ABSTRACT. We document that localized policies designed to mitigate climate risk can lead to regulatory arbitrage by firms, resulting in unintended consequences. Using detailed plant level data, we investigate the impact of the most extensive regional climate policy in the United States, the California cap-and-trade program, on corporate real activities such as greenhouse gas emissions and plant ownership. We show that industrial plants governed by the policy reduce emissions in California when the parent company is financially constrained, but that these firms internally reallocate their emissions to plants located in other states. Similarly, constrained firms are more likely to reduce ownership in Californian plants and increase ownership in plants outside California. In contrast, unconstrained firms generally do not adjust plant emissions and ownership either in California or in other states. Overall, firms do not reduce their total emissions when part of their assets are affected by the regulation, but in fact increase them if financially constrained. The results document real spillover effects stemming from resource reallocations by constrained firms to avoid regulatory costs, undermining the effectiveness of localized policies. Our study has important implications for the current debate on global climate policy agreements.

ABSTRACT. We estimate income elasticities of demand for metal and energy commodities both at aggregate levels and individually based on an autoregressive distributed lag model and 1965-2017 data. The countries’ growth orientation, technological change, population concentration, and ex-change rate movements were captured through control variables. We find that as income grows, aggregate commodity consumption increases at a decelerating rate, thus rendering support to the plateauing hypothesis. While the income elasticities, evaluated at median incomes, are close to unity for both groups, adjustment rates to long-run equilibrium are higher for metals than energy. At an individual commodity level, the results were mixed, consistent with the literature, most likely reflecting the presence of strong substitutability among commodities. The control variables (investment to GDP ratio, population density, and time trend as a proxy for technology) matter for energy but not for metals. Robustness checks on data and estimation procedure did not alter the key results of the paper.

ABSTRACT. We examine the effects of political uncertainty on commodity markets from both theoretical and empirical points of view. Our theoretical model shows that political uncertainty on the demand (supply) side has a negative (positive) impact on commodity prices, and a positive (negative) impact on convenience yields and risk premiums. To test our model, we construct a comprehensive sample of 87 commodities across 12 countries over the 1960-2017 period. Consistent with our model predictions, commodity prices decline by 6.4% and convenience yields increase by 1.8% in the quarter leading up to U.S. presidential elections as the proxy for political uncertainty on the demand side. An opposite result is obtained for political uncertainty on the supply side.

ABSTRACT. We consider the discrete time optimal control problem for a battery in the intraday market which is linked to a photovoltaic power station. We discuss recent Regression Monte Carlo (RMC) methods to solve the stochastic backward dynamic programming equations arising in the problem. Introducing two exogenous processes, the electricity price and the photovoltaic infeed process and a two dimensional control which determines the endogenous state of charge process, the resulting three-dimensional state space restricts the range of RMC methods to approaches including multivariate regression of the value function using all state variables. This leads to two problems: Firstly, the curse of dimensionality makes a suitable choice of the approximation space difficult. Secondly, the control affects the endogenous state of charge process which is delicate in backward induction approaches. Solutions for this comprise resimulation and backward construction of the endogenous process. While the former is numerically expensive the latter introduces some bias leading to suboptimal control. We propose adaptive model selection methods from the statistical learning literature for the optimal choice of basis functions which reduces the complexity of the control optimization problem and significantly reduces the running time of the algorithm while improving the accuracy for small simulation budgets. Furthermore, we introduce a control map approximation method using Multivariate Adaptive Regression Splines (MARS), which further reduces the running time of the algorithm with negligible loss of accuracy and allow for an efficient nested simulations approach.

ABSTRACT. Optimal management in the energy sector is getting increasingly sophisticated. Technical progress has recently demonstrated that Battery Energy Storage Systems (BESS) represent a technology which is becoming an increasingly viable option in the power sector. Nevertheless, the optimal operational management of BESS is crucial for the economical justification of their installation due to high costs. In an ideal situation, they are recovered by positive effects resulting from load shift / shaving of price spikes. However, this presumes an optimized integration of the energy trade (usually managed with forward contracts) and BESS operations, which takes into account the most important costs factor, the so-called deep discharge costs. In this picture a significant complication is the (increasing) presence of intermittent power generation, represented by Renewable Energy Sources(RES). In this paper, we address an algorithmic methodology for an optimized integration of BESS and energy trade in the presence of RES. In our approach, we apply a novel subgradient-based method to solve stochastic optimal control problems arising in this context. Furthermore, we show how a duality-based technique can be used to confirm the quality and accuracy of the numerical solution.

