To forecast accurately the welfare gains and effects on demand from introducing seasonal electricity rates, we use a variable response model with data from the Los Angeles Electricity Rate Study. The model exploits the time series and cross sectional variation in the data and permits household-specific responses to hot and cold weather and in the "permanent" level of consumption. The full information maximum likelihood estimates permit more flexible and accurate forecasts as well as appropriate test statistics. We find that price elasticity is small but significant and that it increases with hotter weather. These are important differences among subpopulations, but the overall welfare gain appears very small in absolute terms and small when compared with the gains from introducing time-of-day pricing estimated from the TOD component of the Los Angeles Study.