, Fiscal Plans, and Fiscal Implementation Errors, 2004-08 Average
Overall SurplusTotal SpendingTotal Revenue
ActualPlanErrorActualPlanErrorActualPlanError
(Change in percent of GDP)
Sub-Saharan Africa
Mean-0.130.24-0.361.07-0.491.420.94-0.251.07
Standard Deviation15.820.013.54.62.44.515.319.813.6
Number of Observations217217217217217217217217217
Oil Exporters
Mean-0.701.23-1.93-0.40-0.710.31-1.100.51-1.61
Standard Deviation11.79.411.93.72.84.010.78.510.4
Number of Observations353535353535353535
Middle-income Countries
Mean-1.81-0.72-1.091.32-0.311.62-0.50-1.030.53
Standard Deviation3.61.93.14.71.65.03.31.73.4
Number of Observations404040404040404040
Low-income Countries
Mean-1.45-1.42-0.031.18-0.211.39-0.27-1.631.36
Standard Deviation11.47.19.22.31.62.711.47.28.9
Number of Observations757575757575757575
Fragile States
Mean2.642.210.561.58-0.811.944.211.402.50
Standard Deviation24.034.920.76.53.36.023.134.621.3
Number of Observations676767676767676767
Advanced Economies
Mean-0.960.09-1.060.63-0.240.87-0.33-0.15-0.18
Standard Deviation2.320.752.201.680.831.711.210.811.44
Number of Observations163163163163163163163163163
Other Developing Economies
Mean-0.620.38-1.000.34-0.681.00-0.28-0.300.00
Standard Deviation4.72.54.63.42.53.63.72.73.6
Number of Observations455455455455455455455455451
Source: IMF, World Economic Outlook, and IMF staff estimates.Notes: Fiscal outcomes (f), that is overall surplus, total spending, and total revenue, are defined in percent of GDP. Change in fiscal outcomes (df) defined as differences in f between years t and t+1 and t' according to data available on year f (i.e. df =f(t+1, f)- f(t, f)). Actual changes (dfa) and planned changes (dfb) based on data available at t+1 and t, respectively (i.e., dfa = f(t+1, t+1) - f(t, t+1)); dfp = f(t+1, t) -f(t, t′))). Implementation error is defined as the difference between actual and planned changes in fiscal outcomes (ef = dfa -dfp).
  • Ambitious plans are subject to large implementation errors. Large fiscal adjustment or fiscal expansions may reflect overambitious fiscal targets at the planning stage owing to weaknesses in budget execution, the need to secure political support for approval, or unrealistic fiscal targets set for political reasons. The magnitude is quite large and in some cases amounts to an additional shortfall of 0.7 percent of GDP for an extra planned adjustment of 1 percent of GDP.

  • Too optimistic or pessimistic real GDP growth projections tend to lead to a significant shortfall in implementation relative to planned fiscal policy. This is probably because revenues are lower than projected; errors in forecasting inflation were not significant.

  • Political competition—not necessarily through elections—seems to matter. Checks on the executive either through formal rules or more political competition were systematically associated with smaller errors, which do not seem to increase or fall significantly during election years.

  • Solid budget procedures help reduce implementation errors, particularly when checks on the executive are sufficient. However, not all budget procedures matter equally. A more transparent budget seems to be more effective in reducing implementation errors than a more top-down one.

Figure 1.Fiscal Policy Implementation in Sub-Saharan Africa: Econometric Evidence

Sources: IMF, World Economic Outlook; IMF staff estimates; Dabla-Norris and others (2010); and Polity IV database.

1 Difference between implemmented and planned changes in the overall balance.

2 Difference between actual and projected real GDP growth rates.

3 Budget procedure index ranges from 0 to 4, with a higher score reflecting better performance.

4 Executive constraints range from 1 (unlimited authority) to 7 (executive parity or subordination).

Note: This box was prepared by Victor Lledó and Marcos Poplawski-Ribeiro.1 The strength of budget institutions is assessed using multidimensional indices of the quality of budget institutions presented by Dabla-Norris and others (2010a). The indices record the quality of budget institutions at various stages (planning, approval, and implementation) and with different characteristics of the budget process (for example, centralization, effective rules and control, sustainability, and transparency).2 Fall WEO projections for a given year are used to better approximate policymakers’ fiscal plans for the following year, which for most countries are prepared on the basis of information available on the last quarter of the preceding year.3 Spending errors are larger in absolute terms than revenue errors.

Figure 2.9.Sub-Saharan Africa: Revenue and Spending Deviations, 2009 Budget Plans vs. Outturns

Source: IMF, African Department survey.

This was particularly true for oil exporters, probably because it is so hard to forecast oil prices, especially in periods of high volatility. However, for fragile states inadequate revenue collection capacity was more pervasive.