ABSTRACT. We model the risk management problem of an electricity storage operator who participates in a wholesale electricity market and hedges risk by trading currency and power futures contracts. Our model considers three types of risks: operational forwards risk due to future supply uncertainty, exchange rate risk for operations and trading in different currencies, and profit risks due to power price variability. We model the problem as a multistage stochastic programming problem and propose a sequential solution approach to handle the high complexity of the optimization problem. Our contribution is three-fold: first, we show how currency risk and currency derivatives can be included in real option models of electricity storage; second, we introduce variables for accurate replication of the cash flow structure from a portfolio of financial contracts; and third, we compare optimization using a risk measure with often used simple hedging strategies. We quantify, for the case of a Norwegian hydropower producer, the reduction in risk through currency hedging where there is currency risk. We find that currency hedging leads to a moderate decrease in profit risk, and that including monthly power futures in a hedging strategy allows precision hedging that can contribute to substantial reductions in risk.

ABSTRACT. We show that the oil market has become closer to “one great pool,” in the sense that price differentials between crude oils of different qualities have generally become smaller over time. We document, in particular, that many of these quality-related differentials experienced a major structural break in or around 2008, after which there was a marked reduction in their means and, in many cases, volatilities. Several factors explain these shifts, including a growing ability of the global refinery sector to process lower-quality crude oil and the U.S. shale boom, which has unexpectedly boosted the supply of high-quality crude oil. Differentials between crude oils of similar quality in general did not experience breaks in or around 2008, although we do find evidence of breaks at other times. We also show that these structural breaks can affect tests of stationarity for many price differentials.

ABSTRACT. Spatial price integration is extensively studied in commodity markets as a means of examining the degree of integration between regions of a geographically diverse market. In this paper, we provide an analysis of price integration as a function of underlying transportation network features. Results include a price decomposition which explicitly isolates the influences of market forces (supply and demand), transportation costs and congestion surcharge on price integration. Using this price decomposition, we develop a novel methodology that captures price shocks indicative of structural disruptions in the underlying network using pricing data alone. Applying the methodology to gasoline prices in the South-Eastern US, we find the methodology is able to capture the effects of well-documented network disruptions, and evidence of capacity-driven price disruptions during normal operations.

ABSTRACT. We investigate the potential convergence of natural gas prices at one hub and several city gates in the United States. First, we estimate the fundamental natural gas price component. Second, we gauge if U.S. regional natural gas prices follow the law of one price. Our findings confirm the difference between West and East natural gas prices, but also highlight discrepancies between West and other regional natural gas prices. Third, we measure the distance between the fundamental gas price component and both Henry hub and city gate prices. The proximity/remoteness of regional gas prices with the fundamental gas price component supports market integration/segmentation. We show that the U.S. natural gas market shifts towards a more integrated structure after August 2004. Such shift results from the delayed impact of recent FERC reforms and the development of interstate pipelines. However, intrastate pipeline-deficient regions require an enforced development and improvement of natural gas infrastructures.

ABSTRACT. We extend the model of Bashiri and Lawryshyn (2018) to the metals market and measure the significance of the relationship between metals' prices and normalized excess supply. We utilize Commodity Exchange prices, the World Bureau of Metal Statistics production and consumption data, as well as inventory data from the London Metal Exchange, Commodity Exchange, and Shanghai Futures Exchange from 1997 to 2017. We find significant relationships for copper, nickel, zinc, lead, and tin, while aluminum exhibits no relationship due to deficient inventory data. We find that during certain intervals of the 2009-2011 period, there exist deviations from the long-term relationship for the metals' prices possibly due to speculation, financialization, price stickiness, and increase of liquidity as a result of central banks' quantitative easing programs.

