Commodity Price Volatility and Inclusive Growth in Low-Income Countries
Chapter

Chapter 11. Commodity Price Volatility, Poverty, and Growth Inclusiveness in Sub-Saharan African Countries

Author(s):
Rabah Arezki, Catherine Pattillo, Marc Quintyn, and Min Zhu
Published Date:
October 2012
Share
  • ShareShare
Show Summary Details
Author(s)
François Bourguignon

Introduction

The comparative advantage of sub-Saharan African countries predominantly lies in the export of commodities, whether they be oil, mineral, or agricultural. Managing development based on this kind of comparative advantage is difficult, however, so much so that some refer to the availability of natural resources in a country as the natural resource curse. Yet, there is little doubt that, on average, high real commodity prices as well as the discovery and exploitation of new resources are associated with faster growth, even though possibly tempo-rarily.1 From that point of view, it is difficult not to relate the sustained growth in sub-Saharan Africa since the turn of the century (and for some countries, really from the mid-1990s) to the sustained surge in international commodity prices, itself most likely fed by the heavy demand of emerging economies. It is also interesting that this increase in commodity prices and acceleration of growth came after a long period (i.e., 15 years) of decline in commodity prices and economic stagnation.

The difficulty is that both descending and then ascending trends in the real prices of the commodities exported by sub-Saharan African countries and therefore in their terms of trade have come with a level of volatility so high that it has been impossible to detect such trends. Part of the inability of sub-Saharan African countries to take advantage of them or to avoid mismanaging them, and therefore part of the commodity curse, lies precisely in this volatility. There are innumerable examples of governments that increased their spending as a response to high commodity prices and had to suddenly stop ambitious investment programs because of a drop in the prices and the impossibility of borrowing on foreign markets—often because of debt accumulated in boom times.

Interestingly enough, the huge fall in international commodity prices that took place during the 2008–09 world crisis may have been the first instance in which most sub-Saharan African countries were able to apply countercyclical policies instead of being caught, as in previous instances, in the general crush. Yet, it is not clear whether they could have gone on for very long with such policies if prices had not vigorously rebounded in 2010.

How much poor people have been affected by this volatility of commodity prices over the last 10 to 15 years and whether governments have been able to take advantage of recent favorable circumstances to engineer a new development regime that would reduce their vulnerability in the future is hard to say. In this respect, one important observation is that the structure of growth in sub-Saharan African countries does not seem to have changed much over the last decade. Faster GDP growth in recent years has been concentrated in the sectors of construction and services, as if the additional income brought by the boom in commodity prices was mostly used to feed additional consumption and possibly to improve infrastructure in urban centers. If African governments have been more cautious than in the past in managing commodity export revenues, they do not seem to have put their countries on a development path that will allow them to progressively free themselves from the volatility of international commodity prices.2

In short, the possible trend reversal in commodity prices that may have taken place at the turn of the century may not yet have had structural effects on sub-Saharan African economies and on their poverty reduction capacity. A possible reason for this is that the natural volatility of commodity prices rationally pushed governments not to take radical action, unlike what they often did in the past.

This chapter focuses on the evolution of poverty in the sub-Saharan African region in connection with changes in the terms of trade faced by the various countries. Of course, the link goes through GDP growth and the way it is affected by commodity prices and volatility. In this chapter, we first discuss the available empirical evidence on poverty, growth, and commodity prices. We then look at the policy instruments available in African countries to cope with the volatility of commodity prices, at both the macro level and the micro level, and also through the help that the international community can contribute to African countries.

Poverty, Inclusiveness, Growth, and Commodity Price Volatility

Sub-Saharan countries are heavily dependent on the export of a few commodities, and the well-being of the population may be strongly affected by fluctuations in commodity prices on international markets. The channel may be a direct one, as seen with agricultural export prices affecting farmers’ incomes. It may be indirect, such as when agricultural exports are handled by a marketing board with an implicit export tax feeding the government’s budget or when the exported commodity is extracted from the ground by national or foreign companies paying royalties and taxes to the government. With the indirect channel, the well-being of the people and, of course, the poverty rate are affected by the way the government spends the revenues it gets from commodity exports.

With perfect capital markets and some stationarity in the stochastic process behind commodity prices, the effect of their fluctuations on the well-being of agents could be smoothed through saving in good times and borrowing in bad times. In the absence of such markets and given the limited stationarity of commodity price fluctuations, the well-being of the agents and the growth rate of the whole economy may depend not only on the current price of commodities, but also on their uncertainty or volatility.

In view of this, the effect of commodity price and commodity price volatility on poverty may go through two channels. At a point of time, poverty may be higher because commodity prices are low but also because commodity prices are highly volatile, and this volatility affects the behavior of agents. Both effects may be direct, as in the case in which poor people’s incomes are directly impacted by commodity prices, such as with agricultural exports without marketing board shielding, or indirectly impacted, such as in the case in which poor people’s incomes are impacted through the government’s policy and its effect on the growth rate of the economy. If this is the case, the volatility of commodity prices might not affect poverty directly, but rather through the volatility they induce in the level of economic activity.

The importance of these various channels linking commodity prices and poverty is explored in Table 11.1, which reports the results of various regressions where the dependent variables is the change in poverty—as measured by the headcount index with the usual threshold of US$1.25 at purchasing power parity—between periods for which this poverty measure is available and the independent variables are the growth of GDP per capita, terms of trade, and measures of their volatility.

