International Monetary Fund, April 26, 2011

1. The Natural Resource Curse I:Pitfalls of Commodity WealthJeffrey FrankelHarpel Professor of Capital Formation & GrowthHarvard University International Monetary Fund, April 26, 2011

2. The Natural Resource Curse • The NRC pertains especially to oil & minerals, but sometimes to timber & agricultural products too. • Seminal references: • Auty (1990, 2001, 07, 09) • Sachs & Warner (1995, 2001) • Other studies find a negative effect of oil in particular, on economic performance: • including Kaldor, Karl & Said (2007); Ross (2001); Sala-i-Martin & Subramanian (2003); and Smith (2004). • Frankel, “The Natural Resource Curse: Survey,” • NBER Working Paper 15836, 2010. • forthcoming in Export Perils, • edited by B.Shaffer (U. of Pennsylvania Press: 2011)

3. Examples • Conspicuously high in oil resources and low in growth: Venezuela & Gabon. • Conspicuously high in growth and low in natural resources: China & other Asian countries. • The overall relationshipon average is slightly negative:

4. Growth falls with fuel&mineral exports

5. Are natural resources necessarily bad? No, of course not. • Commodity wealth neednot necessarily lead to inferior economic or political development. • Rather, it is a double-edged sword, with both benefits and dangers. • It can be used for ill as easily as for good. • The priority for any country should be on identifying ways to sidestep the pitfalls that have afflicted other mineral producers in the past, to find the path of success.

6. The goal is to enjoy the success of • Chile, vs. Bolivia • Botswana, vs. Congo • Norway, vs. Sudan. • The last section of my paper explores some of the policies & institutional innovations that might help avoid the natural resource curse and achieve natural resource blessings instead.

7. How could abundance of commodity wealth be a curse? • What is the mechanism for this counter-intuitive relationship? • At least 7 channels have been suggested:

8. 7 Possible Natural Resource Curse Channels • Downward price trend • Price volatility • Crowding-out manufacturing • Inhibited development of institutions • Unsustainably rapid depletion as a result of unenforceable property rights • Proclivity for armed conflict • The Dutch Disease

9. The 7 NRC Channels Elaborated • World commodity price trendcould be downward (Prebisch-Singer); • High volatility of oil prices could be problematic ; • Natural resources could be dead-end sectors (Matsuyama):they may crowd out manufacturing, • which may be the home of dynamic benefits & spillovers. • “Industrialization” could be the essence of development.

10. The 7 NRC Channelscontinued 4. Countries where physical command of mineral deposits by the government or a hereditary elite automatically confers wealth on the holders may be less likely to develop the institutions that are conducive to economic development (Engerman-Sokoloff …), • e.g., rule of law & decentralization of decision-making, • as compared to countries where moderate taxation of a thriving market economy is the only way to finance government.

11. The 7 NRC Channelscontinued 5. Non-renewableresources are depleted too fast, where it is difficult to enforce property rights,as under frontier conditions. 6. Countries that are endowed with minerals may have a proclivity for armed conflict, which is inimical to economic growth. 7. Swings in commodity prices can engender macroeconomic instability(Dutch Disease), via the real exchange rate and government spending.

12. (7)The Dutch Disease and Procyclicality • Developing countries have historically been prone to procyclicality: • Especially procyclical government spending • “Procyclical” means destabilizing. • This is particularly true of commodity producers. • The Dutch Disease describes unwanted side-effects from a strong, but perhaps temporary, upward swing in the world price of the export commodity.

13. Procyclicality Volatility in developing countries • arises both from foreign shocks, • including export commodity price fluctuations, • and from domestic shocks • including macroeconomic & political instability.

14. Procyclicality Volatility in developing countries • Most developing countries in the 1990s brought under control the chronic runaway budget deficits, money creation, & inflation, that they experienced, • but many still showed monetary & fiscal policy that was procyclical rather than countercyclical: • They tend to be expansionary in booms • and contractionary in recessions, • thereby exacerbating the magnitudes of the swings. • The aim should be to moderate swings • -- the countercyclical pattern that economists, after the Great Depression, originally hoped discretionary policy would take.

15. Procyclicality in developing countries The procyclicality of fiscal policy • Many authors have shown that fiscal policy has tended to be procyclical in developing countries, • especially in comparison with industrialized countries.[1] • A major reason for procyclical public spending: receipts from taxes or royalties rise in booms.The government cannot resist the temptation or political pressure to increase spending proportionately, or more. [1] Cuddington (1989), Tornell & Lane (1999), Kaminsky, Reinhart, & Vegh (2004), Talvi & Végh (2005), Alesina, Campante & Tabellini (2008), Mendoza & Oviedo (2006), Ilzetski & Vegh (2008), Medas & Zakharova (2009) and Gavin & Perotti (1997).

