Most studies measuring asymmetric adjustments in vertical price transmissions fail to provide empirical support to explain such behavior. The literature invokes theoretical models, which derive asymmetric behavior based on variables that are difficult to measure such as oligopolies' coordination policies, market imperfections or menu costs. Therefore, with no empirical support explaining the asymmetries, these studies leave no room for policy implementation. In this paper I relate asymmetric price responses to a theory of behavior under risk driven by the perishable rate of the goods. Retailers of a perishable good facing an increment in the wholesale price may decide not to increase their prices for fear of being left with a spoiled product. Using three agricultural products with different perishable rates I reject the null hypothesis of symmetric adjustments in the most perishable product but fail to reject for the less perishable goods. The nonlinear responses are consistent with the prediction of the model. The test for asymmetries uses a threshold cointegration technique where the threshold level and the cointegration vector are estimated from the data instead of being imposed by the econometrician.