In electricity markets, the price paid by retail customers during periods of peak demand is far below the cost of supply. This leads to overconsumption during peak periods, requiring the construction of excess generation capacity compared to first-best prices that adjust at short time intervals to reflect changing marginal cost. In this paper, I investigate a second-best policy designed to address this distortion, and compare its effectiveness to the first-best. The policy allows the electricity provider to raise retail price by a set amount (usually 3 to 5 times) during the afternoon hours of a limited number of summer days (usually 9 to 15). Using a quasi-experimental research design and high-frequency electricity consumption data, I test the extent to which small commercial and industrial establishments respond to this temporary increase in retail electricity prices. I find that establishments reduce their peak usage by 13.4% during peak hours. Using a model of capacity investment decisions, these reductions yield $154 million in welfare benefits, driven largely by reduced expenditures on power plant construction. I find the current policy provides 43% of the first-best benefits but that, with improvements in targeting just the days with the highest demand, a modified peak pricing program could achieve 80% welfare gains relative to the first-best pricing policy.