ABSTRACT. Financial and operational risk management are central concepts at the intersection of Finance, Operations, and Risk Management. Prior research has examined the effects of either financial or operational risk management, yet evidence on the implications for a joint financial and operational risk management strategy is lacking. We use a fine-grained data set comprising the financial and operational risk management decisions of 82 gold miners from 2003 to 2011 to provide empirical evidence for the effects of risk management on operational performance. Gold miners may vary the grade of gold they extract and process in times of volatile gold prices, yet this strategy increases inventory and profit variance. Financial risk management through forwards and calls can be similarly used to mitigate gold price risk, yet this reduces inventory and profit variance. More importantly, we find that in some cases, financial and operational risk management may be implemented as complementary strategies. The results highlight that managers should consider the implications on both operations and financials when executing risk management.

ABSTRACT. We develop a general equilibrium asset pricing model in which the emissions of fossil-fuel intensive firms lead to climate change, which negatively affects general economic output. The model allows us to analyze the effects of climate productivity risk and climate policy risk and to characterize related risk premia. We confront the model with the data by considering the stock market performance of climate sensitive vs. robust industries and dirty vs. clean industries. Our results are consistent with an increasing awareness of investors for climate change risks since the beginning of the 2000s. For the oil sector, the commodity price boom of the last decade temporarily masked the negative impact of climate risks.

ABSTRACT. We investigate the uncertainty dynamics surrounding extreme weather events, partic- ularly hurricanes, through the lens of stock and option markets. We combine firm establishment-level data with novel hurricane forecast and damage data to identify market responses to both the uncertainty regarding potential landfall and subsequent economic impact. In the days following landfall, stock options on firms exposed to the landfall region exhibit increases in implied volatility of up to 10 percent, reflecting the impact uncertainty. Using hurricane forecasts, we show that a reduction in landfall uncertainty leads to the stronger incorporation of impact uncertainty in prices, consistent with investors paying attention to the forecasts. We find no evidence that markets forecast hurricanes better than the National Oceanic and Atmospheric Administration. Improvements to hurricane forecasts could have economically significant effects in financial markets.

ABSTRACT. Elementary portfolio theory implies that environmentalists optimally hold more shares of polluting firms than non-environmentalists, and that polluting firms attract more investment than otherwise identical non-polluting firms. These results reflect the demand to hedge against high pollution states. Pigouvian taxation can reverse the aggregate investment results, but environmentalists still overweight polluters. We introduce countervailing motives for environmentalists to underweight polluters, comparing the implications when environmentalists coordinate to internalize pollution, or have nonpecuniary disutility from holding polluter stock. With nonpecuniary disutility, introducing a green derivative product may dramatically alter who invests most in polluters, but has no impact on aggregate pollution.

ABSTRACT. Natural gas storage valuation is an important business problem in energy trading, yet most valuation approaches are based on heuristics or ignore that gas markets are incomplete. We propose an exact valuation model for incomplete markets based on multistage stochastic programming. The model requires analysis of a combined control problem of storage operation and futures trading that takes an agent's risk preferences into account. As the problem is subject to the curse of dimensionality, we propose a second-order learning algorithm that discretizes the price process to a scenario lattice and then approximate the problem using stochastic dual dynamic programming. We prove convergence of the learning algorithm and show that our solution is near-optimal by comparing it with a known upper bound from the literature. We also show that the intrinsic value of storage corresponds to the value under perfect risk aversion and that the rolling intrinsic value - which is popular among practitioners - is not suitable for valuation in incomplete markets.

ABSTRACT. This paper formulates dynamic density functions, based upon skewed-t and similar representations, to model and forecast electricity price spreads between different hours of the day. This supports an optimal day ahead storage and discharge schedule, and thereby facilitates a bidding strategy for a merchant arbitrage facility into the day-ahead auctions for wholesale electricity. The four latent moments of the density functions are dynamic and conditional upon exogenous drivers, thereby permitting the mean, variance, skewness and kurtosis of the densities to respond hourly to such factors as weather and demand forecasts. The best specification for each spread is selected based on the Pinball Loss function, following the closed form analytical solutions of the cumulative density functions. Those analytical properties also allow the calculation of risk associated with the spread arbitrages. From these spread densities, the optimal daily operation of a battery storage facility is determined.

