52822 The World Bank notes January PREM Economic Policy 2010 NUMBER 147 Natural Resources and Development Strategy after the Crisis Milan Brahmbhatt and Otaviano Canuto Recent events have rekindled interest in the development strategy for commodity-de- role of primary commodities in develop- pendent, low-income countries (LICs).1 In ment. Was the boom in commodity prices this note we briefly review four questions: from around 2003 through 2008 just a cycli- How dependent are developing countries cal event, or does it suggest that prices have on primary commodity exports? What is entered on a period of secular strength, the outlook for primary commodity prices? driven by factors such as demand in big, fast- Is there a natural resource "curse" (or bless- growing developing countries like China? It ing)? What policies can help poor countries is notable that, while commodity prices fell best manage commodity resources for long- sharply from their peak in 2008 with the run development? onset of the global recession, they gener- ally remained much higher than previous How dependent are developing recession lows, often as high as in 2005­07, a countries on primary period of robust world growth. Furthermore, commodity exports? prices have also rebounded smartly over the If we view developing countries as a single course of 2009 (figures 1 and 2). aggregate then only about 40 percent of If a period of sustained commodity their merchandise exports were primary strength is imminent, what are the implica- commodities by value in 2005­07, down tions for development policies? Develop- from around 50 percent in the early 1990s. ment economists have long debated the This aggregate measure can be misleading, problems associated with the traditionally however, because it is dominated by a few big high specialization in production and export economies like China that are almost entirely of primary commodities of most developing exporters of manufactures. countries. Many argue that dependence on A different picture emerges if we take a primary commodities has proved to be a simple average across developing countries poisoned chalice or curse for development, (that is, giving each country an equal weight). which, given this view, necessarily entails Commodities still comprised a little over 60 structural change and rapid industrializa- percent of the merchandise exports of the av- tion. Others, however, suggest that sustained erage developing country in the middle part high commodity prices could reduce the of this decade, although this was down from relevance of an industrialization-focused over 90 percent in the late 1960s. Looking at FROM THE POVERTY REDUCTION AND ECONOMIC MANAGEMENT NETWORK Figure 1. Commodity Price Indexes, January Figure 2. Commodity Price Indexes, January 2000 to November 2009 (US$ 2000 = 100) 2000 to November 2009 (US$ 2000 = 100) 500 400 Food Oil 400 Agr. raw materials 300 Metals & minerals Non-energy 300 Index Index 200 200 100 100 0 0 00 1 2 03 ov 04 5 07 8 9 00 M 001 2 03 N 004 5 07 M 008 9 00 00 00 00 00 00 00 00 20 20 20 20 20 20 20 .2 .2 .2 .2 .2 .2 .2 2 .2 .2 .2 n. l. p. n. n. l. p. n. ay ay ar ar ay ov ay ar ar Ju Ju Ja Se Ja Ja Se Ja M M M M M N M Source: World Bank data and staff estimates Source: World Bank data and staff estimates the median, half of developing countries still Based on econometric study of long have commodity export dependence of over time series, the present consensus appears 70 percent. Among LICs, commodity export to be that real commodity prices do not dependence averages around 75 percent. display any permanent trend or drift over Viewed by regions, Africa, Latin America time. Figure 4 shows the Grilli and Yang and the Caribbean, and the Middle East time series of real non-energy commodity and North Africa are the most commodity prices (updated by other researchers) for dependent, while South Asia, East Asia, and the period 1900­2008. (Grilli and Yang, Europe and Central Asia are the least (fig- 1988; Pfaffenzeller et al., 2007). The series ure 3). So, although declining, commodity is a weighted index of the nominal prices or natural resource dependence remains of 24 non-energy commodities, divided by a fact of life for a majority of developing an index of the unit values of manufactured countries. Figure 3. Developing Countries: Commodity Exports Share (%) in 2003­07 What is the outlook for primary commodity prices? 80 In the 1950s the famous Prebisch-Singer thesis argued that real primary commodity 70 prices (for example, relative to manufactures 60 prices) displayed a long-run declining trend. Percent Faced with a resulting steady decline in their 50 terms of trade, developing countries should foster industrialization, following, accord- 40 ing to the thinking of the time, an import 30 substitution strategy. During the commodity price spike of the 1970s, on the other hand, 20 AFR LAC MNA All ECA EAP SAS many analysts argued that permanent natu- ral resource scarcity would result in steadily Source: World Development Indicators, World Bank. *As % of merchandise exports. Simple averages for country rising real commodity prices. groups. 