The Wall Street Journal had an interesting article yesterday entitled "Want to Kill Innovation at Your Company? Go Public", which discusses new research that suggests that after a company goes public, innovation goes down. The researchers compared companies that went through an IPO (Initial Public Offering) with those that had planned IPO's that never occurred. There were able to "quantify" innovation by looking at the patents both before and after the IPO. Patents that were heavily cited by later patents were deemed more innovative than those that were less cited. They also looked at the key inventors and saw that many of them left the companies in the post-IPO period.
This methodology obviously has some serious flaws. First, it is limited to small companies that haven't gone public. While there are the occasional giant corporations that have IPO's (Facebook and Google being recent examples of mega-IPO's), most IPO's are with smaller firms, and as such, they are more likely to be one-trick ponies - they have one good technology in the bag and that's it. Innovative companies with a whole portfolio of technologies very seldom go public, as they most likely have a long time ago. Think about the IBM's, Samsung's, Apple's and 3M's of the world (just to name a few).
In considering the small innovative tech companies, the reasons for the non-IPO's can be manifold. The timing may be wrong, other funding may have become available, or maybe the technology is found to have serious shortcomings and more research/development is needed, meaning more "high quality" patents may result.
With the small companies that do have their IPO's, it's easy to say that "Increased scrutiny and accountability to shareholders may also affect the kind of research and development a newly public firm chooses to pursue" and there is a certain element of truth to that. Having worked for a private company that underwent change of ownership, (although this is quite similar to when a company has a new CEO) there's a period of time where it takes new management and the employees a while to reach an understanding of how the company will be run going forward. Productivity is reduced, but only temporarily. How much of the observed results here is ephemeral in nature?
And lastly, academic researchers and journal publishers all know the general uselessness of counting citations of papers as a way to quantify the "value" of the research. Doing it with patents is just as crazy.