To address natural resource scarcity and externalities, economists emphasize property rights and prices. In contrast, engineers and policymakers often emphasize resource-conserving technologies, such as energy-efficient or water-efficient technologies, and input-efficient (precision) agriculture and forestry. Proponents of public programs that encourage adoption of these technologies have identified numerous product adoption "puzzles," in which adoption [disadoption] rates are low [high] despite engineering estimates that imply both users and the environment would benefit. Economists have been skeptical of such puzzles, but have relied on observational designs in which unbiased estimation of effects is challenging, or on experimental designs with low adoption rates.
To shed light on this debate, we report results from two randomized trials using water-efficient technologies at the household level. In a first RCT, we confirm prior claims that engineering estimates of input reductions are substantially exaggerated (more than three-fold). We demonstrate that the divergence in impact estimates can be attributed to engineering and behavioral reasons other than the "rebound effect" that has attracted the most attention from economists. Moreover, by combining our experimental estimator, detailed cost information, and experimentally elicited and jointly estimated time and risk preferences from the target population, we demonstrate the private welfare gains from technology adoption are approximately zero, implying no "efficiency paradox."
We then move to explore whether disadoption of pro-social goods can be reduced when longer exposure enhances the perceived net benefits of using the goods. In a second RCT we find that incentivized exposure increased the likelihood that a household was still using low-flow water fixtures one year after subsidies were no longer offered. Longer exposure provided households with better information on water bill savings and increased households’ taste for the fixtures. We use a conceptual model to develop intuition about the optimal time distribution of subsidies and show that short-term exposure subsidies for pro-social goods can outperform purchase subsidies and use subsidies.
Francisco Alpízar (jointly with Maria Bernedo, Paul Ferraro, and Ben Meiselman)