Many households lack meters to measure their electricity or water consumption. I show that provision of electricity meters leads to a large reduction in electricity consumption over the first four months following meter installation. This is consistent with previous overconsumption by unmetered users facing a zero marginal price. However, it also reflects underconsumption by metered customers paying a marginal price exceeding marginal cost. I quantify these welfare effects using billing data from a large sample of mostly rural counties in Colombia. I show that the efficiency effects are relatively small compared to the distributional effect of metering. Very poor households, whose electricity consumption is low, would particularly benefit from meter provision.