16. Procyclicality in developing countries The procyclicality of fiscal policy, continued • Procyclicality is especially pronounced in countries with natural resources and where income from those resources tends to dominate the business cycle. • Cuddington (1989) and Sinnott (2009) • An important recent development: some developing countries, including commodity producers, have been able to break the historic pattern in the most recent cycle: • Taking advantage of the boom of 2002-2008 • to run budget surpluses & build reserves, • thereby earning the ability to expand fiscally in the 2008-09 crisis. • Chile is the outstanding model.

17. (i) Public investment projects • Two large budget items that account for much of the increased spending from oil booms: • (i) investment projects and • (ii) the government wage bill. • Regarding the 1st budget item, investment in infrastructure can have large long-term pay-off if it is well designed; too often in practice, however, it takes the form of white elephant projects, which are stranded without funds for completion or maintenance, when the oil price goes back down. • Gelb (1986) .

18. (ii) Public sector wage bills • Regarding the 2nd budget item, oil windfalls have often been spent on higher public sector wages -- Medas & Zakharova(2009) . • They can also go to increasing the number of workers employed by the government. • Either way, they raise the total public sector wage bill, which is hard to reverse when oil prices go back down. • Figures 2 & 3 plot the public sector wage bill, for two oil producers, Iran & Indonesia. • against primary product prices over the preceding 3 years.

19. Iran’s Government Wage Bill Is Influenced by Oil Prices Over Preceding 3 Years(1974, 1977-1997.) Source: Frankel (2005b)

20. Indonesia’s Government Wage Bill Is Influenced by Oil Prices Over Preceding 3 Years(1974, 1977-1997.) Source: Frankel (2005b)

21. Public sector wage bills,continued • There is a clear positive relationship. • That the relationship is strong with a 3-year lag shows the problem: oil prices may have fallen over 3 years, but public sector wages cannot easily be cut nor workers laid off. • Arezki & Ismail (2010) find that current government spending increases in boom times, but is downward-sticky.

22. The Dutch Disease: 5 side-effects of a commodity boom • 1) A real appreciation in the currency • 2) A rise in government spending • 3) A rise in nontraded goods prices • 4) A resultant shift of resources out of non-export-commodity traded goods • 5) A current account deficit

23. The Dutch Disease: The 5 effects elaborated • 1) A real appreciation in the currency • taking the form of nominal currency appreciation if the exchange rate floats • e.g., floating-rate oil exporters • Kazakhstan, Mexico, Norway, & Russia. • or the form of money inflows & inflation if the exchange rate is fixed [1] ; • e.g. fixed-rate oil-exporters, the UAE & Saudi Arabia. • 2) A rise in government spending • in response to increased availability of tax receipts or royalties.

24. The Dutch Disease: 5 side-effects of a commodity boom • 3) An increase in nontraded goods prices (goods & services such as housing that are not internationally traded), • relative to traded goods(manufactures & other internationally traded goods other than the export commodity). • 4) A resultant shift of resources out of non-export-commodity traded goods • pulled by the more attractive returns in the export commodity and in non-traded goods.

25. The Dutch Disease: 5 side-effects of a commodity boom • 5) A current account deficit • thereby incurring international debt that may be difficult to service when the boom ends [2]. • Most developing countries avoided it in 2003-10. • [2] Manzano & Rigobon (2008): the negative Sachs-Warner effect of resource dependence on growth rates during 1970-1990 was mediated through international debt incurred when commodity prices were high. • Arezki & Brückner (2010a): commodity price booms lead to increased government spending, external debt & default risk in autocracies. • Arezki & Brückner (2010b): the dichotomy extends also to effects on sovereign spreads paid by autocratic vs democratic commodity producers.

26. Summary: Channels of the NRC • (1) Commodity price volatility is high, imposing risk & costs. • (2) Specialization can crowd out the manufacturing sector. • (3) Depletion can be unsustainably rapid, • especially if property rights are not adequately protected. • (4) Mineral riches can lead to civil war. • (5) Mineral endowments can lead to poor institutions,such as corruption, inequality, class structure, chronic power struggles, and absence of rule of law and property rights. • (6) The Dutch Disease. A commodity boom: => real currency appreciation and increased government spending, => which expand nontraded sector and render uncompetitive non-commodity export sectors such as manufactures.