Table 11.1Poverty Reduction Determinants in Sub-Saharan Countries: Alternative Regression Results1

Dependent variable: poverty change during a development spell 1

(i)(ii)(iii)(iv)(v)(vi)(vii)
Initial poverty−0.53−0.63−1.16−1.35−1.55−1.37−1.62
−3.34−3.78−4.39−4.96−5.74−4.90−5.56
GDP per capita growth−0.92−2.31−0.76−2.42−2.722.43−2.81
−5.12−2.75−3.94−2.75−3.25−2.71−3.05
Initial poverty*GDP per capita growth0.030.030.040.030.04
1.691.932.451.922.27
Standard deviation of GDP per capita growth20.440.48
2.302.05
Change in terms of trade (%)−0.03

−0.38
−0.04

−0.44
Coefficient of variation of terms of trade20.04

0.58
Fixed country effectsNoNoYesYesYesYesYes
R20.390.420.690.720.760.720.77
Number of observations153535353535353
Source: United Nations Development Programme Millennium Development Goals database; World Bank, World Development Indicators and author’s calculations.Note: Poverty data come from the United Nations Development Programme Millenium Development Goals database, which seems to be the most consistent source for low-income countries Gross domestic product (GDP) per capita figures are from the World Development Indicators released by the World Bank, the same being true of the terms of trade indices, which for most countries summarize the impact of both exported and imported commodity prices on the domestic economy. The analysis is restricted to 53 growth spells in 25 sub-Saharan countries.

An observation is a period between two dates at which poverty data are available in the United Nations Development Programme Millennium Development Goals database. All explanatory variables are defined over the same time interval; t-statistics are in italics.

Multiplied by the number of years in the development spell for consistency with the rest of the specification.

Source: United Nations Development Programme Millennium Development Goals database; World Bank, World Development Indicators and author’s calculations.Note: Poverty data come from the United Nations Development Programme Millenium Development Goals database, which seems to be the most consistent source for low-income countries Gross domestic product (GDP) per capita figures are from the World Development Indicators released by the World Bank, the same being true of the terms of trade indices, which for most countries summarize the impact of both exported and imported commodity prices on the domestic economy. The analysis is restricted to 53 growth spells in 25 sub-Saharan countries.

An observation is a period between two dates at which poverty data are available in the United Nations Development Programme Millennium Development Goals database. All explanatory variables are defined over the same time interval; t-statistics are in italics.

Multiplied by the number of years in the development spell for consistency with the rest of the specification.

The story told by Table 11.1 is interesting. As could be expected, GDP per capita growth turns out to be a strong explanatory factor of poverty change, with an elasticity that depends on the initial level of poverty. The elasticity of poverty with respect to GDP found in column (i) of the table is a bit below unity, an order of magnitude consistent with earlier work, such as Bourguignon (2003) and Ravallion (2001). Column (ii) also shows that the relationship between poverty reduction and GDP growth is not linear, as found again in previous studies. All these results are robust with respect to fixed country effects as can be seen in columns (ii) and (iv).

The other columns of Table 11.1 show two other interesting results. On the one hand, the poverty change seems to be independent from the terms of trade, whether one considers the change in terms of trade during a growth spell or the volatility of that change. On the other hand, the change in poverty depends on the volatility of GDP growth during the same period. Other things being the same, a given growth rate over some period of time generates less poverty reduction if it has been obtained through an irregular process than with a constant annual growth rate.

The first result suggests that if poverty reduction reacts to changes in terms of trade, it does not do so directly, but rather through the impact of terms of trade or their volatility on GDP, assuming, of course, that terms of trade directly affect GDP, a point that will be discussed later in this chapter. This would seem reasonable in countries exporting mineral commodities, the extraction of which involves a tiny number of workers and the revenues of which accrue directly to the state. It may be seem more surprising for agricultural commodity exporters, since those commodities are much more labor intensive and often originate in small farms more vulnerable to poverty. Sudden falls in coffee or cocoa prices in countries like Côte d’Ivoire or Ghana are indeed known to have caused an acute increase in poverty.3 However, controlling for the nature of exported commodities does not modify the independence between poverty reduction and terms of trade.

A first explanation of this result is that producers of exported agricultural commodities often are shielded from fluctuations in international prices through administered prices set by marketing boards, which does not prevent drastic adjustments in the latter from time to time. Another possible interpretation of the lack of significance of terms of trade in explaining poverty changes lies in the denominator of the terms of trade. In many countries, oil is an important component of import, and its price has fluctuated widely over the recent past. It is not clear that poor people who are potentially affected by fluctuations in agricultural commodity prices would be equally affected by fluctuation in oil prices. Thus, terms of trade may not be the best indicator to describe the impact of export prices on poverty.

Unfortunately, there are few sub-Saharan countries for which full series of export price indices are available. The alternative to using the terms of trade would be to construct original data based on international commodity price series, export taxes, exchange rates, and consumer price indices. However, this proved to be feasible only for a small number of countries.

This possible weakness of the terms of trade for catching the impact of commodity price changes on poverty may explain why GDP growth is a powerful predictor of poverty reduction. Indeed, a fall in the international price of exported agricultural commodities is likely to produce both an increase in poverty and a slowdown or even a reversal in GDP growth. The GDP variable in the regressions reported in Table 11.1 may thus catch both effects at the same time.

