Commodity Booms and Busts - Gordon Rausser

ANNUAL REVIEWS

Further

Click here for quick links to Annual Reviews content online, including:

Annu. Rev. Resour. Econ. 2011.3:87-118. Downloaded from www.annualreviews.org by University of California - Berkeley on 10/31/11. For personal use only.

• Other articles in this volume • Top cited articles • Top downloaded articles • Our comprehensive search

Commodity Booms and Busts Colin A. Carter,1,2 Gordon C. Rausser,2,3 and Aaron Smith1,2 1 Department of Agricultural and Resource Economics, University of California, Davis, California 95616; email: [email protected], [email protected] 2

Giannini Foundation of Agricultural Economics, University of California

3

Department of Agricultural and Resource Economics, University of California, Berkeley, California 94720; email: [email protected]

Annu. Rev. Resour. Econ. 2011. 3:87–118

Keywords

First published online as a Review in Advance on June 13, 2011

commodity markets, asset bubbles, oil prices, food price crisis, speculation, inventories

The Annual Review of Resource Economics is online at resource.annualreviews.org This article’s doi: 10.1146/annurev.resource.012809.104220 Copyright © 2011 by Annual Reviews. All rights reserved JEL: E31, G12, G13, Q02, Q11 1941-1340/11/1010-0087$20.00

Abstract Periodically, the global economy experiences great commodity booms and busts, characterized by a broad and sharp comovement of commodity prices. There have been two such episodes since the Korean War. The first event peaked in 1974 and the second in 2008, 34 years apart. Both created major economic and political shocks, including fallen governments and human suffering due to high food prices. Each occurrence raised serious concerns over food and energy security and led to more government intervention in the commodity markets. Although there is no simple explanation for what causes such complex events, they do share similar characteristics. We find at the core of these cycles a set of contemporaneous supply and demand surprises that coincided with low inventories and that were magnified by macroeconomic shocks and policy responses. In the next few decades, the world faces the prospect of continued increases in the demand for commodities and greater uncertainty about supply. However, because market participants are likely to respond by increasing inventory holdings and investing in new technologies, we see no reason to expect an increase in the frequency of dramatic commodity booms and busts.

87

1. INTRODUCTION


Commodity markets occasionally exhibit broadly based massive booms and busts. These events affect the poor’s ability to purchase the most basic necessities such as food and energy, and they often cause political unrest. Prominent riots generated by commodity price spikes include the Peterloo Massacre in Manchester, England, in 1819; the Southern U.S. bread riots in 1863; and unrest in Haiti, West Africa, and South Asia in 2008. Commodity booms and busts thus resonate with the populace and affect social welfare in a way that other asset price spikes do not. Commodity booms and busts have the greatest economic and social impact on developing nations, where most of the world’s population resides. Agriculture accounts for a sizeable portion of economic activity in these countries, and households there spend a large share of disposable income on food commodities. In addition, booms and busts can have dire macroeconomic effects in developing countries because many of these economies are highly dependent on commodity trade. In rich countries, booms and busts in energy and industrial metal prices are more often salient than food price spikes. Most food in rich countries is heavily processed, so the price of the raw commodity makes up a small fraction of the retail price. In contrast, energy prices have large effects on the retail cost of transportation, heating, and cooling. Moreover, energy price spikes portend macroeconomic recessions (Hamilton 2009). Asset price booms and busts are not unique to commodity markets, although one of the most widely cited examples of a price boom followed by a large crash took place in a market for an agricultural commodity: the Dutch tulip mania of the 1630s. Other famous asset market booms and busts include the South Sea Company stock market crisis of 1719–1720, the great stock market crash of 1929–1932, the dot-com mania of 1999–2000, the crash following Japan’s asset price boom of 1986–1991, and the global real estate boom and bust of 2003–2008. The term bubble is often used to describe price booms and busts, especially in the popular press. Most economists define an asset bubble as a period when prices are driven by trader beliefs peripheral to underlying supply and demand factors. For example, Stiglitz (1990, p. 13) defined an asset price bubble as follows: “[I]f the reason that the price is high today is only because investors believe that the selling price will be high tomorrow—when “fundamental” factors do not seem to justify such a price—then a bubble exists.” Garber (1989, 1990) studies spot and futures prices for rare tulip bulbs during the Dutch tulip mania and finds that market fundamentals, and not irrational behavior, were the most important factor driving prices at that time. However, Garber also concludes that during the last month of the tulip bulb speculation, the rise and fall of common bulb prices were possibly a bubble. His conflicting findings for the two periods of tulip bulb prices are typical of the literature: Economists cannot easily distinguish bubbles from a major change in market fundamentals. Kindleberger (1978) finds common links across asset market booms and busts: Price peaks often occur after an exogenous shock that creates new incentives for participants to purchase assets. Debt accumulation accelerates the process, but then prices overshoot, and finally asset prices tumble. Although some investors may purchase commodities for speculative reasons, consumers and firms continue to demand physical commodities during commodity booms. These buyers are not speculating on the future value of the commodity, so they would not pay a price in excess of the marginal consumption value of the commodity. The only way to slide 88

Carter

Rausser

Smith


up the demand curve and raise the market price to this group of buyers is to reduce supply. A commodity bubble therefore implies that speculators hold some inventory off the market with the expectation that they can sell it at a higher price in the future, thereby reducing available supply and raising the current price (Hamilton 2009). It follows that stockholding behavior provides an important clue to explaining commodity booms and busts, along with other fundamental factors such as supply and demand shocks, macroeconomic shocks, and policy responses. Do we expect to see more or fewer commodity booms and busts in the coming years as the world economy evolves? The long history of commodity price booms and busts suggests they are inevitable. They occur in agrarian economies and industrial economies. The events of 2007–2008 suggest that the transition toward a knowledge-based economy (Romer 1986) has not reduced the world’s vulnerability to commodity booms and busts. In this article, we assess the likely size and frequency of future booms and busts, given economic changes such as continued globalization, population growth, urbanization, increased regional specialization of agricultural production and trade, biofuel demand, and climate change. We explore the economics of commodity booms and busts using as examples the two largest and most dramatic events since World War II:1 the booms and busts of 1973–1974 and 2007–2008. Broadly speaking, both occasions experienced similar and sharp upward movements in commodity prices and subsequent declines. We find no simple explanations for either event; each had at its core a sequence of supply and demand shocks that reduced inventories to low levels. Macroeconomic events, cross-commodity linkages, and policy responses with unintended consequences exacerbated these fundamental shocks. Viewed in this light, these two events are much more similar than different, and they provide a context to assess possible future booms and busts. The article proceeds as follows. In Section 2, we describe the 1973–1974 and 2007–2008 commodity booms and subsequent busts, and we characterize the magnitude of the price variability. Market structures across commodity groups are explored in Section 3, where we isolate the contributions of supply and demand differences across the various commodity systems. Section 4 outlines the important role of commodity stockholding, which normally serves to smooth price fluctuations. Section 5 assesses macroeconomic linkages to commodity prices, including the important role of exchange rates and interest rates. In Section 6 we examine the importance of general equilibrium or crosscommodity linkages, including links through factor substitution and input costs. Temporary policy responses to commodity booms are described in Section 7, where we explain why policies such as export quotas often aggravate the volatility of world commodity prices and thereby send the wrong price signals to domestic markets. Section 8 concludes the paper.

2. TWO MAJOR COMMODITY BOOMS AND BUSTS: 1973–1974 AND 2007–2008 China. . .is a big force in the extraordinary boom in commodities. Its voracious appetite for everything from corn and wheat to copper and oil has helped push up U.S. commodities prices by some 50% over the past 12 months. But China

1 There were smaller commodity booms during the Korean War and in 1979–1980, but we do not analyze these events in this paper.

www.annualreviews.org

Commodity Booms and Busts

89

is by no means the whole story. Speculators—including small investors—are also playing a huge role. . . . —Epstein (2008) [T]he commodity price boom is rooted in the Fed’s weak dollar policy, and not in a change in “relative prices” due to rising global demand. —Wall Street Journal (2008)


The above quotes from the financial press in the spring of 2008 summarize opposing views as to what caused the most recent commodity boom and bust. When commodity prices exploded in 2008, Barron’s financial magazine attributed the boom to global supply and demand fundamentals (see also IOSCO 2009), but at the same time the Wall Street Journal argued that the boom was due to low interest rates and a weak U.S. dollar (see also Frankel 2008b). Others argued that the boom constituted a speculative bubble (Khan 2009) and that index fund speculation played a key role (Baffes & Haniotis 2010, Gilbert 2010), arguments that were refuted convincingly by Irwin & Sanders (2011). Similarly opposing views were promoted during and after the 1973–1974 boom and bust. In this section, we briefly describe these two major events. Commodities are typically placed into several categories, depending on their physical characteristics and end use. These categories are energy (e.g., crude oil and natural gas), cereal grains (e.g., corn, wheat, and rice), vegetable oils (e.g., soybeans and palm oil), softs (e.g., sugar, coffee, cocoa, and cotton), metals (e.g., gold, silver, aluminum, and copper),2 and livestock (e.g., hogs and cattle). Figure 1 (see color insert) shows the real prices for each commodity category during the two boom-bust cycles. The 2008 price boom was characterized by price increases comparable to those in 1974. Crude oil prices increased approximately fourfold between 1972 and 1974 and tripled between 2007 and mid-2008. Prices of cereal grains (e.g., corn, rice, and wheat) more than tripled in the early 1970s before declining and then did almost the same thing from 2006–2008. Both of these boom-bust cycles exhibited a general sharp upward comovement in the prices of many commodities (especially food and energy), which calls for a common explanation. However, there are also some notable differences between the two episodes. For instance, agricultural commodities led the 1973–1974 commodity boom, but they moved concurrently with energy prices in 2007–2008. Cereal grains, vegetable oils, and energy accounted for most of the 1973–1974 commodity price spike, whereas in 2007–2008 metals joined these three groups to create four price leaders. In fact, most metals peaked earlier than the other commodities in 2007. The soft commodities (such as coffee, sugar, cocoa, and cotton) played a much smaller role in the 2007–2008 commodity boom compared with their huge price spike in the early 1970s.3 The livestock index exhibited an 80% increase in 1973–1974 but very little change in 2007–2008. Finally, the bust was much faster and more coordinated in 2008 than that following the 1973–1974 boom. The emergence of the global financial crisis in September 2008 and the associated macroeconomic slowdown were the catalysts for the bust; no such event occurred in 1974. 2

Other categorizations exist. For example, metals are sometimes divided into precious metals (e.g., gold and silver) and into industrial or base metals (e.g., aluminum and copper).

3

The November 1974 spike in softs was driven mostly by sugar. The other soft commodities exhibited mild booms in both episodes.

90

Carter

Rausser

Smith


The high price of crude oil was the poster child of the market frenzy in 1973 and again in 2008. In October 1973, OPEC imposed an oil embargo on the United States and on many parts of Europe in response to several countries’ support of Israel during the Yom Kippur War. Coupled with price controls imposed by President Nixon, the embargo led to gasoline shortages in the United States, which prompted long lines at gas stations and consumer violence (Frum 2000, p. 320). The 2007–2008 boom again saw oil prices in the headlines. According to Google, oil was the sixth most common economics-related search term of 2008,4 trailing only items associated with the financial crisis. Moreover, the U.S. Congress held several hearings in 2008 that focused on factors contributing to the high price of crude oil. In the early 1970s, the Club of Rome (Meadows et al. 1972), a high-profile global think tank, predicted a worldwide catastrophe within a generation due to a food shortage. The Club was concerned mostly that densely populated countries like China and India would face food shortages and induce panic in the rest of the world. The Club’s ideas were motivated by the food crisis at the time, and its projections were loosely based on the writings of British economist Thomas Malthus, who 200 years ago said the world would eventually face a large-scale famine because population growth would outstrip the food supply. In a 1990 book entitled The Population Explosion, Paul and Anne Ehrlich built on this Malthusian theme and argued that humans are on a “collision course with massive famine.” More recently, environmentalist Lester Brown predicted in a 1995 book entitled Who Will Feed China? that demand from China would soon push food prices so high as to cause mass starvation. These doom-and-gloom predictions all proved wrong. Figure 2 (see color insert) shows that the real price of most commodities declined through the 1980s and 1990s. During this period, great advances were made in reducing malnutrition in poor countries, as food production outpaced population growth in developing countries outside Sub-Saharan Africa. If anything, there was concern in the 1980s and 1990s over agricultural commodity prices being too low, discouraging farm production in developing countries. Generous government subsidies in rich Organization for Economic Cooperation and Development (OECD) countries were blamed for oversupply and depressed food commodity prices in the 1980s and 1990s (Anderson & Martin 2006). As Figure 2 reveals, in 2008 nonenergy commodity prices did not reach the levels experienced in 1973–1974. For instance, in late 2007 and early 2008, long-grain rice futures prices increased more than that for any other grain on the Chicago Board of Trade futures market and at the peak in April 2008 reached $24 per hundredweight (more than $500 per metric ton), but adjusted for inflation this was 50% less than the peak of U.S. long-grain rice prices in late 1973. Energy prices stand out in Figure 2 because the 1979 oil crisis caused real prices to double at a time when other commodities were not booming. The cumulative energy price increase in the period from 1998–2005 was also much greater than for the other commodities, so energy prices entered the 2007–2008 boom-bust cycle at a relatively high level. Real food prices stopped declining in the early 2000s, several years before the boom occurred in 2007–2008. As Piesse & Thirtle (2009) correctly point out, when food commodity prices started to rise in late 2006, the trend was not an abrupt reversal of declining real prices but instead more of a change from stable to rising prices. The real prices of energy and metals actually started increasing in approximately 2002. For example, crude oil prices increased from $25 per barrel in 2002 to $70 in mid-2007, an increase that was attributed to supply and demand fundamentals, such as strong economic growth in China 4

See http://www.google.com/intl/en/press/zeitgeist2008/mind.html.