ABSTRACT. We present a joint calibration-estimation risk measurement methodology, extending recent literature, which incorporates both market calibration and historical estimation risk within a meaningful distributional assessment of parameter risk. Extending the emerging literature on model risk issues in energy markets, we apply our technique to the problem of natural gas storage valuation, using a flexible multifactor Mean Reverting Variance Gamma model specification that is both forward curve consistent and calibrated to market traded options. Realistic models of the natural gas forward curve cannot be calibrated to benchmark instruments alone due to the lack of a liquid time-spread options market and thus the correlation structure is typically estimated from historical data. We additionally devise an accessible model selection technique based on our distributional assessment of parameter risk. For a basic one-year 20in/20out storage contract, we show that the parameter risk of our two-factor Mean Reverting Variance Gamma model is higher relative to single-factor Mean Reverting Variance Gamma and Mean Reverting Jump-Diffusion benchmarks, with very different distributional characteristics. Formally pricing the parameter risk, shows the model based bid-ask spread to be over five times that of the benchmarks. The greater flexibility of the two-factor Mean Reverting Variance Gamma model in capturing more extrinsic value therefore comes at the cost of greater uncertainty. Our novel model selection technique shows, however, this increased uncertainty to be bearable, concluding that the two-factor Mean Reverting Variance Gamma model is an acceptable choice over its one-factor counterpart.

ABSTRACT. We develop a model of Cournot competition between capacity-constrained firms that sell a single good to multiple regions. We characterize the unique equilibrium allocation of the good across regions and provide a convergent algorithm for computing it. We show that a reduction in transportation costs may negatively impact the overall consumer welfare if the impacted firm is capacity constrained. Our results imply that policies promoting free trade may have unintended consequences and reduce consumer surplus in capacity-constrained industries. We calibrate our model to the international market of fertilizers and show that the model accurately predicts prices across regions and over time.

ABSTRACT. This paper studies the effect of storage announcement news on natural gas returns. We find that returns on announcement days account for more than 53% of the annual risk premium. The significant return difference between announcement and non-announcement days cannot be explained by the information content of the event. Demand and supply measures or spill-overs from the options or crude oil market can only partially explain the excess return. Explanations based on commodity trading signals, coinciding macroeconomic announcements or liquidity driven channels also fail to fully explain the differences. Intraday data suggests an asymmetric effect between positive and negative announcement surprises resulting from a pre-announcement drift for positive surprises.

ABSTRACT. We offer an equilibrium model of input and output prices and profit dynamics in commodity processing industries that links operational features, such as capacity utilization, production constraints, and adjustment costs, to equilibrium processing margins. The main driver of time-variation in processing margins in our model is the heterogeneity in the efficiency of processing plants and the degree of aggregate capacity utilization. When the demand for the final product is weak, or the supply of input is constrained, the processing industry uses its most efficient units and processing profits all small. When the demand is sufficiently strong, or when the supply of input is favorable, the processing industry utilizes inefficient units, and processors earn a large efficiency rent. In the presence of adjustment costs, the flexibility of the processing industry to respond to supply/demand shocks is reduced, which can even result in transitory negative processing profits. Empirical evidence from the oil refinery industry supports the broad predictions of the model.

ABSTRACT. Using novel data on WTI crude oil swaps and futures positions by individual dealers, we relate dealer hedging and liquidity provision to changes in risk appetite over the 2007-2015 period. We measure risk appetite using balance sheet and trading VaR measures. Consistent with a theoretical model of dealer activity, we find that swap dealers engage in fewer swaps, and hedge those swaps more tightly, when dealer risk appetite decreases. We also find evidence that dealers have larger single commodity swap books if they have a larger index book, suggesting that increased commodity index activity enhances liquidity provision.

ABSTRACT. The value of USDA reports in commodity futures markets has been intensively researched, but statistical hypothesis tests have been limited by the sheer number of reports and the consequent need to limit parameters to be estimated. This has led most tests of USDA report announcement effects to be based on single coefficients for each report series. We relax the implicit assumption that a report series has a constant impact on futures price volatility or returns in two ways in order to introduce more flexible tests for announcement effects. First, we introduce a time trend into the impact of announcements on futures price volatility to see if USDA reports are becoming more or less influential over time. Then we allow each report to have a different impact on futures price returns using Theil-Goldberger mixed estimation. The results show that many, but not all, USDA reports have significant effects on corn and soybean futures market returns or volatility.