2 PREMNOTE JANUARY 2010 Figure 4. Real Non Energy Commodity positive or negative. We find essentially the Prices, 1900­2015* same results for the Grilli-Yang series over 350 the period 1900­2008. So, based on statistical properties alone, 300 we have little reason to expect real commod- ity prices to trend either up or down in the Index (2000 = 100) 250 long run. It is a feature of unit root processes, 200 however, that series with this property are highly correlated over time. So it is quite 150 possible for commodity prices to move significantly lower or higher for substantial 100 Historical 1900­2008 periods even in the absence of a long-run Forecast 2009­2015 trend or drift, such as, for example, the long 50 period of unusually low prices from the mid- 05 15 25 35 45 55 65 75 85 95 05 15 19 19 19 19 19 19 19 19 19 19 20 20 1980s through the 1990s. Again, based on Sources: Grilli and Yang (1988); Pfaffenzeller et al. (2007). World Bank estimates 2004­08, forecasts 2009­15. statistical properties alone, one would not * Indexes, 2000 = 100. Deflated by unit value of be surprised to see a sustained period of manufactured exports. high prices following the low prices of the goods exported from developed to develop- 1980s and 1990s. ing countries. Visual inspection of figure 4 Are there plausible fundamental eco- suggests a definite downward trend, and this nomic factors to support such an outlook? appears to be confirmed by regression of the The price of commodities relative to the log of the Grilli-Yang series on a determin- price of manufactures can be usefully ana- istic time trend (modeling the error process lyzed in terms of demand and supply: the as a first-order AR1 process) over the whole demand for primary commodities relative to period 1900­2008, which yields an estimate the demand for manufactures, and the sup- that real commodity prices fall on average by ply of commodities relative to the supply of 0.5 percent per year, apparently confirming manufactures. the Prebisch-Singer hypothesis. On the supply side, if long-term pro- However, it is now well understood that ductivity growth in agriculture and minerals attempts to assess long-run trends on the is less than in manufacturing then, other basis of visual inspection and simple time things equal, one would expect agricultural series models can be misleading, espe- and mineral prices to rise relative to those cially if the series in question are so-called of manufactures. But there is little evidence unit root processes. In this case processes to suggest that productivity growth in com- without any deterministic trend can yield modities sectors is significantly different apparently significant but actually spuri- from that in manufactures, so this is unlikely ous regression results. Cuddington et al. to influence relative prices either way (World (2007) carefully survey econometric stud- Bank, 2009). It is true, however, that invest- ies of the Grilli-Yang series through 1998. ment in new capacity in energy and minerals Their overall conclusion is that, although was cut substantially when prices were low there is clear evidence of a structural break in the 1980s and 1990s and is recovering in 1921, it is not possible to reject the unit only slowly due to skill shortages, techni- root hypothesis for real commodity prices. cal difficulties in developing new reserves There is also no evidence of drift, either (for example, deep offshore), and political JANUARY 2010 PREMNOTE uncertainty in regions with new reserves. scenario, projecting only a gradual easing in Bio-fuel subsidies have also helped switch real commodity prices from existing levels by grain acreage away from food to fuel use, 2015. Forecast real prices in this period are providing a major reason for the steep grain in fact squarely in the range that prevailed price hikes from 2005 through the early from the 1920s through the early 1980s (fig- part of 2008. Over the longer term, though, ure 4). If correct, this means that commodity one would expect a more copious supply exporters are likely to face a more benign response, as skill shortages and technical medium-term price environment than in the difficulties are overcome, and new reserves 1980s and 1990s. and acreage are brought into production. Relative demand for commodities could Is there a natural resource also rise in the medium term to the extent "curse" (or blessing)? world growth after the financial crisis is more The short answer is "no" or rather "it de- dependent on developing countries and de- pends." A survey of the large and rapidly mand in these countries is more commodity growing empirical research in this area sug- intensive than elsewhere. In the longer term, gests that, in the words of a recent World however, production processes in develop- Bank report, natural resources are "neither ing countries will continue to become more curse nor destiny" (Lederman and Maloney, efficient in terms of raw material consump- 2007). Studies of the relationship between tion, approaching