Association for Public Policy Analysis & Management

Many policies and incentive programs are implemented to encourage the adoption of energy efficiency. However, in certain cases even without these policy interventions, consumers on increasing block rate might have an incentive to over-invest in energy efficiency because the marginal prices paid by consumers are higher than the marginal social cost of electricity supply. In order to help determine whether existing policies over-incentivize energy-efficiency investment, we also need to quantify the social and private benefits of energy efficiency under other rate structures as well, particularly time-variant electricity pricing. In this study, we provide empirical evidence of the private and social benefits of energy efficiency using hourly smart meter data under Time-of-use (TOU) pricing and compare with those under increasing block rate.

Data: Our main data comes from the utility company Salt River Project (SRP) in Arizona. We obtained the energy efficient AC replacements recorded by SRP’s rebate program in 2016-2017 and a Residential Equipment and Technology survey in 2014. We have data on enrollment in electricity rate plans, building characteristics, and socio-demographics. The customer-level smart metering data spans from 2013 to 2017. The final dataset compiles the information for about 16,000 households.

Methods: We conduct propensity score matching to find a group of control (without energy efficient ACs) and treatment (with energy efficient ACs) customers, using a rich set of building and household characteristics as well as pre-treatment energy consumption patterns. We then run the fixed effects panel regression for the TOU and non-TOU (increasing block rate) consumers separately. In addition, we conduct several robustness checks, including coarsened exact matching, building-by-year fixed effects, and also a machine learning approach in which we use LASSO and regression tree to predict counterfactuals.

Results: In summer months, electricity savings occur across the day for both TOU and non-TOU consumers. The largest electricity savings occur during the late afternoon and evening, which are usually the peak hours. During the winter months, for the non-TOU consumers, there are electricity savings in the evening hours and early morning hours. Furthermore, the electricity savings vary across households with different characteristics.

Based on the hourly electricity savings, we find that for the non-TOU consumers, the social benefits are greater than the private benefits by 7% while for the TOU consumers, the social benefits exceed the private benefits by 16%. Moreover, energy efficiency makes the consumers more elastic to price changes during the summer months.

Conclusion: This study provides the first empirical evidence of a comparison of the social versus private benefits from energy efficiency under TOU rate. We show that in Arizona for both TOU and non-TOU consumers, the social benefits are larger than their private benefits. Our results have several policy implications. First, the deviations between private and social benefits are different for different price plans, implying potentially different policy incentives for encouraging energy efficiency for consumers under different price plans. Second, policymakers can improve the design of the demand response programs by combining energy efficiency and TOU pricing due to larger price elasticities under TOU.