Electricity tariff reforms will be an essential part of the clean energy transition. Existing tariffs rely on average cost pricing and often set a price per unit that exceeds marginal cost. The higher price encourages over-adoption of residential solar panels and under-adoption of electric alternatives to fossil fuels. However, an efficient tariff based on fixed charges and marginal cost pricing may harm low-income households. Households with low demand for electricity may prefer to disconnect from the grid rather than pay an equal share of fixed costs. We propose an alternative methodology for setting fixed charges based on the predicted willingness-to-pay of each household. Using household data from Colombia, we show the fiscal burden and economic inefficiency of the existing tariffs. We then show how our new tariff methodology could improve economic efficiency, create incentives for the adoption of clean energy technologies, while still leaving low-income households better off.