This article develops a method to estimate and simulate the adoption of a network good. I estimate demand for mobile phones as a function of individuals’ social networks, coverage, and prices, using transaction data from nearly the entire network of Rwandan mobile phone subscribers at the time, over 4.5 years. I estimate the utility of adopting a phone based on its eventual usage: subscribers pay on the margin, so calls reveal the value of communicating with each contact. I use this structural model to simulate the effects of two policies. A requirement to serve rural areas lowered operator profits but increased net social welfare. Developing countries heavily tax mobile phones, but standard metrics that neglect network effects grossly understate the true welfare cost in a growing network, which is up to 3.12 times the revenue raised. Shifting from handset to usage taxes would have increased the surplus of poorer users by at least 26%.

An increasing number of decisions are guided by machine learning algorithms. But when consequential decisions are encoded in algorithms, individuals may strategically alter their behavior to achieve desired outcomes. This paper develops an empirical approach that adjusts decision algorithms to anticipate manipulation. By explicitly modeling incentives to manipulate, our approach produces decision rules that are stable under manipulation, even when the rules are fully transparent. We stress-test this approach through a large field experiment in Kenya. When implemented, linear strategy-robust decision rules outperform standard linear models such as LASSO.

When a policy prioritizes one person over another, is it because they benefit more, or because they are preferred? This paper develops a method to uncover the values consistent with observed allocation decisions. We estimate how much each person benefits from an intervention, and then reconcile the allocation with (i) the welfare weights assigned to different people; (ii) heterogeneous treatment effects of the intervention; and (iii) weights on different outcomes. We demonstrate this approach by analyzing Mexico’s PROGRESA anti-poverty programme. The analysis reveals that while the programme prioritized certain subgroups—such as indigenous households—the fact that those groups benefited more implies that the programme did not actually assign them a higher welfare weight. We also find evidence that the policy valued outcomes differently from households. The PROGRESA case illustrates how the method makes it possible to audit existing policies, and to design future policies that better align with values.