Panel data of M&A's and patenting activity

[Jelle Smits]

Hello,

I have a question that is a bit general, but hopefully somebody can help me since I’m quite stuck on this. It’s regarding my thesis.

My dependent variable is the amount of patents assigned by an acquirer after an M&A. I want to measure this 1, 2, 3 and 4 years after the M&A year. To illustrate, the amount of the acquirers patents in 2009 could be influenced by M&As in 2008, 2007, 2006 and 2005.

The independent variables are (1) the amount of patents that the target obtained up to 5 years prior to the M&A. And (2) the relatedness between these patents and the acquirer’s patents. These are both count data. So then I will measure the effect that acquiring targets with a lot of patents has on generating more patents of the acquirer.
I want to measure this over multiple years. So I will not only look at 2009, but also 2010 and 2011.

The part that I find complicated is that one firm can acquire several targets in one year. Could I make an identifier per M&A event for each acquirer? I saw some mentions of this on the forum but I don’t know if and how I could apply it here.

Any suggestions are welcome.
Thanks in advance,

Jelle

[David Benson]

Jelle,

Welcome to Statalist!

So as an interim step you are likely to have 3 datasets (that will get merged into the final dataset that you run regressions on)

1) Firm-year data on the acquirer (likely starting with Compustat) - This will have the number of patents each year the company applied for, as well as financial data (R&D spending, assets, etc)

2) A dataset with the acquisitions and info about the targets

3) A dataset with patents (that matches them to firms). But this will have the technology classes of the patents, etc You will use this to create the measures of overlap, relatedness, tech distance, etc for both the target firms and the acquiring firms.

Given your description, I suspect you will run your regressions on dataset #1, and that dataset will include data from datasets #2 & #3.

Start here:
A great paper to read to help you implement this is Ahuja & Katila (2001). Technological Acquisitions and the Innovation Performance of Acquiring Firms. Strategic Management Journal, Vol. 22, No. 3 (Mar., 2001), pp. 197-220. Stable URL: https://www.jstor.org/stable/3094458

In particular, see how they construct the following independent variables: Absolute size of acquired knowledge base, relative size of acquired knowledge base, and Relatedness of acquired knowledge base. They do a great job of explaining how they constructed their measures and why they constructed them that way.


Other papers to take a look at (for how they handled similar issues to yours):

• "Experience Effects and Collaborative Returns in R&D Alliances" LINK. Though this paper deals with the effect of alliances on subsequent patent performance, she has to deal with all the issues you do (multiple alliances per year, tech relatedness between the alliance partner and the focal firm, etc).
• What They Know vs. What They Do: How Acquirers Leverage Technology Acquisitions LINK
• Mergers and acquisitions: Their effect on the innovative performance of companies in high-tech industries LINK

Hope that helps!
--David

[Jelle Smits]

Hi David,

Thank you very much for your comprehensive answer!
I will look into the papers that you sent. Those of Ahuja and Katila and Cloodt et al are familiar to me. I have gathered nearly all information on my variables now.
Yesterday it was also really not clear to me how to put everything into one "master dataset" to run the regression. I also talked to someone who is more competent with Stata than me which made that more clear. Combined with the information that you give, I now have a better idea what steps to take from here.

Best,
Jelle