« Back to Results

Water Conservation Policy

Paper Session

Friday, Jan. 4, 2019 2:30 PM - 4:30 PM

Atlanta Marriott Marquis, A701
Hosted By: Association of Environmental and Resource Economists
  • Chair: Laura Grant, Claremont McKenna College

Determinants of Water Conservation: Evidence from the Recent California Drought

Oliver Roman Browne
University of Chicago
Ludovica Gazze
University of Chicago
Michael Greenstone
University of Chicago


Managing demand for residential water use is increasingly important in light of increased climate
variability, growing population, depleted groundwater, and the high costs of developing new supply.
During the recent 2012-2016 drought in California, many utilities achieved state-mandated goals to
reduce their residential water use by over 25%, representing a huge public policy success. However, it is
unclear which particular policies drove this conservation. This paper seeks to disentangle the impacts
different state and municipal policies had on residential water use. We use hourly micro-data from over
86,000 single family households between 2013 and 2016 in a large Californian city.
First, we find that a 10% increase in marginal rates is associated with a decrease in use of 20
gallons/day. Over our sample period, rate changes are responsible for saving 25 gallons/day. Second,
reducing the number of days households are allowed to use water outdoors results in households
substituting water use from banned to non-banned days. However, water use persistently decreases by
6% (30 gal/day) after this policy change, specifically during hours when outdoor use was never
permitted, suggesting households may be reacting to the announcement of the change. Consistently,
water use declines by 74 and 44 gallons/day after the announcement of “State of Emergency” and
“Mandatory 25% Water Use Reductions,” respectively. Indeed, these major state-level announcements
appear to induce interest in the drought, as measured by Google searches. A mediation analysis shows
that our measure of drought awareness is highly correlated with water use, but, after controlling for city
and state policies, this correlation disappears. Finally, while we see large savings (55 gallons/day) from
toilet and lawn rebates, total impacts are negligible due to low-take up rates.

Exposure-Enhanced Goods and Technology Disadoption: Evidence from a Randomized Controlled Trial with Resource-Conserving Technologies

Paul Ferraro
Johns Hopkins University
Francisco Alpizar
Tropical Agricultural Research and Higher Education Center (CATIE)
Maria Bernedo
University of Maryland-Baltimore County
Ben Meiselman
Johns Hopkins University


Pro‐environmental technologies are widely promoted for conserving resources and mitigating negative
externalities. However, the impacts from their adoption will be blunted if they are subsequently
disadopted. Disadoption is a problem in many environmental contexts (e.g., agriculture, energy and
water efficiency, cook stoves, sanitation). Yet in comparison to the myriad studies exploring the
mechanisms by which firms and consumers adopt technologies, much less theoretical and empirical
attention has been paid to the process of disadoption.
We focus on a mechanism that may contribute to disadoption and can be directly influenced by policy
interventions. We posit that the perceived net benefits from some technologies are enhanced by
exposure to the technologies. Exposure affects adopters’ perceived net benefits through habit
formation, learning‐by‐doing, or consumption experience (for experience goods). These mechanisms
have been extensively studied in separate literatures, but all have the same implication: exposure to a
technology can increase the perceived benefits or decrease the perceived costs of use, which implies
that the longer adopters are exposed to a technology, the less likely they are to disadopt it.
We present a unified conceptual framework for exposure‐enhanced goods and test whether short‐term
cash incentives that increase exposure to such goods are effective at reducing longer‐term disadoption.
In a randomized controlled trial in rural Costa Rica that assigned water‐efficient technologies to nearly
one thousand households, some households were offered a cash bonus if, during an unannounced audit
four to six months later, they were still using the technologies. Studying the effect of exposure on
disadoption is typically complicated by endogeneity, but the randomized bonus provides an exogenous
shock to short‐term exposure. A year after the bonus was discontinued, disadoption rates were
measured again in a surprise audit. Consistent with the hypothesized mechanisms for disadoption of
exposure‐enhanced goods, longer exposure induced by the bonus reduced disadoption.

Energy Efficiency in Agriculture: Electricity Use, Water Demand, and Farm Profits

Fiona Burlig
University of Chicago
Louis Preonas
University of California-Berkeley
Matthew Woerman
University of California-Berkeley


Subsidies for energy efficiency are a key component of global climate policy, yet a
growing body of empirical research finds that energy efficiency investments underdeliver
relative to ex ante engineering projections. However, relatively little is
known about how energy efficiency performs in industrial settings—where high
energy costs may increase both the salience and savings potential of energy efficient
technologies. In this paper, we study an important industrial setting: California
agriculture. California’s electric utilities spend nearly $50 million annually to
improve energy efficiency in the agricultural sector, and thousands of farms have
received subsidies for energy efficient pump replacements. If these subsidies
succeed in reducing energy consumed in groundwater pumping, this would lower
the marginal cost of irrigation and could potentially increase consumption of water,
a scarce common-pool resource.
We plan to quantify the private welfare impacts of agricultural pump upgrades.
First, we will leverage an arbitrary eligibility cutoff to estimate the causal effect of
pump upgrades on electricity use, and compare our ex post savings estimates to ex
ante engineering projections. Then, we will convert electricity consumption to water
use to estimate the causal effect of pump upgrades on water extraction.
Incorporating satellite data on agricultural yields and data from a farm survey of
non-energy input costs, we will be able to estimate the effect of pump upgrades on
farm profits. This will contribute to the energy efficiency literature, where credible
welfare estimates are rare. We will also contribute to the agricultural literature with
well-identified quasi-experimental estimates of the demand elasticity of water.

The Roles of Price and Persuasion on Consumer Behavior: Experimental Evidence from Water Consumption

Daniel Brent
Louisiana State University
Casey Wichman
Resources for the Future


Social comparisons are a popular behavioral nudge to promote conservation of energy and
water, partially because raising the price of these resources is politically difficult. Nudges may
interact with prevailing prices, however, potentially crowding out intrinsic motivation to
conserve or increasing the salience of prices.
To test the mechanism through which consumer respond to behavioral nudges, we explore
water conservation behavior in response to social messaging experiments and large changes
in marginal prices for water use. Our analysis produces causal effects by design: first, we
evaluate the effects of two independent, randomized messaging experiments implemented by
WaterSmart Software at different points in time for neighboring water utilities in Southern
California. Second, we exploit two sources of variation that introduce price changes at the
household level. One source of price variation comes from arbitrary lot-size thresholds
within nonlinear water rate structures. The second source of price variation arises from the
utilities' rate-setting practices, included in an instrumental variables framework. Our
methodology cleanly identifies the separate impacts of the social comparison treatment and
price on consumer behavior as well as their joint effect.
Within our empirical approach, we answer two questions. First, do customers facing higher
price levels respond more strongly to norm-based conservation campaigns? We identify this
price-level effect by comparing responsiveness to behavioral treatments for otherwise identical
households on either side of a price discontinuity introduced by arbitrary lot-size thresholds
within utility rate structures. Second, do norm-based conservation campaigns increase
customers' price sensitivity? We identify this price-sensitivity effect by estimating consumer
demand and observing whether our randomized behavioral treatment significantly alters our
price elasticity estimates.
Our results suggest that responsiveness to social comparisons are unrelated to price,
providing strong evidence that social comparison nudges operate through a “moral tax”
mechanism rather than providing private economic benefits via internality correction.
Laura Grant
Claremont McKenna College
Nick Magnan
University of Georgia
Katrina Jessoe
University of California-Davis
Michael Price
University of Alabama
JEL Classifications
  • Q2 - Renewable Resources and Conservation
  • D9 - Micro-Based Behavioral Economics