Hedonic trait valuation methods can be described as follows:
where: Vy is some measure of the economic value of a variety i (or a set of varieties) in location j; T1 „, T2i Tny is an index of the possession of traits 1, 2 n
(e.g. resistance to gall midge) of the variety or set of varieties in location j; Zy is a vector of economic and ecological factors. The function itself may be linear or non-linear.
Crop-loss data are typically available for a location and time period. In Indonesia, losses by type (insect and disease) are measured by province and year. A matching set of trait indexes is required. This requires data on varieties planted and trait ratings by variety. For Indonesia it was possible to compute the percentage of area planted in each region and period with specific traits.
Equation (1) does not explicitly deal with several econometric problems. The first, as noted above, is that there is a natural incidence factor for pests and diseases. And this may vary by both location and time period. If a good measure of this factor were included in Zy, equation (1) could be regarded as a 'technical' relationship between losses and traits. The fact that the traits, i.e. the adoption of varieties with the traits, may be endogenous (e.g. it may respond to the Zy vector and to economic factors) can be set aside if the Zy vector is complete and controls for differences in natural incidence. But if natural incidence is not well measured, endogeneity cannot be set aside.
A related problem lies behind this specification; reduced crop losses (as measured by per cent of the crop actually lost) may be a poor measure of the value of a trait. Ideally one would like a measure of average variable cost. Pesticides, herbicides, etc., as well as farm practices, can be and are used to reduce crop losses. The incorporation of traits may reduce the costs of these chemicals and practices. As noted earlier, these traits may also have the effect of enabling the adoption of modern high-yielding varieties (i.e. with high-yielding quantitative traits) to be adopted in locations where they otherwise would not be adopted (the ecological stress tolerance traits would be particularly likely to have this effect).
In this study, the problem of endogeneity of traits will not be addressed directly (see Evenson, 1994, for a treatment of this problem). However, three measures of this will be utilized: crop losses, pesticide use and TFP (which is an index of changes in average variable cost).
Was this article helpful?