closer to developed- natural resource abundance and growth country levels, while relative final demand have, however, often tended to generate dis- for commodities like food will continue to parate and sometimes contradictory results. decline due to low income elasticity relative The influential study by Sachs and Warner to things like services. There is also evidence (1995) is representative of results which find that real commodity prices are affected by that natural resource abundance has a strong monetary conditions (Frankel, 2008). Since negative impact on growth. Lederman and commodities are traded in flexible price Maloney (2007), on the other hand, chal- markets, their prices tend to overshoot in lenge the Sachs and Warner findings on response to monetary changes relative to measurement and econometric grounds and general manufactures and services prices, find natural resource abundance to have a which adjust more sluggishly. Commodity positive effect on growth. prices will tend to be high when real interest A recent effort to reconcile such appar- rates are low and monetary conditions lax, ently disparate research findings (Collier and as at present, since inventory carrying costs Goderis, 2007) observes that, first, negative are low and there is more incentive to leave long-run growth effects are mostly related depletable natural resources in the ground. to oil and minerals --concentrated "point In the longer term, however, general price source" resources that can easily become levels and real interest rates can be expected the object of rent-seeking and redistributive to rise, removing the overshooting in real struggles (including armed conflict). On commodity prices. the other hand, there is little evidence of So there are both supply and demand negative growth effects related to high prices factors that could support the present rela- for agricultural commodities, which are tively high level of real commodity prices in generally more open to competitive entry. the medium term, although these factors will Second, high oil and mineral prices mostly tend to dissipate in the longer term. Current have a negative impact on long-run growth World Bank forecasts are consistent with this in exporting countries with bad governance. 4 PREMNOTE JANUARY 2010 They have a significant positive impact on goods (as well as imports), leading to higher growth in exporters with good governance. prices and output in the non-tradables sec- This finding suggests that continued high tor. Wages in the economy also tend to rise, commodity prices in the next few years squeezing profits in sectors of the economy could provide valuable resources to accel- that are internationally tradable but which erate economic and social development in are not based on natural resources, such commodity exporting countries with good as manufacturing, where prices are largely policies and governance. fixed at international levels. With increased There are several considerations to inflation in non-tradables prices there is keep in mind when evaluating the ways in an appreciation of the real exchange rate which natural resource abundance can lead and an output contraction in non-resource to worse economic performance, especially tradables sectors like manufacturing. These under conditions of poor governance. adjustments are of concern if one believes First, because of political economy rea- that sectors like manufacturing have some sons, countries with weak governance are special characteristics that stimulate long- more likely to adopt poor economic policies run growth, for example increasing returns to manage commodity booms, contributing to scale, learning by doing, or abundant to significant misallocation and mismanage- technological spillovers. Evidence that ment of resources. For example politicians manufacturing possesses these special char- may expand public spending and employ- acteristics is mixed, but there is fairly robust ment excessively and too rapidly, with the evidence for the more general proposition aim of increasing their patronage networks of a negative relation between real exchange and improving their chances of staying in rate overvaluation and growth (Rodrik, power, while resources shift out of produc- 2007; Aguirre and Calderon, 2005.) tive activity into unproductive rent-seeking There are also problems because vola- activity (Mehlum et al., 2006; Robinson et tility of primary commodity prices and al., 2006.) Poor fiscal policy indeed appears revenues can drive volatility in government to be at the heart of economic mismanage- spending and real exchange rates, with the ment in the wake of natural resource booms. resulting uncertainty damaging investment Studying natural resource boom episodes in and growth. Another related way in which the 1970s and 1980s Gelb (1988) concluded commodity price volatility may affect growth that "the most important recommendation is by fostering over-borrowing. High com- to emerge from this study is that spending modity prices in the 1970s encouraged many levels should have been adjusted to