ABSTRACT. We use non-public data regarding all trader-level futures positions, reported to the U.S. grain and oilseed derivatives market regulator (the CFTC), in order to describe the nature of market participants, the maturity structure of their positions, and the aggregate positioning patterns for nine categories of traders that we differentiate based on their main lines of business. We provide novel evidence about the true extent of calendar spreading and the contribution of commercial traders to the overall spreading activity.

Our sample’s 3,854 traders account for 86 to 93 percent of the total futures open interest at the end of an average day in 2015–2018. Well over 90 percent of their positions have maturities of less than a year. Among our nine trader categories, just three (hedge funds and commercial dealers/merchants, plus commodity index traders on the long side) account for about four fifths of all large trader positions. In fact, 197 “permanent” large traders (overwhelmingly from these three categories) make up the bulk of the daily open interest in the four largest agricultural futures markets.

In the aggregate, the positions of commercial dealers and hedge funds (including commodity pool operators, commodity trading advisors, managed money traders, and associated persons) are highly negatively correlated. This pattern holds for long, short, and spread positions. As a result, the sum total of commercial dealers and hedge funds’ open-interest shares fluctuates rather little over time.

We show, for the first time, that calendar spreads account for over a quarter of all large trader positions; that much of the variation in the total futures open interest can be tied to the extent of calendar spreading; that about half of all spread positions involve contracts expiring in 4 to 12 months (either spreading with shorter-dated contracts, or involving only maturities of 4 to 12 months); and that commercial traders who are not swap dealers (mostly commercial dealers and merchants) make up from a quarter to two fifths of all calendar spread positions. These patterns cannot be inferred from public data, as the CFTC’s Commitments of Traders Reports (COT) do not break out spreads for “traditional” commercial traders.

ABSTRACT. For the valuation and implementation of renewable energy investments, the issue of providing private investors with a financial incentive to accelerate their investment is frequently a critical component. We apply this principle to the Chinese context. This paper focuses on using the binomial model to compute the required subsidy that would incentivize investors to optimal immediate exercise of the American-style option embedded at the launching phase of the projects for Chinese renewable energy investments. In addition, this paper also aims at contrasting the binomial model with the more-laborious Monte-Carlo simulation previously used to evaluate the proper subsidy. By using the same data but a different method, and reducing the number of uncertain factors to one, it is suggested these two methods have similar outcomes but the binomial method requires substantially less computation and is more self-explanatory. This paper thus provides government with an easy-to-implement alternative way to compute the required subsidy.

ABSTRACT. I provide a novel theoretical approach to value wind energy investments. It allows to adjust for a number of risk parameters, including wind speeds, electricity price forecasts, discount rates, and uncertainty in subsidies. I use this approach to model wind energy investments under two different subsidy schemes in Denmark through a numerical Monte Carlo simulation. Moreover, I model wind energy investment under the assumption of a subsidy-free asset class. I compare the three systems and expose them to various sources of uncertainty through which I provide more clarity on which risk parameters matter most to wind energy investors and how the three systems compare to each other.

ABSTRACT. We offer a modeling framework to determine the price dynamics of renewable identification numbers (RINs) -a floor-and-trade market-based mechanism for enforcing renewable energy standards. Using a continuous-time stochastic control formulation, we explicitly model the option value embedded in the RINs prices as an American spread option by taking into account the specific institutional constraints. We derive a closed-form solution of the RINs prices when underlying commodity prices are geometric Brownian motion (GBM). We also characterize the solution for setups with mean-reverting and jump specifications for the underlying prices, which need to be solved numerically. We propose a tight numerical approximation using duality methods. Among other results, we show that the price of RINs has a U-shape relationship with the volatility of ethanol and gasoline prices and a \itshape negative \upshape relationship with the correlation between the two price processes. We examine the empirical performance of the model versus real historical data and find that our model allows a significant improvement over naive pricing models. We also find the mean-reverting assumption for price processes delivers the best performance.

