December 10, 2008, Aart Kraay and Luis Servén[*]
Developing countries contemplating fiscal policy responses to the current crisis should heed lessons from past experience. Successful episodes of countercyclical discretionary fiscal policy in developing countries have been rare. Too often fiscal responses are too late and too difficult to reverse, with the perverse effect of increasing volatility and contributing to unsustainable debt accumulation. The note concludes with recommendations for the future to avoid these past pitfalls and to ensure that short-term crisis alleviation is done in a way that supports longer-term growth.
The current financial crisis in rich countries has propagated as an adverse external shock to many developing countries. Slower growth in rich countries has reduced demand for developing country exports and lowered remittances, and the seizure of credit markets has led to a sudden stop in external financing for many market-access emerging economies. The prospect of a potentially deep global slump has brought attention to discretionary fiscal policy as one of the tools potentially available to developing countries to mitigate the severity of the growth slowdown.
This note first reviews the state of the empirical evidence on the effectiveness of discretionary countercyclical fiscal policy in developing countries. On the whole, successful fiscal interventions of this type have been rare in the developing world. The note also reviews briefly a few contrasting experiences of success and failure in industrial and developing countries. It concludes with several recommendations motivated by past experience that policymakers should consider before adopting any fiscal responses to the current crisis.
Lessons from the Past
The effects of fiscal policy on short-run economic fluctuations as well as on long-run growth in industrial and developing countries have been extensively studied. This literature points to the following conclusions regarding stabilizing role of countercyclical fiscal policy in developing countries:
There is strong evidence that fiscal policy is procyclical
. Fiscal expansions tend to take place in good times, and not during bad times when they might play some role in smoothing output declines. This applies to a variety of measures of fiscal policy – including total expenditure, the share of total expenditure in GDP, public consumption and public investment (with the latter being the most procyclical of all). While there has been debate over how to interpret this correlation, recent evidence suggests that it mostly reflects a mis-timing of discretionary fiscal interventions.[1
Fiscal procyclicality in developing countries arises from both the weakness of automatic stabilizers and the procyclical bias of discretionary policies.
While in industrial countries countercyclical discretionary policy contributes to dampen aggregate fluctuations, in developing economies discretionary policy is usually procyclical. In addition, in most developing countries automatic fiscal stabilizers – such as income taxes and transfer programs built into the fiscal system – are too small to have a significant smoothing effect on aggregate fluctuations.[2
The procyclical bias in fiscal policy reflects underlying fundamental challenges facing developing countries.
Research suggests that two main sets of factors account for this procyclicality of discretionary fiscal policy: (i) the inability of developing countries to access external finance to pay for fiscal expansions during downturns,[3
] and (ii) political economy problems that contribute to an overspending of public revenues when they are abundant in good times.[4
] Such fundamental factors are difficult to overcome in the short run, suggesting deep underlying limits on the ability of most countries to generate successful fiscal responses to the current crisis.[5
Regardless of whether they occur in booms or recessions, the evidence is mixed that fiscal expansions can actually stimulate aggregate demand and output in developing countries in the short run.
Some recent studies conclude that there are at most small positive immediate effects but overall negative effects in the medium-term, while others find some positive effect in the medium-term as well. Most studies also find that the expansionary effect of fiscal policy tends to be much smaller in developing countries than in developed ones, and so extrapolating from past successful experiences with countercyclical fiscal policy in the latter to policy recommendations for the former is risky.[6
To the extent that fiscal expansions do boost output, the procyclicality of fiscal policy means that discretionary fiscal interventions tend to increase output volatility rather than reduce it
. This finding suggests that fiscal policy has been very unsuccessful as a tool for stabilization. In fact, the additional volatility induced by procyclical fiscal policy has been found to undermine growth in the long run as well.[7
Fiscal expansions are difficult to reverse
. If governments commit to unsustainably large spending programs during recessions as a countercyclical device, these may be very difficult to reverse when times improve, threatening fiscal sustainability in the long run.[8
] This is why automatic stabilizers in the form of lower tax takes and larger transfer payments during recessions are viewed as a more sustainable approach to countercyclical fiscal policy in industrial countries.[9
] But, as already noted, the difficulty for developing countries is that most have very weak tax and transfer schemes and so the effectiveness of automatic stabilizers is limited.
