In the standard specification of intertemporal EU maximization, it is common to assume
an additive and homogeneous von Neumann-Morgenstern utility index. Such a specification,
however, intertwines two distinct aspects of preference, intertemporal substitutability and
relative risk aversion (Epstein and Zin, 1989). Additionally, these models did not perform well
in empirical examinations (Hansen and Singleton, 1983; Mehra and Prescott, 1985). As a more
general framework, the GEU model adds extra flexibility in identifying intertemporal
substitution and is able to disentangle the intertemporal substitution from the risk aversion.
With the possible and testable separability for risk preference and intertemporal
substitutability, it is possible to use the GEU model to estimate preference parameters separately
and examine the form of the objective function. Continuing on from their theoretical paper,
Epstein and Zin (1991) empirically investigated the parameter estimation and the testable
restrictions. They got favorable and theoretically consistent estimates. Lence (2000) used 1936-
1994 U.S. farm data to study the fitness of a GEU framework. He found the estimated farmers’
utility parameters satisfy the theoretical restrictions of the GEU model. Furthermore, the EU
model is rejected in favor of the GEU model. The empirical results from resource economics
studies using the GEU model (Knapp and Olson,1996; Howitt et al. 2002) underscore the
importance of using the more general specification of intertemporal preferences.
On the other hand, studies on agricultural risk management strategies have been
extended from the earlier one-element models to portfolio models, and focus more on the
interactions and relative impacts of the instruments within a portfolio. Among them are
portfolios of crop yield insurance and futures contracts (Myers, 1988), futures market and
government farm programs (Crain and Lee, 1996), crop yield insurance, futures, options and