In short, there is some ambiguity in the model being estimated. Due to data limitations, it is not clear whether commodity prices affect poverty essentially through their effect on aggregate activity, which is likely for mineral commodities, or also affect poverty directly.

The second interesting result in Table 11.1 is the fact that poverty reduction significantly depends on the volatility of growth during the spell being observed. This result is consistent with the view that poor people are somehow penalized more than the rest of the population in case of a big drop in GDP per capita and benefit less from very fast growth. In other words, they are more vulnerable to negative shocks and less favorably affected by positive shocks. On the other hand, in line with a previous argument, this nonlinearity may also capture the fact that the volatility of GDP growth is itself linked to that of commodity prices that also affect poverty directly.

Because the only significant determinant of poverty changes is GDP growth and GDP growth volatility rather than changes in terms of trade and its volatility, the extent to which GDP growth and volatility are affected by changes in terms of trade and their volatility remains to be seen.

First, it must be recalled that GDP growth over a period and its volatility over the same period are strongly and negatively correlated. In cross-sections of countries and over different time periods, it has been observed that countries where annual GDP growth rates are highly volatile tend to grow at a slower rate over time (Ramey and Ramey, 1995; Martin and Rogers, 2000; Hnatkovska and Loayza, 2005). An illustration of that relationship is given in Figure 11.1, which plots the mean and the standard deviation of annual growth rates of a sample of 20 LICs (dark markers, mostly sub-Saharan countries) from 1980 to 2008. As can be seen from the trend line, a 1 percentage point increase in the standard deviation of GDP reduces the mean growth rate by 0.6 percentage points, a rather sizable effect.4 For comparison, the figure also includes observations for a small sample of 10 middle-income countries (MICs; light markers). It can be seen that their distribution is analogous to the distribution of LIC observations.

Figure 11.1The Negative Long-Term Relationship between Growth and Growth Volatility: Selected Low-Income and Middle-Income Countries, 1990 to 2008

Sources: World Bank, World Development Indicators; and author’s calculations.

Note: BKN: Burkina Faso; BRA: Brazil; BUR: Burundi; CAR: Central African Republic; CHL: Chile; COL: Colombia; DRC: The Democratic Republic of Congo; EGY: Egypt; GAM: The Gambia; GHA: Ghana; HAI: Haiti; KEN: Kenya; MAD: Madagascar; MAL: Malaysia; MAU: Mauritania; MLW: Malawi; MOR: Morocco; MOZ: Mozambique; NEP: Nepal; NIG: Niger; PER: Peru; PHI: Philippines; SEN: Senegal; THA: Thailand; TOG: Togo; TUN: Tunisia; UGA: Uganda; ZIM: Zimbabwe; ZMB: Zambia.

The next step is to understand the channels through which GDP growth volatility affects the average rate of GDP growth as well as to identify the sources of volatility. An important literature has developed in this respect that we do not intend to summarize here.5 It makes a distinction between exogenous causes of GDP volatility such as foreign prices, trade partners’ activity, contagion in times of international financial crisis (e.g., flight for quality and sudden stops in capital flows), and endogenous factors like self-inflicted economic crises or inadequate response to exogenous shocks. Channels through which exogenous factors affect overall growth may indeed be ill-adapted policies or expectations about future prices on export and import markets or about future capital flows. They may also be of a structural nature like the degree of openness of the economy, its access to international credit markets, or simply its governance and the capacity of policymakers to make appropriate decisions.

Turning now to the effect of terms of trade on growth, the literature on growth regressions offers plenty of evidence. Most growth panel regressions that are run on five-year periods find that the terms of trade significantly affect growth performances in developing countries (Loayza and Soto, 2002; World Bank, 2010). From that point of view, it is thus likely that the effect of GDP growth on poverty in the regressions reported in Table 11.1 implicitly includes the effect of changes in terms of trade and commodity prices on the level of economic activity.

Less emphasis has been given in the literature on the impact of the volatility of terms of trade on the average rate of growth. Yet this influence is not marginal. Based on the view that past volatility is an indicator of future uncertainty regarding export and import prices, Mendoza (1997) showed that a substantial part of cross-country differences in consumption growth could be explained by terms of trade volatility. Bleaney and Greenaway (2001) found a significant impact of the volatility of the terms of trade on growth in a small sample of sub-Saharan countries. Recently, Furth (2010) showed, on the same basic World Development Indicators data as those used above, that one-quarter of the variance in growth rates among 51 developing countries observed between 1980 and 2007 could be explained by the standard deviation of the terms of trade with respect to their trend and that very little could be explained by this trend itself or even by the volatility of GDP. The lack of relationship with the trend in the terms of trade in that study may seem surprising in view of the strong relationship found in some standard growth regressions, as mentioned previously. The difference lies in the time horizon. Growth regression studies describe the short-term impact of fluctuations in terms of trade, whereas the analysis in Furth (2010) refers to the long-term effect, with the implicit assumption that over various decades, countries had the time to adjust their growth strategy to long-term trends in export and import prices.

