Generate matching group

[Kai Rieck]

Good Morning/Afternoon,

after conducting an event-study on two different groups of bonds (Green and non-Green Bonds) I want to match a non-Green bond with similar characteristics to my given green bond based on two observables I have in my data (time dimension in trading days as date variable (announcement date) and amount issued). The dates of announcement of the respective bond should not be apart more than one or two years ideally. For the amount issued the minimum of the amount issued as a greenbond and as non-green Bond shall be matched (min(amount issued_GB-amount issued_regular)).
To disentangle both bonds I used a greenbond dummy == 1 if the respective bond is a green bond.

After this matching has been done, I want to re-run my Event-Study and figure out if there is differences in returns for the created pairs and conduct additional Analysis.

Is this possible in Stata? If so, which function do you suggest?
If you need further information (i.e. data etc.) please let me know!

Thank you for your help already!

Best

Kai

[Mike Lacy]

Yes, what you want to do is almost certainly possible in Stata.

1) The built-in-teffects- command gives quite a few options for matching in various ways.
2) -search match- in Stata yields many hits, including user-written matching commands (-calipmatch- is one I like.)
3) There are a large number of postings of questions on this list about matching, including matching with dates. I just used a search engine on /site:statalist.org match "event study"/ and got 70+ hits.

So, there's not one way to do this, but many ways, and choosing would depend on the particulars of your problem, research traditions in your area, and so forth. Some of these might offer a better fit with "event studies" (about which I know more or less nothing) than others.