The Economics of Commodity Markets
May 15 and 16, 2015
Ke Tang, Tsinghua University, and Haoxiang Zhu, MIT and NBER
This paper proposes and tests a theory of using commodities as collateral for financing. In the presence of capital control and financial frictions, financial investors import commodities and pledge them as collateral to capture the higher expected return in the importing country. The collateral demand for commodities increases commodity prices globally; it also increases futures risk premium in the importing country but reduces that in the exporting country. Tang and Zhu test the theoretical predictions on eight commodities in China and developed markets. The evidence supports their theory. The results suggest that collateral demands can explain up to 11.9%-15.0% of the price increase of major industrial metals since 2007. Overall, the researchers' theory and evidence complement the theory of storage and provide new insights to the financialization of commodity markets.
Darien Huang, Wharton School, University of Pennsylvania
Huang shows that the ratio of gold to platinum prices (GP) reveals variation in risk and proxies for an important economic state variable. GP predicts future stock returns in the time-series and explains variation in average stock returns in the cross-section. GP outperforms existing predictors and similar patterns are found in international markets. Contrary to conventional views of gold as a hedge, gold prices fall in recessions, albeit by less than platinum prices. GP is persistent and significantly correlated with option-implied tail risk measures. An equilibrium model featuring recursive preferences, time-varying tail risk, and shocks to preferences for gold and platinum can quantitatively account for the asset pricing dynamics of equity, gold, and platinum markets, rationalize the return predictability, and explain why gold prices fall in bad times.
Itay Goldstein, University of Pennsylvania, and Liyan Yang, University of Toronto
Goldstein and Yang theoretically study how commodity financialization affects trading behavior, prices and welfare through affecting risk sharing and price discovery in futures markets. In their model, the general equilibrium feature makes financial traders either provide or demand liquidity in the futures market, depending on the information environment. Consistent with recent evidence, commodity financialization reduces the futures price bias through broadening risk sharing and injecting information into the market. Each financial trader loses and final-end commodity consumers benefit in the process of commodity financialization. Commercial hedgers can either lose or win, and their welfare improves with commodity financialization only when the number of active financial traders takes intermediate values.
Akshaya Jha, Stanford University, and Frank A. Wolak, Stanford University and NBER
With risk neutral traders and zero transactions costs, the expected value of the difference between the current forward price and the spot price of a commodity at the delivery date of the forward contract should be zero. Accounting for the transactions costs associated with trading in these two markets invalidates this result. Jha and Wolak develop statistical tests of the null hypothesis that profitable trading strategies exploiting systematic differences between spot and forward market prices exist in the presence of trading costs. The researchers implement these tests using the day-ahead forward and real-time locational marginal prices from California's wholesale electricity market and use them to construct an estimate of the variable cost of trading in this market. During their sample period, the authors observe the introduction of convergence bidding, which was aimed at reducing the costs associated with exploiting differences between forward and spot prices. Jha and Wolak's measures of trading costs are significantly smaller after the introduction of convergence bidding. Estimated trading costs are lower for generation nodes relative to non-generation nodes before explicit virtual bidding and trading costs fell more for non-generation nodes after explicit virtual bidding, eliminating any difference in trading costs across the two types of nodes. The researchers also present evidence that the introduction of convergence bidding reduced the total amount of input fossil fuel energy required to generate the thermal-based electricity produced in California and the total variable of costs of producing this electrical energy. Taken together, these results demonstrate that purely financial forward market trading can improve the operating efficiency of short-term commodity markets.
Thomas Covert, University of Chicago
Little is known about how firms learn to use new technologies. Using novel data on inputs, profits, and information sets, Covert studies how oil companies learned to use hydraulic fracturing technology in North Dakota between 2005-12. Firms only partially learned to make profitable input choices, capturing just 60% of possible profits in 2012. To understand why, the researcher estimates a model of input use under technology uncertainty. Firms chose fracking inputs with higher expectations but lower uncertainty about profits, consistent with passive learning but not active experimentation. Most firms over-weighed their own information. These results provide evidence of impediments to learning.