91


and India (Hamilton 2009). Oil consumption in China increased by 50% from 5.2 to 7.6 million barrels per day from 2002 to 2007. Even though the decline in real-commodity prices stopped well before the 2007–2008 boom, the price surge caught many by surprise (especially for agricultural commodities). For this reason, and reflecting the fact that no major single world event marked its beginning, The Economist magazine referred to the 2007–2008 commodity boom as a “silent tsunami.” The sharp rise in food prices created a food crisis; sharp price increases gave rise to concerns about inflation in rich countries and worries over increased hunger and political instability in poor countries. The mass media was abuzz with nonstop descriptions of food hoarding and attempts by governments to manipulate the market to calm the panic-like situation in many countries. An unprecedented World Food Crisis Summit of political leaders was held in Rome in June 2008, organized by the UN Food and Agriculture Organization (FAO). One objective of the summit was to try to find some common ground for ways to alleviate the problem. The summit focused on what gave rise to the surge in agricultural commodity prices and its relationship to the concurrent surge in nonagricultural commodity prices. Many experts (FAO 2008, IFPRI 2008) predicted that this was the beginning of a new permanent long-term upward shift in commodity and food prices. After the dramatic price crash in late 2008, these dire predictions appear just as misguided as their Malthusian counterparts from the 1970s. It would thus appear on the surface that the 2007–2008 boom was not fundamentally different from the 1973–1974 event; both entailed temporary price spikes. What underlies claims that these events were fundamentally different? One argument is that commodities have been “financialized” in the past decade through the development of investment vehicles like commodity index funds and the increased participation of hedge funds in commodity markets. Although commodity index funds are indeed new, the hypothesis that futures market speculative trading may exacerbate booms and busts is not. For example, Labys & Thomas (1975) analyze the effect of futures market trading on the 1973–1974 boom by quantifying “the degree to which this speculation rose and fell with the switch of speculative funds away from traditional asset placements and towards commodity futures contracts.” Irwin & Sanders (2011) provide strong evidence against this hypothesis. Another hypothesis underlying claims that the 2007–2008 boom was likely to be permanent is that commodity demand will soon outstrip supply. Proponents of the peak oil hypothesis (e.g., Simmons 2005) claimed that global production of crude oil had peaked or was about to peak.5 A recent slowdown in the growth of crop yields also raised concern that agricultural supply would be unable to grow fast enough to keep real prices from rising (Alston et al. 2009). Moreover, the rise of biofuels led the FAO and others to claim that commodities are now more closely tied with the prices of fossil-based fuels than they have ever been. As the world economy evolves, the race between technology and scarcity will determine whether commodity prices continue their long-run decline or increase according to the Hotelling (1931) rule as scarcity bites.

3. MARKET STRUCTURE In this section, we describe the importance of the underlying fundamentals of commodity supply and demand. We present a broad framework that encompasses the commodity 5

In August 2005, Simmons bet John Tierney and Rita Simon, the widow of Julian Simon, $2,500 each that the price of oil averaged over the entire calendar year of 2010 would be at least $200 per barrel (in 2005 dollars).

92

Carter

Rausser

Smith


groups included in Figure 1. With the exception of livestock, these commodities are storable. After presenting the economics of consumption and production in this section, we address storage in the next section. All these commodities are supported by market structures that include physical or spot markets, futures markets, and forward markets. Much of the volatility and the potential for booms and busts in each of these various commodity markets can be traced to the inherent market structure characteristics. In terms of the microstructure, we describe the behavior of four separate groups: consumers, speculators, producers, and processors. Consumers purchase and use processed commodities according to their preferences; they do not typically trade in futures markets or engage in forward contracting. Speculators trade in futures markets, but they do not generally handle the physical commodity. Producers include farmers, miners, and drilling and extraction firms. Processors are intermediaries who convert the raw commodity into a good for final sale and may include oil refiners, food processors, grain mills, and exporters. Both producers and processors engage in futures markets and forward contracting. The behavior of the four groups gives rise to three markets for which equilibrium conditions must be satisfied. Sudden supply shocks often hit commodity markets. These shocks may emanate from climatic conditions; adverse weather, such as droughts, floods, or hurricanes; labor strikes; pests and plant disease; or geopolitical events, such as wars and trade disputes. Such nonmarket supply shocks dominate short-run volatility in many commodity markets. However, demand shocks also arise, as demonstrated in 2007–2008 by the jump in demand for grains as a feedstock for use in producing biofuel. Moreover, from the perspective of a particular producing country, supply shocks in other countries appear as shocks to export demand. With these stylized facts in mind, we follow numerous authors (e.g., Hirshleifer 1988) and consider a two-period decision process for market participants. In the first period, producers choose inputs to production, and producers, processors, and speculators take hedging positions in futures and forward markets. At the time of these first-period decisions, the final demand is uncertain, as is the final amount of production. In the second period, output and demand are realized. Processors choose how much to process, consumers choose how much to purchase, and the participants in futures markets realize their profits or losses. To provide a concrete framework for understanding supply and demand of a diverse set of commodities, our characterization of supply and demand makes several simplifications. The conceptual framework is represented in Figure 3. Without material loss of generality, we assume that individuals do not migrate among groups due to asset fixities. We also assume symmetry of information across market participants, and we ignore transactions costs. Because we focus on boom-and-bust cycles, we do not consider long-run supply and demand. Thus, we abstract away from the fact that commodity demand schedules tend to shift slowly over time with evolving technology, wealth, and tastes, as exemplified by the changes in commodity flows generated in recent years by strong economic growth in Asia. Supply also exhibits a slowly moving component as, for example, new seed technology improves crop yields or advances in mining and drilling technology lower energy and mineral extraction costs. Risk aversion and uncertainty play important roles in this framework. Together, they create an incentive for firms to hedge and thereby make the futures and forward markets relevant. As pointed out by Rausser (1980) and Moschini & Hennessy (2001), the classic hedging literature neglects basis risk and production uncertainty (e.g., Telser 1958, www.annualreviews.org


93

First Period

Second Period

SPECULATORS choose - Number of futures contracts

SPECULATORS realize - Futures profits or losses

futures price

forward price

PROCESSORS choose - Number of futures contracts - Number of forward contracts

PRODUCERS realize - Futures profits or losses PRODUCERS choose - Quantity supplied

Supply Shocks Demand Shocks


PRODUCERS choose - Inputs to production - Number of futures contracts - Number of forward contracts

raw commodity spot price

PROCESSORS realize - Futures profits or losses PROCESSORS choose - Quantity demanded (raw commodity) - Quantity supplied (final good)

final good spot price

CONSUMERS choose - Quantity demanded (final good) Figure 3 Commodity supply and demand framework.

Anderson & Danthine 1980, Feder et al. 1980). By treating production as fixed, these studies produce separation between the hedging and production decisions, which implies that the cost of hedging and risk aversion do not affect supply. Other authors model production and price risk jointly (e.g., McKinnon 1967, Newbery & Stiglitz 1981, Britto 1984, Moschini & Lapan 1995). Joint modeling captures the fact that price tends to be negatively correlated with production because supply shocks have a negative effect on price. Hirshleifer (1988) generalizes this framework by modeling jointly the behavior of the four types of market participants described above. Futures markets and forward contracting are the two main hedging tools used by market participants to insure their exposure against unfavorable price movements. In general, a forward contract is a bilateral agreement between two market participants to 94

Carter

Rausser

Smith


transact a stipulated quantity and grade of the commodity at a specific price on a stated future date. In our framework, producers may enter forward contracts with processors. Futures contracts also specify the quantity, grade, time, and price of a future transaction of the commodity. However, futures contracts are traded anonymously on exchanges rather than bilaterally. If both deliver the same location and grade at the same time and if daily interest rates are nonstochastic, then futures and forward prices must be identical (Cox et al. 1981). For this reason, much of the commodity pricing literature treats them as identical. In practice, forward contracts are customizable (i.e., have little or no basis risk), which explains why farm contracts with merchants commonly use them. Also, futures require participants to post margin, which may not be costless for firms that are credit constrained. For instance, the 2008 cotton price spike brought down three large U.S. cotton merchants who could not meet their futures contract margin calls (Carter & Janzen 2009). The participation of speculators in futures markets enhances liquidity, making it easier for hedgers to transact at current market prices without encountering large bid-ask spreads. The precise trade-off between liquidity, margin costs, and basis risk varies widely across firms, which explains why both futures markets contracting and forward contracting are actively used. The fact that many participants use forward contracts suggests that the cost of doing so is lower for them than for futures contracts and/or that they value the basis risk insurance that they achieve with a forward contract relative to a futures contract. To the extent that forward contracting reduces the risk faced by producers, production is encouraged, which in turn mitigates boom-bust cycles.6 This is the only channel by which futures market speculation affects the price paid by consumers; i.e., futures market speculation affects cash prices only to the extent that the quantity supplied is affected (Hamilton 2009, Irwin & Sanders 2011). Consumers enter this framework only in the second period, so they are represented by a market demand schedule. In the very short run, most of the commodity markets feature highly inelastic demands [exceptions would be meats and, less so, cotton (Rausser 1982)]. Shifts in supply that move along a short-run inelastic demand schedule can clearly be a major source of booms and busts. In the medium run, demand can adjust as people buy fuel-efficient cars and processors use new ingredients, for example, high-fructose corn syrup instead of sugar. There is a vast literature on estimating demand elasticities. For food and agricultural commodities, two well-documented sources are presented by Iowa State University7 and the U.S. Department of Agriculture (USDA).8 Roberts & Schlenker (2010) estimate world supply and demand calories derived from corn, soybeans, wheat, and rice. These four crops make up approximately three-quarters of the caloric content in global food production. They estimate a short-run demand elasticity of 0.04. The Energy Information Administration of the U.S. Department of Energy has compiled results from a large number of studies of energy commodity systems (see also Rausser et al. 2004; http://www.eia.doe. gov/analysis/). Hughes et al. (2008) estimate that the price elasticity of demand for

6 In the United States, forward contracting is threatened by the recent Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010. This law seeks to push more hedging activity onto exchanges and to clear trades through central clearinghouses. This change will likely raise the cost of hedging to firms and have a dampening impact on the production of risk-averse firms. 7

See http://www.fapri.iastate.edu/tools/elasticity.aspx.

8

See http://www.ers.usda.gov/Data/Elasticities/data/DemandElasData092507.xls.