sharp resource-abundant countries to borrow rises in income levels more cautiously than heavily from abroad, to finance large invest- they actually were."2 ment projects and high public consumption. Second, natural resource booms create When prices plunged in the 1980s these complicated problems in macroeconomic countries were left with balance of payments management that are challenging even crises and unsustainable external debt levels. in economies with good governance and Again, it is critical to note that the actual capable institutions, and much more so in extent of Dutch Disease effects, volatility, economies without these advantages. One and over-borrowing will depend to a large of these problems is the so-called Dutch extent on policies--for example, on the ex- Disease effect: increased domestic income tent to which cautious fiscal policies are able from the booming natural resource sector to moderate aggregate demand pressures, generates higher spending on domestic JANUARY 2010 PREMNOTE smooth volatility in government revenues, What policies can help poor and curb external over-borrowing. countries best manage Lastly, in addition to problems of commodity resources short-run economic management, natural for development? resource­abundant countries also face First, given the evidence that problems with important longer-run questions about the governance are at the root of economic optimal pace at which to deplete their re- problems associated with natural resource sources and about how much of the proceeds abundance, efforts to enhance transpar- to consume today and how much to save for ency and strengthen checks and balances the welfare of future generations. An impor- concerning all aspects of natural resource tant metric here is whether the country's extraction and use are clearly vital for ensur- economic strategy is sustainable, meaning ing accountability. These aspects include the one that transfers sufficient capital to fu- terms of contracts with companies engaged ture generations to allow them to achieve in resource extraction or operation, ongoing at least the same level of welfare as current monitoring of operations, and the collec- generations.3 From this perspective natural tion and use of government taxes and other resources can be viewed as part of a country's revenues from natural resources. Broad overall capital stock, alongside its physical global efforts like the Extractive Industries capital stock (such as existing machinery and Transparency Initiative can play a part, as, buildings) and intangible capital (including at the domestic level, can anti-corruption human capital, social capital, and other fac- reforms, measures to improve transparency tors such as the quality of its institutions). To and scrutiny by civil society and media, pro- increase its overall capital stock, a country's curement reforms, strengthening of formal investment in its physical, human, and other audit, parliamentary scrutiny, and so on. capital must be larger than the depreciation Equitable sharing of benefits across regions, of its physical and other capital, including ethnic groups, and so forth can help reduce the depletion of its natural resources. This the danger of civil strife over resources. measure of countries' adjusted net savings rates is shown on the vertical axis Figure 5. Natural Resource Depletion and of figure 5. The horizontal Net Savings Rates, 2007 axis shows countries' annu- al depletion of their natural 60 resources (principally oil Adjusted net savings rate (% of GNI) and minerals, together with 40 a measure of forest deple- tion). Figure 5 suggests that y = ­0.579x + 11.921 countries with high rates of 20 R˛ = 0.3317 natural resource depletion are often on unsustainable 0 development paths: they 0 10 20 30 40 50 60 are not saving enough to ­20 cover the depletion of their natural resources, resulting ­40 in negative adjusted net Depletion of natural resources (% of GNI) savings rates. Source: World Development Indicators, World Bank. PREMNOTE JANUARY 2010 An institutional innovation that has at- ral resources on the country's longer-run tracted much recent attention is the use a growth and poverty reduction efforts. Figure separate (extra-budgetary) Natural Resource 6 provides a schematic of basic choices open Fund (NRF) to facilitate good management to the government, for example whether of revenues. However, experience suggests to return revenues to private citizens (via that the establishment of such funds is no tax cuts or transfers, which will then be substitute for sound overall fiscal and eco- reflected in increased private consumption nomic management, although in certain and investment), or to retain resource rev- circumstances it may buttress the right policy enues in public hands, which then need to mix. While Natural Resource Funds are be allocated between public consumption sometimes created to protect resource rev- and various kinds of public investment (or enue from political pressure and potential net asset accumulation). waste and corruption, and this argument has At a very