Harrison Hong, Princeton University and NBER; Áureo de Paula, University College London; and Vishal Singh, New York University
Hoarding by large speculators is often blamed for contributing to commodity market panics and bubbles. Using supermarket scanner data on U.S. household purchases during the 2008 Rice Bubble, Hong, de Paula, and Singh show that hoarding is in fact more systemic, affecting even households who have no resale motive. Export bans led to a spike in prices worldwide in the first half of 2008, which spilled over into U.S. markets. Anticipating shortages, U.S. households with previous purchases of rice, especially those of Asian ethnicity, nearly doubled their buying around the peak of the bubble. The researchers document transmission mechanisms through over-extrapolation from high prices and contagion, as many households bought rice for the first and last time during the bubble.
Davidson Heath, University of Southern California
In this paper, Heath constructs a macro-finance model for commodity futures. Model estimates suggest a feedback relationship between crude oil prices and U.S. real activity. Moreover, the channel from real activity to oil prices is unspanned - meaning not identified in current futures prices - consistent with oil futures as a hedge asset against supply shocks. Relative to a benchmark spanned-risk model, incorporating unspanned real activity raises the volatility of the estimated oil risk premium tenfold and raises the value of real options by 35 to 400%.
Sylvain Leduc, Federal Reserve Bank of San Francisco; Kevin Moran, Department of Economics, Universite Laval; and Robert Vigfusson, Federal Reserve Board
Leduc, Moran, and Vigfusson examine the role of learning in accounting for the movements in oil price futures during the 2000s, a period during which the oil market experienced important changes. The researchers show that a simple unobserved component model in which investors must form beliefs about whether the source of oil price movements is transitory or permanent accounts remarkably well for the fluctuations in oil price futures. The authors' simple framework notably accounts for the relatively slow increase in futures prices at the beginning of the past decade and their unprecedented run-up between 2004 and 2008. Even during the first half of 2008, a period during which oil prices reached historic highs, the model predicts a level of futures prices that is broadly in line with the data. The researchers' estimates suggest that, through learning, investors revised up the contribution of permanent shocks to the variance of oil prices throughout this period. Using a DSGE model in which oil is storable and used in production, they then show that this learning process may have significantly muted the effects of oil shocks on the economy during that period.
Erik P. Gilje, University of Pennsylvania; Robert C. Ready, University of Rochester; and Nikolai Roussanov, University of Pennsylvania and NBER
Gilje, Ready, and Roussanov use evidence from asset price data to quantify the contribution of shale oil to the U.S. economy. Equity market valuations of firms engaged in shale oil extraction reflect the market's expectations about the future growth in shale oil supply and its potential for raising aggregate productivity in the U.S. economy. Authors pursue two complementary methods to estimate the value from shale oil discoveries. First, they examine returns on an index of shale oil producers orthogonalized with respect to oil prices and industry-wide return controls to extract an empirical measure of shale-specific productivity innovations. Second, they use the cross-section of stock returns that captures the variation in industries' responses to shale-specific news announcements to construct a shale-factor mimicking portfolio. While the two methods produce a wide range of estimates, overall they suggest that approximately 20%-25% of the increase in aggregate U.S. equity market capitalization since 2009 can be attributed to shale oil, corresponding to $2.5 to $3 trillion in value.
Rabah Arezki, International Monetary Fund; Valerie A. Ramey, University of California, San Diego and NBER; and Liugang Sheng, Chinese University of Hong Kong
This paper explores the effect of news shocks on the current account and other macroeconomic variables using worldwide giant oil discoveries as a directly observable measure of news shocks about future output the delay between a discovery and production is on average four to six years. Arezki, Ramey, and Sheng first present a two-sector small open economy model in order to predict the responses of macroeconomic aggregates to news of an oil discovery. They then estimate the effects of giant oil discoveries on a large panel of countries. The researchers' empirical estimates are consistent with the predictions of the model. After an oil discovery, the current account and saving rate decline for the first five years and then rise sharply during the ensuing years. Investment rises robustly soon after the news arrives, while GDP does not increase until after five years. Employment rates fall slightly for a sustained period of time.