Source: SOUTH DAKOTA STATE UNIVERSITY submitted to NRP
THE PRICING AND RISK MANAGEMENT OF DERIVATIVES ON AGRICULTURAL COMMODITIES
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
HATCH
Reporting Frequency
Annual
Accession No.
0221959
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Apr 1, 2010
Project End Date
Sep 30, 2015
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Recipient Organization
SOUTH DAKOTA STATE UNIVERSITY
PO BOX 2275A
BROOKINGS,SD 57007
Performing Department
Economics
Non Technical Summary
Farmers, elevators and food processors (hereafter referred as market participants in general) are increasingly aware of the importance of using agricultural commodity derivatives to manage the risks embedded in agricultural commodities. The primary risks include price risk and variance risk. The former involves uncertainty in future commodity price movements, whereas the latter involves variation in the uncertainty of price movements. Market participants need to understand these risks and know how to manage them when participating in the commodity futures and their derivatives markets. Market participants can buy or sell futures contracts to lock in a price for some future date, and protect themselves against price fluctuations, or price risk. Farmers who raise corn may sell futures contracts in anticipation of harvest. Similarly, food processors concerned about price increases can use futures contracts to lock in their costs of input purchase. Market participants may also buy or sell options on commodity futures' contracts to mitigate the adverse effect of volatility. In order to protect from significantly increasing variance, they can form a long straddle, namely a combination of long positions in both a call option and a put option with the same strike price and expiration date. If a price decrease is accompanied by an increasing variance, the effect of the price decrease will be offset by the gain from the combinatory positions in options. Conversely, if a price decrease is accompanied with a decreasing variance, a combination of short positions in a call option and a put option with the same strike price and expiration date (a short straddle) will yield a positive gain to offset the loss from the price decrease. In implementing these options strategies, transaction costs should be taken into consideration. The use of the above risk management tools calls for a thorough understanding of the dynamics of both the price of the underlying commodity and its variance. The unexpected high commodity prices during 2007 and 2008 alerts us to the importance of extreme events that happen more often than are perceived by all market participants. Such a changing market dynamic highlights the need for updating price forecasting and risk management tools.
Animal Health Component
(N/A)
Research Effort Categories
Basic
(N/A)
Applied
(N/A)
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
6021510301010%
6021549301010%
6021820301010%
6022410301010%
6041510301010%
6041549301010%
6041820301010%
6091510301010%
6091549301010%
6091820301010%
Knowledge Area
602 - Business Management, Finance, and Taxation; 609 - Economic Theory and Methods; 604 - Marketing and Distribution Practices;

Subject Of Investigation
1510 - Corn; 1549 - Wheat, general/other; 1820 - Soybean; 2410 - Cross-commodity research--multiple crops;

Field Of Science
3010 - Economics;
Goals / Objectives
The overall goal of this project is to analyze agricultural commodity prices, and to increase the understanding of the use of their derivatives for risk management. The initial focus will be on corn and soybean prices. Previous studies in this area are based on stochastic models of commodity prices that often do not adequately account for drastic price changes. The price dynamics of commodities and their derivatives will be analyzed with model-independent and model-dependent approaches, the latter of which considers extreme price movements in addition to normal price movements which have become more important in recent years. Strategies to hedge against price fluctuations (price risk) and time-varying uncertainty of the fluctuation (variance risk) are proposed based on the assumed price dynamics. The project has four objectives: to describe the statistics of commodity price movement based on historical data; to quantify the realized variances from historical futures data and future variances implied in option prices and investigate the dynamics of the difference between them; to propose structural models which can explain the historical price movement of commodities and derive the pricing implication for their derivatives (futures and options); to derive the implications for risk management practices and analyze the relevance of these practices to farmers and agricultural businesses in South Dakota. The first stage of the project targets the first two objectives. The second stage, built on the findings from the first stage, will focus on the modeling and forecasting of commodity futures and options prices, including research on a single commodity and inter-commodity relationships. The risk management aspects will be emphasized throughout the project. The following outcomes are expected: a full description of statistical properties of corn and soybeans futures prices; computation of two different measures of variance: realized variances from corn and soybeans futures prices and implied variances from their respective futures options prices; development of volatility indices for corn and soybeans; Description and statistical testing of the significance of the variance risk premium over time; estimation of proposed models of spot commodity price and convenience yield; evaluation of the forecasting performance of estimated models on commodity futures (and options); evaluation of the hedging performance of estimated models on commodity futures (and options); summary of risk management implications and development of risk management tools for market participants.
Project Methods
This research explores variance dynamics of the underlying commodities, and the price and hedging of their futures and options. More specifically, this research employs both model-independent and model-dependent approaches for the analysis of variance dynamics, whereas the evaluation on the pricing and hedging of futures and options is done through a model-dependent framework. Regarding variance dynamics, one can directly calculate realized variance from historical commodity futures prices. The formula for realized variance is available from any standard statistics textbook. Similarly, variance (volatility) information can be extracted from an option for its remaining life, often referred to as forward implied variance (volatility). In calculating this forward variance, one can rely on a parametric model (model-dependent), such as Black (1976), or employ a model-independent approach to back out variance, such as Carr and Wu (2009). The model-dependent approach is widely used in research on agricultural commodities ((Manfredo and Sanders (2004), Egelkraut, Garcia and Sherrick (2007)). The model-independent approach is pioneered by Bakshi and Madan (2000), Britten-Jones and Neuberger (2000), Carr and Wu (2009). Implied variance can be inferred (synthesized) from a portfolio of options. The synthesizing procedure does not assume any specific model of the underlying price, hence it is referred to as a model-independent approach. Given historical realized variance and forward implied variance, the ex post difference between them is commonly defined as the variance risk premium as in Carr and Wu (2009). This premium shows how market participants assess variance risk. This research will investigate: (1) whether there exists a significant variance risk premium across different periods, (2) whether the premium is constant or time varying, and (3) what the risk management implications are in the presence of non-zero variance risk premium. In fact, a higher implied variance than realized variance means that market participants are willing to pay a variance risk premium as protection against excessive price volatility. A non-zero premium will highlight the need for understanding the driving forces for variance risk and managing such risk. Regarding pricing and hedging of commodity futures and options, the research considers five one-factor and two-factor models for commodity futures prices. A one-factor model only specifies the stochastic process for spot prices, whereas two-factor models describe the stochastic processes for both spot prices and convenience yield (see Gibson and Schwartz (1990), Schwartz (1997), Casassus and Collin-Dufresne (2007) among others). The proposed models deviate from the current existing research by explicitly incorporating discontinuity in price movements. Given the model specification, the parameters in the models are estimated using historical commodity futures prices. The out-of-sample pricing and hedging performance are compared among these models. Based on the model performance on futures, one can select the models that best describe the underlying commodity prices and apply them to options pricing and hedging.

