An Empirical Analysis of Farm Adaptation Response to Water Security Using a Natural Policy Experiment

The capacity of farm-level adaptation to mitigate the impacts of climate change in arid regions dominated by irrigated agriculture fed by snowpack is a critical challenge for developing accurate integrated engineering-economic modeling tools. Economic optimization models provide a valuable benchmark for the theoretical limit of adaptation given a set of clear objectives, conditions, and constraints. However, a major limitation to specifying tractable and accurate models is the large number of potential adaptation strategies that can be taken. There is a need for more empirical research that reveals preferred adaptation strategies in a way that identifies causal relationships.

This research seeks to add to the empirical literature on adaptation by exploiting what in the econometric literature is called a “natural experiment” where a policy has isapplied to one group but not another in a way that is random relative to the variables of interest so as to reduce problems of bias in coefficient estimates caused by unobserved heterogeneity.

The region of study is the Yakima Basin in Washington State which is a highly diverse region in terms of crop and irrigation technology. This creates significant complication for modeling adaptation since farmers have a wide array of choices including changing cropping patterns and irrigation technologies. Other strategies including water trading and deficit irrigation.

Two irrigation districts in the Yakima Basin, Roza and Sunnyside, are adjacent to each other and are nearly identical in growing conditions. The difference is that Roza is severely curtailed during droughts while Sunnyside is not. With the availability of detailed field level data on cropping patterns, irrigation technologies, and land ownership this presents an opportunity to identify the effect of water security risk on farm-level decision making. Preliminary results show that after controlling for other features, a field in Roza is 5% more likely to have an efficient irrigation technology than Sunnyside. Roza irrigators are more likely to incorporate wine grapes which are tolerant to water stress and use less water than most other crops. Other findings with respect to farm and owner characteristics help inform model calibration.