Overall, empirical evidence from the past 30 years suggests that developing countries have not been overly successful in using discretionary fiscal policy to stabilize output fluctuations -- although some successes do exist. Box 1 reviews a few contrasting experiences in industrial and developing countries.
Box 1.Examples of Fiscal Stimulus: The Good, the Bad, and the Ugly
The Great Depression in the U.S. has been cited by some observers as the only crisis in the last century of greater magnitude than the current one. To combat the Depression, in the early 1930s the U.S. government adopted a broad set of policies, including deposit insurance, abandoning the gold standard, monetary expansion, and a discretionary fiscal expansion. Prior to that time, there had been virtually no use of fiscal policy for macroeconomic stabilization: except in time of war, governments simply aimed to balance the federal budget. In cyclically-adjusted terms, the budget went from a surplus of 1 percent of GNP under the Hoover administration, to a deficit of around 2 percent under Roosevelt’s. (DeLong 1998). While the magnitude of the change may seem modest, it has to be compared with the very small size of the budget at the time (around 5% of GNP).
Under the New Deal policies, the expenditure increases consisted mainly in federal grants for work relief associated with public works, and transfers to farmers. Most observers agree that the fiscal expansion helped the recovery substantially, although the extent of its contribution – as opposed to that of other policy changes -- remains debated. Closer analysis reveals that the two major spending items had quite different effects: public works grants probably raised income almost one-for-one, while transfers to farm owners – intended to compensate them for taking land out of production – had little or no effect on spending and income, owing partly to their regressive redistributive impact on rural incomes (Fishback et al 2005).
China's response to the East Asian financial crisis of 1997-1998 is viewed as an unusual example of a successful countercyclical intervention by a developing country in the face of an external shock. China's fiscal response was prompt, with the cyclically-adjusted budget deficit increasing quickly from 1% of GDP prior to the crisis to 1.9% in 1999 and 2.4% in 2000, and remaining in the vicinity of 2% through 2003 (Kuijs and Xu 2008). This period of expanded deficits coincided with a growth slowdown from 9.5% per year in the five years before the crisis to an average of 7.9% in the five years following the crisis. Unusually relative to the experience of most developing countries, the fiscal stimulus was reversed, with cyclically-adjusted budget deficits returning to 1% of GDP after 2003. Finally, the fiscal stimulus was achieved primarily through increases in infrastructure spending.
Several unique factors contributed to China's ability to carry out this policy response. First, China entered the East Asian crisis with very low public debt, and so there was considerable scope to engage in deficit spending. Second, China's prohibition on subnational government borrowing combined with many years of very rapid growth ensured that local governments had strong demand for infrastructure investment projects but were effectively credit-constrained from carrying them out, and so relaxing these constraints through a fiscal expansion generated a rapid response. Third, although the budgetary size of the fiscal stimulus was modest at around 1% of GDP, it was leveraged into significantly larger investment spending -- the central government typically provided only cofinancing for infrastructure projects, which together with an implicit loan guarantee, encouraged bank lending to finance the remainder of these projects. Finally, this fiscal expansion was implemented during a period of still-very-rapid growth -- even during growth slowdowns growth in China is faster than in virtually any other developing country!
In contrast to the case of China, the risks of a mistimed expansionary change in fiscal policy are illustrated by the experience of Argentina in the mid-1990s. After undergoing a recession in 1995, largely due to the propagation of Mexico’s Tequila crisis, a recovery took hold in 1996 and strengthened in 1997. In the midst of the post-Tequila boom, Argentina’s fiscal policy took an expansionary turn that fueled the boom further: cyclically-adjusted fiscal indicators (that is, after removing the endogenous response of fiscal variables to the cycle) reveal a major fiscal impulse from the second half of 1996 to late 1998 (Servén and Perry 2005). Much of the expansion was due to declining revenue collection, partly associated with a pension system reform whose adverse short-term consequences for public revenue were not offset by increases in other taxes or reductions in spending. As a result, public debt as a ratio to GDP kept a rising trend in spite of the economic bonanza. When the Russia crisis erupted in 1998, with the ensuing global financial turmoil, Argentina’s fiscal authorities had no room for maneuver left to mitigate the recession, and had to engage instead in a severe fiscal contraction that deepened the slump. This eventually became one of the leading factors behind the eventual collapse of the Convertibility regime.