Progress 04/01/10 to 09/30/15

Outputs
Target Audience:The target audiences for the analysis of pricing and risk management of agricultural commodities are farmers, agribusinesses, investors, policy makers, students and researchers in South Dakota and other states that produce similar commodities. These audiences are reached through SDSU Economics Staff Papers, Economics Commentators, presentations in classrooms and professional conferences. Financial and economic researchers are reached through professional papers and journal articles. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?A training workshop entitled "A Brief Introduction to Matlab in Economics and Finance" was provided for South Dakota State University Economics Graduate Student Association on March 22, 2013. Graduate students at Department of Economics at South Dakota State University attended the workshop. How have the results been disseminated to communities of interest?The results have been published in 6 research articles at high impact peer-reviewed academic journals including Agricultural Economics, Canadian Journal of Agricultural Economics, Journal of Economics, Journal of Empirical Finance, and Journal of Futures Markets. The results are also presented at 11 professional conferences across the United States. Some select results are also shared with South Dakota farmers and agribusinesses through the extension publication: Economics Commentator at South Dakota State University. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? My AES research during the current evaluation period revolved around three themes: (1) the pricing of agricultural commodities and their derivatives (four projects); (2) the evaluation of price and volatility risk in agricultural commodity markets (four projects); (3) the price and risk connections among agricultural markets (two projects). Four projects on Theme 1 help farmers and agribusinesses with modern approaches to decide what price they should pay or take for protection again price risk, volatility risk over time. Four projects on Theme 2 provide guidance to farmers/ranchers and agribusinesses with the most appropriate measure of risk when predicting future grain price changes, and a decision framework as to how to market their fed cattle, by pen or on the grid. Two projects on Theme 3 provide an important perspective for agribusinesses to understand the spillover of risks across markets, instead of risk within each market independently. The new perspective makes their risk management more effective. The topics on the first theme include the theoretical and empirical pricing of futures and options on corn, soybeans and wheat using a comprehensive model, calendar spread options on corn, soybeans and wheat. We employed end-of-day futures, options, and calendar spread options price data since 1987. We developed and tested various mathematical models and found appropriate ones for agricultural commodities to inform farmers and agribusinesses of (a) what level to lock in grain prices for the future (i.e. commodity futures); (b) what premium to pay for downside risk protection (i.e. commodity options) or mitigate adverse price change over time (calendar spread options). The topics on the second theme include the evaluations of (1) the premium for taking volatility risk in agricultural commodities; (2) the effect of informational uncertainty and financial risk on fed cattle marketing. Based on the options data from 1987 to 2010, we proposed and computed a model-free measure of volatility risk, termed "Corn/Soybean/Wheat VIX", which is the first in the agricultural commodity literature and outperforms other traditional measures. The VIX-like volatility measure for agricultural commodities has been an important reference for risk management in the agricultural commodity industry. Based on the fed cattle revenue data from 2001 to 2008, we found empirical evidence for the coexistence of two marketing channels (by pen and on the grid), and financial risk along with other factors (quality and time trend) drives the price differential between the two marketing channels. The topics on the third theme include (1) the market connection between grain inputs and food companies; (2) the risk connection among the soybean complex, namely soybeans, soybean oil, and soybean meal. We examined price volatility transmission between publicly traded food companies (processed and packaged goods, meat, dairy, and farm products sectors) and grain markets in the United States. Using the end-of-day price data from 2008 to 2013, we found the volatility spillover between grain markets and food companies is bi-directional and the degree of connectedness ranks as follows: Processed food > Meat > Farm > Dairy. Understanding the degree of volatility spillover will help both farmers and agribusinesses effectively evaluate and manage risks in grain markets. Using the same options data from 1990 to 2009, we conducted empirical research to determine how much market participants are willing to pay to reduce the risk due to volatile prices, i.e. variance risk premium (VRP), and how VRPs in the soybean complex impact each other. Our empirical results can help farmers, grain elevators, and users of soybean-related options manage volatility risk in the markets of soybeans, soybean oil and soybean meal.