As far as the effect of terms of trade volatility on growth is concerned, it is not unlikely that the results in the various studies mentioned above significantly depend on the sample being used, the period of analysis, and the cross-section versus panel nature of the methodology. Whether working on the sample of LICs as in Figure 11.1 or on the extended sample that also includes some MICs, it turns out that results are not as clear-cut as suggested in the previous studies. Differences in growth rates over the last three decades (1980–2008) or the last two decades (1990–2008) do not seem to depend significantly on terms of trade trends and to depend only very weakly on their volatility during the period. This variability of regression results across samples would be consistent with a relationship between growth and terms of trade volatility that would be strong for some countries but weaker or even nonexistent in other countries. Some work remains to be done to identify which types of countries would belong to the first set, particularly whether they are more predominantly specialized in mineral or oil exports.

We will now examine the relationship between the volatility of the terms of trade and the overall volatility of GDP growth. This is less problematic as there is something mechanical in that relationship. If it is the case that the terms of trade are a significant determinant of growth in the short term, as recalled above, then the volatility of GDP growth should automatically increase with that of the terms of trade, other things being equal. This is illustrated in Figure 11.2, which plots the volatility of GDP growth against that of the terms of trade for a sample of countries relatively similar to that in Figure 11.1.

Figure 11.2The Contribution of Terms of Trade Volatility to GDP Volatility: Selected Low-Income and Middle-Income Countries, 1990 to 2008

Sources: World Bank, World Development Indicators; and author’s calculations.

Note: BEN: Benin; BGL: Bangladesh; BKN: Burkina Faso; BRA: Brazil; BUR: Burundi; CAR: Central African Republic; CHL: Chile; CHN: China; COL: Colombia; DRC: The Democratic Republic of Congo; EGY: Egypt; GAM: The Gambia; GHA: Ghana; HAI: Haiti; KEN: Kenya; MAD: Madagascar; MAL: Malaysia; MAU: Mauritania; MLW: Malawi; MOR: Morocco; MOZ: Mozambique; NEP: Nepal; NIG: Niger; PER: Peru; PHI: Philippines; SEN: Senegal; THA: Thailand; TOG: Togo; TUN: Tunisia; UGA: Uganda; ZMB: Zambia.

In summary, the relationship between poverty and commodity prices in LICs as it appears in the cross-country analysis undertaken in this chapter is quite simple. The first point is that commodity prices, somewhat awkwardly approximated by the terms of trade, seem to affect poverty mostly through GDP growth and, more interestingly, through GDP growth volatility. This is confirmed by standard cross-country analysis of GDP growth, in which indeed terms of trade play a significant role in the short term, and their volatility significantly affects overall GDP growth volatility. Of course, it is most likely that direct effects of commodity prices are missed in such an aggregate analysis because of a lack of detail in the econometric specification being used and because of the cross-country nature of the exercise. Unfortunately, more precise data are unavailable for a large enough number of countries.

It is well-known that growth is a major determinant of poverty reduction. The preceding results simply confirm this very simple fact, adding to it the interesting idea that the relationship between poverty and growth might be nonlinear, with poor people more affected than others in big recessions and benefiting less in exceptionally good times.

Repeating the preceding exercise with inequality rather than poverty so as to evaluate the effect of commodity prices and their volatility on the inclusiveness of growth did not yield interesting results. GDP growth and volatility turned out to be nonsignificant, the same being true of the terms of trade. This confirms that if terms of trade and their volatility play any role in poverty, it is through growth and not through changes in the distribution of income.

The charts in Figure 11.3 show the evolution of the terms of trade and the inequality in consumption expenditures in four African countries, the exports of which consist mostly of agricultural commodities, the price of which is expected to directly affect some specific groups in the population. According to the distributional data available, it can be seen that the severe cocoa-led worsening of the terms of trade in Côte d’Ivoire and Ghana in the late 1980s and the first half of the 1990s did not produce any big change in inequality. It increased a bit in Ghana and decreased a bit in Côte d’Ivoire. Then inequality increased in both countries, although more in Côte d’Ivoire, whereas the terms of trade were recovering in Ghana and remained more or less stable in Côte d’Ivoire. In Madagascar, the improvement in the terms of trade in the second half of the 1990s has been associated with an increase in inequality, but the opposite evolution in the early 2000s did not produce any noticeable change. Out of the four countries in Figure 11.3, Kenya seems the only one to show a systematic pattern in which inequality varies in the opposite direction of the terms of trade.

Figure 11.3Terms of Trade and Consumption Inequality (Gini) in Selected Sub-Saharan Countries

Sources: World Bank, World Development Indicators; and author’s calculations.

Of course, there are simple explanations behind the curves of Figure 11.3 that show that the statistical relationship between inequality and terms of trade is necessarily a complex one. In the case of Côte d’Ivoire and Ghana, for instance, the role of the cocoa (and coffee) marketing boards has been crucial either in isolating domestic producers from international price fluctuations or in smoothing their effects or postponing them. In both countries, there is indeed very little effect of the worsening of the terms of trade that took place in the 1980s and early 1990s on inequality. In Côte d’Ivoire, the administered price of cocoa was drastically lowered in 1990 by the marketing board, Caisstab, which was dismantled 10 years later. Presumably, this should have produced a shock on inequality if cocoa producers had been the only people hit. However, the crisis was truly national. GDP per capita dropped by 5 percent that year so that most agents were hit at the same time, and no big change was observed in the degree of inequality. Interestingly enough, inequality increased much later, when Caisstab had practically ceased its interventions.

