1
$\begingroup$

I apologize if this isn't exactly the right forum to pose this question. I am trying to use an event study to evaluate the adoption of a specific policy in several different states that happen at different times. I had a professor suggest using an event study but I am struggling to find resources on the use of event studies in policy research. I understand the mechanics of how to formulate the equation and structure the data and come up with a graph. However, I feel like I'm lacking knowledge that would allow me to fully interpret results. I feel like I don't know all the requirements of an event study or what it means to have a significant divergence from 0 following the treatment period.

Does anyone know of any resource that explains event studies in the policy context? I'm aware it's used more often in finance literature but those works don't follow the work I'm doing. This paper is an example of policy event studies but I'm looking for a more in-depth breakdown of when and how to use this technique.

A good example of a confusing result for me would be the following:Pre-transformation

The treatment, in this case, is the implementation of a law that requires those under domestic violence restraining orders turn in their firearms and the outcome is the number of intimate partner homicides per 100,000 people. There are an additional set of policy and sociodemographic controls. I spoke to my advisor and he said to shift the graph down and center the year before treatment at 0 (but to do so transparently). I get the following when I do so: Post-transformation.

The problem is that I don't know how to fully interpret this or why it would be okay to make this transformation. If anyone could shed any light on why someone might do this, that would be very helpful. However, more broadly, I'd appreciate any resources that would help me help myself a bit more with this topic.

$\endgroup$
1
  • 2
    $\begingroup$ "Mostly Harmless Econometrics". $\endgroup$
    – Michael
    Aug 13, 2020 at 18:00

1 Answer 1

0
$\begingroup$

It sounds like you are really looking at differences in differences, not event studies.

Diff in diff compares outcomes before/after a policy intervention (as does an event study) (the first diff) and between treated and untreated units (the second diff).

The underlying assumption is that, in the absence of the policy, treated and untreated units (here: states) would have experienced the same changes over time. A good reference is Ch 5 of Mastering Metrics (by Angrist and Pischke), but the previously suggested "Harmless Econometrics" by the same authors is also very good.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.