Each year, 1.6 million children die from diarrheal diseases; unsafe drinking water is a major cause. This paper reviews evidence from randomized trials on domestic water access and quality in developing countries, interpreting the results through a public economics framework. It argues that subsidies for water treatment are likely warranted, while more evidence is needed to assess the case for subsidizing programs to improve access to water. Multiple randomized trials show that water treatment can cost-effectively reduce reported diarrhea. However, many consumers have low willingness to pay for cleaner water, with less than 10% of households purchasing household water treatment under existing retail models. Provision of information on water quality can increase demand, but only modestly. Free point of collection water treatment systems designed to make water treatment convenient, salient, and public, combined with a local promoter, can generate take up of more than 60 percent. The projected cost is as low as \$20 per year of life saved, comparable to vaccines. In contrast, the limited existing evidence suggests many consumers are willing to pay for better access to water, but it does not yet demonstrate that this improves health. Randomized impact evaluations have also generated a number of methodological insights, suggesting that: at least in some contexts, merely surveying households can lead them to change their behavior; separately randomizing offer and transaction prices does not yield evidence of sunk cost effects; revealed preference measures of clean water valuation are much lower than contingent valuation estimates; and randomized evaluations can be used to estimate parameters for structural policy simulations.