Publications

  • Type: Journal Articles Status: Awaiting Publication Year Published: 2015 Citation: Wu, F., Meyers, R., Guan, Z. and Wang, Z. (2015). Risk-adjusted implied volatility and its performance in forecasting realized volatility in corn futures prices, Journal of Empirical Finance, forthcoming.
  • Type: Journal Articles Status: Published Year Published: 2015 Citation: Osei, M., and Wang, Z. (2015). Seasonality and Stochastic Volatility in Wheat Options, Journal of Economics, XLI(1), 1-28.
  • Type: Journal Articles Status: Published Year Published: 2014 Citation: Fausti, S.W., Wang, Z., Qasmi, B.A. and Diersen, M.A. (2014). Risk and Marketing Behavior: Pricing Fed Cattle on a Grid, Agricultural Economics, 45(5), 601-612.
  • Type: Journal Articles Status: Published Year Published: 2014 Citation: Schmitz, A., Wang, Z. and Kimn, J. (2014). A Jump Diffusion Model for Agricultural Commodities with Bayesian Analysis. Journal of Futures Markets, 34(3), 235-260.
  • Type: Journal Articles Status: Published Year Published: 2012 Citation: Fausti, S.W., Wang, Z. and Lange, B. (2012). Uncertainty and Producer Fed Cattle Marketing Decisions: Theory and Evidence, Canadian Journal of Agricultural Economics, 61(3):371-395.
  • Type: Journal Articles Status: Published Year Published: 2012 Citation: Wang, Z., Fausti, S.W. and Qasmi, B.A. (2012). Variance Risk Premiums and Predictive Power of Alternative Forward Variances in the Corn Market, Journal of Futures Markets, 32(6): 587-608.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Wang, Z., Fausti, S., and B. Qasmi (2014). The Behavior of the Variance Risk Premium in the Soybean Complex. Agricultural & Applied Economics Association Annual Conference, Minneapolis, MN, July 29, 2014.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Schmitz, A., Wang, Z. and Kimn, J. (2014). Parallel Bayesian Analysis of a Comprehensive Model of Agricultural Futures. 2014 NCCC-134 Annual Meeting, St. Louis, MO.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2013 Citation: Schmitz, A., Wang, Z. and Kimn, J. (2013). Pricing Calendar Spread Options on Agricultural Commodities. NCCC-134 2013 Annual Meeting, St. Louis, MO; 2013 Financial Management Association Annual Conference 2013, Chicago, IL.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2012 Citation: Schmitz, A., Wang, Z. and Kimn, J. (2012). A Jump Diffusion Model for Agricultural Commodities with Bayesian Analysis. 2012 Midwest Finance Association annual conference, New Orleans, LA; NCCC-134 2012 Annual Meeting, St. Louis, MO; 2012 Financial Management Association Annual Conference 2012, Atlanta, GA.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2012 Citation: Fausti, S., Wang, Z., Diersen, M. and Qasmi, B. 2012. Pricing Fed Cattle on a Grid: An Analysis of the Incentive Mechanism over Time. 2012 Southern Agricultural Economics Association Annual Meeting, Birmingham, AL.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2011 Citation: Pokharel, K. and Wang, Z. 2011. Measuring and Forecasting Variance and Variance Risk Premium in the Agricultural Market. 2011 Missouri Valley Economic Association annual conference, St. Louis, MO.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2010 Citation: Wang, Z., Fausti, S. and Qasmi, B. 2010. Variance Risk Premiums and Predictive Power of Alternative Forward Variances in the Corn Market. 2010 Annual meeting of Western Economic Association, Portland, OR.; Financial Management Association annual conference 2010, Denver, CO.
  • Type: Theses/Dissertations Status: Other Year Published: 2014 Citation: Schmitz, A. (Fall 2010-Fall 2014). Two Essays on Agricultural Commodity Option Pricing. PhD in Math/Statistics, South Dakota State University.
  • Type: Theses/Dissertations Status: Other Year Published: 2014 Citation: Graham, A. (Fall 2012-Summer 2014). Volatility Transmission: A Linkage between Grain Markets and Food Companies. MS Thesis in Economics, South Dakota State University.
  • Type: Theses/Dissertations Status: Other Year Published: 2013 Citation: Michael Osei (Fall 2011-Summer 2013). Seasonality and Stochastic Volatility in Wheat Options. MS Thesis in Economics. South Dakota State University.
  • Type: Theses/Dissertations Status: Other Year Published: 2011 Citation: Krishna Pokharel (Fall 2009-Summer 2011). Measuring and Forecasting Variance and Variance Risk Premium in the Soybean Market. MS Thesis in Economics. South Dakota State University.


Progress 10/01/13 to 09/30/14

Outputs
Target Audience: The target audiences for the analysis of pricing and risk management of agricultural commodities are farmers, agribusinesses, investors, policy makers, students and researchers in South Dakota and other states that produce similar commodities. These audiences are reached through SDSU Economics Staff Papers, Economics Commentators, presentations in classrooms and professional conferences. Financial and economic researchers are reached through professional papers and journal articles. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? The paper entitled "Pricing and Hedging Calendar Spread Options on Agricultural Grain Commodities" was presented at 2014 International Mathematical Finance Conference, Miami, FL. March 20, 2014. The paper entitled "Parallel Bayesian Analysis of a Comprehensive Model of Agricultural Futures" was presented at 2014 NCCC-134 Annual Meeting, St. Louis, MO. April 22, 2014. The paper entitled "The Behavior of the Variance Risk Premium in the Soybean Complex" was presented at Agricultural & Applied Economics Association Annual Conference, Minneapolis, MN. July 29, 2014. The paper entitled "Volatility Transmission: A Linkage between Grain Markets and Food Companies" was presented at South Dakota State University Department of Economics Seminar on July 3, 2014 by Angela Graham, a graduate student under my supervision. What do you plan to do during the next reporting period to accomplish the goals? Plan to revise the research manuscripts under review at peer reviewed journals and seek possible publication; Present research results to both academic conference and seminar/workshop that are of interest to stakeholders; Advise students on agricultural economics and finance research.