In Ghana, floor producer prices remained in force despite the liberalization of the Cocobod, the equivalent of Caisstab. They were progressively adjusted proportionally to international prices at the same time as GDP per capita started stagnating. There too, the impact on inequality, although positive, was limited. Comparable country-specific explanations are available for other countries as well. For instance, the loose inverse relationship between inequality and terms of trade in the case of Kenya might have to be associated with the weak regulation of the main agricultural exports in that country.

This discussion of the elusive relationship between the terms of trade and inequality helps one to understand the results obtained earlier on the evolution of poverty that seemed to depend on commodity prices mostly though GDP growth and its volatility. It was already stressed that this was logically the case for exporters of mineral commodities and oil. In the presence of domestic price–smoothing mechanisms, it is clear that international fluctuations in the price of agricultural commodities also affect national economy agents only through aggregate mechanisms and therefore GDP fluctuations. It is therefore sufficient to have enough heterogeneity in the nature of the commodities being exported and in the marketing of agricultural exports for no discernable effect of commodity price fluctuations on poverty or the distribution of standards of living to appear in statistical cross-section analysis. This does not mean that commodity prices do not have a direct impact on poverty and inequality in some countries or under particular circumstances. The result from the statistical analysis in this chapter simply suggests that this is not the most frequent case as commodity price shocks in countries that heavily depend on commodity exports have almost an immediate macro-economic effect and hit most agents rather than only the commodity producers.

Instruments to Cope with the Effects of Commodity Price Volatility

The factual analysis of the relationship between poverty, inequality, and commodity price volatility suggests that the main instruments for coping with the adverse effects of that volatility are more of a macroeconomic than a microeconomic nature. At the same time, the presence of instruments that a would shield some vulnerable groups in the society from commodity shocks, whether they are hit directly or indirectly, through macroeconomic spillovers, may enhance the efficiency of macroeconomic policies in reducing the social cost of commodity price fluctuations. From that point of view, considerable progress has been made lately in developing countries in implementing redistribution instruments that could possibly be used to cushion macroeconomic shocks on poor people. The implications of the availability of these new instruments will be examined next, followed by a discussion of macroeconomic instruments and the potential role of the international community.

Micro Instruments to Help Poor People Cope With Income Shocks in LICs

The conditional cash transfer programs Progresa and Bolsa Familia were launched in Mexico and Brazil, respectively, and have attracted a lot of attention in the development community. They have shown in particular that it is possible to manage huge cash transfers in developing countries at a rather low administrative cost and to reduce poverty and inequality substantially through these instruments, both in the short term by targeting poor people and in the long term through the obligation conferred on beneficiaries to send their children to school and health clinics. Programs of this type have now spread to many other Latin American countries and to other developing regions. They are now considered an important part of social protection and of the “social safety net” in those countries.

Such a view calls for several important remarks, with respect first to the actual “income insurance” content of these instruments and, second, to their applicability in a low-income African context.

Most cash transfer programs are targeted toward poor people who are identified through permanent household characteristics rather than current market income because the latter is seldom observed. The consequence is that the programs cannot really insure against income shocks of the type triggered by price fluctuation on international commodity markets and should not be considered a true safety net. Yet, as they provide limited resources on a regular basis, cash transfers reduce the volatility of market incomes, and when the liquidity constraint is binding, they make adjustment to a worsening economic environment less painful. It must be noted, moreover, that this is true of all cash transfer programs, whether conditional or unconditional. For instance, transfers paid to elderly people without formal pension payments play this uncertainty-reducing role when the elderly live within larger households, as the latter can help alleviate overall households’ liquidity constraints.

Some other social protection instruments in developing countries may play a more direct income insurance role. Although unemployment insurance schemes are still infrequent (and are mostly limited to the formal sector of the economy), public employment guarantee schemes at some arbitrary (low) wage have been used in periods of crisis and are permanent in some countries, such as India and South Africa. Also, microcredit programs that allow poor people to borrow at a reasonable rate help them cope with the consequences of shocks, provided that they allow people to borrow for consumption purposes.

As most of these programs have been developed in MICs, the issue arises regarding their applicability in sub-Saharan LICs and their capacity to provide a safety net against commodity price fluctuations.6 Concerning conditional cash transfer programs, numerous pilots have been launched in various countries (e.g., Burkina Faso, Ghana, Kenya, Malawi, Nigeria, Tanzania, Zambia). However, the emphasis in these experiments is more on the impact of these programs on the behavior of beneficiaries in terms of schooling and health checks of children than on the role of these transfers in reducing income volatility. With poverty rates very often close to or even above 50 percent, it is not clear whether it would not be extremely costly, in relative terms, to scale up these pilots so as to cover all poor people in a country.

Of course, more modest programs could be envisaged that would target the poorest rather than all poor, although it might often be difficult to distinguish people according to the severity of poverty they face. Targeting specific groups of poor people who are easily identifiable might also be considered. Social pension programs, like the Old Persons Grant in South Africa, for instance, might not cost as much and could already do much in reducing poverty and vulnerability, not only for the elderly, but for the households in which they live (Bertrand, Mullainathan, and Miller, 2003; Duflo, 2003).

Even if the programs are relatively modest and target only a segment of the poor population, such cash transfer programs offer an interesting possibility in the context of commodity price shocks. They provide effective channels through which governments, and possibly foreign donors, can transfer more or less purchasing power to poor people depending on the strength of the macroeconomic shocks they have to face due to commodity price variations. In other words, once in place, those programs provide a convenient way for governments to provide some kind of insurance.

