In PWP-017, Train and Mehrez specify an econometric model of tariff choice and time-of-use (TOU) consumption under optional TOU tariffs. The specification is consistent with the theory of self-selecting tariffs, in that each customer's choice among tariffs is based on its demand parameters, which vary in the population. The model is estimated on data from a specially designed experiment under which the optional TOU rates varied over customers and the TOU consumption of customers who remained on standard rates was recorded. With the estimated model, Train and Mehrez compare the TOU consumption and price elasticities of customers who chose TOU rates with those who chose standard rates. They also estimate the impact of the experimental TOU rates on consumer surplus and the utility's profits. The analysis suggests that, with one possible exception, the offering of the optional TOU rates did not Pareto dominate standard rates. However, total surplus, excluding measurement costs, is estimated to have risen by $1.41 to $2.25 per month per customer who chose the TOU rates. For measurement costs below these amounts, the optional TOU rates are therefore found to increase total welfare, with customers who chose the TOU rates being hurt. For higher measurement costs, the optional TOU rates cannot be justified on surplus grounds.