Impacts
What was accomplished under these goals? My AES research during the current evaluation period revolved around three themes: (1) the pricing connection between grain markets and food companies, (2) the risk connection among the soybean complex, namely soybeans, soybean oil, and soybean meal, and (3) the pricing of agricultural commodity futures. Regarding the first theme, we study price volatility transmission between publicly traded food companies (processed and packaged goods, meat, dairy, and farm products sectors) and grain markets in the United States. Our empirical research provides a valuable perspective on whether and how much the change in grain prices will impact food companies and vice versa. Understanding the degree of volatility spillover will help both farmers and agribusinesses effectively evaluate and manage risks in grain markets. Regarding the second theme, we conduct empirical research to determine how much market participants are willing to pay to reduce the risk due to volatile prices, i.e. variance risk premium (VRP), and how VRPs in the soybean complex impact each other. Our empirical results can help farmers, grain elevators, and users of soybean-related options manage volatility risk in the markets of soybeans, soybean oil and soybean meal. Regarding the third theme, we propose and test a comprehensive model for corn futures. The modeling technique can help all users of corn futures determine its fair price. In the research, entitled "volatility transmission: a linkage between grain markets and food companies," we employ two versions of Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models: Baba, Engle, Kraft, and Kroner (BEKK) model and the Constant Conditional Correlation (CCC) models to explain volatility transmission between grain markets and top food companies in the United States. We estimate the SV model parameters via maximum likelihood method and evaluate the degree of volatility transmission in both directions (from grain to food and the reverse). We collect daily grain index and food stock price data from April 2008 to December 2013 from Yahoo Finance for empirical estimation. Three grain indexes and eleven food stocks with top market capitalization are chosen to represent the grain markets and four food sectors (processed and packaged goods, meat, dairy, and farm products) that directly or indirectly utilize grains as inputs to their production. Our results show evidence of bi-directional volatility spillover effects with strong emphasis running from the grain markets to the food companies. Meanwhile, volatility spillover from the grain markets to food sectors has the following rank in terms of magnitude: processed and packaged goods sector > meat sector > farm sector > dairy sector, from the strongest to the weakest. In the research, entitled "The Behavior of the Variance Risk Premium in the Soybean Complex," we examine the determination and cross-market influence of variance risk premiums in soybean, soybean oil and soybean meal markets. We employ daily prices of futures and options on CBOT soybeans, soybean oil and soybean meal from January 1990 to November 2009 as the basis of our empirical analysis. We first compute 60-day realized variance from daily historical futures return and infer 60-day risk-neutral variance from options using the CME/CBOE VIX calculation method. Variance risk premium for each commodity is measured by the difference between realized and risk-neutral variances. We document negative variance risk premiums that are statistically significant for all three commodities in the soybean complex, more specifically average of -168.63 for soybeans, -68.27 for soybean oil and -47.10 for soybean meal. We further investigate the dynamics of variance risk premiums in the soybean complex using a seemingly unrelated regression model. We find that (1) VRP in the soybean market leads that in the oil and meal markets at the daily interval; (2) VRPs in the soybean complex are relatively transient with half time of 1.5-2 months; (3) higher commodity price, higher volatility risk level, higher crush margin and mismatching of contract expiration in the soybean complex help drive their respective VRP while recession have the opposite effect on VRPs. The empirical finding across the three markets can help farmers and grain elevators effectively manage volatility risk. In the research, entitled "Bayesian Analysis of a Comprehensive Model for Agricultural Futures," we propose a single comprehensive model that includes stochastic volatility, seasonal spot price volatility, and stochastic cost-of-carry. We apply the proposed model to analyze the corn futures market from January 3rd, 1989, to December 31st, 2008. We conduct parameter estimation using Markov chain Monte Carlo (MCMC) with a novel dynamic tuning scheme. We also employ a parallel MCMC scheme for state variable estimation. Parameter estimates and model errors indicate the comprehensive model to be effective for modeling corn futures. The working paper that was reported in the previous report, entitled "Seasonality and Stochastic Volatility in Wheat Options," has undergone the revision and has been accepted for publication in Journal of Economics. The working paper that was also reported in the previous report, entitled "Risk and marketing behavior: pricing fed cattle on a grid," has undergone two rounds of revisions and has been published in Agricultural Economics in 2014.

Publications

  • Type: Journal Articles Status: Published Year Published: 2014 Citation: Fausti, S., Wang, Z., Qasmi, B. and Diersen, M. (2014). Risk and marketing behavior: pricing fed cattle on a grid. Agricultural Economics, 45(5), 601-612.
  • Type: Journal Articles Status: Awaiting Publication Year Published: 2014 Citation: Osei, M., and Z. Wang (2014). Seasonality and Stochastic Volatility in Wheat Options. Journal of Economics, in print
  • Type: Journal Articles Status: Published Year Published: 2014 Citation: Schmitz, A., Wang, Z. and Kimn, J. (2014). A Jump Diffusion Model for Agricultural Commodities with Bayesian Analysts. Journal of Futures Markets, 34(3), 235-260.
  • Type: Theses/Dissertations Status: Published Year Published: 2014 Citation: Graham, A. (2014). Volatility Transmission: A Linkage between Grain Markets and Food Companies. Masters Thesis, Department of Economics, South Dakota State University
  • Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Schmitz, A., Wang, Z. and Kimn, J. (2014). Parallel Bayesian Analysis of a Comprehensive Model of Agricultural Futures. 2014 NCCC-134 Annual Meeting, St. Louis, MO.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Schmitz, A., Wang, Z. and Kimn, J. (2014). Pricing and Hedging Calendar Spread Options on Agricultural Grain Commodities. 2014 International Mathematical Finance Conference, Miami, FL. 3/19-21/2014.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Wang, Z., Fausti, S., and Qasmi, B. (2014). The Behavior of the Variance Risk Premium in the Soybean Complex. Agricultural & Applied Economics Association Annual Conference, Minneapolis, MN, July 29, 2014.


