feeds coefficient for cows is negative across the production systems albeit only statistically
significant for the Kenyan production systems revealing the reluctance of the cattle keepers to
have cows that require externally purchased feed inputs. The purchase price coefficients are
not statistically significant for both cows and bulls traits. The Independence of Irrelevant
Alternatives (IIA) test procedure developed by Hausman and McFadden showed IIA
violations for both bulls and cows at the 1 percent level. Consequently, mixed logit, a less
restrictive model, was estimated.
The results of the simulated maximum likelihood estimates for the mixed logit model
are presented in Table 3. The overall model is statistically significant and fit the data slightly
better than the fixed, multinomial logit model, with a Pseudo R-squared of 31% for bulls and
22% for cows. The likelihood ratio tests also result in rejection of the null hypothesis that the
multinomial logit models fit the data significantly better than the mixed logit models. The
mean coefficients of the random parameters are statistically significant, with the expected
signs. The standard deviations of the random parameters are also statistically significant
indicating significant preference heterogeneity within the sampled population. The non-
random parameter, low watering frequency is positive and highly significant for bulls,
implying that there is preference for bulls that are drought tolerant (need to water only once in
two days), it is however, not significant for cows. The constant variable in tables 2 and 3
represent the “none” choice option and is the base for the choice model, as it is associated
with “zero” utility. It takes a value of one if the option is “none” and zero otherwise. The
results indicate a strong negative preference for this option, implying that the respondents
preferred to select the other two choice options associated with various trait levels.
The marginal rate of substitution between the traits and the purchase price coefficient
(jβk /βp ) provides an estimation of implicit prices of the traits, also known as willingness to