general level these decisions its merits, an NRF of itself will not prevent need to be guided by a comparison of the such waste and abuse unless it is part of a government's social discount rate (which broader effort to strengthen governance measures the value it puts on consumption and integrate the fund with an overall fiscal today versus consumption at later dates) policy framework. with the rates of return available on various Second, attention also needs to be paid kinds of investment, for example the return to the actual substance of economic policy on foreign assets, the return from reducing decisions about the allocation of natural re- foreign debt (not generally the same thing source revenues between consumption and in developing countries), and returns to savings of various kinds. These decisions will domestic public investments. A commonly help determine how well the country is able used benchmark for fiscal policy in a natural to handle the macro management problems resource­rich economy is the permanent in- associated with natural resource abundance, come rule. Under this rule the country should such as the Dutch Disease and commodity save all resource revenues over and above a price volatility, as well as the impact of natu- certain permanently sustainable increase in Figure 6: Government choices in allocating resource revenues Allocate to Private consumption private sector via tax cuts, transfers, etc. Private investment Natural resource revenues Public consumption Public real domestic investment Allocate to Public Domestic lending or public sector investment reduce domestic debt Buy foreign assets or reduce foreign debt JANUARY 2010 PREMNOTE 7 the level of consumption, which is equal to scarce public infrastructure and to improve the annuity value of the country's natural the investment climate so as to raise returns resource wealth.4 In practice the rule often on private investment. Under these circum- leads to a recommendation to establish a stances a more optimal strategy would be to Natural Resource or Sovereign Wealth Fund devote a larger portion of resource revenues that invests in foreign assets, the returns to high-return public domestic investments, from which can support spending on the leading to higher growth and, ultimately, a government's non-natural-resource fiscal higher value of consumption than under the budget. permanent income strategy. It is worth noting that the permanent Evidently, much of the success of a income approach addresses several of the strategy oriented more toward domestic key issues associated with natural resource investment will depend on how efficiently fiscal management. It is by definition a sus- public investment funds can be allocated tainable policy in that converts a temporary, and managed to achieve high returns in exhaustible stock of natural resources into practice. So, thirdly, reforms to strengthen a stock of financial assets that generates a public investment management, cost benefit permanent income stream. Since the policy analysis, monitoring and evaluation, and calls for saving a substantial proportion of budget processes and institutions provide natural resource revenues, it reduces the another crucial element of a successful pressure of rising domestic demand that natural resource­based development strat- leads to real exchange rate appreciation egy. To the extent that it will take time to and Dutch Disease effects. By smoothing develop a pipeline of good projects and to expenditures, the policy also moderates strengthen public investment management the problems caused by volatility in natural capacity, it may be prudent for the country prices and revenues. to initially continue to invest most of its rev- There is nevertheless something anoma- enues in foreign assets, but to then increase lous about viewing the permanent income the proportion invested domestically, in rule as a long-run development strategy, line with domestic absorptive capacity. The with poor capital-scarce countries financ- country will also likely continue to flexibly ing investments in rich countries through shift resources into or out of its natural sovereign wealth funds. Several analysts have resources fund, depending on the need to recently argued that the permanent income address price volatility and Dutch Disease rule is optimal only under special circum- type pressures. However, investment climate stances that do not apply to most developing reforms, support for innovation, and high- countries; essentially these conditions are return domestic infrastructure investments the ability to freely borrow and lend at the can also help alleviate Dutch Disease pres- world rate of interest, which would result in sures by increasing the supply capacity of foreign and domestic rates of return becom- the economy. ing aligned (Collier and Venables, 2008; Van We conclude that booming commodity der Ploeg and Venables, 2009). Most devel- revenues raise difficult challenges that, if not oping countries, however, are characterized adequately addressed, can harm long-run by restricted access to world capital markets, development. However, with good policies, capital scarcity, and potentially high rates of governance, and management, such rev- return on domestic investment, especially if enues can also be a valuable resource that the government is able to efficiently supply PREMNOTE JANUARY 2010 helps accelerate overall economic and social Africa Economic Research Consortium development. 