More direct income insurance instruments include public employment programs offering below–market wage jobs in public works to people who cannot earn a living on the market. Most often, such programs are created in periods of crisis. This has been the case in several sub-Saharan countries (e.g., Senegal, Tanzania, Uganda, Zambia). The difficulty of such temporary programs is the cost and the delay of putting them in place, which often makes them not very effective. There are strong arguments for making such programs permanent, with some monitoring of their intensity precisely to address the problems arising from fluctuations in economic activity. South Africa’s Expanded Public Work Program, created in 2004, and Ethiopia’s Productive Safety Nets Program, launched in 2005 and financed by foreign aid, are examples of such permanent programs. Yet there is little evidence on the way they can be scaled up or down as a response to the impact of macroeconomic shocks on individual standards of living.

Microcredit, and more generally microfinance, would seem to allow poor people to effectively smooth out their consumption when hit by positive and negative shocks. Yet the evidence on the overall effects of microfinance on poverty and vulnerability to poverty is, at this stage, somewhat mixed. Rigorous impact evaluation is difficult to conduct because of the difficulty of designing truly random experiments and because of the considerable heterogeneity of microfinance operators, from nongovernmental organizations to commercial banks to informal operators, as well as in the definition of operations. The common view has long been that microfinance contributes to an alleviation of poverty and consumption smoothing but also that it does not always reach out to the poorest (Morduch, 2002). The few randomized controlled trials available to date tell a more cautious story (Bauchet and others, 2011). In a sub-Saharan context, in which the large majority of poor people do not have access to any financial operator, it is unlikely that microfinance could be an effective way of coping with market income volatility today. At the same time, things are changing rapidly, and this might be an area for policy intervention that deserves special attention.

Fiscal Policy Management to Cope with Terms of Trade Volatility in LICs

The micro instruments listed above are important for addressing idiosyncratic risks that poor households in developing countries are badly equipped to deal with. Of course, they also permit systemic risks of the type generated by terms of trade volatility to be addressed. Doing so, however, requires that governments have the capacity to scale up those transfers and programs through fiscal policies able to go against the direct budget effects of changes in commodity prices.

Economies that are price takers on all foreign markets and have limited access to the international credit market have few degrees of freedom to counteract the effects of adverse changes in the foreign price of their exports or imports. Failures to satisfactorily handle this uncertainty have led to dramatic experiences in the past. It is quite clear, for instance, that the structural adjustment period in sub-Saharan African countries and the so-called lost decade of development in the 1980s are the direct consequences of a mismanagement, both at the country and at the world level, of the commodity price cycle that started with the oil price boom in the mid-1970s.

Since then, lessons have been learned, the most important one being the need to apply prudent fiscal policies, even in times of improving terms of trade. Accumulating reserves, under one form or another, during favorable times and de-accumulating them in a countercyclical way when hard times hit is the basic recipe for reducing the volatility of GDP and for minimizing the impact of commodity price volatility on poverty and inequality.

Although simple, this basic principle is not necessarily easy to implement.7 Political pressures for increasing fiscal spending in good times are strong and get stronger the longer the bonanza lasts. The right institutions must be set up to resist this type of pressure. At the same time, it is clear that in the obvious absence of stationarity in the stochastic process that governs commodity prices, no automatic rule can be set, which leaves room for discretion in the conduct of policy. At the same time, excessive prudence may be costly because it slows down expected growth. Finally, given the unpredictability of commodity price behavior, it is also clear that countercyclical policies will be impossible to implement if prices remain low for long enough.

Thus, prudent fiscal policies are a necessary condition for escaping the natural resource curse brought about by the volatility of commodity prices, but they are not sufficient. There will always be situations in which LICs stricken by a long spell of low commodity prices will need to rely on foreign credit or foreign aid.

The 2008–09 crisis was a good example of how the accumulation of reserves during the previous period of favorable terms of trade helped some sub-Saharan LICs weather the global shock thanks to countercyclical fiscal policies. But such policies could not have been maintained if commodity prices had not quickly recovered in 2010.

Reserve accumulation and fiscal adjustments are short- or medium-term policies to deal with commodity price volatility. Other aspects of macropolicies (i.e., monetary policy and exchange rate management) are important, too. In the long term, however, it may be more important to seek to reduce the overall importance of commodity exports within sub-Saharan countries through the diversification of economic activity. This may be accomplished through more trade or more regional integration, but also through the adequate structure of public spending. It is thus not only the size, but also the structure, of public spending that matters in regard to dampening the effect of commodity price volatility on growth and inclusiveness.

Role of the International Community and Donors

The preceding discussion points to two obvious areas for the intervention of the international development community and donors to mitigate the negative effect of commodity price volatility on growth and poverty reduction in LICs. The first one concerns the liquidity constraint of LICs that have gone through a sequence of negative shocks on commodity markets. The second one concerns the help that can be given to these countries to diversify their economies and make them less dependent on commodity exports.

On the first point, credit facilities, such as the Standby Credit Facilities provided by the IMF to LICs with short-term balance of payment needs should be made more accessible. More generally, more innovative lending products that would alleviate the liquidity constraint faced by LICs in periods of stress should be considered. This includes, for instance, the idea of mobilizing private lending more effectively, such as through partial guarantees provided by donors.

