Poster Paper:
Difference-in-Differences with Synthetic Controls: Federally Funded Housing's Impact on School District Performance
Friday, November 3, 2017
Regency Ballroom (Hyatt Regency Chicago)
*Names in bold indicate Presenter
Difference-in-Differences (DID) is a statistical technique that is often used to estimate the impact of a policy change. DID mimics a natural experiment in which individuals are exposed to a treatment, such as a policy change, and the occurrence of exposure depends on factors outside of the researcher's control. The control group is purposefully chosen by the researcher based on similarities to the treatment group affected by the policy. A critical assumption underlying the DID framework is that the treatment and control groups trend together over time, apart from the impact of the policy. In practice, this assumption is often violated limiting the validity of the results. The use of synthetic control methods for generating control groups may improve the validity of estimates in situations where the parallel trends assumption is violated. The synthetic control method uses information from the pre-treatment observations of both the control and treatment group along with post-treatment observations from the treatment group to generate a synthetic control group.
An example of when the synthetic controls may be beneficial is in analyzing the impacts of policy on school districts. School districts commonly differ in size, funding, population demographics and other relevant characteristics that create difficulties in matching treatment and control groups to satisfy the parallel trends assumption. In this study, we utilize the DID model and the generalized synthetic control method to estimate the impact of federally funded affordable housing projects on school district performance.
