Spatially Correlated Unbalanced Longitudinal Energy Billing Data Analysis

• Activity Number: 475
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Business and Economic Statistics Section
• Author(s): Min Niu*+ and Zhengjun Zhang
• Companies: University of Wisconsin at Madison and University of Wisconsin
• Keywords: spatial panel data ; energy bills

Abstract

Residential energy bills are featured with unequally spaced longitudinal data which are collected by utilities. The collected data also demonstrate spatial correlation. In energy efficiency program evaluations, billing data based analysis is one of the major methods, yet it has received little discussion in the recent past. Motivated by this billing analysis, this study proposes a varying-coefficient multilevel mixed-effects model with spatial correlation introduced through the community level effects, and an auto-correlation structure within each household. In the proposed model, the efficiency of regression coefficient estimators depends on how accurately spatial and time-wise covariance parameters are estimated. The widely used restricted maximum likelihood method is not applicable to the proposed model as optimization is a problem. A two step estimation method is employed for estimating covariance matrices. The proposed estimators are consistent, which provide a feasible way to solve the optimization problem with satisfied precision. Sampling properties and finite-sample performance are presented. Real energy billing data is analyzed.