Predicting Poverty and Wealth from Mobile Phone Metadata Joshua Blumenstock University of Washington Accurate and timely estimates of population characteristics are a critical input to social and economic research and policy. Here, I describe how an individual’s past history of phone use can be used to infer his or her socioeconomic status, using a method for feature engineering based on a deterministic finite automaton. The fitted model generates predictions of the attributes of millions of individuals, which in turn can be used to accurately reconstruct the distribution of wealth of an entire nation, or to infer the welfare of micro-regions comprised of just a few households. In resource-constrained environments where censuses and household surveys are rare, this creates an option for gathering localized and timely information at a fraction of the cost of traditional methods. I will also discuss ongoing work that extends the method to a dynamic setting, which makes it possible to detect changes in an individual's welfare over time, and which enables new approaches to policy monitoring and impact evaluation.