2
$\begingroup$

Dasgupta,2019, p.2598 documented that they

consider all nonfinancial firms in Worldscope, Compustat Global, and Compustat North America data sets from 1990 to 2012

However, they used the [-2;+5] event window (page 2600) around the law implementation day to examine the impact of laws on dependent variables. In their country list on Table 2 page 2599, the last countries that passed the law are Ukraine, Taiwan in 2012. Therefore, based on their identification, their data sample should be from 1991 to 2017 rather than 1990-2012.

Apart from that, the first country applied this law is US in 1993, therefore, the data source should start from 1991 (two years before the implementation day).

$\endgroup$
2
$\begingroup$

Dasgupta in the study says he is using also other event windows in dynamic estimations:

To explore the dynamics of the issuance activities and leverage change, we create dummy variables corresponding to the following windows around the treatment year: from 1 to 4 years before the treatment; the treatment year and the 2 years after treatment; the next 3 years; and the years beyond.

with window 1 to 4 years before treatment it makes sense.

Dasgupta also uses diff-in-diff and for that you need to test for pre-treatment trends so that also can be a reason.

Dasgupta also says that cartel data are from some 2013 publication so probably there was no point in getting more data from compustat. If they miss any data for their main variable of interest after 2012 then they can't run regression on longer than 2012.

$\endgroup$
6
  • $\begingroup$ Thank you for your answer, csilvia. However, "Dasgupta also uses diff-in-diff and for that you need to test for pre-treatment trends so that also can be a reason.", I did not see that Dasgupta tested pre-treatment trend, can I ask why he did not do that in his work? $\endgroup$ – BeautifulMindset May 26 at 4:04
  • $\begingroup$ And I am curious, I agree with you about the dynamic testing, but he said "from 1 to 4 years before treatment", so it means that regarding the US case, it should start from 1989 rather than 1990. I hope I do not misunderstand your points $\endgroup$ – BeautifulMindset May 26 at 4:06
  • 1
    $\begingroup$ @BeautifulMindset but they mention that they were testing for per-treatment trends in chapter 4, they just don't show those tests, not all tests are always shown some are just mentioned. 1-4 years means between 1 and 4 years so for the country where those laws started for 1993 they might have just done 3 years. Also maybe they don't have data before 1990 $\endgroup$ – csilvia May 26 at 13:05
  • $\begingroup$ Do you mean the sentence "We find that firms first start growing and issuing.....taken into account" at the end of 4.2? $\endgroup$ – BeautifulMindset May 26 at 19:53
  • 1
    $\begingroup$ @BeautifulMindset I mean the paragraph starting with Using this definition of treatment year.. they discuss it briefly later in that paragraph $\endgroup$ – csilvia May 27 at 9:30

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.