UC Berkeley

This dissertation consists of three essays that provide insights into the economics of water across different dimensions of the resource and its role in the state of California. The first essay examines the social welfare impacts of variation in irrigation supplies that are available through major public projects. Discussion of the value and significance of the irrigation services that are made available through the Sacramento-San Joaquin Bay Delta often focus on the immediate impacts to agricultural production and direct farm jobs. This essay, however, considers the reach of these impacts by evaluating how agriculturally based communities are effected by shortages in irrigation supplies. The second and third essays shift attention toward urban water usage. Methods of forecasting urban water demand are reconsidered and a suggestion is made for an alternative approach to evaluating the predictive power of demand models in the second essay. Finally, the third essay measures the effect of consumption analytics and social norm messaging on household decision-making around water usage. Taken together, these three essays address some of the key features of California's water economy.

While there exists much research that measures the impact of precipitation shocks on agricultural regions, whether in production or in other outcomes, less research is available that specifically focuses on the impacts of variation in developed irrigation supplies. Given that developing irrigation infrastructure is oft regarded as an adaptation strategy for climate change, it is worth understanding how shocks in the supply of managed water effect individual and regional outcomes. The first essay exploits exogenous variation in the availability of California's Sacramento-San Joaquin Delta irrigation water to estimate the impact on crime rates for the agricultural counties that use this input. This research provides suggestive evidence in support of the hypothesis that reductions in the availability of this irrigation supply lead to a socially and economically significant increase in both property and violent crime rates. Empirical results support the argument that farm jobs is the mostly likely mechanism, with suggestive evidence that demographic changes are also important.

Urban water managers rely heavily on forecasts of water consumption to determine management decisions and investment choices. Typical forecasts rely on simple models whose criteria for selection has little to do with their performance in predicting out-of-sample consumption levels. This essay demonstrate this issue by comparing forecast models selected on the basis of their ability to perform well in-sample versus out-of-sample. Results highlight the benefits of developing out-of-sample evaluation criteria to ascertain model performance. Using annual data on single-family residential water consumption in Southern California, this research illustrates how prediction ability varies according to model evaluation method. Using a training dataset, this analysis finds that models ranking highly on in-sample performance significantly over-estimated consumption $(10\%-25\%)$ five years out from the end of the training dataset relative to observed demands five years out from the end of the training dataset. Whereas, the top models selected using the out-of-sample criteria came within 1\% of the actual total consumption. Notably, projections of future demand for the in-sample models indicate increasing aggregate water consumption over a 25-year period, which contrasts the downward trend predicted by the out-of-sample models.

The third essay estimates how household-level water consumption may be impacted by the distribution of Home Water Use Reports (HWURs) by Dropcountr (DC), a digital and web-based consumption analytics platform. Similar to Opower in the energy sector, DC offers social comparison, consumption analytics, and conservation information to residential accounts, primarily through digital communications. Having initiated relationships with several California utilities, as well as major Texas and Colorado providers, the effect of these programs may be measured and will contribute to three areas of academic literature: 1) the study of social norms and moral suasion on consumption behavior, in general; 2) the effects of such methods in the water sector, in specific; and 3) understanding alternatives to price mechanisms in demand-side management of water resources. This research discusses the potential of this type of information to generate measurable effects of interest, both to researchers and to water managers alike. Particular focus will be given to results with a mid-sized California utility and a major Texas provider. Early results indicate an economically and statistically significant $5-8\%$ and $3-4\%$ reduction in average monthly household water consumption for the California and the Texas utility, respectively, for the typical household under treatment of the DC program.