Social comparisons and groundwater use: Evidence from Colorado and Kansas

In the United States, agriculture is responsible for the majority of consumptive water use. To reduce consumptive use in water scarce regions, policymakers have implemented a number of costly interventions. These interventions range from land retirement to subsidies that encourage the adoption of efficient irrigation technologies. In nonagricultural contexts, costly policy interventions have been complemented by low-cost interventions inspired by behavioral economics. Whether these behavioral interventions are effective in the context of commercial farming is not well understood. In a preregistered, randomized field intervention, we estimate the impact of social (peer) comparisons on agricultural groundwater users in Colorado and Kansas. More than three thousand irrigators were randomized to receive either an annual peer comparison or no comparison. The peer comparison contrasted each irrigator's groundwater use to the distribution of use by neighboring irrigators. The comparison intervention reduced average annual groundwater use by 4.05% [95% CI (-5.87%, - 2.21%)], resulting in an aggregate reduction of more than 21,000 acre-feet per year at a cost less than $1.31 per acre-foot conserved. The estimated treatment effect was larger among irrigators with lower pre-intervention water use. In the 3-year experiment, we observed no evidence that the treatment effect substantially attenuated over time. We did, however, detect within-irrigator spillovers in the treatment group: groundwater use also declined among wells that were not included in the peer comparisons (peer comparisons included a maximum of three wells). The results imply that social comparisons can be a cost-effective tool, alongside other policy interventions, aimed at reducing agricultural water use.