Progress 01/01/13 to 09/30/13

Outputs
Target Audience: The target audiences for the analysis of pricing and risk management of agricultural commodities are farmers, agribusinesses, investors, policy makers, students and researchers in South Dakota and other states that produce similar commodities. These audiences are reached through SDSU Economics Staff Papers, Economics Commentators, presentations in classrooms and professional conferences. Financial and economic researchers are reached through professional papers and journal articles. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? A training workshop entitled “A Brief Introduction to Matlab in Economics and Finance” was provided to South Dakota State University Economics Graduate Student Association on March 22, 2013. Graduate students who worked on my research projects attended the workshop. How have the results been disseminated to communities of interest? The paper entitled “Pricing and Hedging Calendar Spread Options on Agricultural Grain Commodities,” was presented in NCCC-134 conference on “Applied Commodity Price Analysis, Forecasting, and Market Risk Management”, St. Louis, MO, on April 16-16, 2013; The research entitled “Seasonality and Stochastic Volatility in the Wheat Options” was presented at South Dakota State University Department of Economics Seminar on July 8, 2013 by M. Osei, a graduate student under my supervision. Applied research results were shared to corn producers in South Dakota at SD Corn Growers Annual Meeting at Sioux Falls on January 19, 2013. What do you plan to do during the next reporting period to accomplish the goals? Plan to revise the research manuscripts under review at peer reviewed journals and seek possible publication; Present research results to both academic conference and seminar/workshop that are of interest to stakeholders; Advise students on agricultural economics and finance research.

Impacts
What was accomplished under these goals? My AES research during the current evaluation period revolved around two themes: pricing options on agricultural commodity and fed cattle marketing. Calendar spread options (CSOs) on grains and wheat options were investigated concerning the first theme, while the marketing behavior of fed cattle between 2001 and 2008 was explored concerning the second theme. As for the first theme about CSOs, consumers (supplies) of grains in agribusinesses face the rollover risk of accepting unfavorable prices when refilling (depleting) their inventories, which can be hedged by the CSOs. Our theoretical and empirical research provides a valuable guidance on how to evaluate CSOs and how to employ CSOs to hedge rollover risk for agribusinesses dealing with grains. As for the first theme about wheat options, our empirical research can help farmers and grain elevators in South Dakota and other states effectively manage price risk. As for the second theme about fed cattle marketing, we conducted empirical research to determine if market signals sent through the grid pricing system indicate an improvement in the grid incentive mechanism over time, and if uncertainty associated with carcass quality has been a barrier to grid price adoption. Our empirical results can help cattle producers and meat packers evaluate the risk in cattle marketing and make appropriate decisions accordingly. In the research entitled “Pricing and Hedging Calendar Spread Options on Agricultural Grain Commodities,” we proposed two concise models, namely geometric Brownian motion (GBM) and stochastic volatility (SV) models to price and hedge CSOs. We estimated the SV model parameters via implied-state-generalized-method-of-moments (IS-GMM) and evaluated the in-sample and out-of-sample performances of both models by employing CSO data from 2009 to 2012. We found that the SV model can price corn CSOs with 0.79% error on average, soybeans with 0.75% for soybeans, and wheat with less than 0.1%. The research enhances our understanding of the pricing and hedging (and risk management) of CSOs for agribusinesses dealing with grains. In the research entitled “Seasonality and Stochastic Volatility in Wheat Options,” we examined the implication of stochastic volatility and seasonality on the pricing performance in the Chicago and Kansas City wheat options markets. A seasonal stochastic volatility model that is consistent with the Samuelson Hypothesis was proposed. The model parameters were estimated using panel data comprised of daily prices of wheat futures and American-style options written on these futures contracts from 2002 to 2010. The seasonal stochastic volatility (SSV) model was compared with the benchmark Black’s (1976) model and stochastic volatility (SV) model to examine the effects of seasonality and stochastic volatility on the pricing performance of the wheat options. The results showed that incorporating seasonality, the Samuelson’s Hypothesis, and stochastic volatility significantly improves the pricing accuracy of wheat options in the two major wheat markets. The theoretical and empirical results can help farmers and elevators with effective price risk management. The research entitled “Risk and Marketing Behavior: Pricing Fed Cattle on a Grid” revised the previous paper entitled "Pricing Fed Cattle on a Grid: An Analysis of the Incentive Mechanism over Time." We determined if market signals sent through the grid pricing system are encouraging producers to market on a grid and discourage marketing by the pen, and if uncertainty associated with carcass quality, when producers market on a grid, has been a barrier to grid price adoption. We thoroughly revised the empirical model and extensively expanded the discussion about the behavior of market participants under risk aversion and informational asymmetry in the eariler version. We showed that the grid premium and discount structure is slowly adjusting carcass quality market signals to encourage marketing on a grid and discourage marketing by the pen, and that carcass quality uncertainty adds financial risk associated with the adoption of grid pricing. The empirical results enhance our knowledge in understanding the marketing behavior of fed cattle over time and the impact of financial risk associated with cattle quality uncertainty.

Publications

  • Type: Journal Articles Status: Published Year Published: 2013 Citation: Fausti,S., Wang, Z. and Lange, B. (2013). Expected Utility, Risk, and Marketing Behavior: Theory and Evidence from the Fed Cattle Market. Canadian Journal of Agricultural Economics, 61(3):371395.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2013 Citation: Schmitz, A., Wang, Z. and Kimn, J. (2013). Pricing and Hedging Calendar Spread Options on Agricultural Grain Commodities. NCCC-134 2013 Annual Meeting, St. Louis, MO. 4/16-17, 2013.
  • Type: Journal Articles Status: Under Review Year Published: 2013 Citation: Fausti, S., Wang, Z., Qasmi, B. and Diersen, M. (2013). Risk and Marketing Behavior: Pricing Fed Cattle on a Grid. Working Paper, South Dakota State University.
  • Type: Theses/Dissertations Status: Other Year Published: 2013 Citation: Osei, M. (2013). Seasonality and Stochastic Volatility in the Wheat Options. Masters Thesis, Department of Economics, South Dakota State University.