ABSTRACT. Although the value of portfolios of real options is often affected by both exogenous and endogenous sources of uncertainty, most existing valuation approaches consider only the former and neglect the latter. In this paper we introduce an approach for valuing portfolios of interdependent real options under both types of uncertainty. In particular, we study a large portfolio of options (deferment, staging, mothballing, abandonment) under conditions of four underlying uncertainties. Two of the uncertainties, decision-dependent cost to completion and state-dependent salvage value, are endogenous, the other two, operating revenues and their growth rate, are exogenous. Assuming that endogenous uncertainties can be exogenised, we formulate the valuation problem as a discrete stochastic dynamic program. To approximate the value of this optimisation problem, we apply a simulation-and-regression-based approach and present an efficient valuation algorithm. The key feature of our algorithm is that it exploits the problem structure to explicitly account for reachability - that is the sample paths in which resource states can be reached. The applicability of the approach is illustrated by valuing an urban infrastructure investment. We conduct a reachability analysis and show that the presence of the decision-dependent uncertainty has adverse computational effects as it increases algorithmic complexity and reduces simulation efficiency. We investigate the way in which the value of the portfolio and its individual options are affected by the initial operating revenues, and by the degrees of exogenous and endogenous uncertainty. The results demonstrate that ignoring endogenous, decision- and state-dependent uncertainty can lead to substantial over- and under-valuation, respectively.

ABSTRACT. We study the merchant operations of energy production modeled as a compound switching and timing option, which gives rise to an intractable Markov decision process. It is common to combine least-squares Monte Carlo (LSM) and information relaxation and duality techniques to compute a feasible operating policy as well as lower and dual (upper) bounds on the optimal policy value. However, on realistic merchant ethanol production instances in the literature, the LSM policy exhibits large optimality gaps, which is conjectured to be due to a weak dual bound. To narrow this gap, we broaden the pathwise optimization (PO) method for optimal stopping to merchant energy production. Our extension entails tackling a novel PO linear program whose size grows linearly with the number of stages but is nevertheless difficult to solve because it is ill-conditioned and large-scale. We overcome the former issue by designing a preconditioning procedure based on principal component analysis (PCA) and address the latter issue by developing a convergent iterative method based on block coordinate descent (BCD). On the known merchant ethanol production instances with large LSM optimality gaps, PO leads to considerably smaller dual bounds compared to LSM, which brings to light the near optimality of the LSM policy, albeit at the cost of a substantially larger computational effort. The PO policy nevertheless improves on the LSM policy and both these state-of-the-art strategies add significant value relative to simpler operating rules.

ABSTRACT. We develop a multi-period model for a cooperative organization (CO) that purchases a commodity (e.g., groundnuts) from its member-farmers. The CO faces random supply in every period, all of which it must purchase. After purchasing the raw commodity, the CO processes it to obtain multiple grades of another commodity (e.g., groundnut seeds). The CO aims to sell the processed commodity in the spot market with the aim to maximize profit for its members. To take advantage of the price dispersion among various grades of the processed commodity, the CO may upgrade to obtain a higher grade or blend two grades to obtain an intermediate grade. The CO may also carry inventory of some or all the grades into the following period if the price is likely to rise. To analyze this problem, we formulate a dynamic program to determine the quantity of raw commodity to process and the quantity of each grade of processed commodity to upgrade and/or blend and subsequently, sell, in every period. We find that in general the optimal value function is not separable in raw inventory and the vector of processed inventories. However, the value function may have a simpler form when either blending is not permitted, supply has a discrete probability distribution or there is no processing capacity constraint. Our analysis shows that the marginal value of any grade forms a super-martingale process over time after accounting for holding cost. We also show that when the spot prices for a grade form a martingale process, it is optimal for the CO to not trade the grade during the planning horizon and trade only at the end of the planning horizon.

ABSTRACT. Should oil-rich countries invest in the oil refinery industry? This is a crucial policy question for such economies. We offer theoretical models for a vertical integration strategy based on a risk-hedging view. The first model highlights the trade-off between return and risk-reduction features of upstream/downstream sectors. The dynamic model demonstrates the volatility of total budgetary revenue of each sector. Our theory-guided empirical analysis shows that though the average markup in the refining sector is significantly smaller than the profits in the upstream, downstream investment can provide some hedging value. In particular, the more stable and mean-reverting refining margins provide a partial revenue cushion when crude oil prices are low. We discuss the risk-hedging feature of the refinery industry when the crude oil market faces supply versus demand shocks.

