Optimal Policy Learning: A Tutorial, at MDRC in New York

“Optimal Policy Learning: A Tutorial” will take place on December 17, from 3:00 PM to 5:00 PM (CET) in hybrid format. The two-hour seminar will be hosted by MDRC and will be led by Giovanni Cerulli, researcher at CNR-IRCrES and scientific coordinator of FOSSR project.

The event is hosted by Richard Hendra, a leading expert in causal evaluation of social policies in the United States and Director of the Center for Data Insights at MDRC.

The MDRC (Manpower Demonstration Research Corporation) in New York is one of the world’s most prestigious independent research institutes in the field of social policy. For over fifty years, it has conducted rigorous evaluations of public programs, often based on randomized experiments and advanced causal methods, with the goal of improving policies on employment, education, welfare, and poverty reduction in the United States.

The seminar will present OPL (Optimal Policy Learning), developed within the framework of FOSSR, the NRRP project that is building the first Italian Open Cloud dedicated to the social sciences. OPL is available in Stata, with complementary implementations in R and Python, developed by contributors from ISTAT and Italian universities.

Optimal Policy Learning provides methodological tools for the optimal assignment of public policies and interventions using observational data. These methods combine advances in causal econometrics and machine learning to estimate heterogeneous treatment effects and to construct decision rules aimed at maximizing welfare and policy impact, even under real-world constraints such as budget limitations or targeting requirements.

Participation is free and open to anyone interested in policy evaluation, causal econometrics, applied machine learning, and policy design.