COMMODITY PRICE DYNAMICS: DISCONTINUITY AND JUMPS
Open Access
- Author:
- Tian, Jiachuan
- Graduate Program:
- Agricultural, Environmental and Regional Economics
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- August 16, 2016
- Committee Members:
- Robert D. Weaver, Dissertation Advisor/Co-Advisor
Robert D. Weaver, Committee Chair/Co-Chair
Katherine Y. Zipp, Committee Member
Scott J. Colby, Committee Member
Alexei Novikov, Outside Member - Keywords:
- Price Dynamics
Jumps
Time Series
Food and Energy Prices - Abstract:
- The variation of energy prices has been a traditional source of shocks to the real economy. In many cases, this variation has manifested in jumps in energy prices that were characterized by some persistence. From another perspective, energy price volatility has historically been noted and its effects on real economy debated. Historically, the importance of the shocks to the real economy has led them to be labeled as energy crises, as they were argued to have resulted in substantial changes in real prices that induced changes in behavior on the demand and supply sides of the many markets. The first chapter re-examines evidence of such a linkage by considering the transmission of energy prices into soft commodity prices. This nexus lies within the core of any real effects as softs include food-related commodities. The paper contributes to the literature by re-examining this linkage with a close eye on the role played by structural breaks within a time series and by considering the question of causality within a nonlinear framework. We find that functional form is a critical specification that conditions inference. Using linear forms, we find no cointegration between energy and food in the full sample under the maintained hypothesis that there are no structural breaks. Using linear nonparametric methods, we examine the series for structural breaks and find evidence of their importance. Based on subdivisions of the sample period as suggested by the structural break examination, within the structural break intervals identified we find evidence of cointegration. We next reconsider the issue within the context of nonlinear functional forms posing the question of whether evidence of structural breaks based on linear methods follow from underlying nonlinearity. Our results confirm the importance of functional form specification and we find evidence of nonlinear causality between energy and soft commodity prices. Empirical studies of transmission of energy prices into the real economy have been challenged by a number of significant specification issues that have resulted in substantial variation in inference drawn from results. Among these issues is the question of completeness of model specification. Chapter 2 examines the question of whether such models need to incorporate macroeconomic indicators. Clearly, macroeconomic factors such as interest rates and exchange rates play a role in the determination of energy and commodity prices, however, considerable specification uncertainty characterizes the question of which macro metrics to incorporate. We examine this issue from the perspective of weak exogeneity and find evidence that the parameter estimates associated with time series models that exclude consideration of macro indicators are not compromised by their exclusion. We examine this issue using Italian, U.S. grain, and Brent crude oil prices. While structural break, threshold and asymmetric cointegration models can allow us to characterize the linear and nonlinear dynamics in price transmission in level,it is of equal interest to differentiate across the type of price change to consider what might be thought of as typical price changes versus extreme price changes associated with either temporary structural change or mean reverting change as in what we call price jumps. In particular, while a “structural break” is a permanent and long-run structural shift in DGM, a “jump” in a series represents a sudden temporary change in the pattern of the observations generated. Such change is temporary in a sense that its effect usually diminishes rather quickly (usually in relatively few periods). That means, intuitively, in relatively short time span after a jump, the price series will revert to its “mean” or its long-run smooth pattern which we call the trend of the series. In Chapter 3, we present a detailed discussion of the proper representation of such price jumps and show that there are price jumps in the real-world economic price series. The last chapter is concerned with the micro-structure specification to identify origins of price jumps that can not be generally characterized by the competitive market models. In particular we propose a rather general model of procurement process where imperfectly informed buyers search for and place bids to suppliers to fulfill procurement demand. We show that in this process, search cost, market structure and market condition are crucial factors in generating price jumps. Later in the simulation part we show that the model proposed in this paper can generate jumps that resemble those in the observed economic price series. We also integrate buyers’ risk aversion in market conditions (though they are risk neutral in payoffs) through their personal belief and search costs. We show that buyers’ risk aversion increases their sensitivity to market conditions,which exaggerates price movements with presence of supply shocks.