This paper explains the conceptual basis for the social rate of time preference (STP) and why it is the appropriate method of choosing the social discount rate (SDR), compared to the most prominent alternative method: the social opportunity cost of capital (SOC). We recommend that for intragenerational projects in the United States, a rate of 3.5 percent is appropriate. For long-term intergenerational effects, we recommend using declining rates. Obtaining the SDR is intrinsically a normative exercise in a second-best world. Policymakers should maximize a social welfare function that equals the present value of current and future utility from per capita consumption. In the presence of economic growth, there will be greater future consumption possibilities. Given the assumption of diminishing marginal utility of income, the consumption of a wealthier, future society should be discounted. Displaced private investment should be accounted for by first multiplying by the shadow price of capital, but this will not generally be necessary as most government interventions mainly affect consumption. Systematic risk should be handled by conversion of expected net benefits into certainty equivalents before discounting at the risk-free SDR, but empirically this effect is typically too small to matter. Among governments there is increased adoption of both the STP method and the use of time-declining rates. Even governments using other approaches are lowering their rates, and most OECD countries now apply rates in the 3 to 5.5 percent range.