Progress 01/01/12 to 12/31/12

Outputs
OUTPUTS: My AES research in 2012 was carried out along two lines: continuation of my previous research on pricing agricultural commodity options using Markov Chain Monte Carlo (MCMC) and exploration of a new research area in cattle marketing. In the paper entitled "A Jump Diffusion Model for Agricultural Commodities with Bayesian Analysis", we proposed a comprehensive model that incorporates four important features: price jumps, stochastic volatility, seasonality and stochastic cost of carry. Such model features have been considered separately, but not collectively, in the literature of pricing agricultural commodity futures and options. We employed a special MCMC algorithm, which is new in the agricultural commodity derivatives pricing literature, to estimate the proposed stochastic volatility (SV) and stochastic volatility with jumps (SVJ) models. We applied the proposed models to options and futures market data on corn, soybeans and wheat. Due to the complexity of the SVJ model with sixteen parameters and six unobserved state variables, the computational challenge necessitates a gradual inclusion of more data. We estimated all the parameters and state variables for only one year 2006 in the initial analysis as reported in 2011 AES report. In 2012, we expanded the data to the two-year period (2006-2007) and finally the five year period (2006-2010), which entails months of computational efforts with the state-of-the-art server SDSU Big Jack. Regarding the research on cattle marketing, two lines of research are carried out. The first paper entitled "Expected Utility, Risk, and Marketing Behavior: Theory and Evidence from the Fed Cattle Market," investigates the effect of carcass quality uncertainty on the structure of the slaughter cattle market. A theoretical extension of the "Theory of Factor Price Disparity" is provided. The theoretical foundation for the coexistence of multiple marketing channels is studied. The second paper entitled "Pricing Fed Cattle on a Grid: An Analysis of the Incentive Mechanism over Time" was conducted to determine if market signals sent through the grid pricing system are encouraging producers to market on a grid and discourage marketing by the pen, and if uncertainty associated with carcass quality, when producers market on a grid, has been a barrier to grid price adoption. PARTICIPANTS: Dr. Wang, Dr. Fausti and Dr. Qasmi collaborated on the research on variance risk premium of corn. Drs. Wang, Fausti, Diersen and Qasmi and Mr. Lange collaborated on research on fed cattle marketing and risk analysis. Mr. Adam Schmitz completed the first paper and is working on the second (revised) paper of his dissertation (PhD in Mathematics) on the pricing of options on agricultural commodities under the joint supervision of Dr. Wang of the Economics Department and Dr. Jung-Han Kimn of the Department of Mathematics and Statistics. The first paper is accepted for publication in the Journal of Futures Markets and has been presented at Midwest Finance Association annual meeting and NCCC134 in 2012. The second paper is also accepted for presentation at NCCC134 in 2013. Mr. Michael Osei, Masters' candidate in Economics, is working toward his thesis on options pricing on Kansas, Illinois and Minnesota wheat under Dr. Wang's supervision. TARGET AUDIENCES: The target audiences for the analysis of pricing and risk management of agricultural commodities are farmers, agribusinesses, investors, policy makers, students and researchers in South Dakota and other states that produce similar commodities. These audiences are reached through SDSU Economics Staff Papers, Economics Commentators, presentations in classrooms and professional conferences. Financial and economic researchers are reached through professional papers and journal articles. PROJECT MODIFICATIONS: Not relevant to this project.

Impacts
The paper on agricultural commodity options pricing has been presented in agricultural and financial academic conferences. It will appear in the Journal of Futures Markets. We find that overall model fitness tests favor the SVJ model consistently for one-year, two-year and five-year data. The in-sample and out-of-sample pricing and hedging results on corn, soybeans and wheat generally, with few exceptions, lend support for the SVJ model vs. the SV model. The results highlight a largely neglected empirical feature: jumps in commodity futures prices. Although the CME group sets daily price limits on agricultural commodity futures, futures prices still exhibit jumps in the daily sample data. The implication of our findings is that agricultural commodity options traders, including farmers and agribusinesses, should be concerned with price jumps and other model features when they manage risks and trade options. Accuracy of options and futures pricing is relevant to agricultural businesses and farmers (buyers) and market makers/dealers (sellers). Our comprehensive SVJ model prices options better than the SV model. For instance, the at-the-money soybean calls had a pricing error of 3.73 cents for the SV model and an error of 1.97 cents, a difference of 1.7 cents. For wheat puts the SV model has a pricing error of 2.47 cents and the SVJ model has an error of 0.48 cents, a difference of 1.99 cents. These small differences have an effect on a grain elevator's bottom line when millions of bushels are bought and sold. The first paper on cattle marketing is forthcoming in Canadian Journal of Agricultural Economics. We find that the coexistence of a risk premium wedge between marketing channel (live weight, dressed weight, and grid) pricing mechanisms, in conjunction with varying degrees of producer risk aversion or producer perception of carcass quality uncertainty, contributes to the coexistence of multiple marketing channels. It is also demonstrated that risk and risk preference provide the linkage between carcass quality uncertainty and producer marketing decisions. We demonstrate how this linkage can affect the structure of the fed cattle market and the variability in slaughter volume across marketing channels. We also confirm the linkage between value-based production techniques that increase seller information on carcass quality and seller increased usage of grid pricing regardless of actual carcass quality. Empirical evidence is provided in support of the supposition that carcass quality uncertainty plays a role in grid market share variability. The policy implication of our findings is that, it is essential to reduce informational uncertainty on carcass quality in order to encourage grid marketing. A working draft of the second paper on cattle marketing was presented in Southern Agricultural Economics Association annual conference. Using a novel EARCH-in-Mean model, we find that the grid premium and discount structure is slowly adjusting carcass quality market signals to encourage marketing on a grid and discourage marketing by the pen. If this trend continues, grid market share of steer and heifer slaughter volume should increase in the future.

