Selection of Representative Crop Activities

Choices of farm activities are determined by the research problem and the scope of agricultural issues identified by the government. For example, if policy-makers are interested in the tax/subsidy impact of government policies on the agricultural sector, one or two representative budgets for each crop should be sufficient. If the research instead focuses on a single crop or technology, a more detailed specification of commodity production is needed and a larger number of representative firms should be used. Sectoral income distribution objectives require commodity systems that highlight the small farm-large farm dichotomy; concerns for regional growth require recognition of region-specific commodity systems.

Within each group of commodity systems, analysts may still desire to characterize production heterogeneity in some detail. Regional classifications are perhaps the most common indicator of heterogeneity, because differences in agroclimatic zones-characterized by soil fertility, topography, and access to water-typically influence the choice of technology and the level of input use. Differences in farm size are a second source of heterogeneity. Small farms often use variable inputs, such as fertilizer and labor, with different intensities than large farms. Large-farm systems are often capital-intensive, and fixed costs account for a more substantial share of total costs. Differential access to re-sources creates heterogeneity in the relative usage of machinery and labor inputs.

The development of a list of representative systems can draw on various information sources. Aggregate production estimates, usually prepared by the ministry of agriculture, can be obtained for particular crops and regions. Sometimes these estimates are decomposed by farm size or technological characteristics (number of animals or amount of machinery, for example); in this circumstance, aggregate data can be used to specify the technological alternatives as well. But the identification of specific technologies usually requires first-hand observation and the assistance of farm management personnel from development projects, universities or ministries of agriculture, and members of local extension services. Initial lists of potentially representative systems for PAM analysis will usually be longer than the research projects can manage, and some reduction in the list will be necessary. Short field trips with expert observers are useful at this stage to give the analyst a better idea of the distinctions among commodity systems. Box 9.1 illustrates the system selection process for agriculture in northwest Mexico.