IDSS Special Seminar: Alberto Abadie - 11/23/15

Title: Synthetic Controls for Policy Evaluation
Abstract: Much of the currently available statistical machinery to evaluate the effects of policy interventions is designed with the goal of estimating average effects over a population from which we have obtained sufficiently large samples of units exposed and units not exposed to the intervention of interest. In practice, however, many of the interventions policy makers care about are implemented at an aggregate level and affect a small number of large entities, such as cities, school districts, or states. In this talk, I will provide an overview of synthetic control methods, a recently developed statistical tool to evaluate the effects of large aggregate interventions. The description of the nature and properties of synthetic controls will be coupled with empirical applications.

[@yanghu3974]

Thank you for sharing, it is very helpful.

[@cindyrojas8278]

Great lecture!

[@dp1490]

i really like this lecture!

[@lexparsimoniae2107]

Great lecture. Thank you.

[@maddoxzain920]

You all prolly dont give a shit but does any of you know a way to log back into an Instagram account? I somehow forgot my password. I would appreciate any tips you can give me

[@crosbylennon6418]

@Maddox Zain instablaster =)

[@maddoxzain920]

@Crosby Lennon Thanks for your reply. I found the site thru google and I'm trying it out atm. Seems to take quite some time so I will reply here later when my account password hopefully is recovered.

[@maddoxzain920]

@Crosby Lennon it did the trick and I finally got access to my account again. Im so happy! Thanks so much you really help me out :D

[@crosbylennon6418]

@Maddox Zain no problem :)

[@akshayabalaji4097]

I have a question. Is it possible to have predictor variables of different frequencies (annual, monthly, etc) within the same synthetic control model?

[@jaredgreathouse3672]

Lots of this depends on your question, but generally, yes. So for example, I'm currently working on a paper for my statistical methods course at Georgia State University. I'm modeling the spread of the Coronavirus, given the implementation of lockdown policies in the United States. 8 total units, 1 treated, 7 controls. My population variable is annual. But my outcome, change in COVID cases, is daily. So take Georgia, where I live. The population is 10 million-ish. We wouldn't expect the population to change that much over time, so the variable is kind of a constant. Whereas the growth rate of COVID, for example, changes almost daily. The issue isn't so much the level of aggregation, the important thing is about the quality of your data.

[@akshayabalaji4097]

@@jaredgreathouse3672 Understood, thank you Jared!

[@jaredgreathouse3672]

@@akshayabalaji4097 you're welcome. What are you studying?

[@akshayabalaji4097]

@@jaredgreathouse3672 Undergraduation in economics

[@jaredgreathouse3672]

@@akshayabalaji4097 Okay. Do you want to use the synthetic control design in your own research?