Cross-Commodity Perspective on Contracting: Evidenc e from Mississippi
Asset specificity does not appear to be significantly influencing contracting decisions.
However, with a p-value equal to 0.13, one may argue that it does play some role. The sign
on the coe¢cient for ASSETSP is as expected. Increases in the level of asset specificity
tend to increase the probability of contracting, but the level of statistical significance is weak.
College graduates are significantly more likely to contract and older producers are less likely
to contract.
A second logisitic regression model was also estimated. In this model, the definition
for contracting was confined to include cash forward contracts, resource providing contracts,
and production management contracts. Marketing pools were not considered contracting
in this model because it may be argued that the motivations for participating in marketing
pools is somewhat different than the motivation for participation in other contract types.
The results of the second regression model are shown in Table 6.
As with the previous model, price risk does not significantly affect contracting deci-
sions. Thus, even with a more restrictive definition of contracting, price risk does not appear
to play a significant role. Also similar to the previous model, transactions cost as reflected
by HOURS, as well as attitudinal variables significantly affect contracting decisions. Unlike
the previous model, asset specificity appears to significantly influence contracting decisions.
Thus, under a more restrictive definition of contracting, asset specificity appears to influence
producer decisions. It should be noted that asset specificity was marginally statistically sig-
nificant in the full model, so this result only strengthens the conclusion that asset specificity
is important.
15