95


gasoline in the United States was between 0.21 and 0.34 from 1975–1980 and between 0.03 and 0.08 from 2001–2006. These estimates are consistent with others in the literature (Hamilton 2009). The total supply function can be naturally decomposed into an asset supply (e.g., exploration and investment in mines in the case of precious and base metals, or land cultivated for various agricultural commodities) and a short-run productivity response (e.g., rate of extraction for existing mines or the yield or productivity of various agricultural commodities; see Chu & Morrison 1986). With respect to supply elasticities, many empirical studies estimate the area, land, or investment in exploration elasticities, but very few studies estimate the productivity or yield elasticities. For the area or land response elasticities, comprehensive sources are again Iowa State University, the USDA (see footnotes 7 and 8), and an earlier study by Askari & Cummings (1976). The latter study estimates more than 600 supply elasticities for different commodities and countries. In accordance with economic theory, this study reveals long-run elasticities that tend to be greater than short-run elasticities. Most of the numerical values of the elasticities for the short run are in the range of 0.0 to 0.3; the second largest frequency falls in the range of 0.34 to 0.67. In addition to the core market parameters and the nonmarket supply factors, the potential for booms and busts depends critically on the role of expectation formation patterns across the various commodity market participants. Generally, authors in the literature impose the expectation formation pattern as part of their maintained hypotheses in their empirical models.9 The most internally consistent empirical representations of commodity price formation are rational expectations, first introduced by Muth (1961). This formulation is applied to commodity futures markets by Bray (1981), Danthine (1978), and Rausser & Walraven (1990), among others. More generally, Tirole (1985) demonstrates that unless agents have different priors about the value of a particular commodity asset or are able to secure insurance in the corresponding market, speculation (gains from trade) is ruled by rational expectations. In all applications of rational expectations in commodity markets (e.g., Miranda & Helmberger 1988) and in macroeconomics (e.g., Lucas & Sargent 1981), rationality is driven only by benefits. The cost of collecting information and data to formulate rationally expected prices is swept under the rug or neglected. Theoretically, for some economic environments, naive expectations are in fact rational. In empirical models, this apparent paradox results from the failure of rational expectations to incorporate the cost of collecting information on critical variables.10 Even though much of the empirical literature imposes an expectation formation pattern as part of the maintained hypothesis, periods of booms, busts, or bubbles are unlikely to arise in a rational world with constant risk aversion. In other words, for each commodity system, the operative expectation pattern depends critically upon the economic environment. 9

However, certainly for those commodities that have both active futures and spot markets, sufficiently rich data sets exist to discriminate across various expectation formation patterns. These patterns range from rational expectation (the core principle in the efficient market hypothesis of Fama) to naive expectations. For further details, see Just & Rausser (2001).

10

In the theoretical landscape, Grossman & Stiglitz (1976) explicitly address this question, demonstrating the impossibility of informationally efficient markets. To our knowledge, there has been no empirical application of this theoretical model, although Stein (1992) shows some evidence that prices on the Sydney Futures Exchange match the predictions of an informationally efficient Bayesian learning model.

96

Carter

Rausser

Smith


For example, the weights appearing in any convex combination depend not only on the expected benefits but also on the cost of information used in forming expectations. To be sure, booms are most likely to emerge when many of the market participants naively form their expectations of the future. In volatile commodity markets, risk must also be recognized and is typically represented in terms of the probabilities of (squared) deviations from an expectation. If the wrong expectation is used to assess deviations, then agents could over- or underestimate the risk they face. Along with expectation formation patterns’ role in explaining booms and busts for our two major events, 1973–1974 and 2007–2008, the market structure was an important component in explaining and facilitating the price spikes that took place. Numerous shocks to either supply or demand caused large price responses because short-run supply and demand elasticities are small. In the case of the 1973–1974 commodity events, a major inward shift in supply resulted in a spike in crude oil prices and all refinery energy products. This shift can be directly traced to OPEC imposing an oil embargo on the United States as punishment for that country’s support of Israel triggered by the start of the Yom Kippur War. In the case of grains, export demand shifted outward as a result of both Russian and Asian demand expansion. The El Nin˜o weather patterns dramatically lowered the anchovy fish catch, with the result of an increased protein demand, which expanded the demand for oilseeds, especially for animal feeds. Policy failures elsewhere (e.g., cheap food policies in poor countries) also reduced food supplies in the early 1970s, but the overall reduction in the grain supply was rather modest leading up to the crisis (Cooper & Lawrence 1975). Shifts in supply and demand also played a crucial role in the events of 2007–2008. Due to the rapid increases in income in many countries, including China, India, and Russia, global energy demand shifted outward. This phenomenon also expanded the demand for many grains, including corn and soybeans. For corn, public sector–orchestrated incentives also expanded the demand for biofuels, which made a major contribution to the spike in feed grain prices. Mitchell (2008) identifies the rise in biofuel production in the United States and the European Union (EU) as the leading cause of higher agricultural commodity prices in 2007–2008. Cross-price elasticities assisted in further expanding the demand for biofuels due to the price spikes in crude oil. (See Section 5 on cross-commodity linkages.) Inward supply shifts due to nonmarket supply factors, particularly weather shocks in Eastern Europe and Australia, contributed to price spikes in food grains. Combined with the unprecedented extension of the multiyear Australian drought that reduced wheat production, transport cost increases made matters worse. For metals, supply also shifted inward, in particular for platinum because of the closing of South African underground mines. This shift created the foundation for new price records being broken almost on a daily basis in 2008 as consumers of this metal panicked over the security of supplies. With respect to the role played by market structure in contribution to higher commodity prices in 2007–2008, there is a strong difference of opinion on the relative importance of market structure versus other forces. For example, the price of corn more than doubled between 2006 and 2008, but the U.S. government suggested that biofuel policies were a relatively minor influence on the higher corn price (Lazaer 2008). Alternatively, Roberts & Schlenker (2010) estimate that biofuel demand has caused a 20–30% increase in the average price of staple food commodities. Both the International Food Policy Research Institute (IFPRI) in Washington, DC, and OECD in Paris found that biofuels explained approximately 30% of the corn price increase in 2007–2008. Some argued that biofuels played an even larger role in explaining agricultural commodity price increases (FAO 2008). www.annualreviews.org


97


In the context of the linkages among spot, futures, and forward markets, many explanations have been sourced with the role of speculators. After accounting for the market structure fundamentals, some economists attribute part of the 2007–2008 oil price spike to speculation driven by concerns about global supply and the development of new investment vehicles such as commodity index funds (Hamilton 2009). Wray (2008) claims that the rise of speculative investments in the commodity futures market (e.g., through index trading) was the largest contributor to the rise in commodity prices before the bust in 2008, but Irwin & Sanders (2011) provide convincing econometric evidence against this claim. For both events, 1973–1974 and 2007–2008, the literature has generally recognized the critical role of commodity stocks in any dynamic extension of Figure 3. Piesse & Thirtle (2009) argue the single most important factor in agricultural prices was low inventories, as the stocks/utilization ratio for grains and oilseeds dropped to 15% in 1972 and 1973 and did not touch such a low level again until 2008. They also identify the importance of rising prices of fuel and fertilizer in 1973–1974 and again in 2007–2008 in driving up the cost of production and transportation. Radetzki (2006) argues that aggregate demand growth played a key role in both the 1973–1974 and 2007–2008 commodity price booms. Trostle (2008) recognizes the importance of fundamental supply and demand factors, emphasizing the decline in stocks-to-use ratio for wheat, rice, and corn grains leading up to the boom witnessed in 2008. Accordingly, we turn in the next section to the evidence on the role of stockholding behavior in explaining booms and busts.

4. STOCKHOLDING Plentiful inventories provide a buffer against supply and demand shocks. In response to such shocks, inventories can be drawn down, mitigating the impact on prices. When inventories are low, the lack of a buffer leaves the markets vulnerable to price spikes. Thus, to explain commodity booms and busts, we need an understanding of what determines stock levels and, in particular, what may cause inventories to be depleted. Stockholders compare the current price of a commodity with the expected price at some future date and the cost of carrying inventories (including the opportunity cost of funds). If the expected profit from holding inventories exceeds the payoff from selling the commodity immediately, then stockholding firms may choose to store the commodity. Conversely, if the expected future price is too low to compensate firms for the cost of holding inventory, then they will not store. It follows that current stockholding as well as future stockholding are determined both by current supply and demand and by expected future supply and demand and are thus inherently dynamic. The staple of the stockholding literature is the competitive rational storage model, which originated with Williams (1936). Gustafson (1958) was the first to solve for the optimal storage rule in this model, and Williams & Wright (1991), Deaton & Laroque (1992, 1996), and Routledge et al. (2000) made further important advances with this model. The competitive storage model specifies that risk-neutral stockholding firms make rational expectations about the future and act to maximize expected profit in a competitive market. Moreover, it provides basic insights necessary for understanding the role of stockholding in commodity booms and busts. In what follows, we examine extensions to the basic model that permit nonrational expectations (Nerlove 1958); risk aversion (Keynes 1930, Newbery & Stiglitz 1981, Gorton et al. 2007); technology or transportation costs (Williams & Wright 1989, Brennan et al. 1997, Carlson et al. 2007); government attempts 98

Carter

Rausser

Smith


at price stabilization (Newbery & Stiglitz 1981, Miranda & Helmberger 1988); and a convenience yield, which is a positive flow of services from stockholding (Kaldor 1939). Throughout, we focus on the features of the model relevant to price booms and busts. Following Williams & Wright (1991) and Deaton & Laroque (1992), suppose that the inverse demand function for current use is Pt ¼ f(Dt), where Dt denotes the quantity demanded and Pt denotes the commodity price in period t. Quantity supplied (St) is determined by an iid (independent and identically distributed) harvest shock. Compared with the two-period framework depicted in Figure 3 of Section 3, this dynamic setting adds an additional type of firm, the speculative storage firm. Such firms interact with producers and processors to determine the raw commodity price. They provide an additional source of demand for the raw commodity in the current period and an additional source of supply in subsequent periods. Without loss of generality and for clarity of exposition, we assume that the current-use demand function is constant over time. This formulation captures the real-world feature of markets dominated by temporary supply shocks, and it abstracts away from slow-moving supply and demand shifts. Speculative storage firms may choose to store It units of the commodity at volumetric cost d per period and face an opportunity cost of capital equal to r. Such firms will store an extra unit of the commodity if the expected price next period, net of interest and warehousing costs, exceeds the current spot price. They will store fewer units if the current price is high relative to the expected price next period. Thus, we have the following intertemporal equilibrium arbitrage condition: Pt ¼

1d Et (Ptþ1 ) 1þr

if It > 0,

Pt

1d Et (Ptþ1 ) 1þr

if It ¼ 0.

Together with the market clearing condition that quantity demanded plus incoming inventories equal quantity supplied plus outgoing inventories, this model produces a stationary rational expectation equilibrium (see Williams & Wright 1991, Deaton & Laroque 1992). This equilibrium is characterized by a downward-sloping inventory demand curve. We depict this equilibrium in Figure 4. In any given period, total demand for the commodity equals the horizontal sum of the inventory demand curve and the demand for current use. Inventory demand in this figure is the difference between total demand and f(Dt), which is demand for current use. When speculative inventories are positive, total demand is relatively elastic because the market can respond to adverse shocks by drawing down inventories. When speculative inventories are zero, total demand is relatively inelastic because there is little capacity to draw down inventories. A key feature of the competitive storage model is the restriction that inventory stocks cannot be negative. It is impossible to borrow stocks from the future. As a result, this fact can be a source of booms and busts. When a negative supply shock drives inventory to zero, i.e., causes a stockout, the price is determined by the inelastic current demand curve. Because this part of the curve is steep, even a small negative supply shock can cause a large price spike. However, such price spikes tend to be short-lived as supply is replenished and inventory begins accumulating again. If consumption demand becomes more elastic in this model, average inventory levels decline, and stockout-induced booms and busts become more frequent and smaller in www.annualreviews.org


99

PRICE

Supply

Pt

Inventory

Consumption

Total Demand Annu. Rev. Resour. Econ. 2011.3:87-118. Downloaded from www.annualreviews.org by University of California - Berkeley on 10/31/11. For personal use only.

f (Dt) Dt

It-1+St

QUANTITY

Figure 4 Equilibrium in the competitive storage model.