More explicitly linked to commodity price fluctuations, commodity price contingent aid contracts are a very attractive instrument. For instance, the French Development Agency (AFD) offers a concessional countercyclical loan (PTCC) to countries that have difficulties in repaying their debt and face an adverse evolution in their terms of trade. Namely, the PTCC may be triggered as soon as the export revenues of a country in year t are more than 5 percent below the average revenue observed during the 5 years before t.8 Clearly, such a facility will not prevent the country from having to adjust if the drop in export revenues is permanent, but it will reduce its cost if such an adjustment must be undertaken and will avert unnecessary social costs in case terms of trade appreciate again.

Concerning the diversification of economic activity in commodity-export-dependent countries, it must be stressed that the initiative relies first of all on the countries themselves rather than on donors. Indeed, what is at stake here is the general development strategy of the country. Yet foreign partners can also be very effective in helping countries achieve such a diversification.

As far as donors are concerned, focusing aid on the development of trade-facilitating infrastructures (“aid for trade”) as often suggested would certainly help. More helpful for African LICs would be to grant them effective trade preferences that would permit them to expand new nontraditional export markets for manufactured or possibly agricultural products.9 Indeed, such preferences already exist—the African Growth Opportunity Act (AGOA) in the United States and the Everything But Arms (EBA) agreement in the European Union. The problem is that they are extremely limited in scope. AGOA refers to a few, mainly textile products, whereas EBA imposes very restrictive rules of origin. Very substantial progress can be made on these two fronts without real danger for domestic producers in developed countries who are not present on these production lines. Such preferences would mostly affect the geographical distribution of their imports, with Asian investors possibly delocalizing part of their production to Africa. These preferences could be defined on a temporary basis so as to simply allow African countries to overcome infant industry handicaps that currently prevent them from entering some international manufacturing markets where they could develop some comparative advantage in the long term.

Another way of diversifying economic activity in African LICs is to push more toward regional integration. Organized in true custom unions (instead of a set of limited, mostly trade-diverting rather than trade-creating preferences), larger blocks of African countries would offer more import substitution opportunities and allow individual countries to improve their competitiveness by operating in larger markets. Of course, this would require these trade-integrated blocks to protect their common markets at a reasonable rate, a measure that is not incompatible with present World Trade Organization (WTO) rules. This would also require building the necessary infrastructure to facilitate trade between countries in the same trade bloc.

Conclusion

In short, there are policy measures that should reduce the cost of commodity price volatility in terms of growth, poverty reduction, and inclusiveness in the sub-Saharan LICs. The first set deals with the effects of shocks to commodity prices on individual incomes. They combine both fiscal policy rules and micro instruments, thereby allowing governments to reach the most vulnerable people and to reduce the volatility of their incomes. Such rules are being applied, and some of the micro instruments are starting to be developed in some countries. Yet considerable work remains to be done before they can become truly effective in mitigating the impact of commodity price shocks on poverty, inequality, and growth. A considerable amount of both national and international effort has to be devoted to that development.

The second set of measures aim at making commodity exports, and therefore commodity price volatility, less important in the development of African LICs. They have to do with the diversification of their economic activity and their foreign trade. The issue here is not only to reduce the individual and collective cost of commodity price volatility in those countries, but also to prepare them for more inclusive and sustainable development in the future. Given their size and the pace of demographic growth, it is indeed not very likely that a development strategy based exclusively on the rent of natural resources, whether mineral or agricultural, would be a long-term option for inclusive growth in Africa.

References

    ArbacheJ. S. and J.Page2010. “How Fragile Is Africa’s Recent Growth?Journal of African Economies Vol. 19 No. 1 pp. 124.

    • Search Google Scholar
    • Export Citation

    BarnichonR.2009The Optimal Level of Reserves for Low-Income Countries: Self-Insurance against External Shocks,IMF Staff Papers Vol. 56 pp. 85275.

    • Search Google Scholar
    • Export Citation

    BauchetJ.C.MarshallL.StaritaJ.Thomas and A.Yalouris2011Latest Findings from Randomized Evaluations of Microfinance,Access to Finance Forum No. 2 (Washington: Consultative Group to Assist the Poor/World Bank).

    • Search Google Scholar
    • Export Citation

    BertrandM.D.Mullainathan and D.Miller2003Public Policies and Extended Families: Evidence from Pensions in South Africa,World Bank Economic Review Vol. 17 No. 1 pp. 2750.

    • Search Google Scholar
    • Export Citation

    BleaneyM. and D.Greenaway2001The Impact of Terms of Trade and Real Exchange Rate Volatility on Investment and Growth in Sub-Saharan Africa,Journal of Development Economics Vol. 65 No. 2 pp. 491500.

    • Search Google Scholar
    • Export Citation

    BourguignonF.2003The Growth Elasticity of Poverty Reduction: Explaining Heterogeneity across Countries and Time Periods,in Growth and Inequalityed. byT.Eicher and S.Turnovsky (Cambridge, Massachusetts: MIT Press).

    • Search Google Scholar
    • Export Citation

    CogneauD. and R.Jedwab2012Commodity Price Shocks and Child Outcomes: the 1990 Cocoa Crisis in Côte d’Ivoire Economic Development and Cultural Change Vol. 60 No. 350734.

    • Search Google Scholar
    • Export Citation

    CohenD.H.Djoufelkit-Cottenet P. Jacquet and C.Valadier2007Prêter aux Pays les Plus Pauvres: Un Nouvel Instrument de Prêt Contracyclique,Afrique Contemporaine Vol. 3 Nos. 223–24 pp. 87111.