Of course, attempts at fiscal stimulus can fail in industrial countries too. The often-cited case of France in the 1980s illustrates the dangers of a fiscal expansion in a context of weak investor confidence and conflicting monetary and fiscal policy objectives (Sachs and Wyplosz 1986). Over 1981-83, the authorities engaged in a major fiscal expansion to combat rising unemployment in the midst of a global recession. The expansion involved an expenditure increase, at first in the form of transfers and subsidies, accompanied later by a rise in public investment as well. The structural (i.e., cyclically-adjusted) fiscal deficit rose by 1.5 percent of GDP between 1980 and 1982. In spite of extensive capital controls, France’s currency peg, part of the EMS arrangement, came immediately under pressure as skeptical investors anticipated higher inflation and/or tax hikes, and the pressure (with the corresponding foreign reserve losses) redoubled as current account deficits mounted. Repeated interest rate hikes, which partly undid the expansionary aggregate demand effects of the fiscal stimulus, could not prevent two devaluations in 1981-82. The modest growth effects of the fiscal expansion quickly dissipated, and eventually the authorities were forced to abandon it and adopt a fiscal austerity package in 1983.
Back to top
Recommendations for the Current Crisis
The global scope and the anticipated magnitude of the current crisis set it apart from the “normal” cyclical downturns and the more localized emerging-market crises of the last two decades. In these dimensions, the current crisis represents a rare event, and caution is needed when applying the lessons from past experience to the present situation. Nevertheless, previous experience suggests that the following issues should be considered carefully when designing potential fiscal responses by developing countries to the current crisis.
Any fiscal response should be commensurate to the shock experienced by the country
. The effect of growth slowdowns in rich countries on developing economies differs considerably across countries, depending on the strength of their links with rich countries through trade, capital flows, and remittances.[10
] Moreover, past experience suggests that the adverse effects of growth slowdowns in rich countries have been modest for many developing countries.[11
] Given risks associated with discretionary fiscal policy it seems prudent to ensure that the global crisis does not trigger unsustainable fiscal expansions in those countries that are relatively unaffected by it.
Consideration needs to be given to the scope for monetary as well as fiscal policy responses, as both entail risks. Policymakers need to consider the scope for coordinated monetary and fiscal interventions, given the uncertain effects of the latter. In many developing countries the central bank’s policy interest rates are still high and inflation is modest, suggesting that there may be scope for traditional easing measures -- unlike in the U.S. where policy rates are approaching their lower bound of zero. But of course a monetary response entails risks as well, including (i) downward pressures on exchange rates, and (ii) a loss of hard-won anti-inflationary credibility in countries with past histories of reckless monetary policy and accompanying high inflation.
Fiscal expansions need to be credibly and sustainably financed. As noted above, lack of access to external finance has been an important impediment to expansionary fiscal policy during downturns in developing countries. This problem is likely to be particularly acute during the crisis, with its accompanying paralysis in international credit markets. Many developing countries have limited capacity for domestic borrowing to finance increased spending, and monetizing fiscal deficits is potentially very risky. Moreover, while investors’ generalized ‘flight to safety’ has lowered the cost of public debt service in the U.S. and other industrial countries, facilitating the financing of expansionary fiscal policies, the opposite has happened in most emerging markets, whose sovereign spreads have risen significantly in recent months. The episode of France in Box 1 shows that a fiscal expansion with weak investor confidence in the sustainability of public finances can be self-defeating. All this suggests that only those developing countries with strong fiscal positions and large reserve stocks can afford to finance a fiscal expansion.
The fiscal expansion must be timely but not rushed. There is consensus that fiscal interventions need to be timely in order to be effective, and that mistimed interventions can be counter-productive (as illustrated by the Argentina episode in Box 1). This has been a challenge in developed countries, and is even more so in developing countries, where data quality (to identify downturns and recoveries in real time) and fiscal institutions (to design and implement any proposed spending increases) are weak. In the case of the current crisis, the magnitude and likely length of the crisis make timing issues easier. But nevertheless there are serious risks to rushing ahead to expand public spending when adequate oversight institutions and capacity to appraise new projects are not in place – as is the case in many developing countries. In this context policymakers should first carefully consider the scope for expanding tested and well-functioning existing projects and financing pre-appraised and 'shovel-ready' new projects before embarking on completely new and untried public spending projects that risk breeding white elephants.