amplitude (Williams & Wright 1991). Serial correlation in prices also declines as demand becomes more elastic. In a set of influential papers, Deaton & Laroque (1992, 1995, 1996) argue that, when calibrated to real data, the competitive storage model with iid supply produces too little serial correlation. To match the data, they propose a competitive storage model with serially correlated shocks, which is also a specification favored by Routledge et al. (2000). Serially correlated shocks generate serial correlation in prices by allowing the inventory demand curve to shift, which occurs because the current shock provides information about likely future shocks and therefore affects willingness to hold inventory. Cafiero et al. (2009) show that Deaton & Laroque’s criticisms were overstated. With more accurate solution techniques, Cafiero et al. estimate a much smaller demand elasticity parameter, which enables the model to more closely match the data by exhibiting strong autocorrelation and infrequent stockout-induced booms and busts even with iid shocks. Carter & Revoredo-Giha (2009) show that strong autocorrelation can also be induced by including processors as holders of inventory. Much of the literature on competitive storage models focuses on agricultural commodities. However, a parallel literature with a similarly long history exists for exhaustible commodities such as metals, oil, and gas. In a world with zero extraction costs and known reserves, the producer’s problem is the same as for the competitive stockholding firm. In the absence of uncertainty, Hotelling (1931) shows that under certain conditions prices should grow at the rate of interest as reserves are depleted. This rule is exactly the intertemporal equilibrium condition with positive inventory shown above. Weitzman & Zeckhauser (1975) and Pindyck (1980) show that this rule continues to hold when demand is stochastic. The price history of exhaustible commodities does not match the predictions of this simple theory. In fact, real resource prices have tended to decline over time. For this reason, the recent literature has moved from treating resource extraction as an inventory problem to treating it as a production problem. The resulting models incorporate features such as technological progress (e.g., Lin & Wagner 2007) and adjustment costs 100

Carter

Rausser

Smith


(e.g., Carlson et al. 2007), and they generate modified versions of Hotelling’s (1931) rule. Thus, as noted by Slade & Thille (2009, p. 256), “the often-cited fact that the Hotelling model is frequently rejected by the data. . .must be interpreted with caution.” A competitive storage model for any resource predicts price spikes when inventories become low. In the presence of production constraints, extraction cannot respond quickly to shocks, and aboveground inventories become important. Aboveground storage of petroleum is limited by its high cost, which implies that aboveground inventory levels are always low relative to those of other commodities, and stockholding behavior affects prices only at horizons of a couple of months. Thus, when applied to aboveground inventories, the rational storage model cannot explain petroleum price variation except at very short horizons. In contrast, storage costs are lower for minerals, suggesting that the rational storage model is more relevant for these commodities. The market price of storage as determined by the futures term spread is often negative during low inventory periods. Thus, not only do firms not run inventories all the way to zero, they appear willing to hold inventory at a loss. Working (1949) first documented this phenomenon in Chicago wheat during the 1920s. Three separate theories, each with a long history, have been proposed to explain the apparent willingness of firms to store inventory under backwardation (distant futures prices lower than current spot or nearby futures prices): (a) convenience yield, (b) spatial aggregation, and (c) risk aversion. We address each of these explanations below, and we also address the implications of market power and nonrational expectations for the applicability of the competitive storage model. Eastham (1939) posed an early version of the model in Figure 4, with the added feature of heterogeneity among stockholders (Carter & Revoredo-Giha 2009). Some firms hold stocks for speculative purposes, whereas others hold stocks as working inventories. This formulation foresaw the development of convenience yield models, in which firms may hold inventory because it produces a flow of services (Kaldor 1939, Brennan 1958, Telser 1958, Brennan & Schwartz 1985). The concept of convenience yield is somewhat vague. It has been described variously in the literature as representing the value of an option to change production at short notice to take advantage of market conditions (Litzenberger & Rabinowitz 1995), an option to avoid shutting down a processing plant if supplies were to become scarce (Brennan 1958); as a reflection of high fixed costs of acquiring or disposing of a batch of inventory (Bobenreith et al. 2004); or as a loss-leading strategy to draw in customers who pay for merchandizing services (Paul 1970). Williams & Wright (1989) and Brennan et al. (1997) challenge the convenience yield theory by pointing out that commodities are stored differentially across space. If transportation costs are significantly large, then inventory at inconvenient locations may be unable to be shipped out in a timely manner. Comparing prices at locations with no inventory with prices at other locations with positive inventory necessarily makes it seem that firms are storing at a loss when they may not be. However, recent work by Carter & Revoredo-Giha (2007) and Franken et al. (2009) demonstrates at the firm level that stockholding firms do hold stocks at an apparent expected loss. These two papers provide support for the existence of convenience yield. Their results imply that stocks are held for reasons other than simple speculation and that the spatial aggregation argument is insufficient to fully explain storage at a speculative loss. Risk aversion may produce storage at an expected loss. Risk-averse processing firms may be willing to hold inventory to avoid uncertainty over the price that they may have to www.annualreviews.org


101


pay to acquire the commodity later. This uncertainty may be greater when inventories are low and volatility is high, which would cause these firms to hold more inventories off the market, thereby exacerbating any price boom. However, the risk aversion literature tends to focus on storage firms rather than on processors and thus finds risk premia of the opposite sign. Gorton et al. (2007) provide some evidence of increased risk premia when inventories are low. They argue that firms need to be encouraged to hold inventories in such periods because of the price risk associated with the future sale of those inventories. Their story is consistent with the normal backwardation hypothesis of Keynes (1930), which was reinforced by Carter et al. (1983), Bessembinder (1992), and De Roon et al. (2000). Thus, although risk aversion could exacerbate booms and busts, the literature tends to find that the risk aversion displayed in the markets acts to dampen them. Although the rational storage model presumes perfectly competitive markets, many commodity markets display characteristics that violate these assumptions. For example, OPEC openly attempts to collude on crude oil production (Smith 2009), many countries set up government-backed entities to control exports and imports, and a small set of merchandising firms handle a high proportion of the international grain trade. Quantifying the impact of noncompetitive behavior on a commodity market is challenging because of strategic behavior and because some government-controlled marketing entities are driven by political rather than economic motivations. However, as first noted by Adam Smith, market power in storage markets implies too little storage on average. A firm with a monopoly on storage space would profit by withholding space and setting prices above competitive levels. Thus, market power does not explain why firms may choose to store at an apparent loss. Nonetheless, with lower average inventory levels, a noncompetitive market would tend to hit the steep part of the total demand curve more frequently, and thus stockout-induced price spikes would occur more often and with larger amplitude (Williams & Wright 1991). Imprecise information about current inventory levels or about future demand and supply can also reduce storage efficiency. Market participants may possess imprecise information because of fundamental uncertainty, such as agricultural yields that are sensitive to weather events. The nature of stockholding also affects information precision. In many less-developed countries, substantial amounts of grain are stored in homes and on small farms, making it difficult for markets to assess the quantity of inventory in storage (Park 2006). Similarly, government entities frequently engage in extensive commodity storage. Although it may not be the purpose, government storage can enhance inventory measurement, as in the strategic petroleum reserve in the United States or the large, publicly owned grain reserves that accumulated in the United States in the 1980s. In contrast, some governments may have a strategic incentive to keep inventory levels secret, as happens for petroleum in OPEC nations and for grain in China. Associated with the incomplete availability of commodity storage data is an information externality. Such an externality means that the social gains to information collection exceed the private gains (Grossman 1977). In many markets, notably grains, government agencies alleviate this potential externality by collecting and publishing inventory information. However, even in the face of accurate inventory information, market participants face substantial uncertainty that may be difficult to quantify. In such circumstances, it may be efficient for market participants to resort to simpler expectation formation rules. Examples of such rules include making output decisions on the basis of the price at the time of planting, which often provides the basis for the booms and busts of a cobweb model (Kaldor 1934). Such expectations may exacerbate booms and busts if they lead firms to 102

Carter

Rausser

Smith


interpret a temporary negative supply shock as a permanent shock. Such firms would respond by hoarding the commodity, driving prices up further than they would under rational expectations. Prices then crash back to earth once it becomes clear that excessive inventory has accumulated. Systematic errors in expectation sometimes arise in asset markets as market participants display “irrational exuberance” (Shiller 2005). In commodities, viewed through the lens of a dynamic storage model, such inflated expectations about future prices raise the demand for storage. Stockholding firms see profit opportunities in holding additional stocks to profit from the expected higher prices. As with any outward shift in the storage plus current use demand curve, we would expect to see increases in both the price of the commodity and the quantity in storage, all else equal. Only if the demand for current use is perfectly inelastic will a price increase not raise inventory levels. Hamilton (2009) argues that because the demand for current use of crude oil may be very inelastic, such increases in inventory could be imperceptible. However, all else equal, the prediction that a speculation-fueled price increase causes inventory accumulation holds. In practice, all else is not equal. As Kindleberger (1978) observes, episodes of irrational exuberance tend to follow fundamental shocks. Market participants observe prices increasing in response to the fundamental shocks and overreact, further bidding up prices and often claiming that “this time is different” (Reinhart & Rogoff 2009). In commodities, the fundamental shock may take the form of a supply disruption, which raises prices and reduces inventory holdings. If market participants overreact and induce increased speculative demand for inventory, then prices rise further and inventories accumulate. Thus, the net change in inventory from the fundamental shock and the irrational exuberance may be positive or negative. Wright (2011) articulates the important role that dynamic stockholding behavior played in the two major events, 1973–1974 and 2007–2008, described in Section 2. To demonstrate this point, Figure 5 (see color insert) plots annual price paths against inventory levels for six commodities. Each panel in the figure shows a scatter plot of real prices against inventories, with the points on the scatter connected to make a path through time. By following the path for a particular commodity, we trace the evolution of prices and inventories. We display in Figure 5 the three major grains (corn, wheat, and rice), the largest energy commodity (crude oil), and an important industrial commodity (copper). We also include cotton because it provides an interesting contrast to the grains: It did not experience a dramatic boom and bust in either 1973–1974 or 2007–2008. All price series span 1960–2010, except crude oil, for which the price series begins in 1974. For the agricultural commodities, we measure price in the middle of the crop year at a time when the size of the previous crop is known and the weather shocks that will affect the size of the upcoming crop have not been realized (March for corn, wheat, and cotton; April for rice). For consistency across commodities, we also measure crude oil and copper prices in March. We use log prices deflated by the U.S. consumer price index, so one unit on the vertical axis corresponds to a 69% real price difference. We measure inventories using the global stocks-to-use ratio where possible.11 For crude oil, we use the longer sample of U.S. stocks in Figure 5, although OECD stocks data produce very similar results. 11 Most rice stocks in China are held on farm and never enter commercial channels, which makes rice data from China unreliable. Moreover, in contrast to corn and wheat, estimated Chinese rice stocks display strong trends that seem unrelated to prices. We therefore omit Chinese rice stocks.



103


Rational storage theory suggests a downward-sloping demand curve that becomes steep when inventories run low. The three grain commodities experienced booms in 1973 and 2008 and exhibited their lowest inventory levels at these times. In contrast, cotton inventories in 1973 and 2008 were approximately average. Copper displays some similarities to the grains, with the 2008 boom and bust occurring in the face of relatively low inventories. However, copper did not experience a 1973 boom. As was the case for cotton, copper inventories were healthy in 1973, so although the real price was above average during this period, the market did not exhibit a large price spike. These cases illustrate the point that plentiful inventories enable commodity markets to cushion the blow of unexpected supply disruptions or jumps in demand by drawing down inventory. In contrast, when stocks are low, current supply and demand shocks must be met by a reduction in current use. Because of the magnitude of the short-run elasticities, the price response can be large, even if the underlying shocks to supply and demand appear relatively small. As noted above, crude oil is expensive to store aboveground, so most storage is largely operational. It follows that inventories are less volatile, which manifests in Figure 5 as a price path that traces out a circle, rather than as one that oscillates wildly, as do the other commodities. The lack of intertemporal profit-based storage also implies that price spikes for crude oil are only weakly related to aboveground inventory levels. All the price spikes displayed in Figure 5 for corn, wheat, rice, and copper follow the same template. The price path follows a clockwise pattern. In the lead up to the spike, stocks get run down, and prices increase slightly. Then, when stocks reach a critical point, the price spikes. In the ensuing year or two, stocks gradually get replenished, and the price declines. The price decline typically occurs more slowly than the spike. Unlike prices of the other commodities, corn prices spiked in 2008 at higher inventory levels compared with 1973, which suggests an outward shift in the demand for inventories. The source of this demand increase was the dramatic growth of the biofuel industry. In the 2008 crop year, more than 30% of U.S. corn supply was diverted into ethanol production, up from just 14% in 2005. This diversion has a significant impact on world corn prices because the United State typically produces approximately 40% of the world’s corn and accounts for 60% or more of global exports. This dramatic increase in U.S. corn ethanol production stemmed from mandates in the Energy Policy Act of 2005. Because of the long lead time in building ethanol plants, 2007 and 2008 ethanol production was essentially known by late 2006 and therefore would have been incorporated into corn prices by late 2006 as stockholding firms sought to increase storage in advance of the upcoming ethanol production boom. In summary, the basic rational storage model predicts rare zero-inventory periods that generate booms and busts. Modifications to the simple storage model that permit convenience yield and account for spatial heterogeneity can explain the absence of stockouts. Allowing for expectation errors due to high information collection costs or irrational exuberance, or incorporating market power in storage, may exacerbate price spikes. However, neither of these modifications changes the results that short-lived booms and busts are relatively rare and that they tend to happen when stocks reach low levels. Consistent with the theory, the 1973–1974 and 2007–2008 commodity boom-and-bust episodes exhibited extreme price spikes only for those commodities with low stocks. 104