Financing constraints do not seem to have been a major obstacle to implementing fiscal policy in most countries. Increased official external financing was facilitated by providing additional fiscal space.14 During 2009, countries that had higher official external financing and better fiscal positions were more likely to be in a good position to loosen their fiscal stance (Figure 2.10). To ease the financial burden for countries, the IMF, for instance, nearly doubled its disbursements to US$2.7 billion in program assistance (some of which was used for budget support) in 2009 compared with US$1.4 billion in 2008. At the same time, to boost the external reserve position, the IMF made a Special Drawing Rights (SDR) allocation of about US$12 billion to sub-Saharan Africa in 2009.

Figure 2.10.Sub-Saharan Africa: Foreign Financing and Fiscal Stance1

Source: IMF, African Department database

1 Change in foreign financing in 2009 compared with the average for 2003–07 (in percent of GDP) vs. whether a country loosened its fiscal stance (overall balance) in 2009 compared with the average for 2003–07.

Even countries that did not have additional official external financing mostly loosened their fiscal stance, mainly through additional domestic borrowing. That said, there are limits to domestic borrowing given the relatively low level of development of domestic financial markets. Moreover, increasing recourse to these markets could raise borrowing costs and lead to crowding out of private investment.

How Did Public Investment and Social Spending Fare during the Crisis?

A major concern in most sub-Saharan African countries previously was the tendency to cut pro-growth and pro-poor spending at times of budgetary pressure. This section looks at the preliminary data on capital spending (as a proxy for pro-growth spending) and health and education spending (as a proxy for pro-poor spending) to see if that concern is still valid. These are of course imperfect proxies, but these numbers are readily available at this stage for most sub-Saharan African countries, although the data are preliminary for 2009. If there is any indication that spending in these areas is being compressed, it would be good to know sooner rather than later. It appears that the higher outlays on public investment and health and education observed before the downturn were sustained in 2009.

To a degree, this is not surprising because aggregate spending continued to grow significantly in real terms. If anything, the emphasis on public investment and health and education spending increased in 2009 (Table 2.4)—across oil exporters, low-income countries, and fragile states. However, in middle-income countries, median health and education spending lost some ground in 2009, although public investment levels were generally stable.

Public Investment since the Crisis

In recent years, most governments in sub-Saharan Africa have been increasing capital spending in percent of GDP (Figure 2.11). This trend was observed in a majority of countries across all groupings, with fragile states appropriately recording the most significant increases. Real growth rates have also been quite high, although there were considerable differences among various country groupings (Table 2.4). In 2009, all country groupings—except for oil exporters who faced sharp declines in revenues—recorded strong real growth.

Figure 2.11.Sub-Saharan Africa: Capital Expenditure, 2003–08

Source: IMF, African Department database.

Preliminary data suggest that despite revenue shortfalls in most countries, capital spending did indeed increase as a ratio to GDP (Figure 2.12). Median capital expenditure rose by about 1½ percentage points to 9.1 percent of GDP in 2009 compared with both 2008 and 2003–07.

Although capital spending by middle-income and fragile states was less than 8 percent of GDP in 2009, for both groups, that is still a sharp increase relative to the median for 2003–07.

Table 2.4.Sub-Saharan Africa: Median Capital and Health and Education Expenditure, 2003–09
Capital ExpenditureHealth and Education
2003-07200820092003-0720082009200720082009200720082009
(Percent of GDP)(Real growth, percent)(Percent of GDP)(Real growth, percent)
Sub-Saharan Africa7.57.49.213.216.011.14.95.56.05.44.86.8
Oil Exporters7.97.99.516.214.9-13.83.13.24.118.514.14.6
Middle-income Countries6.28.37.96.230.314.67.88.88.5-0.113.40.2
Low-income Countries9.99.09.513.20.611.15.56.15.97.51.66.8
Fragile States4.84.67.116.425.739.14.04.45.87.82.517.8
Source: IMF, African Department database.

Figure 2.12.Sub-Saharan Africa: Capital Expenditure, 2008–09 vs. 2003–07

Source: IMF, African Department database.

Despite increases in observed capital spending, capacity issues continue to make it difficult for many African governments to execute their capital budgets as planned (Figure 2.13). Among low-income countries, observed capital spending averaged just 76 percent of appropriations. However, in Benin, Burkina Faso, Malawi, and Tanzania, capital spending in 2009 exceeded the amounts budgeted. In contrast, Uganda and Zambia spent less than planned, although execution of the capital budget was higher than in 2008. In middle-income countries, by contrast, the execution rate was lower in 2009.

Figure 2.13.Sub-Saharan Africa: Planned vs. Observed Capital Expenditure, 2008 and 2009

Source: IMF, African Department database.

