Under what conditions does cost-of-service regulation lead firms to distort costs? This paper analyzes changes in fuel procurement practices by coal- and natural gas-fired electricity generating plants in the United States following state-level legislation that ended cost-of-service regulation among investor-owned electric utilities in the late 1990s. I construct a detailed dataset that links confidential, shipment-level data on the price of virtually all of the fuel delivered to coal- and gas-fired electricity plants in the United States from 1990-2009, with plant-level data on operations and regulatory status. Using a matched difference-in-difference estimation strategy to account for confounding shipping costs, I found the price of coal drops by 12% at deregulated plants relative to matched plants that were not subject to any regulatory change, whereas there was no relative drop in the price of gas. Deregulated plants disproportionately switch to burning low-sulfur coal rather than install capital-intensive abatement equipment to comply with environmental regulations, and expand imports from out of state by 25% if they were initially burning in-state coal. I show how these results lend support to theories of asymmetric information between generators and regulators, regulatory capture, and capital-bias as important sources of distortion under cost-of-service regulation. I then show that the drop in the price of coal is associated with a reallocation of purchases to more productive mines, rather than simply a transfer of regulatory rents from coal producers to electricity generators. Although only one quarter of U.S. coal-fired capacity has been deregulated, the end of cost-of-service regulation has reduced the price of fuel by about one billion dollars per year for these plants.