ABSTRACT. This article explains contracting preferences in an automotive supply chain in which an automaker purchases raw-material-intensive parts from suppliers. The basic premises behind our explanations are that raw material costs fluctuate over time and that automakers respond to shifting costs by adjusting output, which is something we document empirically. The extent of this adjustment depends on the contracts automakers use to procure parts. The auto industry uses two types of supply agreements: the price adjustment agreement (PA) and the material plus surcharge agreement (MPS). The marginal price with either contract is \emph{cost plus surcharge}. The difference is that the ``cost'' with the MPS (PA) contract is indexed to the supplier's actual (expected) cost. Automakers offer potential suppliers the contract type and award the contracts through auctions. When all firms can see the supplier's actual raw material cost, we show that the MPS contract is a better choice for the automaker because it maximizes the total surplus. Often, however, automakers are not privy to the costs that suppliers pay for raw materials, i.e., cost information is asymmetric. In such cases, the comparison between the two contracts comes down to efficiency that the automaker loses by choosing the PA contract and information rent that she must pay by selecting the MPS contract. The following factors emerge as critical determinants in an automaker's contract choice: raw material cost variability, the supplier's reservation payoff, the elasticity of consumer demand, and information asymmetry. We conclude by empirically testing our theoretical predictions using a unique data set provided to us by BMW.

ABSTRACT. Traditionally, hedging pressure is measured by the amount by which short hedging exceeds long hedging, while speculative pressure is measured by the amount by which long speculation exceeds hedging pressure. I offer new measures of hedging and speculative pressure which explicitly recognize that not all of long hedging equals balancing hedging contracts. When open short hedging is greater (less) than or equal to open long hedging, I define hedging pressure as the difference between short (long) hedging and balancing hedging contracts and speculative pressure as the difference between long (short) speculation and hedging pressure. I estimate hedging pressure and speculative pressure for 21 different futures contracts in 7 different groups. I show that, when balancing hedging contracts are explicitly accounted for, hedging pressure is higher and speculative pressure is lower in magnitude, than when estimated by the traditional measures of hedging and speculative pressure, respectively. I investigate the effects of the financialization of futures markets and the financial crisis upon the measures of hedging and speculative pressure, and the effect of the measures upon spot price volatility. The results indicate that hedging pressure has a destabilizing influence, while speculative pressure has a destabilizing or stabilizing influence, upon volatility in the spot market.

ABSTRACT. We find that commodity futures returns contain relevant information for stock market returns and macroeconomic fundamentals for a large number of countries. Commodity futures returns predict stock market returns in 65 out of 70 countries and macroeconomic fundamentals in 62 countries. This predictability is not concentrated in the energy and industrial metals sectors, being economically and statistically significant across all sectors. Surprisingly, we find a rather limited role of countries' dependence on commodity trade in explaining this predictability. This holds even when considering new measures that take into account indirect exposures through financial and trade linkages among countries. We find much stronger evidence for predictability being related to the ability of commodities to forecast inflation rates. Overall, our evidence is consistent with commodity markets having a truly global information discovery role for financial markets and the real economy.

ABSTRACT. Due to deliverable-specific supply-demand conditions, or a plan to execute a manipulation, holders of some fraction of the long open interest prevailing at the onset of the liquidation period do not liquidate their positions, but retain them in order to take delivery.This fraction is not known to the market at large, because the determinants of delivery are private to a subset of traders. However, the rate of liquidation of the long side of expiring contract (which occurs usually in the month leading up to the delivery period) provides information about how many contracts will be closed by delivery. Given the mean and variance of the rate of liquidation by traders not standing for delivery Bayesian market participants make inferences about the number of deliveries. These estimates become more precise as the delivery process proceeds. Since market participants extract information about the ultimate number of deliveries from the liquidation process, prices prior to expiration vary as market participants update their estimates of the number of deliveries. One implication of the model is that prices can become distorted as a result of manipulation before the manipulator actually stands for delivery.