Health and Education Spending during the Downturn

In most sub-Saharan African countries, spending on health and education was trending up until the downturn. Outlays averaged about 5½ percent of GDP in 2006–07 and rose to about 7 percent in 2008—accounting for almost one-third of all primary spending in the region. Middle-income countries spent the most on health and education in 2006–07 at about 8 percent of GDP, and oil exporters spent the least at less than 3 percent of GDP.

And it appears that outlays on health and education were preserved in most countries in 2009 (Figure 2.14). According to preliminary outturn data, median spending on health and education was above 2006–07 levels for all country groups, especially for fragile states, which sharply increased their outlays (Table 2.4). As a result, health and education spending in fragile states as a percentage of GDP is now in line with low-income countries. Except for middle-income countries in 2009, real growth rates of health and education spending have been quite robust.

Figure 2.14.Sub-Saharan Africa: Health and Education Spending, 2008–2010 vs. 2006–07

Source: IMF, African Department database.

A growing number of countries in sub-Saharan Africa also have programs designed to provide social protection in bad times and social promotion in good times. Social safety nets are heterogeneous and extremely complex, and unfortunately timely information on them is not available (Box 2.3). A handful of African countries have successfully established poverty-related cash transfer programs, and many others are taking note of the modest cost and relative effectiveness of such programs. The challenge will be to preserve the sense of urgency to strengthen social safety nets in Africa after growth resumes, so that Africans can be better protected when the next shock comes their way.

Based on budgets and outturns, the fiscal policy intention seems clear: ramp up public investment and protect health and education spending. However, keeping health and education spending in line with the trend before the crisis will be challenging.

Conclusions

Early indications then are that the fiscal response to the crisis has been appropriately countercyclical, while protecting social and capital spending. More countries than in the past seem to have had the economic stability and fiscal space to pursue countercyclial fiscal policies. Most of them did this by increasing or sustaining spending despite declining revenues. In general, social and capital spending has been protected during the downturn. Nonetheless, execution problems meant that not all budgeted spending, especially capital spending, was achieved. In some cases, this was also a result of unexpectedly dramatic reductions in GDP growth and, as a result, in tax revenues.

Box 2.3.Social Protection and Promotion Programs in Sub-Saharan Africa

Cash transfer programs have been an important part of the crisis response for some developing countries. They provide a minimum income for vulnerable households and protect their access to the basic consumption basket in the short term. Their relatively effective targeting mechanisms offer high impact at low costs. While a growing number of African countries have some kind of poverty-related cash transfer program in place (Weigand and Grosh, 2008), most of the programs are still in the pilot stage and too small to be macroeconomically significant. Of the handful of poverty transfer programs captured in our survey, all reported planned and actual increases in spending, providing some evidence that established programs were well insulated from the effects of the crisis. Angola and South Africa, which have the two largest programs as a percent of GDP, both increased appropriations in nominal terms for their cash transfers in 2009 and again in 2010 (Table 1).

Table 1.Social Spending, 2006–2010
20062007200820092010
(Percent of GDP)
Cash Transfer Programs
Angola2.32.11.82.31.9
South Africa4.54.64.85.25.1
Poverty-related Transfer
Angola0.80.90.71.00.8
South Africa1.61.91.92.22.3

Social pensions, also known as poverty-related transfers, are increasingly popular among policymakers looking for cost-effective ways to alleviate poverty. In South Africa, poverty-related transfers as budgeted for 2010 represent more than 2 percent of GDP and are expected to account for 7 percent of government expenditures. In countries where data Source: IMF, African Department database. are available, it appears that these programs have escaped largely unscathed from the global crisis.

Countries that lean on a community development approach to social protection are becoming more prevalent. Both Côte d’Ivoire and Kenya significantly increased funding for their public works employment programs in 2009. Funding for Kenya’s youth employment program tripled as a percent of GDP owing to increased coverage. Malawi and Zambia both have extensive fertilizer subsidy programs. The subsidy is expected to cost Malawi 2.4 percent of GDP in 2010 (but down from a peak of 5.6 percent in 2008—induced by high petroleum prices). In Zambia, the fertilizer subsidy has grown to 0.7 percent of GDP and 2.7 percent of total expenditures.

The landscape of social protection and promotion programs in sub-Saharan Africa is heterogeneous and complex. Programs are in a large number of sectors and are funded by domestic and foreign resources and managed by a variety of government agencies. A detailed look at some of the most prominent and closely watched programs provides valuable insight into pro-poor spending in times of crisis:

Kenya—Cash Transfers for Orphans and Vulnerable Children. Preliminary evidence based on monitoring reports and the first round of independent evaluations suggest that the impact is positive despite administrative challenges and disbursement delays (Bryant, 2009). The budget, largely financed by donors, escaped significant cuts during the crisis. Recently, the government launched an electronic cash transfer program for vulnerable urban households, which is expected to cover 40,000 households in 2010.

