Query Regarding Unbalanced Panel Data (How to Merge SEO's data with Quarter Financial Observations)

[Ammar Hussain]

Respected Researchers. Hopefully, all of you are doing well.

I am working on the role of Seasoned Equity Offerings (SEO's) with some financial outcomes. I selects the time span from 2001 to 2022 (with approximately 2500 firms) and during this time span nearly 5000 SEO's events are occurred (at random dates during this span). While I use quarter financial outcome variables (like Assets, Liabilities and among others) which are having 100000 firm-quarter observations. Many Papers used this type of data in their studies, but I am much confused how to merge the 5000 SEO's offer price variable with firms financial outcome variables 100000 firm-quarter observations because I need to calculate the SEO timing variable which is equals to latest SEO's offer price / previous quarter closing stock price.

Kindly, 1) needed your guidance on this how to merge 5000 SEO's with 100000 observations in STATA; 2) How to calculate SEO timing variable for all quarters using 5000 SEO's and 100000 firms-quarter stock price variable.

Waiting for the kind response and valuable suggestions from you dear researchers.

Thanks. Stay Blessed.

[George Ford]

why not assign a quarter to the SEO data and then joinby quarter?

[Ammar Hussain]

I assign the quarter according to the date to SEO's, but still confused how to run the panel because SEO' are 5000 and firm quarter observations are nearly 100000. The previous studies which used this approach they also have 3000 to 5000 SEO's with large number of firm quarter observations and have shown thousand's number of observations in descriptive statistics. So, confused if a company offers an once SEO during selected time span then whether I use same SEO' offer price in all quarters of their respective company??

Note:
As per Dittmar, Duchin, and Zhang (2020) "SEO Timing (Price Ratio, defined as the ratio between the average stock price of the previous quarter and the firm’s most recent equity offer price.)"

[George Ford]

I don't know these models, but was responding to your question about merging data. It sounds like they compute l.stockprice/seoprice. The rest are 0?

[Ammar Hussain]

Please see the attached paper especially page 2 (paragraph 3) starts from "we begin our empirical analysis.........equity offer price". Then see section 2. "Data sources and sample construction". Actually, I am confused about this section how they merge 3021 SEO's offer price data with nearly 100000 COMPUSTAT Financial quarter observations.
Waiting for your kind response.
Thanks.

[George Ford]

Clear as mud.

I'd contact the authors.

[Ammar Hussain]

I were also contact the authors many times through email, but doesn't received any response.

Waiting for your kind response.
Thanks.

[Ammar Hussain]

george Ford......Have you received any information from Authors'.

[Ammar Hussain]

george Ford..... Have you received any information from Authors'.

[George Ford]

I'm not that interested to contact the authors, sorry.

Amy Dittmar is now at Rice, so not the same email as in the paper. ([email protected]).

I think what they are doing is filling all quarters between SEOs with the price from the prior SEO. The means suggest that the case (near 1, which can't be true if there's a bunch of 0s), and it makes sense (to the extent I understand what they're after). As the price gets closer to the SEO price, they issue a new one.

So, you'll need to take the SEO dta and expand to all quarters and make sure the price is the last SEO price until a new SEO is issued. Merge. Then computed the Price Ratio as L_price/SEO_P.

[Ammar Hussain]

George Ford Thank you so much for your kind help and giving your prestigious time. Just last thing, please tell me is there any command in STATA to repeat the same SEO price until the new SEO is issued. Thanks

[George Ford]

stfill might work.