ON WEATHER NORMALIZING CUSTOMER-LEVEL BILLING DATA

This paper develops a conceptual framework for estimating both nonline ar and time-varying weather impacts from customer-level billing histor ies. The framework is flexible insofar as multiple model specification s can be derived from it for real-world application. Next, through sta tistical tests the paper evaluates the performance of 10 progressively complex weather specifications applied to a sample of billing data. T hese tests reject many of the simple and commonly used weather specifi cations, demonstrating the importance of accurate weather modeling.