27. The Natural Resource Curse should not be interpreted as a rule that resource-rich countries are doomed to failure. • The question is what policies to adopt to improve the chances of prosperity. • Destruction or renunciation of resource endowments, to avoid dangers such as the corruption of leaders, will not be one of these policies. • The survey concludes with ideas for policies/institutions designed to address aspects of the resource curse and thereby increase the chance of economic success.

29. Appendices:1) The possible NRC channels in detail2) Procyclical capital flows3 cycles of flows to developing countries3) Skeptics of the NRC

30. Appendix 1: The possible NRC channels in detail (1)The claim of a negative trend in commodity prices on world markets was already dealt with: the data do not suggest a robust long-term trend, certainly not a negative one if updated to 2010. 30

31. (1) Long-term world price trend (i) Determination of the price on world markets (ii) The old “structuralist school” (Prebisch-Singer): The hypothesis of a declining commodity price trend (iii) Hypotheses of a rising price trend Hotelling Malthus (iv) Empirical evidence Statistical time series studies 31

32. (i) The determination of the export price on world markets Developing countries tend to be smaller economically than major industrialized countries, and more likely to specialize in the exports of basic commodities. As a result, they are more likely to fit the “small open economy” model: they can be regarded as price-takers, That is, the prices of their export goods are generally taken as given on world markets. 32

33. (ii) The old “structuralist school”Raul Prebisch (1950) & Hans Singer (1950) The hypothesis: a declining long run trend inprices of mineral & agricultural products relative to the prices of manufactured goods. The theoretical reasoning: world demand for primary products is inelastic with respect to world income. That is, for every 1 % increase in income, raw materials demand rises by less than 1%. Engel’s Law, an (older) proposition: households spend a lower fraction of their income on basic necessities as they get richer. Demand => P oil 33

34. (iii) Hypotheses of rising trendsHotelling on depletable resources;Malthus on geometric population growth. Persuasive theoretical arguments that we should expect oil prices to showan upward trend in the long run. 34

35. Assumptions for Hotelling model (1) Non-perishable non-renewable resources: Deposits in the earth’s crust are fixed in total supply and are gradually being depleted. (2) Secure property rights: Whoever currently has claim to the resource can be confident that it will retain possession, unless it sells to someone else, who then has equally safe property rights. This assumption excludes cases where warlords compete over physical possession of the resource. It also excludes cases where private mining companies fear that their contracts might be abrogated or their holdings nationalized. 35

36. One more assumption, to keep the Hotelling model simple: (3) The fixed deposits are easily accessible: the costs of exploration & extraction are small compared to the value of the mineral. Hotelling (1931) deduced from these assumptions the theoretical principle: the price of oil in the long run should rise at a rate equal to the interest rate. 36

37. The Hotelling logic: The owner chooses how much mineral to extract and how much to leave in the ground. Whatever is mined can be sold at today’s price (price-taker assumption) and the proceeds invested in bank deposits or US Treasury bills, which earn the current interest rate. If the value of the commodity in the ground is not expected to rise in the future, then the owner has an incentive to extract more of it today, so that he earns interest on the proceeds. 37

38. The Hotelling logic,continued: As minng companies worldwide react in this way, they drive down the price today, below its perceived long-run level. When the current price is below its long-run level, companies will expect the price to rise in the future. Only when the expectation of future appreciation is sufficient to offset the interest rate will the commodity market be in equilibrium. Only then will mining companies be close to indifferent between extracting at a faster rate and a slower rate. 38

39. The complication: supply is not fixed. True, at any point in time there is a certain stock of reserves that have been discovered. But the historical pattern has long been that, as that stock is depleted, new reserves are found. When the price goes up, it makes exploration & development profitable for deposits farther under the surface. …especially as new technologies are developed for exploration & extraction. 39

40. What is the overall statistical trend in commodity prices in the long run? Some authors find a slight upward trend, some a slight downward trend.[1] The answer seems to depend, more than anything else, on the date of the end of the sample: Studies written after the 1970s boom found an upward trend, but those written after the 1980s found a downward trend, even when both went back to the early 20th century. [1] Cuddington (1992), Cuddington, Ludema & Jayasuriya (2007), Cuddington & Urzua (1989), Grilli & Yang (1988), Pindyck (1999), Hadass & Williamson (2003), Reinhart & Wickham (1994), Kellard & Wohar (2005), Balagtas & Holt (2009) and Harvey, Kellard, Madsen & Wohar (2010). 40

