While researchers have suggested that WDs may be useful for hedging systemic
risk, the use of WDs by producers is questionable. For example, Vedenov and Barnett
(2004; hereafter VB) analyze the efficiency of WDs as primary hedging instruments for
corn, soybeans, and cotton in the U.S. at the CRD level of aggregation. Based on
relatively complex non-linear combinations of monthly (June, July, and August)
precipitation and temperature indexes, VB’s results suggest only the limited efficacy of
WDs in hedging disaggregated production exposures.4
This study builds on earlier research in two important dimensions. First, hedging
effectiveness of WDs are investigated at varying levels of spatial aggregation (i.e., the
state and CRD level). Yields evaluated at low levels of aggregation (e.g., farm or CRD
level) are likely much riskier than those at higher levels (e.g., state level) because the
potential degree to which idiosyncratic risks self-diversify increases as the level of
aggregation increases. Yet, high temperature spatial correlations induce significant
correlations among low-level yield exposures. Thus, relatively more risk may be left in
the form of systemic weather risk and the hedging effectiveness of WDs may increase as
the level of aggregation is increased. Analysis of aggregated yields may also be more
relevant from the re/insurers viewpoint as aggregate yield risk more accurately embodies
their systemic risk.
Second, we investigate straight forward seasonal temperature WDs in lieu of
complex monthly temperature and precipitation WDs. Persistence in weather conditions
may induce a high degree of collinearity among precipitation and temperature (Namias
1986). This, along with the fact that weather conditions in the U.S. during the summer
tend to be autocorrelated (Jewson and Brix 2005), increases the probability of