top of page



Machine Learning for Prediction of Tax Evasion (with Marco Battaglini and Douglas Miller)


We develop a machine-learning prediction model aiming at improving the targeting of auditing resources. Using millions of tax filings, the model will produce recommendations on filing selection into an audit.  In collaboration with the Italian Tax Authority, the model will be tested and eventually updated. The Tax Authority will collect a small sample of audits randomly, and another small sample of audits that are predicted to be “high yield” audits. These audits can be compared against the predictions of the model. The predictive model, alongside novel data collection on manager-office moves over time, will also be used to construct a novel measure of manager productivity in the government service sector.


NSF project 2049207 - years 2021-2023.  Total amount awarded to Cornell: $240,000.

Cross-SES Friendship and Socio-Economic Attainment in the US (with Michael W. Macy and Benjamin Rosche)


The alarming increase in social inequality corrodes the national ethos of a “land of opportunity” where rich and poor can co-mingle at work, at play, and at school. This study investigates the long-term consequences of friendships that bridge socioeconomic boundaries for socioeconomic attainment of disadvantaged youth. We will use data collected over the past 30 years that tracks friendships between school children. Do the attributes that help children overcome disadvantaged backgrounds also promote friendships across socio-economic boundaries? Do these friendships have a lasting impact on life trajectories that would not have otherwise been likely? We answer these questions by integrating network dynamics with models of social stratification. Our analysis will inform social policy by identifying opportunities to improve the educational and occupational mobility of disadvantaged youth through the formation of social ties that bridge the deepening gulf of economic inequality.

GRANT REVIEW ACTIVITY:  National Science Foundation (NSF),  European Research Council (ERC),  Luxembourg National Research Fund (FNR),  Israel Science Foundation (ISF), Austrian Science Fund (FWF), Research Council of Canada (SSHRC),  Ministry of Education, Universities, and Research- Italy (VQR)

bottom of page