    • Search Google Scholar
    • Export Citation

    CollierP. and B.Goderis2007Commodity Prices, Growth, and the Natural Resource Curse: Reconciling a Conundrum,Working Paper Series No. 2007–15 (Oxford, U.K.: Centre for the Study of African Economies, University of Oxford).

    • Search Google Scholar
    • Export Citation

    CollierP. and A.Venables2007Rethinking Trade Preferences: How Africa Can Diversify Its Exports,World Economy Vol. 30 No. 8 pp. 132645.

    • Search Google Scholar
    • Export Citation

    DeatonA.1999. “Commodity Prices and Growth in Africa,Journal of Economic Perspectives Vol. 13 No. 3 pp. 2340.

    DufloE.2003Grandmothers and Granddaughters: Old-Age Pensions and Intrahousehold Allocation in South Africa,World Bank Economic Review Vol. 17 No. 1 pp. 125.

    • Search Google Scholar
    • Export Citation

    EasterlyW.R.Islam and J.Stiglitz2001Shaken and Stirred: Explaining Growth Volatility,in Annual World Bank Conference on Development Economicsed. byB.Pleskovic and N.Stern (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    European Commission2010European Report on Development 2010: Social Protection for Inclusive Development—A New Perspective in EU Co-operation with Africa (Brussels).

    • Search Google Scholar
    • Export Citation

    FurthS.2010Terms of Trade Volatility and Precautionary Savings in Developing Countries,Conference Paper No. c015_013 (Kiel, Germany: Dynamics, Economic Growth, and International Trade [DEGIT]).

    • Search Google Scholar
    • Export Citation

    HnatkovskaV. and N.Loayza2005Volatility and Growth,in Managing Economic Volatility and Crises: A Practitioner’s Guideed. byJ.Aizenman and B.Pinto (Cambridge, U.K.: Cambridge University Press) pp. 65100.

    • Search Google Scholar
    • Export Citation

    LoayzaN.R.RancièreL.Serven and J.Ventura2007Macroeconomic Volatility and Welfare in Developing Countries: An Introduction,World Bank Economic Review Vol. 21 No. 3 pp. 34357.

    • Search Google Scholar
    • Export Citation

    LoayzaN. and R.Soto2002The Sources of Economic Growth: An Overview,in Economic Growth: Sources Trends and Cyclesed. byN.Loayza and R.Soto (Santiago: Central Bank of Chile) pp. 140.

    • Search Google Scholar
    • Export Citation

    MartinP. and C.A. Rogers2000Long-Term Growth and Short-Term Economic Instability,European Economic Review Vol. 44 No. 2 pp. 35981.

    • Search Google Scholar
    • Export Citation

    McKinsey Global Institute2010Lions on the Move: The Progress and Potential of African Economies (New York).

    MendozaE. G.1997Terms-of-Trade Uncertainty and Economic Growth,Journal of Development Economics Vol. 54 No. 2 pp. 32356.

    • Search Google Scholar
    • Export Citation

    MorduchJ.2002Analysis of the Effects of Microfinance on Poverty Reduction” (unpublished; New York: New York University).

    RameyG. and V.Ramey1995Cross-Country Evidence on the Link between Volatility and Growth,American Economic Review Vol. 85 No. 5 pp. 113850.

    • Search Google Scholar
    • Export Citation

    RavallionM.2001Growth, Inequality and Poverty: Looking beyond Averages,World Development Vol. 29 No. 11 pp. 180315.

    SchubertB. and R.Slater2006Social Cash Transfers in Low-Income African Countries: Conditional or Unconditional?Development Policy Review Vol. 24 No. 5 pp. 57178.

    • Search Google Scholar
    • Export Citation

    SpataforaN. and I.Tytell2009Commodity Terms of Trade: The History of Booms and Busts,” IMF Working Paper 09/205 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    World Bank2010Global Monitoring Report 2010: The MDGs after the Crisis (Washington).

François Bourguignon is Director of the Paris School of Economics.
1There is a rather large literature on the growth impact of commodity prices in sub-Saharan Africa. See in particular Deaton (1999), Spatafora and Tytell (2009), and Collier and Goderis (2007) for the apparent contradiction between short-term and long-term effects of commodity prices.
2For a contrarian and more optimistic view, see McKinsey Global Institute (2010). This view about recent development policies in sub-Saharan countries is shared by other observers; see Arbache and Page (2010) and Collier and Goderis (2007) in particular.
3See for instance, Cogneau and Jedwab (2012) regarding the effect of the fall of cocoa prices in Côte d’Ivoire in the 1980s.
4This elasticity has the same order of magnitude as the estimates found in the literature. See in particular Hnatkovska and Loayza (2005).
5On the size, channels, and policy implications of volatility in developing countries, see, for instance, Loayza and others (2007). On the structural causes of volatility, see Easterly, Islam, and Stiglitz (2001).
6For a general analysis of social protection in sub-Saharan countries, see European Commission (2010); on the issue of the replicability of Latin American conditional cash transfer programs in Africa, see the early analysis of Schubert and Slater (2006).
7For an explicit optimal reserve accumulation policy in front of uncertainty on the terms of trade, see Barnichon (2009).
8The basic principles for the design of this facility were set in Cohen and others (2007).
9This argument was first put forward by Collier and Venables (2007).

    Other Resources Citing This Publication