Carter

Rausser

Smith


5. MACROECONOMICS AND COMMODITY PRICES Macroeconomic linkages are often used as the foundation for popular press assessments of the causal source of booms in commodity prices. To determine the quantitative impact of such linkages, it must be recognized that there are both forward and backward linkages (Rausser 1985). The backward linkages relate shocks to a sector of the economy to real macroeconomic performance measures such as GNP, national income, trade balance, and public sector expenditures. In rich countries, the immediate impact of backward linkages from food commodity price booms is at best minimal, but some empirical evidence suggests that energy price booms tend to lead to macroeconomic recessions (Hamilton 2009).12 Forward linkages as well as possible feedback effects are a potential critical source in explaining the magnitude of commodity booms and busts. These forward linkages directly relate money markets to commodity market prices through two principal channels: interest rates and exchange rates. As noted by Rausser (1985), these linkages became evident in the 1970s with the move to flexible exchange rates, the rapid expansion of international markets, and the emergence of a well-integrated international capital market. The forward linkages between money markets and commodity markets follow directly from a number of causal phenomena. First, the production processes for many of the commodities is extremely capital intensive, with the result that movements in real interest rates have significant effects on the cost structure of producing and supplying such commodities to their respective markets. Second, stock carrying in storable commodity systems is sensitive to interest rates, whereas for nonstorable commodities (for example, live cattle and hogs) breeding stocks are interest rate sensitive. If the influence of interest rates on the value of currencies is taken into account, many countries’ fiscal and monetary policies can exert pressure on both the demand side (export demand, stockholding demand) as well as the cost side and thus on supply to the relevant markets. For our 1973–1974 events, the boom in commodity prices was in part the result of the performance of money markets and foreign exchange rate markets. In particular, during the early 1970s the Federal Reserve Bank expanded the U.S. money supply, with the effective objective of accommodating increases in the real price of energy; other countries also attempted to inflate away energy price shocks. The resulting inflation continued to evolve until the Federal Reserve in October of 1979 adopted a policy of attempting to control money supply directly, rejecting its previous policy of targeting interest rates (Rausser et al. 1986). The rapid inflation reflected by a subset of the commodities during the 1970s ultimately led to tight U.S. monetary policies, which partially explained the subsequent collapse of all commodity prices through much of the 1980s and 1990s (Rausser et al. 1986). This collapse of prices was influenced by real interest rates that reached all-time highs (measured in ex post real terms), helping to reverse the decline of the U.S. dollar that occurred through much of the 1970s. This second-round effect on money markets was triggered by the inflation that can be traced to the 1973–1974 commodity price boom (Blinder & Rudd 2008; see Figure 6, see color insert).13

12 For resource-based economies, particularly mining and energy, a hypothesis has emerged in the literature that the expansion of resource-based commodities actually does harm to the macroeconomy. Wick & Bulte (2009) evaluate this so-called resource curse hypothesis. On the basis of their assessment, they conclude that this curse hypothesis should be rejected. 13 Booms in commodity prices often result in cost-push inflation (Phelps 1978), which does on occasion have a tendency to induce a tightening of monetary and fiscal policies.



105


Given the increasing integration of international markets, it is no surprise that exchange rates have played a role in the movements of commodity prices since at least the early 1970s (Schuh 1974). Chen et al. (2010) show that the dollar exchange rates of commodityproducing countries strongly forecast commodity prices during this period. The dominant exchange rate relates to the U.S. dollar, which is a reserve currency and is also the currency in which most international commodity transactions are denominated.14 Leading up to our second major event (2007–2008), the U.S. dollar lost almost 40% of its value from 2002 through 2008. The fall in the value of the dollar was in part caused by the steady decline in U.S. short-term interest rates, declining from 5% to slightly more than 2% over the short period of time from September 2007 to May 2008. Here again, as in the period of 1973–1974, the forward linkages from money markets to commodity markets were supportive as a result of not only exchange rate movements but also the movement in real interest rates. The seminal theoretical paper that provides the lens for explaining the forward linkages from money markets is Dornbusch (1976). The empirical underpinnings for this work can be traced back to Hicks (1974) and Okun (1975). Hicks and Okun were the first to identify the macroeconomy as being composed of two types of markets: flex price, or what they characterize as auction markets, and fixed price, or what they characterize as customer markets. Given that the latter markets have sticky prices, the speed of adjustment in such markets is much slower than it is for flex-price markets following changes in monetary policy. As a result, disequilibrium in the real rate of interest orchestrated by the central bank (too low relative to the long-run equilibrium real interest rates) means that the flex-price markets composed largely of commodities will overadjust and fixed-price markets will underadjust to expansionary monetary policies. Tight monetary policy will have the opposite effect: Commodity or flex-price markets overadjust on the downside, resulting in contributions to bust cycles, and once again fixed-price markets underadjust. Dornbusch used this basic framework to introduce the notion of overshooting in exchange rate markets. Exchange rates overreact to a monetary shock to compensate for the disequilibrium arising in a more slowly adjusting goods market. In the Dornbusch (1976) formulation, the long-run steady state remains unchanged, whereas the exchange rate equates (temporarily) demand and supply in both the exchange and goods markets. The overshooting in exchange rate markets was extended to storable commodities by Frankel (1986), Rausser et al. (1986), Sephton (1988), and Stamoulis & Rausser (1988). Following the theoretical introduction of commodity price overshooting, sourced with both real interest rate disequilibrium and exchange rate overshooting, a number of empirical studies were conducted to test the theory. Frankel & Hardouvelis (1985) found that overshooting results in real increases in commodity prices, even if expectations are formed rationally. Rausser et al. (1986), Stamoulis & Rausser (1988), and Ardeni & Rausser (1995) found that such overshooting phenomena were insufficient to explain the price spikes in 1973–1974 and the low commodity prices that occurred in the 1981–1982 through 1985 period. Their empirical analysis illustrates that stockholding, basic market structure, cross-commodity linkages, and public policies were also required to explain the booms and busts that took place in many commodity markets. 14

For instance, if Venezuela exports oil to China, payments will most likely be made in U.S. dollars, the currency of settlement.

106

Carter

Rausser

Smith


Their empirical results reject the hypothesis advanced by Frankel (1995, 2008a) that money market linkages are the principal explanation for booms or busts in commodity prices.15 In the context of commodity futures markets, Rausser & Walraven (1990) investigated the implications of different degrees of flexibility across markets by quantifying the linkages among three groups of markets: interest rates, exchange rates, and commodity prices (corn, wheat, and cotton). Their results show that, although all these markets overreact to a shock, commodity markets do so to a much greater degree. Owing to their much greater size, however, the welfare loss arising from overshooting is much larger for interest rate and exchange rate markets. In the context of spot markets, similar questions were investigated by Lai et al. (1996) and Saghaian et al. (2002). Both the 1973–1974 and 2007–2008 commodity booms were preceded by unusually high world economic growth, which no doubt led to strong aggregate demand for commodities (see Figure 6). In particular, income growth was very strong in lowermiddle-income countries.16 The vertical line on the left-hand side of Figure 6 indicates the approximate beginning of the 1973–1974 boom, and the vertical line on the right marks the beginning of the 2007–2008 boom. For the five years leading up to the first boom (1969–1973), real GDP grew by 6.6% per year in middle-income countries. Similarly, for the five years leading up to the second boom (2003–2007), middle-income real GDP grew by 7.2% annually. In no year between 1973 and 2003 did middle-income GDP growth exceed 6%, and the average over this interim period was 3.8%. Following each commodity boom, economic growth slowed dramatically. One apparent macroeconomic difference between the periods following the two booms has been the path of inflation. The 1973–1974 boom was followed by a prolonged period of inflation in OECD countries (see Figure 6). Blinder (1982) argues that commodity price shocks, along with the removal of the Nixon price controls in the United States, were the most important factors in producing this inflation. In contrast, the 2007–2008 boom was followed by the deepest recession since the Great Depression. The resulting contraction in aggregate demand eliminated any inflationary pressure and caused deflation to become the principal concern. However, inflation had been a concern of policy makers during the boom period. In the August 25, 2008, Federal Open Market Committee meeting—less than a month before the emergence of the financial crisis—participants “expressed significant concerns about the upside risks to inflation.” Thus, without the collapse in global aggregate demand beginning in September 2008, there may well have been significant inflation in consumer prices. In summary, strong global demand, especially in lower-middle-income countries, helped set the stage for the 1973–1974 and 2007–2008 commodity booms. This strong demand was reflected in low real interest rates, a declining U.S. dollar, and strong GDP growth, and it contributed to the reduction in inventory levels that made commodity markets vulnerable to supply and demand shocks.

15 For a survey of forward linkages and feedback effects between money markets and commodity prices, particularly agricultural prices, see Ardeni & Freebairn (2002). 16 The category of lower middle income is as defined by the World Bank. In 2008, per-capita income in middleincome countries was $6,227, measured in purchasing-power-parity 2010 dollars.



107

6. CROSS-COMMODITY LINKAGES


The economics of substitution and complementarity in supply and demand generate crosscommodity linkages that create cascading effects of a boom or bust in one commodity market on another. Supply substitutability is most relevant for agricultural commodities because producers can choose which crop to plant on a fixed-acreage base. In contrast to agriculture, production of one mineral or energy commodity typically does not generally crowd out production of another. Commodity complements are also an important feature of agriculture, giving rise to strong cross-price elasticity relationships. For instance, increases in the price of corn affect the profitability, supply, and price of pork because feed accounts for more than 70% of the cost of raising a pig. In agriculture, petroleum commodities are important inputs to production directly through fuel use and indirectly through nitrogen fertilizer and pesticides produced with natural gas. During both the 1973–1974 and 2007–2008 commodity booms, prices of ammonia, nitrogen, potash, and phosphate fertilizers more than doubled, placing strong pressure on the cost of agricultural commodity production. The price of energy is also linked to all other commodities through transportation costs. Figure 7 shows the ocean freight index for bulk commodities constructed by Kilian (2009). It is evident that both the 1973–1974 and 2007–2008 commodity booms were associated with significant increases in freight rates, which Kilian interprets as indicating strong real economic activity (aggregate demand). The higher freight rates contributed to surges in delivered commodity prices. The use of food crops for biofuels is a relatively recent phenomenon and underscores the importance of cross-commodity linkages. Important countries in this regard are the United States, Brazil, and those of the EU. These countries account for 89% of global biofuel production, including ethanol and biodiesel. Policies in the United States and the EU have been criticized in particular because they promote the inefficient production of biofuels (primarily from corn in the United States and from rapeseed in the EU) through 100 80 60

Index

40 20 0 -20 -40 -60 -80 1973

1978

1983

1988

1993

1998

2003

2008

Figure 7 Real dry cargo index. From Kilian (2009). The index is calculated as deviations from the trend. 108

Carter

Rausser

Smith


subsidies, trade barriers, and mandated blending requirements. These policies draw corn and vegetable oils out of the food market channels and divert them into ethanol and biodiesel on a large scale. Another EU scheme, termed energy cropping, is growing in importance as a result of large subsidies. In this scheme, land is taken out of food production, is planted to maize or sugar beets, and is then harvested for biofuel use. The United States currently mandates 15.2 billion gallons of (ethanol equivalent) biofuel use in 2012, rising to 36 billion gallons by 2022. At the same time, the EU has mandated a 10% blend of biofuels into its fuel supply by 2020 and that 20% of its total energy mix come from renewable sources. Statistics for 200917 indicate that the United States produced approximately 10.6 billion gallons of ethanol and 544.2 million gallons of biodiesel, Brazil produced 6.6 billion gallons of ethanol and 405.5 million gallons of biodiesel, and the EU produced 978.2 million gallons of ethanol and 2.7 billion gallons of biodiesel. In the case of the United States and the EU, biofuels represent a relatively minor share of the overall domestic fuel demand: approximately 4% in the United States and 2% in the EU.18 Almost all U.S. ethanol is currently produced with corn, and by 2022, approximately 50% of the ethanol production is forecast to be based on corn. In 2010, more than one-third of the U.S. corn supply will be diverted into ethanol production. As noted above, this diversion has a significant impact on world corn prices. According to the FAO (2008), the increase in global corn demand in 2007 was approximately 40 million metric tons, and 75% of the growth in demand was attributable to ethanol production. The significant shift of corn into ethanol not only has drawn acreage out of wheat and soybeans but also has reduced available corn inventories. In crop year 2007–2008, U.S. corn acreage jumped 19% from the previous year (rising to 93.6 million acres), and at the same time, soybean acreage fell by 16% from 75.5 to 63.6 million acres. It is no wonder that soybean prices then surged, an indirect effect of U.S. ethanol policy on corn demand. Related markets also experienced huge increases in prices, particularly for fertilizers, which resulted in expanded production costs, justifying still higher commodity prices. With few exceptions, historical prices of grains have not been highly correlated with petroleum. But the FAO and others have recognized that commodities are now more closely tied than they have ever been, suggesting that agricultural commodity prices now move up and down with the prices of fossil-based fuels (Mallory et al. 2010). This perspective is based partly on the fact that the changing price of fuel could provide an incentive to move sugar and corn production into fuel channels. Ethanol in Brazil is produced primarily from sugar cane, and the new-linkage argument makes more sense for sugar cane than, say, for corn in the United States. For instance, if the world price of oil rises, pulling up the price of ethanol as a substitute, Brazil can divert sugar from the world sugar market into ethanol production. This will raise the price of sugar. The International Monetary Fund estimated that increased production of biofuels was perhaps the biggest factor in rising food prices in 2008 (Helbling et al. 2008). This effect occurred first through the direct effect on corn prices and second through the supply substitution linkages with other commodities such as soybeans and wheat. In congressional testimony, the Chairman of the Council of Economic Advisors reported that the global

17

See http://www.plateforme-biocarburants.ch/en/home/.