Publications

  • 1. Wang, Z., Fausti, S.W. and Qasmi, B.A. (2012). Variance risk premiums and predictive power of alternative forward variances in the corn market. Journal of Futures Market, 32(6): 587-608.
  • 2. Fausti, S., Wang, Z. and Lange, B. (2012). Expected Utility, Risk, and Marketing Behavior: Theory and Evidence from the Fed Cattle Market. Canadian Journal of Agricultural Economics. DOI: 10.1111/j.1744-7976.2012.01261.x. In Press.
  • 3. Schmitz, A., Wang, Z. and Kimn, J. (2012). A Jump Diffusion Model for Agricultural Commodities with Bayesian Analysis. Journal of Futures Markets. In press.
  • 4. Fausti, S., Wang, Z., Diersen, M. and Qasmi, B. (2012). Pricing Fed Cattle on a Grid: An Analysis of the Incentive Mechanism over Time. Southern Agricultural Economics Association Annual Meeting, Birmingham, AL.
  • 5. Schmitz, A., Wang, Z. and Kimn, J. (2012). A Jump Diffusion Model for Agricultural Commodities with Bayesian Analysis. NCCC-134 2012 Annual Meeting, St. Louis, MO. 2/24-25, 2012.
  • 6. Schmitz, A., Wang, Z. and Kimn, J. (2012). A Jump Diffusion Model for Agricultural Commodities with Bayesian Analysis. Midwest Finance Association 2012 Annual Meeting, New Orleans, LA. 4/16-17, 2012.
  • 7. Schmitz, A., Wang, Z. and Kimn, J. (2012). A Jump Diffusion Model for Agricultural Commodities with Bayesian Analysis. Financial Management Association Annual Conference 2012, Atlanta, GA. 10/18-19, 2012.
  • 8. Wang, Z. and Daigler, R.T. (2012). The Option SKEW Index and the Volatility of Volatility. World Finance and Banking Symposium, Shanghai, China. 12/17-18, 2012.


Progress 01/01/11 to 12/31/11

Outputs
OUTPUTS: My AES research in 2011 was carried out along two lines: expansion of my previous research on variance risk premium in agricultural commodity markets and exploration of pricing agricultural commodity options using Markov Chain Monte Carlo (MCMC). Regarding the research on variance risk premium, in an extension to the published results for corn, we measured and forecasted variance based on realized variance, seasonal GARCH, implied variance based on Black (1976), and implied variance based on variance swap rate (or alternatively referred to as VIX) for two other top grains produced in South Dakota: wheat and soybeans. The existence of variance risk premiums and the predictive power of alternative variance measures were investigated. Adding to the previous literature, the seasonality of variance and variance risk premium was also examined. Regarding the research on pricing agricultural commodity options, we proposed a comprehensive model that incorporates four important features: price jumps, stochastic volatility, seasonality and stochastic cost of carry. Such model features have been considered separately, but not collectively, in the literature of pricing agricultural commodity futures and options. We employed a special MCMC algorithm, which is new in the agricultural commodity derivatives pricing literature, to estimate the proposed stochastic volatility (SV) and stochastic volatility with jumps (SVJ) models. We applied the proposed models to real options and futures data on corn, soybeans and wheat. PARTICIPANTS: Dr. Wang, Dr. Fausti and Dr. Qasmi collaborated on the research on variance risk premium of corn. Drs. Wang, Fausti, Diersen and Qasmi collaborated on research on fed cattle marketing and risk analysis. Mr. Krishna Pokharel, graduate student in Economics, completed his thesis in Summer 2011. Dr. Wang and Mr. Pokharel collaborated on the research on variance risk premium in the agricultural commodity markets. Mr. Adam Schmitz completed the first paper and is working on the second paper of his dissertation (PhD in Mathematics) on the pricing of options on agricultural commodities under the joint supervision of Dr. Wang of the Economics Department and Dr. Jung-Han Kimn of the Department of Mathematics and Statistics. The first paper is accepted for presentation at Midwest Finance Association annual meeting and NCCC134 in 2012. The second paper is also accepted for presentation at NCCC134 in 2012. TARGET AUDIENCES: The target audiences for the analysis of pricing and risk management of agricultural commodities are farmers, agribusinesses, investors, policy makers, students and researchers in South Dakota and other states that produce similar commodities. These audiences are reached through SDSU Economics Staff Papers, Economics Commentators, presentations in classrooms and professional conferences. Financial and economic researchers are reached through professional papers and journal articles. PROJECT MODIFICATIONS: Not relevant to this project.

Impacts
In the research on variance risk premium, we document a clear seasonality in variance and variance risk premium in all three grain markets for the sample spanning 1987 to 2009, as well as verify the existence of variance risk premium. More specifically, variances are higher during the growing season and peak in June; variance risk premiums based on variance swap are negative and usually peak in June or July. In other words, grain commodity options buyers, such as grain elevators and agricultural food processors, pay a high premium to insure against volatile prices, especially in June and July. In terms of predictive power, we confirm the clear dominance of variance-swap-based variance over seasonal GJR-GARCH in wheat and soybeans; on the other hand, we only find a modest advantage of variance-swap-based variance over Black (1976) implied variance in soybeans and wheat compared to a clear advantage in corn. The findings generally support the original conclusion that agricultural commodity option users need to take volatility risk in to consideration when making pricing and risk management decisions. In the research on agricultural commodity options pricing, we find that overall model fitness tests favor the SVJ model. The in-sample and out-of-sample pricing and hedging results on corn, soybeans and wheat generally, with few exceptions, lend support for the SVJ model vs. the SV model. The results highlight a largely neglected empirical feature: jumps in commodity futures prices. Although the CME group sets daily price limits on agricultural commodity futures, futures prices still exhibit jumps in the daily sample data. The implication of our findings is that agricultural commodity options traders should be concerned with price jumps and other model features when they manage risks and trade options. Accuracy of options and futures pricing is relevant to agricultural businesses and farmers (buyers) and market makers/dealers (sellers). Our comprehensive SVJ model prices options better than the SV model. For instance, the at-the-money soybean calls had a pricing error of 3.73 cents for the SV model and an error of 1.97 cents, a difference of 1.7 cents. For wheat puts the SV model has a pricing error of 2.47 cents and the SVJ model has an error of 0.48 cents, a difference of 1.99 cents. These small differences have an effect on a grain elevator's bottom line when millions of bushels are bought and sold.