2008 Annual Conference. Cuddington, John T., Rodney Ludema, and Notes Shamila A. Jayasuriya. 2007. "Prebisch- 1. See, for example, Oxford Analytica Inter- Singer Redux." In Daniel Lederman and national, 2009, "Commodities Force Re-think William F. Maloney, eds., Natural Re- on Growth," August 18. sources: Neither Curse Nor Destiny. Stanford 2. A stronger version of the political University Press and the World Bank. economy channel argues that natural resource Davis, J.M., R. Ossowsky, and A. Fedilino, booms can even lead to a worsening of gov- eds. 2003. Fiscal Policy Formulation and ernance, for example a "voracity effect" as Implementation in Oil Producing Countries. political actors race to seize and spend natural Washington, DC: IMF. resource revenues before others do, provok- Frankel, Jeffrey. 2008. "The Effect of Mon- ing more intense political, bureaucratic, and etary Policy on Real Commodity Prices." even violent conflicts for control of natural In John Campbell, ed., Asset Prices and resource revenues (Tornell and Lane, 1999). Monetary Policy. University of Chicago The evidence for this hypothesis is mixed. Press. 3. Heal (1996) discusses alternative inter- Gelb, Alan and Associates. 1988. Oil Wind- pretations of sustainability. falls: Blessing or Curse? World Bank and 4. The permanent income approach to Oxford University Press. fiscal policy in natural resource­abundant Grilli, Enzo R., and M.C. Yang. 1988. "Pri- economies is studied in more detail in Van mary Commodity Prices, Manufactured Wijnbergen (2008), Davis et al. (2003), and Goods Prices, and the Terms of Trade of Barnett and Ossowski (2002). Developing Countries: What the Long Run Shows." World Bank Economic Review References 2(1): 1­47. Aguirre, Alvaro, and Cesar Calderon. 2005. Heal, Geoffrey.1996. "Interpreting Sustain- "Real Exchange Rate Misalignments and ability." Paine Webber Working Paper Economic Performance." April. Central Series in Money, Economics and Finance Bank of Chile Economic Research Divi- PW-95-24. May. Columbia Business sion. School, New York, NY. Barnett, Steven, and Rolando Ossowski. Lederman, Daniel, and William F. Maloney, 2002. "Operational Aspects of Fiscal eds. 2007. Natural Resources: Neither Curse Policy in Oil-Producing Countries." IMF nor Destiny. World Bank and Stanford Working Paper WP/02/177. IMF, Wash- University Press. ington, DC. Mehlum, Halvor, Karl Moene, and Ragn- Collier, Paul, and Benedikt Goderis. 2007. arTorvik. 2006. "Institutions and the "Commodity Prices, Growth and the Resource Curse." Economic Journal Natural Resources Curse: Reconciling 116(508): 1­20. a Conundrum." August. Working Paper Pfaffenzeller, Stephan, Paul Newbold, and 276. Centre for the Study of African Anthony Rayner. 2007. "A Short Note Economies. on Updating the Grilli and Yang Com- Collier, Paul, and Anthony Venables. 2008. modity Price Index." World Bank Economic "Managing Resource Revenues: Lessons Review 21(1): 151­163. for Low Income Countries." Paper for the JANUARY 2010 PREMNOTE Robinson, James A., Ragnar Torvik, and Tornell, Aaron, and Philip R. Lane. 1999. Thierry Verdier. "Political Foundations of "The Voracity Effect." American Economic the Resource Curse." Journal of Develop- Review 89(1): 22­46. ment Economics 79(2): 447­68. Van Der Ploeg, Frederick, and Anthony Rodrik, Dani. 2007. "The Real Exchange Venables. 2009. "Harnessing Windfall Rate and Economic Growth: Theory Revenues: Optimal Policies for Resource- and Evidence." August. John F. Kennedy Rich Developing Economies." Working School of Government, Harvard Univer- Paper No. 2571. CESifo Group, Munich, sity, Cambridge, MA. Germany. Sachs, Jeffrey D., and Andrew M. Warner. Van Wijnberger, Sweder. 2008. "The Perma- 1995. "Natural Resource Abundance and nent Income Approach in Practice." Un- Economic Growth." NBER Working Pa- published paper. March. PREM Network, per 5398. National Bureau of Economic World Bank, Washington, DC. Research, Cambridge, MA. World Bank. 2009. Global Economic Prospects and the Developing Countries. Washington, DC: World Bank. This note series is intended to summarize good practices and key policy findings on PREM-related topics. The views expressed in the notes are those of the authors and do not necessarily reflect those of the World Bank. PREMnotes are widely distributed to Bank staff and are also available on the PREM Web site (http://www.worldbank.org/prem). If you are interested in writing a PREMnote, email your idea to Madjiguene Seck at mseck@worldbank.org. For additional copies of this PREMnote please contact the PREM Advisory Service at x87736. PREMnotes are edited and laid out by Grammarians, Inc. Prepared for World Bank staff