41. (2) Effects of Volatility Is volatility per se bad for economic growth? Cyclical shifts of resources back & forth across sectors may incur needless transaction costs. A diversified country may indeed be betterthan one 100% specialized in minerals. On the other hand, the private sector dislikes risk as much as the government does, and will take steps to mitigate it; thus one must think where the market failure lies before assuming that a policy of deliberate diversification is necessarily justified. 41

42. Effects of volatility, continued Policy-makers may not be better than individual private agents at discerning whether a commodity boom is temporary or not. But the government cannot ignore the issue of volatility: When it comes to exchange rate or fiscal policy, governments must necessarily make judgments about the likely permanence of shocks. More on medium-term cycles when we get to the Dutch Disease 42

43. (3) Do natural resources crowd out manufacturing? Matsuyama (1992) provided an influential model: the manufacturing sector is assumed to be characterized by learning by doing, while the primary sector (agriculture, in his paper) is not. Also van Wijnbergen(1984)and Gylfason, Herbertsson & Zoega(1999). The implication: deliberate policy-induced diversification out of primary products into manufacturing is justified, and a permanent commodity boom that crowds out manufacturing can indeed be harmful. 43

44. Counterarguments There is no reason why learning by doing should occur only in manufacturing tradables. Nontradable sectors can enjoy learning by doing. [1] E.g., construction… The mineral sector can as well. The USA is one example of a country that has enjoyed big productivity growth in commodity sectors. Productivity gains have been aided by American public investment, since the late 19th century, in such knowledge infrastructure institutions as the U.S. Geological Survey, School of Mines, and Land-Grant Colleges. [2] [1] Torvik(2001) and Matsen & Torvik (2005). [2] Wright & Czelusta(2003, p.6, 25; 18-21). 44

45. Counterarguments, continued Public investment in knowledge infrastructure ≠government subsidy or ownership of the resources themselves. In Latin America, e.g., public monopoly ownership and prohibition on importing foreign expertise or capital has often stunted development of the mineral sector, whereas privatization has set it free. Attempts by governments to force linkages between the mineral sector and processing industries have often failed. 45

46. (4) Institutions Recent thinking in economic development: The quality of institutions is the deep fundamental factor that determines which countries experience good performance.[1] It is futile (e.g., for the IMF & World Bank)to recommend good macroeconomic or microeconomic policies if the institutional structure is not there to support them. [1] Barro (1991) and North (1994). 46

47. What are weak institutions? A typical list: inequality, corruption, insecure property rights, intermittent dictatorship, ineffective judiciary branch, and lack of any constraints to prevent elites & politicians from plundering the country. “Quality of institutions” has been quantified by World Bank, Freedom House, Transparency International, and others. Rodrik, Subramanian & Trebbi (2003) use a rule of law indicator and protection of property rights (taken from Kaufmann, Kraay & Zoido-Lobaton, 2002). Acemoglu, Johnson, & Robinson (2001) use a measure of expropriation risk to investors. Acemoglu, Johnson, Robinson, & Thaicharoen (2003) use the extent of constraints on the executive. 47

48. Institutions can be endogenous: the result of economic growth rather than the cause. The same problem is encountered with other proposed fundamental determinants of growth, e.g., openness to trade and freedom from tropical diseases. Many institutions tend to evolve endogenously, in response to the level of income, such as the structure of financial markets, mechanisms of income redistribution & social safety nets, tax systems, and intellectual property rules… 48

49. Addressing endogeneity of institutions statistically Econometricians address the problem of endogeneity by means of the technique of instrumental variables. What is a good instrumental variable for institutions, an exogenous determinant? Acemoglu, Johnson & Robinson (2001) introduced the mortality rates of colonial settlers. The theory is that, out of all the lands that Europeans colonized, only those where Europeans actually settled were given good European institutions. Acemoglu et al figured that initial settler mortality determined whether Europeans settled in large numbers.[1] [1] Glaeser, et al, (2004) argue against the settler variable. Hall & Jones (1999) consider latitude and the speaking of English or other European languages as proxies for European institutions. 49

50. Institutions: Econometric findings The finding is the same, regardless of IV: “Institutions trump everything else” – Rodrik et al (2002) Acemoglu et al (2002) Easterly & Levine (2002) Hall & Jones (1999) Geography and history matter mainly as determinants of institutions; which is not to say that institutions don’t also have other important determinants. In any case, institutions are important. 50