Nigeria—Conditional Cash Transfer Pilot Program. The program began in 2008 as the food and fuel crisis was ending. It is too new to provide a significant response to the latest global crisis, and it has been plagued with implementation issues, especially related to its payment delivery system (Nwadinobi, 2009).

Sierra Leone—National Commission for Social Action (NaCSA). NaCSA, an organization dedicated to community development, is front and center in the formulation of the social protection strategy. NaCSA focuses on rehabilitating infrastructure and public services using labor obtained through a cash-for-work program. Far from facing budget cuts because of the crisis, the program has been scaled up (Ngebeh, 2009).

South Africa—Cash Transfer Program. South Africa has the continent’s oldest and largest cash transfer program. At an annual cost of about 5 percent of GDP (Table 1), it has had a measurable impact on poverty reduction.

Note: This box was prepared by Irene Yackovlev.

Going forward, a reassessment of fiscal policies in most countries will be in order. Since the growth outlook is generally firming up, a progressive withdrawal of the stimulus will be necessary in order to avoid rapid debt accumulation and ensure that policies remain countercyclical in “good times.”

Obviously, the appropriate pace of withdrawal will depend on country-specific circumstances, notably how far a country is from its medium-term output growth trend, the public debt situation, and overall strength of its economy. In any case, it will be important for governments to identify “good times” when they are happening (not later), and during those years to set expenditure growth somewhat slower than the countries’ medium-term output

growth to ensure that countercyclical fiscal policies are credibly implemented.

The ultimate result of the various strategies has yet to be determined. There is no doubt that depressed growth combined with high food and fuel costs has had a deleterious impact on the poor, endangering hard-won gains in poverty reduction. While cash transfers have a good track record of delivering results, they have been implemented in only a few countries, and even there the coverage is relatively low and in some places there have been considerable administrative hurdles. Ultimately, a recovery to precrisis growth or better is the only sure way of producing a lasting reduction in poverty.

Notes: This chapter was prepared by Montfort Mlachila, Victor Lledó, and Irene Yackovlev, with research assistance from Duval Guimarães and Gustavo Ramírez.

The survey collected both quantitative and qualitative information as of March 2010 on country authorities’ announced and implemented measures to mitigate the impact of the global financial crisis. The quantitative part asked for data on government budgets, fiscal outcomes and projections, and social spending covering the period 2000–10. The qualitative part requested information on, and country teams’ assessment of, authorities’ fiscal responses to the global crisis. There was sufficient information for 41 countries. Guinea, Eritrea, and Zimbabwe are excluded from the sample due to a paucity of reliable data.

This chapter builds on previous staff analysis of the fiscal space available to African countries and its appropriate use (IMF, 2009a,2009b; Berg and others, 2009). Stocktaking has also begun regarding fiscal stimulus for G20 and program countries (IMF, 2009c,2009d).

Measured as the difference between actual and potential growth.

Throughout the chapter, the term “2009 budgets” refers to budgets approved by central governments for their 2009 fiscal years. For countries where the fiscal year differs from the calendar year, the 2009/10 budget was used. Where the budget was revised or a supplementary budget issued, the 2009 budget refers to the revised or supplementary budget or, for countries with IMF-supported programs, the budget that better reflects agreed program numbers.

Unless otherwise indicated, all tables and charts are based on purchasing power parity-weighted averages.

Policy intentions were countercyclical in more than two-thirds of countries, including all middle-income countries, anticipating a growth slowdown on the basis of these two measures.

Tax reductions were limited to a few countries scattered across different groups and often targeted to specific sectors, as in Namibia’s corporate tax reductions for mining companies, Zambia’s elimination of windfall taxes on mining, and Gabon’s reduction of export taxes and VAT on timber.

As discussed in Box 2.1, such results should also be interpreted with caution as they could also reflect appropriate countercyclical responses if trend growth in some countries are underestimated.

Fiscal policy remained countercyclical in about two-thirds of cases when measured by changes to non-oil primary and overall balances, but dropped to about half of the observed negative growth gap cases when only real primary spending increases beyond recent (2003–07) medium-term spending growth are taken into account.

Including Burkina Faso, Mozambique, Rwanda, Togo, Uganda, and Zambia.

This may be partially compensated for by the presence of automatic stabilizers on the revenue side.

One has to be careful, however, to separate appropriate declines in revenue collection reflecting the work of automatic stabilizers from unanticipated weaknesses in revenue collection, which may or may not be related to or exacerbated by the business cycle.

The fact that project execution was the most common problem among middle-income countries may reflect absorptive capacity issues. Some middle-income countries had the fiscal space to increase spending but lacked “shovel-ready” projects or clear plans on how to expand existing projects.

    Other Resources Citing This Publication