18

See http://www.eia.gov/emeu/international/contents.html.



109


increase in corn-based ethanol production accounted for approximately one-third of the increase in corn prices in 2007–2008 (Lazear 2008). Pindyck & Rotemberg (1990) claim that prices of seemingly unrelated commodities tend to move together. They propose but reject a common macroeconomic model to explain the observed comovement. As an alternative hypothesis, they suggest that speculators or irrational expectations may be able to explain such comovements. In later work Leybourne et al. (1994) and Ai et al. (2006) show that Pindyck & Rotemberg’s results can be accounted for without appeal to speculative notions or macroeconomics. In particular, Ai et al. attribute comovements to common tendencies in demand and supply factors (Section 3) that partially reflect cross-commodity linkages. They conclude that supply-side factors play the largest role in the observed price comovements, including inventory stockholding (Section 4).

7. POLICY RESPONSES Johnson (1975) and Chisholm & Tyers (1985) demonstrate that border restrictions by large trading countries have significant impacts on world commodity prices. They show that if a large country restricts trade to stabilize domestic prices, then world prices become more volatile. Such policies are thought to have aggravated commodity price volatility in the early 1970s (Vousden 1990) and again in 2007–2008 (Anderson & Nelgen 2010, Martin & Anderson 2011). There are many interesting examples of attempts by commodity-trading countries to restrict trade in the face of a volatile world commodity market. Both rich and poor countries have engaged in such conduct. The famous Russian grain robbery took place in 1972, when the Soviets purchased 30 million metric tons of grain from the United States over a short time period and the U.S. government willingly provided export subsidies on those grain sales. The USSR was trying to maintain its livestock population, even though it was experiencing poor harvests. Immediately following the Russian purchases, the price of wheat, feed grains, and soybeans escalated. The U.S. government then reacted by imposing export controls on soybeans in 1973, with the aim of trying to contain U.S. food price inflation. This move enraged Japanese soybean importers and reignited food security fears in Japan. In 1974 and 1975, the United States also imposed volume restrictions on exports of corn and wheat to the USSR. The 1973 soybean export embargo was a costly policy mistake because, although it benefited U.S. consumers in the short run, it hurt U.S. farmers by quickly lowering the domestic price and reducing incentives for expanded soybean production. The embargo also gave Japan (a major soybean importer) justification to keep its agricultural protectionism in place for many years. One of the arguments made by the Japanese was that they could no longer depend on the United States as a reliable exporter of food. In other words, the export controls had a lasting reputation effect. This effect was also felt domestically, as after the embargo U.S. farmers could no longer trust their government not to impose export controls, reducing access of U.S. farmers to world markets. At the same time, the U.S. export embargo raised world prices above their equilibrium and provided an incentive for expanded soybean production in Brazil. Now Brazil is the number one export competitor to the United States in soybeans. The 2007–2008 price boom produced similar trade policy mistakes. Governments around the world implemented measures to try to isolate domestic prices from world prices, 110

Carter

Rausser

Smith


exacerbating the problem. These policy measures included export bans, export taxes, and import subsidies. For instance, Argentina’s government led the world in terms of deliberately restricting agricultural exports in the face of higher commodity prices in 2007–2008. Argentina raised export taxes on soybeans from 35% to 45% (by tying the size of the tax rate to world prices) and restricted wheat exports. Farmers in that country responded with a series of nationwide strikes that closed highways for moving grain to export and caused scattered food shortages. These strikes ensured that exports were largely curtailed and that food distribution in Argentina was disrupted on a grand scale. As a result, world prices went even higher because Argentina is the third largest exporter of soybeans (behind the United States and Brazil) and a significant wheat exporter. The higher world prices (caused by Argentina’s export controls) served to further raise farmer opposition to the export controls in Argentina. The market distortion introduced by the export controls discouraged domestic production, and there were reported meat and dairy shortages in Argentina. In the face of high commodity prices, many other countries introduced similar policies to Argentina’s. The International Monetary Fund (2008) reported that at least 30 countries imposed bans or restrictions on food exports. For example, India banned the export of wheat, nonbasmati rice, and pulses, and imposed duties on exports of basmati rice, to try to contain domestic price inflation. Kazakhstan, one of the world’s top wheat exporters, banned exports. Vietnam restricted rice exports and announced plans to build a stateowned buffer stock of rice. India and Vietnam are the second and third largest rice exporters in the world. These export restrictions, and worries that Thailand would follow suit, were a significant source of skyrocketing rice prices in 2007–2008, threatening food security in large rice-importing nations such as the Philippines. Because a relatively small share of world rice production is traded internationally—approximately 7%—world prices were highly sensitive to the distortions imposed by India and Vietnam. India’s rice exports fell from 6.3 million metric tons in 2007 to an estimated 2.5 million metric tons in 2008, a dramatic decline. Martin & Anderson (2011) find that from 2005–2008 approximately 30% of the jump in rice prices and 25% of the rise in wheat prices were due to border controls aimed to stabilize domestic prices. This finding is consistent with the empirical evidence evaluated in Slayton (2009) pertaining to the rice market. Typically, the stated purpose of commodity export bans is to try to contain domestic price inflation. But this type of policy intervention can often make the problem worse because the policy is not directly targeted to the perceived problem. For instance, if the problem is domestic food price inflation for the very poor, then a policy targeted directly at that segment of the population is the most efficient (Corden 1971). This could involve food subsidies to the poor, targeting the problem directly. If an export ban is used instead, this introduces new and costly distortions on the production side. Domestic consumers may enjoy lower food prices in the short run, but production incentives are distorted away from the true value of the commodity; consumers ultimately suffer in the long run. Politicians are attracted to export control policies because they may offer short-term price relief to domestic consumers. However, they inevitably harm domestic farmers and lower the domestic supply response. In the case of many developing countries, these farmers are often among the poorest groups in the nation and are not as politically powerful as urban consumers. Moreover, these policies aggravate the global commodity markets by exporting additional price instability to world markets, providing a further foundation for world commodity price booms. www.annualreviews.org


111

So why do so many governments think it is a good idea to restrict exports? The answer is clearly a political economic one. Food has a special status, and governments in developing countries will go to great lengths to protect domestic (urban) consumers at the expense of rural farmers. Producers are typically not well enough organized to generate sufficient political opposition to these schemes.

8. CONCLUSION


Our analysis of two epic commodity booms and busts leads to five key conclusions. First, at the core of commodity booms and busts is a set of supply and demand shocks. Broad price booms do not occur without fundamental increases in demand or reductions in supply across a wide variety of commodities. Second, if an unexpected sequence of negative shocks causes inventories to be drawn down, then any subsequent shocks must be met by contemporaneous supply and demand adjustments. Even if these shocks appear minimal, they may cause large price responses because short-run demand and supply elasticities are small in absolute value for most commodities. Third, both the 1973–1974 and 2007–2008 booms were preceded by strong economic growth in lower-middle-income countries and by low real interest rates in rich countries and a resulting weak currency. These factors contributed to the tight supply-demand balance and drawdown in inventories that made markets vulnerable to shocks. Fourth, cross-commodity linkages through input substitutability and complementarity cause spillovers to a broader set of commodities than those affected directly by fundamental shock. Finally, policy responses such as export embargoes often exacerbate rather than mitigate booms and busts. We find little support for the argument that the 2007–2008 commodity boom was fundamentally different from the 1973–1974 experience (FAO 2008, Stoeckel 2008, Trostle 2008). Both events arose from temporary market discrepancies. Although a commodity boom and bust may not be predictable ex ante, we maintain that core factors should be recognizable when they occur. Furthermore, each of the three forces we identify (supply/demand, inventories, and the macroeconomic environment) and the linkages and policy responses that amplify them must be evaluated and assessed together in any attempt to explain booms and busts. Do we expect to see more or fewer commodity booms and busts in the coming years? By spring of 2011, many commodities were again approaching the peaks they had reached in 2008, which suggests that the answer to this question should be yes. Certainly, the market conditions that drove the 2007–2008 boom are again present: low grain inventories, strong growth in emerging economies, massive growth in the subsidized biofuel industry, low real rates of interest, relatively weak value of the dollar, and some degree of flexible commodity price overshooting. Predicting the frequency of future booms and busts thus requires information regarding the persistence of these market conditions. Relevant factors include globalization and increased international trade, increased regional specialization of agricultural production, increased biofuel production, supply constraints, growing population and urbanization, monetary policy, and global climate change. Supply and demand shocks are inevitable, so the key to answering this question lies in the implications of these factors for stockholding behavior. Globalization helps to mitigate booms and busts if trade is relatively free. When one producing country is hit by a supply disruption, other producers can usually fill the gap in world demand. However, as we discuss in Section 7, trade can easily be interrupted by the 112

Carter

Rausser

Smith


stroke of a pen. At precisely the time when the market would benefit from unrestricted trade, exporting countries may restrict trade to appease domestic consumers. For instance, Russia imposed a ban on grain exports in 2010 in reaction to a domestic drought. Under free trade, average inventory holdings would decrease with globalization. Because supply can be replenished by imports, there is less need for precautionary stocks. However, the threat of export embargoes from some countries may induce other exporters as well as importing countries to increase storage to mitigate booms and busts. Overall, globalization makes commodity markets less prone to booms and busts, but breakdowns in trade tend to worsen those booms and busts that will inevitably occur. Regional specialization in production may increase production efficiency, but it also increases volatility, as local weather shocks can have a larger effect on world production when the natural diversification benefit of production on many different continents is lowered. This increased volatility induces more stockholding, so it may not increase the frequency of booms and busts. Politics are again relevant, as producing countries may have an incentive to exert market power and withhold supply from world markets, exacerbating booms and busts. In turn, the possibility of trade restrictions would cause importing countries to expand their own production and hold more inventories, again mitigating the boom-and-bust cycles. Concerns about peak oil and the ability of agricultural production to keep up with demand will be recurring issues. What will happen to energy prices when the scarcity of fossil fuel reserves begins to bite? What will happen to food prices when subsidized biofuel production continues to expand or when water availability for agricultural use starts to decline due to increased demand arising from growing urbanization? The simple answer is that average commodity prices will increase. But a tight supply-demand balance will inevitably be temporary. Once markets see an imbalance between supply and demand, inventory holdings will increase, producers will invest in new technology, and consumers will substitute away from high-priced commodities. These actions will restore balance, which gives us no reason to expect more frequent large booms and busts. In spite of this conclusion, we are mindful that higher average prices may have devastating effects on many low-income and developing country consumers.

DISCLOSURE STATEMENT A.S. has been a consultant for the Commodity Futures Trading Commission since June 2009. The other authors are not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.

