Optimal capacity in the anhydrous ammonia industry
Abstract
This thesis extends the real option value framework developed by Pindyck to define optimal capacity for the U.S. anhydrous ammonia industry from 1971-1991. The model is well suited to the characteristics of ammonia capacity investment. It recognizes the role of uncertainty in the demand for ammonia net of the cost of natural gas. This "net demand" series is an important gauge of the industry's situation and has at times been quite volatile. The model incorporates the implicit value of being able to shut down capacity to cut losses in periods of adverse economic conditions. Finally, the model recognizes the ability of firms to delay irreversible capacity investments. This provides the firm with a valuable option that must be exercised at the time of investment. The real option value framework has proven useful in modeling several topics of theoretical importance. To date, however, only a few studies have brought this method to bear in empirical work. This study represents a first attempt at modeling an industry with multiple option characteristics. The model assumes that the net demand curve intercept follows either mean-reverting or constant drift, geometric Brownian motion. In each year, only slight differences in optimal capacity were observed between these two model specifications. If only the optimal level of investment is of interest, then constant drift appears to be a reasonable assumption. The decisions by ammonia firms to expand capacity after 1973 came during a period of high uncertainty. Because increasing uncertainty dampens investment, the model prescribed a much more moderate increase in optimal capacity than what actually occurred. Had analysts had access to the real option value model in the early seventies, they could have warned against such rapid expansion. If increases in the wholesale price of ammonia had been sustained past 1975, the real option model would have called for a gradual increase in capacity in response to lower variance. Confidence intervals for optimal capacity in 1978, 1985, and 1991 were estimated. The results indicate that, in each year, parametric uncertainty could not justify the higher than optimal states of actual capacity.
Degree
Ph.D.
Advisors
Hertel, Purdue University.
Subject Area
Agricultural economics
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.