Most empirical policy work requires the aggregation of policies. Trade policy aggregation exemplifies the aggregation problem poignantly, with thousands of highly dispersed trade barriers. This paper provides methods of policy aggregation that are consistent with two common objectives of empirical work. One is to preserve real income. The other is to preserve the real volume of activity in the parts of the economy being aggregated. Both objectives must be achieved for consistent multi-country policy modeling. An application to India shows that the standard atheoretic method of aggregation overstates India's real income by around 3 times the global gains from free trade.JEL Classification: C43, D58, F13, F17 *I am grateful to Will Martin for suggesting the topic of this paper to me, and for helpful comments, to G. Chris Rodrigo for assistance with preparing the data and for helpful comments. I am also grateful to the World Bank for research support. I thank the referees for their helpful comments.Aggregation of policy variables is unavoidable in empirical work. Data limitations and computational burdens dictate that extremely detailed policies must be combined into summary indexes. The focus of this paper is on trade policy aggregation, a particularly poignant form of the problem, 1 but the discussion applies to other policy aggregation