LITERATURE CITED Abbott P. 2011. Export restrictions as stabilization responses to food crises. Presented at Allied Soc. Sci. Assoc. Meet., Denver, CO Ai C, Chatrath A, Song F. 2006. On the comovement of commodity prices. Am. J. Agric. Econ. 88:574–88 Alston JM, Beddow JM, Pardey PG. 2009. Agricultural research, productivity, and food prices in the long run. Science 325:1209–10 Anderson K, Martin W, eds. 2006. Agricultural Trade Reform and the Doha Development Agenda. London/Washington, DC: Palgrave Macmillan/World Bank Anderson K, Martin W. 2011. Export restrictions and price insulation during commodity price booms. Am J. Agric. Econ. In press www.annualreviews.org


113


Anderson K, Nelgen S. 2010. How do governments respond to food price spikes? Lessons from the past. J. Int. Commer. Econ. Policy 1:1–21 Anderson RW, Danthine JP. 1980. Hedging and joint production: theory and illustrations. J. Finance 35:487–98 Ardeni PG, Freebairn J. 2002. The macroeconomics of agriculture. In Handbook of Agricultural Economics, Vol. 2A, ed. B Gardner, G Rausser, pp. 1455–85. Amsterdam: Elsevier/North-Holland Ardeni PG, Rausser GC. 1995. Alternative subsidy reduction paths: the role of fiscal and monetary policy linkages. In GATT Negotiations and the Political Economy of Policy Reform, ed. G Rausser, pp. 315–45. Berlin: Springer-Verlag Askari H, Cummings J. 1976. Agriculture Supply Response: A Survey of the Econometric Evidence. New York: Praeger Baffes J, Haniotis T. 2010. Placing the 2006/08 commodity price boom into perspective. Policy Res. Work. Pap. No. 5,371, World Bank Bessembinder H. 1992. Systematic risk, hedging pressure, and risk premiums in futures markets. Rev. Financ. Stud. 5:637–67 Blinder AS. 1982. The anatomy of double-digit inflation in the 1970s. In Inflation: Causes and Effects, ed. RE Hall, pp. 261–82. Chicago: Univ. Chicago Press Blinder AS, Rudd JB. 2008. The supply-shock explanation of the Great Stagflation revisited. Work. Pap. No. 14,563, Natl. Bur. Econ. Res. Bobenreith E, Bobenreith J, Wright BD. 2004. A model of supply of storage. Econ. Dev. Cult. Change 52:605–16 Bray M. 1981. Futures trading, rational expectations and the efficient markets hypothesis. Econometrica 49:575–96 Brennan D, Williams J, Wright BD. 1997. Convenience yield without the convenience: a spatialtemporal interpretation of storage under backwardation. Econ. J. 107:1009–22 Brennan MJ. 1958. The supply of storage. Am. Econ. Rev. 48:50–72 Brennan MJ, Schwartz ES. 1985. Evaluating natural resource investments. J. Bus. 58:135–57 Britto R. 1984. The simultaneous determination of spot and futures prices in a simple model with production risk. Q. J. Econ. 99:351–65 Brown LR. 1995. Who Will Feed China? Wake-Up Call for a Small Planet. New York: WW Norton & Co Cafiero C, Bobenrieth E, Bobenrieth J, Wright BD. 2009. The empirical relevance of the competitive storage model. J. Econom. 162:44–54 Carlson M, Khokher Z, Titman S. 2007. Equilibrium exhaustible resource price dynamics. J. Finance 62:1663–703 Carter CA, Janzen JP. 2009. The 2008 cotton price spike and extraordinary hedging costs. ARE Update 13(2):9–11. http://agecon.ucdavis.edu/extension/update/articles/v13n2_3.pdf Carter CA, Rausser GC, Schmitz A. 1983. Efficient asset portfolios and the theory of normal backwardation. J. Polit. Econ. 91:319–31 Carter CA, Revoredo-Giha CL. 2007. The Working curve and commodity storage under backwardation. Am. J. Agric. Econ. 89:864–72 Carter CA, Revoredo-Giha CL. 2009. Eastham’s commodity storage model in a modern context. Oxf. Econ. Pap. 61:801–22 Chen YC, Rogoff KS, Rossi B. 2010. Can exchange rates forecast commodity prices? Q. J. Econ. 125:1145–94 Chisholm AH, Tyers R. 1985. Agricultural protection and market insulation policies: applications of a dynamic multisectoral model. In New Developments in Applied General Equilibrium Analysis, ed. J Piggott, J Whalley, pp. 189–220. Cambridge, UK: Cambridge Univ. Press Chu K, Morrison TK. 1986. World non-oil primary commodity markets: a medium-term framework of analysis. Int. Monet. Fund Staff Pap. 33:139–84 114

Carter

Rausser

Smith


Cooper RN, Lawrence RZ. 1975. The 1972–75 commodity boom. Brookings Pap. Econ. Act. 1:671–723 Corden WM. 1971. The Theory of Protection. Oxford: Oxford Univ. Press Cox JC, Ingersoll JE, Ross SE. 1981. The relation between forward prices and futures prices. J. Financ. Econ. 9:321–46 Danthine J-P. 1978. Information, futures prices, and stabilizing speculation. J. Econ. Theory 17:79–98 De Roon FA, Nijman TE, Veld C. 2000. Hedging pressure effects in futures markets. J. Finance 55:1437–56 Deaton A, Laroque G. 1992. On the behavior of commodity prices. Rev. Econ. Stud. 59:1–23 Deaton A, Laroque G. 1995. Estimating a nonlinear rational expectations commodity price model with unobservable state variables. J. Appl. Econ. 10(Suppl. 1):9–40 Deaton A, Laroque G. 1996. Competitive storage and commodity price dynamics. J. Polit. Econ. 104:896–923 Dornbusch R. 1976. Expectations and exchange rate dynamics. J. Polit. Econ. 84:1161–76 Eastham JK. 1939. Commodity stocks and prices. Rev. Econ. Stud. 6:100–10 Ehrlich P, Erhlich A. 1990. The Population Explosion. New York: Simon and Schuster Epstein G. 2008. Commodities: Who’s behind the boom? Barron’s, March 31 Feder G, Just RE, Schmitz A. 1980. Futures markets and the theory of the firm under price uncertainty. Q. J. Econ. 94:317–28 Food Agric. Organ. (FAO). 2008. Soaring food prices: facts, perspectives, impacts and actions required. Presented at High-Level Conference on World Food Security: The Challenges of Climate Change and Bioenergy, Rome Frankel JA. 1986. Expectations and commodity price dynamics: the overshooting model. Am. J. Agric. Econ. 68:344–48 Frankel JA. 1995. Financial Markets and Monetary Policy. Cambridge, MA: Mass. Inst. Technol. Press Frankel JA. 2008a. The effect of monetary policy on real commodity prices. In Asset Prices and Monetary Policy, ed. J Campbell, pp. 291–327. Chicago: Univ. Chicago Press Frankel JA. 2008b. An explanation for soaring commodity prices. VOX, March 25. http://www. voxeu.org/index.php?q=node/1002 Frankel JA, Hardouvelis GA. 1985. Commodity prices, monetary supplies and Fed credibility. J. Money Credit Bank. 17:425–38 Franken J, Garcia P, Irwin SH. 2009. Is storage at a loss merely an illusion of spatial aggregation? J. Agribus. 27:65–84 Frum D. 2000. How We Got Here: The ‘70s. New York: Basic Books Garber PM. 1989. Tulipmania. J. Polit. Econ. 97:535–60 Garber PM. 1990. Famous first bubbles. J. Econ. Perspect. 4:35–54 Gilbert CL. 2010. How to understand high food prices. J. Agric. Econ. 61:398–425 Gorton G, Hayashi F, Rouwenhorst G. 2007. The fundamentals of commodity futures returns. Work. Pap. 13,249, Natl. Bur. Econ. Res. Grossman SJ. 1977. The existence of futures markets, noisy rational expectations and informational externalities. Rev. Econ. Stud. 64:431–49 Grossman SJ, Stiglitz J. 1976. Information and competitive price systems. Am. Econ. Rev. 66:246–53 Gustafson RL. 1958. Carryover levels for grains: a method for determining amounts that are optimal under specified conditions. Tech. Bull. No. 1,178, U.S. Dep. Agric. Hamilton JD. 2009. Causes and consequences of the oil shock of 2007–2008. Brookings Pap. Econ. Act. pp. 215–61 Helbling T, Mercer-Blackman V, Cheng K. 2008. Riding a wave. Finance Dev. 45:1 Hicks J. 1974. The Crisis in Keynesian Economics. Oxford: Basil Blackwell Hirshleifer D. 1988. Risk, futures pricing, and the organization of production in commodity markets. J. Polit. Econ. 96:1206–20 www.annualreviews.org


115


Hotelling H. 1931. Economics of exhaustible resources. J. Polit. Econ. 39:137–75 Hughes JE. Knittel CR. Sperling D. 2008. Evidence of a shift in the short-run price elasticity of gasoline demand. Energy J. 29:113–34 Int. Food Policy Res. Inst. (IFPRI). 2008. High food prices: the what, who, and how of proposed policy actions. Policy Brief, IFPRI Int. Monet. Fund. 2008. Food and fuel prices—recent developments, macroeconomic impact, and policy responses. http://www.imf.org/external/np/pp/eng/2008/063008.pdf Int. Organ. Secur. Comm. (IOSCO). 2009. Task force on commodity futures markets: final report. https://www.iosco.org/library/pubdocs/pdf/IOSCOPD285.pdf Irwin SH, Sanders DR. 2011. Index funds, financialization, and commodity futures markets. Appl. Econ. Perspect. Policy 33:1-31 Johnson DG. 1975. World agriculture, commodity policy and price variability. Am. J. Agric. Econ. 57:823–28 Just RE, Rausser GC. 2001. Conceptual foundations of expectations and implications for estimation of risk behavior. In A Comprehensive Assessment of the Role of Risk in U.S. Agriculture, ed. RE Just, RD Pope, pp. 53–80. Boston: Kluwer Acad. Kaldor N. 1934. The equilibrium of the firm. Econ. J. 44:60–76 Kaldor N. 1939. Speculation and economic stability. Rev. Econ. Stud. 7:1–27 Keynes JM. 1930. A Treatise on Money. New York: Harcourt, Brace Khan MS. 2009. The 2008 oil price “bubble.” Policy Brief PB09-19, Peterson Inst. Int. Econ. Kilian L. 2009. Not all oil price shocks are alike: disentangling demand and supply shocks in the crude oil market. Am. Econ. Rev. 99:1053–69 Kindleberger CP. 1978. Manias, Panics and Crashes: A History of Financial Crises. London: Macmillan. 3rd ed. Labys WC, Thomas HC. 1975. Speculation, hedging and commodity price behavior: an international comparison. Appl. Econ. 7:287–301 Lai C, Hu S, Wang V. 1996. Commodity price dynamics and anticipated shocks. Am. J. Agric. Econ. 78:982–90 Lazaer E. 2008. Testimony before Senate Foreign Relations Committee Hearing on “Responding to the Global Food Crisis,” May 14, http://georgewbush-whitehouse.archives.gov/cea/lazear20080514. html Leybourne SY, Lloyd TA, Reed GV. 1994. The excess comovement of commodity prices revisited. World Dev. 22:1747–58 Lin C, Wagner G. 2007. Steady-state growth in a Hotelling model of resource extraction. J. Environ. Econ. Manag. 54:68–83 Litzenberger R, Rabinowitz N. 1995. Backwardation in oil futures markets: theory and empirical evidence. J. Finance 50:151745 Lucas RE, Sargent TJ. 1981. Rational Expectations and Econometric Practice. London: Allen & Unwin Mallory ML, Hayes DJ, Irwin SH. 2010. How market efficiency and the theory of storage link corn and ethanol markets. Work. Pap., Univ. Ill. McKinnon RI. 1967. Futures markets, buffer stocks, and income stability for primary producers. J. Polit. Econ. 75:844–61 Meadows DH, Meadows DL, Randers J, Behrens WW III. 1972. The Limits to Growth. A Report for the Club of Rome’s Project on the Predicament of Mankind. New York: Universe Books Miranda MJ, Helmberger PG. 1988. The effects of commodity price stabilization programs. Am. Econ. Rev. 78:46–58 Mitchell D. 2008. A note on rising food prices. Policy Res. Work. Pap. No. 4,682, World Bank Moschini G, Hennessy DA. 2001. Uncertainty, risk aversion, and risk management for agricultural producers. In Handbook of Agricultural Economics, Vol. 1, ed. B Gardner, G Rausser, pp. 88–153. Amsterdam: Elsevier/North-Holland 116

