Experimental Evaluation of Residential Demand Response in California

We evaluate the causal effect of hour-ahead price interventions on the reduction of residential electricity consumption, using a large-scale experiment on 7,000 households in California. In addition to this experimental approach, we also develop a non-experimental framework that allows for an estimation of the desired treatment effect on an individual level by estimating user-level counterfactuals using time-series prediction. This approach crucially eliminates the need for a randomized experiment. Both approaches estimate a reduction of ≈0.10 kWh (11%) per Demand Response event and household. Using different incentive levels, we find a weak price elasticity of reduction. We also evaluate the effect of an adaptive targeting scheme, which discriminates users based on their estimated responses in order to increase the per-dollar reduction ratio by 30%. Lastly, we find that households with smart home automation devices reduce significantly more than households without, namely 0.28 kWh (37%).

[1]  S. Borenstein.  The Long-Run Efficiency of Real-Time Electricity Pricing , 2005 .

[2]  Zhou Datong,et al.  A Bayesian perspective on Residential Demand Response using smart meter data , 2016 .

[3]  T. Funabashi,et al.  One-Hour-Ahead Load Forecasting Using Neural Networks , 2002 .

[4]  Maximilian Balandat.  New Tools for Econometric Analysis of High-Frequency Time Series Data - Application to Demand-Side Management in Electricity Markets , 2016 .

[5]  Susan Athey,et al.  Recursive partitioning for heterogeneous causal effects , 2015, Proceedings of the National Academy of Sciences.

[6]  Peter Palensky,et al.  Demand Side Management: Demand Response, Intelligent Energy Systems, and Smart Loads , 2011, IEEE Transactions on Industrial Informatics.

[7]  Tin Kam Ho,et al.  Demand forecasting in smart grids , 2014, Bell Labs Technical Journal.

[8]  Prashant J. Shenoy,et al.  Private memoirs of a smart meter , 2010, BuildSys '10.

[9]  Steven L. Scott,et al.  Inferring causal impact using Bayesian structural time-series models , 2015, 1506.00356.

[10]  D. Rubin,et al.  The central role of the propensity score in observational studies for causal effects , 1983 .

[11]  C. Glymour,et al.  STATISTICS AND CAUSAL INFERENCE , 1985 .

[12]  Chang-Tien Lu,et al.  Activity analysis based on low sample rate smart meters , 2011, KDD.

[13]  Wesley R. Hartmann,et al.  Welfare Effects of Home Automation Technology with Dynamic Pricing , 2015 .

[14]  D. Rubin.  Estimating causal effects of treatments in randomized and nonrandomized studies. , 1974 .

[15]  P. McSharry,et al.  Short-Term Load Forecasting Methods: An Evaluation Based on European Data , 2007, IEEE Transactions on Power Systems.

[16]  Robert Tibshirani,et al.  An Introduction to the Bootstrap , 1994 .

[17]  P. Diggle.  Analysis of Longitudinal Data , 1995 .

[18]  T. Vincenty.  Geodetic inverse solution between antipodal points , 1975 .

[19]  H. Allcott,et al.  Rethinking Real Time Electricity Pricing , 2011 .

[20]  Leo Breiman,et al.  Classification and Regression Trees , 1984 .

[21]  Q. Henry Wu,et al.  Electric Load Forecasting Based on Locally Weighted Support Vector Regression , 2010, IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews).

[22]  Trevor Hastie,et al.  The Elements of Statistical Learning , 2001 .