The balance of tax cuts versus spending increases needs to be tailored to country circumstances. Assuming that there is a role for discretionary expansionary fiscal policy that can be sustainably financed, conventional wisdom is that spending increases are more effective at stimulating aggregate demand, since households might simply save tax cuts or direct transfers. However, this tradeoff is complicated by two factors in many developing countries. On the one hand, in countries with weak fiscal and oversight institutions the risk is greater that large new public spending programs will be wasteful, captured, and hard to reverse -- even in the U.S., witness the likely success of begathon for public money by the auto industry! On the other hand, tax cuts and/or social insurance transfers will not reach many of the poorest households and firms in the informal sector. In fact, there is a risk that such biased automatic stabilizers might exacerbate inequality during the downturn.
To prevent a weakening of public finances, spending increases should concentrate on areas where the expenditures are either reversible or likely to increase growth in the future. This is crucial to ensure that long-run fiscal and debt sustainability is not jeopardized by the countercyclical spending increase. Concretely, this can be done by focusing spending on projects that act as automatic stabilizers. For example, expansion of means-tested social benefit programs will naturally occur and should be financed during downturns as more and more people fall below the eligibility thresholds, and this will reverse as the economy recovers. Similarly, workfare programs with a clearly below-market wage offer will attract participants in downturns but will not be appealing once the economy recovers. The risks of unsustainable public debt accumulation are also reduced to the extent that increases in spending fall in areas such as infrastructure (as in the China episode in Box 1) where there are reasonably expectations of longer-term growth benefits (and thereby expanding tax bases) as well as direct cost-recovery through future user fees. Moreover, aside from this long-run growth contribution, infrastructure projects – e.g., road construction and maintenance – can also provide the basis for employment programs that reach poor workers.
In summary, historical experience suggests that expansionary fiscal policy has not been an effective tool for most developing countries in responding to economic downturns. But this does not mean that fiscal policy can play no role in mitigating the effects of the current crisis; rather, it implies that recommendations for countercyclical fiscal measures should take into account the (sobering) lessons from past experience.
Overall, developing countries should consider two priorities in the use of the limited scope they may have for expansionary fiscal policy – to ensure that short-term crisis alleviation is aligned with long-run development. First, strengthening social safety nets is key in order to help the most vulnerable and those most affected by the crisis to cope. This is particularly important in areas where short-term coping mechanisms (such as selling assets or cutting back on caloric intake) can have severe long-term impacts (inability to produce, stunting and reduced cognitive abilities). A side benefit of strengthening such safety nets is that it also strengthens automatic stabilizers that are widely viewed as a more effective form of countercyclical fiscal policy. Second, spending increases should concentrate in areas such as infrastructure that are likely to contribute to growth in the long term. Even here however there are obvious risks to proceeding quickly with new infrastructure spending, and expansions in this area could best be used to finance "shovel-ready" projects that have already been appraised and found viable.
Back to top
* Helpful inputs were provided by Claudio Raddatz (DECRG), Hans Timmer (DECPG), Vivek Suri and Eduardo Ley (PRMED), Louis Kuijs (EASPR), Bert Hofman (EAPCF) and Antonio Fatás (INSEAD).
1. See Gavin and Perotti (1997), Kaminsky, Reinhart and Végh (2004), Ilzetzki and Végh (2008), and Calderón and Schmidt-Hebbel (2008).
2. Suescún (2007) shows that automatic tax stabilizers are much weaker in Latin America than in industrial countries.
3. Kaminsky, Reinhart and Végh (2004) refer to this phenomenon as "when it rains, it pours".
4. See for example Tornell and Lane (1999) for a tragedy-of-the-commons argument for overspending bias when many powerful interest groups compete over public spending. Fatas and Mihov (2002, 2007) find that the volatility of fiscal policy is substantially higher in countries with weak institutions.
5. Calderón and Schmidt-Hebbel (2008) find strong empirical support for both institutional failures and credit constraints as determinants of countercyclical fiscal policy in emerging economies.
6. Blanchard and Perotti (2002) provide widely-cited estimates of a statistically significant and economically meaningful effect of fiscal expansions on output for the United States, but the relevance of such findings for developing countries is less clear. Some authors (e.g., Perotti 2007) have even suggested that for developing countries data and methodological problems are sufficiently severe that the effects of fiscal policy on short-term fluctuations are unknowable.
7. Fatás and Mihov (2003).
8. In this light, IMF (2008) concludes that for emerging economies "discretionary fiscal policy does indeed appear to do more harm than good".
