Farmer Production Decisions

The third subproject assesses whether and how farmers adapt to variability in green and blue water, and how farmer decisions about production affect their vulnerability to changes in the rainfall. We are interested in (1) the implications of changing availability of blue and green water for the outcomes of these decisions, (2) whether farmer choices are influenced by changing water availability, and (3) how individual farmer decisions are influenced by market expansion and experiences of peers and will test if intensification of agricultural production is a function of blue and green water, market prices, farmer and farm characteristics, and Forest Code enforcement.


We will collect survey data on (1) costs and revenues from all production activities; (2) household physical and human capital; (3) land use, including specifically in riparian zones and headwaters; and (4) perceptions of the current status and changes over the past decade in blue and green water (including the water table, surface water flows, and water stress affecting pastures). Our survey sample will be 1,500 household farming properties in our three field sites (500 in each site) using double stratification based on rural population (of the municipality) and watershed size. Our survey was informed by focus groups conducted 3 months prior to the survey.

Methods and Activities

We will establish a typology of farm production systems using cluster analysis, and estimate the choice of system and the impacts of green and blue water on farm income conditional on that choice using a multinomial simultaneous equations model with endogenous switching regression. The dynamics of changes in production systems will be estimated with a subset of farms in the sample for which we have over 20 years of panel data on water availability, household and land characteristics, and production and revenues. We also have detailed data on the relationships that these farmers have with local dairies and with agricultural associations. This analysis allows us to go beyond the equilibrium responses estimated using cross-section data to compare short-term and long-term responses to changes in water inputs, and to estimate dynamic feedbacks between the choice of production system and farm incomes.