Publications

  • Wang, Z. and Daigler, R.T. 2011. The performance of VIX option pricing models: Empirical evidence beyond simulation. Journal of Futures Markets. 31(3): pp251-281, March 2011.
  • Wang, Z., Fausti, S.W. and Qasmi, B.A. 2011. Variance risk premiums and predictive power of alternative forward variances in the corn market. Journal of Futures Markets. DOI: 10.1002/fut.20527, In Press.
  • Wang, Z., Fausti, S. and Qasmi, B.A. 2011. Variance risk premiums and predictive power of alternative forward variances in the corn market. Presentation at 2011 meeting of Financial Management Association, Denver, CO. October 20-21.
  • Wang, Z. 2011. Volatility in the Commodity Markets with an Emphasis on Grain Products. Invited presentation at the Sewrey Colloquium, South Dakota State University. February 15.
  • Pokharel, K. and Wang, Z. 2011. Measuring and Forecasting Variance and Variance Risk Premium in the Agricultural Market. Missouri Valley Economic Association 2011, Kansas City, Missouri. October 20.
  • Pokharel, K. 2011. Measuring and Forecasting Variance and Variance Risk Premium in the Soybean Market. Masters' Thesis in Economics, South Dakota State University.


Progress 01/01/10 to 12/31/10

Outputs
OUTPUTS: Realized variances based on corn futures prices and implied variances derived from options prices are computed from 1987 to 2009 at the daily frequency. During the 23-year period, realized variances average 572.4 (or equivalently 23.9% for annualized realized volatility) with standard deviation of 534.0. Implied variances have an average of 764.4 (or 27.6% for annualized implied volatility) with standard deviation of 522.1. It is clearly seen that implied variance is higher than realized variance on average. This observation is called "negative variance risk premium". This implies that, market participants are willing to pay a premium via options to reduce the future uncertainty in the corn market. Additional statistical features of all variance measures are investigated for the sample period. The same measures are also compared between growing season (defined as between May and September of a calendar year) and non-growing season (defined as between October of the same year and April of the following year). Realized variance and implied variance are 778.5 and 981.8 during the growing season of corn, and 468.2 and 654.6 during the non-growing season, respectively. Both measures point to a higher uncertainty in corn prices during the growing season than during the non-growing season. A GARCH variance model with seasonality in corn futures prices is then proposed and estimated. Econometrical analysis on the statistical significance of the difference between realized and implied variances is performed for the whole sample and for subsamples based on growing and non-growing seasons. The difference between the historically realized variance and future implied variance motivates further analysis of the predictive performance of the future variance. Three alternatives for future variance: Black-implied variance, model-free implied variance and seasonal-GARCH variance are proposed and compared based on their predictive performance. PARTICIPANTS: Dr. Wang, Dr. Fausti and Dr. Qasmi collaborated on the research on variance risk premium of corn and the best predictor for future volatility in corn market. Mr. Krishna Pokharel has completed empirical analysis of his thesis (MS in Economics) on variance risk premium of soybeans and predicting volatility in soybeans market. He has completed the first draft of his thesis. Mr. Adam Schmitz is starting his dissertation (PhD in Mathematics) on the pricing of options on agricultural commodities involving soybeans and corn under the joint supervision of Dr. Wang of the Economics Department and Dr. Jung-Han Kimn of the Department of Mathematics and Statistics. They collaborated on research papers on pricing agricultural commodity options. TARGET AUDIENCES: The target audiences for the analysis of pricing and risk management of agricultural commodities are farmers, agribusinesses, investors, policy makers, students and researchers in South Dakota and other states that produce similar commodities. These audiences are reached through SDSU Economics Staff Papers, Economics Commentators, presentations in classrooms and professional conferences. Financial and economic researchers are reached through professional papers and journal articles. PROJECT MODIFICATIONS: Not relevant to this project.

Impacts
Realized variances from corn futures prices are consistently lower than implied variances from options prices from 1987 to 2009, hence there is a negative variance risk premium. The premium varies over time. The results are robust for both growing and non-growing seasons. Seasonality in corn futures price is identified empirically through all measures of variances. Among the three candidates for predicting future variance, the model free implied variance, called "CornVIX", encompasses the most information at the time of prediction, and provides the most superior forecasting performance. Daily realized and implied variances computed from the research are available for farmers, agribusinesses and other participants in the corn market to accurately gauge the volatility risk. Our model-free implied volatility index, CornVIX, predates a similar volatility index product that is planned to be introduced to the market in 2011 by the CME group. Corn VIX contains rich information on how the market participants perceive the uncertainty in the future, and therefore can guide farmers in their marketing decision and agribusinesses in hedging against the volatility risk.

Publications

  • Wang, Z., Fausti, S.W. and Qasmi, B.A. 2010. Variance risk premiums and predictive power of alternative forward variances in the corn market. Econ Staff Paper 2010- 1.
  • Wang, Z., Fausti, S. and Qasmi, B. 2010. Variance risk premiums and predictive power of alternative forward variances in the corn market. Presentation at 2010 meeting of Western Economic Association, Portland, OR. June 30-July 2.
  • Wang, Z. 2009. Volatility risk. SDSU Economics Commentator, No.513.
  • Wang, Z. and Daigler, R.T. 2010. The performance of VIX option pricing models: Empirical evidence beyond simulation. Journal of Futures Markets, doi: 10.1002/fut.20466.
  • Wang, Z. and Bidarkota, P.V. 2010. Risk premia in forward foreign exchange rates: a comparison of signal extraction and regression methods. Empirical Economics. In press.