Second-Order Effects of Research Funding Sources: Is the U.S. Innovation Pipeline Headed for a Hiccup?

9/10/07Follow @rodneyabrooks

Back in the good old days we academic-lab or research directors used to worry that the sources of research funding for our faculty and their research groups might bias their research in ways that were not scientifically pure. Nowadays we have something new to worry about. Today we need to worry about whether the research going on in our laboratories will find its way into vibrant startups and thence to American companies which themselves were startups only a few years ago.

First, a little history. Back in 1980 the Bayh-Dole act was passed which lets U.S. universities keep the IP rights to technology generated via federally funded research and license those inventions as they see fit. The government retained rights for its own purposes but really that had little effect in practice. A big result of Bayh-Dole’s passage was that Universities started to license out the IP developed on their campuses. But more important was the fact that the people they often licensed it to were the people who knew the IP best; the faculty and students who had developed new ideas within universities began spinning out companies to develop those ideas commercially.

This led to the explosion of spinoff companies that has occurred over the last two decades. [Editor's note: for a round-up of the latest crop of such companies, click here.] In many cases, the founders of those spinoffs negotiated exclusive licenses with the universities—after all the technology often seemed to be obscure and still early in its practicality, and existing companies were unlikely to want to license it directly. This exclusivity gave the VCs something to hang on to as real value, and so the small companies would be able to get going right next door to the university where the original work had been done.

Now a little bit about accounting rules in the U.S. When a company pays for basic research—either in-house or via a contract with a university—it comes directly off the firm’s bottom line, negatively affecting its earnings per share (EPS). The market values EPS above almost all else, and so therefore does the board, especially when it sets the CEO’s compensation package. In contrast, an acquisition of another entity is off the balance sheet—it has no impact on EPS. CEOs of publicly traded U.S. companies would much prefer to make an acquisition than to pay for research. That tradeoff has a tangible immediate effect on their own pockets.

So a lot of large U.S. companies (some who were small startups just a little while ago) get new technology more by buying early stage companies that have developed their technology to the level where it is just hitting the market place, rather than by paying university researchers (or even internal research groups) to do the basic research.

So now we have the confluence of Bayh-Dole and U.S. accounting rules. The system that has developed in the last 25 years in the U.S. is that federal government pays for basic research. The VCs take the risk in trolling that research for the pieces that are relevant to U.S. industry, then help academics groom those, develop them, and get them read for market. Some of the small companies formed along the way decide to go for broke and do IPOs, but the majority of successful ones get bought by larger U.S. companies. Those companies (perhaps products of IPOs only a few years before) pay a premium for new technology, but they avoid most of the early stage risk. The VCs get handsomely rewarded for their successful bets, but are successful because they are prepared to take on a risk profile where they will have many failures as well.

Everything works fine for U.S. industry as long as the federal government pays for university research.

But over the last few years that has changed. For instance at MIT CSAIL, where I was lab director until a couple of months ago, the portion of research funded by DARPA dropped over the last 12 years from more than 90% to less than 25%. The National Science Foundation has stepped up its research funding but has not come anywhere close in closing the gap. U.S. companies give small gifts (and in fairness some companies are real champs on this front, but most of the big high tech household names are pretty miserly) but overall there has been a short fall. In to this fray have come foreign companies. Many of them are willing to make sizable research investments in U.S. universities. They are what is keeping the IT research endeavor afloat in the U.S.

But they want IP results in return for their investment. That means no more IP exclusivity for companies spun off by faculty, and tough decisions for VCs when a big international player already has rights to the IP on which the Series A they are looking at is basing its business.

Are we in danger of the innovation pipeline from universities to U.S. businesses having a real hiccup?

Rod Brooks is Panasonic Professor of Robotics at MIT and Chief Technical Officer of iRobot. For a full bio and an archive of Rod’s previous posts and comments, click here.

By posting a comment, you agree to our terms and conditions.