Carter

Rausser

Smith


Moschini G, Lapan H. 1995. The hedging role of options and futures under joint price, basis, and production risk. Int. Econ. Rev. 36:1025–49 Muth JF. 1961. Rational expectations and the theory of price movements. Econometrica 29:315–35 Nerlove M. 1958. The Dynamics of Supply: Estimation of Farmers’ Response to Price. Baltimore: Johns Hopkins Univ. Press Newbery JM, Stiglitz JE. 1981. The Theory of Commodity Price Stabilization. Oxford: Oxford Univ. Press Newbery JM, Stiglitz JE. 1982. The choice of techniques and the optimality of market equilibrium with rational expectations. J. Polit. Econ. 90:223–46 Okun AM. 1975. Inflation: its mechanics and welfare costs. Brookings Pap. Econ. Act. pp. 351–401 Park A. 2006. Risk and household grain management in developing countries. Econ. J. 116:1088–115 Paul AB. 1970. The pricing of binspace: a contribution to the theory of storage. Am. J. Agric. Econ. 52:1–12 Phelps ES. 1978. Commodity-supply shocks and full employment monetary policy. J. Money Credit Bank. 10:206–21 Piesse J, Thirtle C. 2009. Three bubbles and a panic: an explanatory review of recent food commodity price events. Food Policy 34:119–29 Pindyck RS. 1980. Uncertainty and exhaustible resource markets. J. Polit. Econ. 88:1203–25 Pindyck RS, Rotemberg JJ. 1990. The excess co-movement of commodity prices. Econ. J. 100:1173–89 Radetzki M. 2006. The anatomy of three commodity booms. Resour. Policy 31:56–64 Rausser GC. 1980. Hedging and joint production. J. Finance 35:498–501 Rausser GC, ed. 1982. New Directions in Econometric Modeling and Forecasting in U.S. Agriculture. New York: Elsevier/North-Holland Rausser GC. 1985. Macroeconomics and U.S. agricultural policy. In U.S. Agricultural Policy: The 1985 Farm Legislation, ed. B Gardner, pp. 207–52. Washington, DC: Am. Enterp. Inst. Rausser GC, Adams G, Montgomery W, Smith A. 2004. The social costs of an MTBE ban in California. Giannini Found. Rep. No. 349 Rausser GC, Chalfant JA, Love HA, Stamoulis KG. 1986. Macroeconomic linkages, taxes, and subsidies in the U.S. agricultural sector. Am. J. Agric. Econ. 68:399–412 Rausser GC, Walraven N. 1990. Linkages among commodity futures markets and dynamic welfare analysis. Rev. Econ. Stat. 72:631–39 Reinhart CM, Rogoff KS. 2009. This Time Is Different: Eight Centuries of Financial Folly. Princeton, NJ: Princeton Univ. Press Roberts MJ, Schlenker W. 2010. Identifying supply and demand elasticities of agricultural commodities: implications for the US ethanol mandate. Work. Pap. No. 15,921, Natl. Bur. Econ. Res. Romer PM. 1986. Increasing returns and long-run growth. J. Polit. Econ. 94:1002–37 Routledge BR, Seppi DJ, Spatt CS. 2000. Equilibrium forward curves for commodities. J. Finance 55:1297–338 Saghaian SH, Reed MR, Marchant MA. 2002. Monetary impacts and overshooting of agricultural prices in an open economy. Am. J. Agric. Econ. 84:90–103 Sephton PS. 1988. World non-oil primary commodity markets: comment on Chu and Morrison. Staff Pap. Int. Monet. Fund 35:198–203 Schuh GE. 1974. The exchange rate and U.S. agriculture. Am. J. Agric. Econ. 56:1–13 Shiller RJ. 2005. Irrational Exuberance. Princeton, NJ: Princeton Univ. Press. 2nd ed. Simmons M. 2005. Twilight in the Desert: The Coming Saudi Oil Shock and the World Economy. New York: Wiley Slade ME, Thille H. 2009. Whither Hotelling: tests of the theory of exhaustible resources. Annu. Rev. Resour. Econ. 1:239–60 Slayton T. 2009. Rice crisis forensics: how Asian governments carelessly set the world rice market on fire. Work. Pap. No. 163, Cent. Global Dev. www.annualreviews.org


117


Smith JL. 2009. World oil: market or mayhem? J. Econ. Perspect. 23:145–64 Stamoulis KG, Rausser GC. 1988. Overshooting of agricultural prices. In Macroeconomics, Agriculture, and Exchange Rates, ed. PL Paarlberg, RG Chambers, pp. 163–89. Boulder, CO: Westview Stein JL. 1992. Price discovery processes. Econ. J. 68(Suppl. 1):34–45 Stiglitz JE. 1990. Symposium on bubbles. J. Econ. Perspect. 4:13–18 Stoeckel A. 2008. High food prices: causes, implications and solutions. Publ. No. 08/100, Rural Ind. Res. Dev. Corp., Aust. Gov. Telser LG. 1958. Futures trading and the storage of cotton and wheat. J. Polit. Econ. 66:233–55 Tirole J. 1985. Asset bubbles and overlapping generations. Econometrica 53:1071–100 Trostle R. 2008. Global agricultural supply and demand: factors contributing to the recent increase in food commodity prices. WRS-0801, Econ. Res. Serv., U.S. Dep. Agric. Vousden N. 1990. The Economics of Trade Protection. Cambridge, UK: Cambridge Univ. Press Wall Str. J. Editorial: commodity break. Wall Str. J., March 24, p. A14 Weitzman MC, Zeckhauser RJ. 1975. The optimal consumption of depletable natural resources. Q. J. Econ. 89:371–92 Wick K, Bulte E. 2009. The curse of natural resources. Annu. Rev. Resour. Econ. 1:139–55 Williams JB. 1936. Speculation and the carryover. Q. J. Econ. 50:436–55 Williams JC, Wright B. 1989. A theory of negative prices for storage. J. Futures Mark. 9:1–13 Williams JC, Wright B. 1991. Storage and Commodity Markets. Cambridge, UK: Cambridge Univ. Press Working H. 1949. The theory of price of storage. Am. Econ. Rev. 39:1254–62 Wray LR. 2008. The commodities market bubble: money manager capitalism and the financialization of commodities. Public Policy Brief No. 96, Levy Econ. Inst. Wright BD. 2011. The economics of grain price volatility. Appl. Econ. Perspect. Policy 33:32–58

118

Carter

Rausser

Smith

a

1973–1974

500


Real Price Index (Jan 72 = 100)

450 400 350 Energy 300

Cereals

250

Veg Oil Softs

200

Metals 150

Livestock

100 50 0 Jan-72

b

Jan-73

Jan-74

Jan-76

Jan-75

2007–2008


300

250

200

Energy Cereals

150

Veg Oil Softs Metals

100

Livestock 50

0 Jan-05

Jan-06

Jan-07

Jan-08

Jan-09

Jan-10

Figure 1 Two booms and busts. Commodity prices were obtained from the International Monetary Fund and were deflated by the U.S. consumer price index, excluding food and energy. The metals index represents industrial metals like copper, lead, tin, nickel, and aluminum. Data from the International Monetary Fund and the Commodity Research Bureau.



C-1

800



700

600

500

Energy Cereals

400

Veg Oil Softs Metals

300

Livestock 200

100

0 Jan-57

Jan-62

Jan-67

Jan-72

Jan-77

Jan-82

Jan-87

Jan-92

Jan-97

Jan-02

Jan-07

Figure 2 Real-commodity prices, 1957–2010. Data from the International Monetary Fund and the Commodity Research Bureau.

C-2

Carter

Rausser

Smith

Wheat 6.5

5.5

6.0

5.0

1971-1976 2004-2010

March Price

March Price

Corn 6.0

4.5

5.5

1971-1977 2004-2010

5.0

4.0

4.5 0

0.1

0.2

0.3

0.4

0.5

0

0.1

Global Stocks/Use

0.2

6.5

6.0

6.0

5.5 1971-1976

5.0

2004-2010

4.5

0.5

5.5 1971-1976

5.0

2004-2010

4.5

4.0

4.0 0

0.1

0.2

0.3

0.4

0.5

0

0.1

Global Stocks/Use

0.2

0.3

0.4

0.5

Global Stocks/Use (excl. China)

Copper

Crude Oil

4.5

3.5 3

3.5 1971-1976

3

2004-2009

March Price

4

March Price

0.4

Rice

6.5

April Price (Thai)

March Price

Coon Annu. Rev. Resour. Econ. 2011.3:87-118. Downloaded from www.annualreviews.org by University of California - Berkeley on 10/31/11. For personal use only.

0.3

Global Stocks/Use

2.5 2 1.5

1974-1976

1

2004-2010

2.5

0.5 2 0

0.1

0.2

0.3

0.4

0

0.5

0

Global Stocks/Use

0.1

0.2

0.3

0.4

0.5

US Stocks/Use

Figure 5 Real prices and ending stocks to use (1960–2010). Prices were measured in March for corn, cotton, and wheat and in April for rice. We use U.S. prices for corn, cotton, wheat, copper, and crude oil. We use Thai prices for rice. For each series, we plot the log of the price minus the log of the all-items consumer price index. The two boom-and-bust periods are highlighted in green for 1971–1976 and in blue for 2004–2010. All stocks-to-use ratios are detrended by a linear trend. We exclude China from rice stocks to use. Data from the Commodity Research Bureau, U.S. Department of Agriculture, and International Monetary Fund.



C-3

2007

1973

14

12


Growth Rate

10 8

6

4

2

0 1965

1970

1975

1980

OECD Inﬂaon

1985 World GDP

1990

1995

2000

2005

2010

Lower Middle Income GDP

Figure 6 World GDP growth and inflation. Inflation is for high-income OECD countries, and GDP growth is for the world economy, in real terms. Original data from World Bank Indicators (http://data.un.org).

C-4

Carter

Rausser

Smith

Annual Review of Resource Economics Volume 3, 2011

Contents


Prefatory Plowing Through the Data Yair Mundlak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Methods for Performance Evaluations and Impact Measurement Green National Income and Green National Product John M. Hartwick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Behavior, Robustness, and Sufficient Statistics in Welfare Measurement Richard E. Just. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 The Challenges of Improving the Economic Analysis of Pending Regulations: The Experience of OMB Circular A-4 Art Fraas and Randall Lutter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 The Economics of Commodity Markets and Food Supply Chains Commodity Booms and Busts Colin A. Carter, Gordon C. Rausser, and Aaron Smith . . . . . . . . . . . . . . . . 87 Food Quality: The Design of Incentive Contracts Rachael E. Goodhue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Nutritional Labeling and Consumer Choices Kristin Kiesel, Jill J. McCluskey, and Sofia B. Villas-Boas. . . . . . . . . . . . . 141 The Economics and Policy of Natural Resources Efficiency Advantages of Grandfathering in Rights-Based Fisheries Management Terry Anderson, Ragnar Arnason, and Gary D. Libecap . . . . . . . . . . . . . 159 Game Theory and Fisheries Ro¨gnvaldur Hannesson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

viii

Natural Resource Management: Challenges and Policy Options Jessica Coria and Thomas Sterner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 The New Economics of Evaluating Water Projects Per-Olov Johansson and Bengt Kristro¨m . . . . . . . . . . . . . . . . . . . . . . . . . 231 The Economics of Human and Environmental Health Risks Management of Hazardous Waste and Contaminated Land Hilary Sigman and Sarah Stafford . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255


The Economics of Infection Control Mark Gersovitz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 The Economics of Natural Disasters Derek Kellenberg and A. Mushfiq Mobarak . . . . . . . . . . . . . . . . . . . . . . . 297 Valuing Mortality Risk Reductions: Progress and Challenges Maureen Cropper, James K. Hammitt, and Lisa A. Robinson. . . . . . . . . . 313 Environmental Economics and Policy Pricing Nature Edward B. Barbier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337 The Economics of Non-Point-Source Pollution Anastasios Xepapadeas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355 Microeconometric Strategies for Dealing with Unobservables and Endogenous Variables in Recreation Demand Models Klaus Moeltner and Roger von Haefen. . . . . . . . . . . . . . . . . . . . . . . . . . . 375 The Environment and Trade Larry Karp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397 The Social Cost of Carbon Richard S.J. Tol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419 Corporate Average Fuel Economy Standards and the Market for New Vehicles Thomas Klier and Joshua Linn. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445 Errata An online log of corrections to Annual Review of Resource Economics articles may be found at http://resource.annualreviews.org

Contents

ix