5
$\begingroup$

Regarding Difference-in-Difference, from this paper of Dasgupta,2019, I am wondering if they use the data of the observations on the event year.

I asked this question because in their paper, they did not mention clearly whether they use the data on event year in their study. What they said from section 3.1 is

Figure 1 plots the mean of the mean change in asset growth of treated firms and control firms in the same industry, 2 years before and 5 years after the adoption of a leniency law in a country. Thus, the control firms are all firms in the same industry in countries that had not passed a leniency law in the 7 years surrounding the event date.

I am wondering why they say "7 years surrounding the event"? Whether "surrounding the event date" includes the event date?

For clarifying the question, for example:

If Korea passed the law in 1997, so the treatment including Korean firms in 1995,1996,1997,1998,1999,2000,2001 and the Korean firms in 1997,1998,1999,2000,2001 will receive the value of 1 (post x treatment)

or

the treatment including Korean firms in 1995,1996,1998,1999,2000,2001,2002 and the Korean firms in 1998,1999,2000,2001,2002 will receive the value of 1 (post x treatment)

or another option as one answer below

So, Treat x Post of Korean firms in 1995,1996 will receive the values of 0 but Korean firms in 1997,1998,1999,2000,2001,2002 will receive the value of 1. That is why I said 8 years in total. "7 years surrounding" here means 2 years before and 5 years after and the event year, totally 8 years.

$\endgroup$
3
$\begingroup$

First, clarifying the definition of "treatment" and "post":

  1. Treatment is the units that were treated, independent of time.

  2. Post is the year when treatment began.

  3. Treat x Post : is your effect and includes treated firms in treated years. Let's say t=0 is the year of the event and t>0 are posterior years.

So the treatment includes only treated firms in all years. Post includes all firms only in treated years when t>=0. So, in your case, it would be the first case

If Korea passed the law in 1997, so the treatment including Korean firms in 1995,1996,1997,1998,1999,2000,2001 and the Korean firms in 1997,1998,1999,2000,2001 will receive the value of 1 (post x treatment)

But I also learnt something from this paper. It seems that, he used 8 years for a treatment in total. Let's say, if Korea passed the law in 1997. So, Treat x Post of Korean firms in 1995,1996 will receive the values of 0 but Korean firms in 1997,1998,1999,2000,2001,2002 will receive the value of 1. That is why I said 8 years in total. "7 years surrounding" here means 2 years before and 5 years after and the event year, totally 8 years.

But it is still ambiguous to me if it is the case (8 years)

$\endgroup$
4
  • 1
    $\begingroup$ Thank you for your clarification, but it seems that we are still ambiguous about it is 7 or 8 years in total for a treatment country. $\endgroup$ – diff-in-diff May 27 at 23:36
  • 2
    $\begingroup$ The post-treatment variable isn't well-defined in this setting. You don't need to define it. The binary treatment variable (i.e., $LL_{kt}$) should just 'turn on' if a country is in the treatment group and is in a post-treatment time period. The interaction term is implicit in the coding of the treatment dummy. Assuming the law is passed in 1997, then the dummy for Korean firms should equal 1 in 1997 and stay on for the duration of the panel. $\endgroup$ – Thomas Bilach May 28 at 0:14
  • 2
    $\begingroup$ By "7 years surrounding the event" they mean the effects in the 2 years before and the 5 years after the event date. The periods are "relative to" the first adoption year. $\endgroup$ – Thomas Bilach May 28 at 0:30
  • 1
    $\begingroup$ Many thanks for your clarification, Thomas. I want to clarify one more thing: if I understand you correctly, if Korea passed the laws in 1997, so 1995 and 1996 is pre-period and the binary treatment variable being turned on in 6 years: 1997,1998,1999,2000,2001,2002 $\endgroup$ – NoviceMindset May 28 at 1:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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