9. See for example Beetsma (2008).
10. Recent DECPG forecasts confirm this heterogeneity. Although growth for the developing world is projected to fall by 2% in 2009 due to the crisis, for nearly 50 developing countries the projected growth reduction is less than 1%, while for 15 countries it is more than 3%.
11. For example, past experience indicates that fluctuations in industrial countries' growth performance accounted for only 1.5% of output volatility in relatively disconnected regions such as Sub-Saharan Africa but 2% and 6% in more connected regions such as Latin America or East Asia (Raddatz (2007, 2008)).
12. For much more detailed recommendations in this area see Ravallion (2008).
Blanchard, Olivier, and Roberto Perotti. 2002. "An Empirical Characterization of the Dynamic Effects of Changes in Government Spending and Taxes on Output." Quarterly Journal of Economics 117(4):1329-68.
Beetsma, Roel. 2008. "A Survey of the Effects of Discretionary Fiscal Policy." University of Amsterdam. Processed.
Calderón, César, and Klaus Schmidt-Hebbel. 2008. "Business Cycle and Fiscal Policy: The Role of Institutions and Financial Markets." Central Bank of Chile and World Bank. Processed.
DeLong, Bradford. 1998.“Fiscal Policy in the Shadow of the Great Depression.” In Michael Bordo, Claudia Goldin, and Eugene White, eds., The Defining Moment: The Great Depression and the American Economy in the Twentieth Century. Chicago and London: University of Chicago Press.
Fatás, Antonio, and Ilian Mihov. 2003. "The Case for Restricting Fiscal Policy Discretion." Quarterly Journal of Economics 118(4):1419-47.
Fatás, Antonio, and Ilian Mihov. 2007. “Fiscal Discipline, Volatility and Growth.” In Guillermo Perry, Luis Servén and Rodrigo Suescún, eds., Fiscal Policy Stabilization and Growth: Prudence or Abstinence? Washington DC: The World Bank.
Fishback, Price V., William C. Horrace, and Shawn Kantor. 2005. “Did New Deal Grant Programs Stimulate Local Economies? A Study of Federal Grants and Retail Sales During the Great Depression.” Journal of Economic History 65: 36-71.
Gavin, Michael, and Roberto Perotti. 1997. "Fiscal Policy in Latin America." NBER Macroeconomics Annual. Cambridge and London: MIT Press. pp. 11-61.
Ilzetzki, Ethan, and Carlos Végh. 2008. "Procylical Fiscal Policy in Developing Countries: Truth or Fiction?". Manuscript, University of Maryland.
International Monetary Fund. 2008. World Economic Outlook. Chapter 5.
Kaminsky, Graciela, Carmen Reinhardt and Carlos Végh. 2004. "When It Rains, It Pours: Procyclical Capital Flows and Macroeconomic Policies." NBER Macroeconomics Annual. Cambridge and London: MIT Press.
Kuijs, Louis, and Gao Xu. 2008. "China's Fiscal Policy: Moving to Center Stage." World Bank Beijing Office. Processed.
Perotti, Roberto. 2007. "Fiscal Policy in Developing Countries: A Framework and Some Questions." World Bank Policy Research Working Paper 4365.
Raddatz, Claudio. 2007. "Are External Shocks Responsible for the Volatility of Output in Low-Income Countries?" Journal of Deveopment Economics.
Raddatz, Claudio. 2008. "Have External Shocks Become More Important for Output Fluctuations in Africa?" Development Research Group, World Bank. Processed.
Ravallion, Martin. 2008. "Bailing Out the World's Poorest." World Bank Policy Research Working Paper 4763.
Sachs, Jeffrey. and Charles Wyplosz. 1986. “The Economic Consequences of President Mitterrand.” Economic Policy 261-322.
Servén, Luis, and Guillermo Perry. 2005. “Argentina’s Macroeconomic Collapse: Causes and Lessons.” In J. Aizenmann and B. Pinto, eds. Managing Volatility and Crises, Cambridge University Press.
Suescún, Rodrigo. 2007. “The Size and Effectiveness of Automatic Fiscal Stabilizers in Latin America.” In Guillermo Perry, Luis Servén and Rodrigo Suescún, eds., Fiscal Policy Stabilization and Growth: Prudence or Abstinence? Washington DC: World Bank.
Tornell, Aaron. and Philip Lane. 1999. "The Voracity Effect." American Economic Review 89(1):22-46.
Back to top