Illustrations: Mark Matcho

Venture capital funds have swelled hugely in the past decade or so—and that's good, isn't it? Venture capital lights fires under scrappy and ambitious start-ups. It can help bring great new ideas to market, some of which go on to disrupt entrenched industries, spawn entirely new ones, perhaps even permanently change the world.

Old established companies rarely do that. They're much better at making incremental innovations, because they generally have more to lose than to gain from disruptive technologies. Yahoo and Google came out of left field, not the R&D labs of Microsoft or IBM. The personal computer as we know it came out of Apple Computer, not Hewlett-Packard, itself the original Silicon Valley start-up. Cryptography was brought to market by new companies like RSA Security and VeriSign, not by AT&T.

In theory, then, venture-capital-backed start-ups are the best engines of innovation. But are they in fact? With venture capital funding an order of magnitude greater today than it was in the early 1990s, now is an excellent time to ask: has all that funding over the past decade brought more innovation or less?

As venture capitalists ourselves, we've had considerable experience watching our colleagues make investment decisions. We had our own theories about how best to turn money into innovation but reserved judgment on the industry as a whole until we could accumulate and analyze the data from what has been the most frenzied decade in technology history.

Our methodology was simple. We examined 1303 electronic high-tech initial public offerings for a 10-year period ending in 2002. We limited ourselves to IPOs from the New York Stock Exchange and Nasdaq, which were ground zero for the telecom and dot-com explosion of the 1990s. We sorted out those that were VC-funded and compared them with those that were not. We rated them on a scale of 1 to 5, with 1 being the most technically innovative. [See sidebar, ""Scoring Innovation"."]

We were shocked by what we found. Overall, the level of innovation during that decade was surprisingly low. Even more dismaying, it did not correlate well with VC funding: the level of innovation actually dropped sharply after 1996, even as venture funding was going through the roof.

To Focus On Truly New Technology, we first excluded spinoffs, such as Lucent Technologies Inc., which broke off from AT&T Corp. in 1996, and recapitalizations, such as Accenture, which, under the name Andersen Consulting, was already a separate business unit when it severed ties with Andersen Worldwide in 2000.

We also eliminated eBay Inc. and other companies that relied on e-commerce business models rather than new technologies. While innovations in e-commerce have created businesses worth billions of dollars, and improved the lives of millions of consumers and retailers alike, finding a new way for them to interact with one another is quite different from coming up with a fundamentally new technology.

There is, of course, a reasonable point to be made about new organizational structures that would be possible only with the Internet. When a garage sale is limited to people who drive past your front yard, you're lucky to find a single person interested in that old weather vane, whereas when the buying universe isn't limited by geography, a bidding war between collectors is not only possible but likely. But the difference between e-commerce businesses and earlier auction or phone- and mail-order companies is largely one of degree, not kind, and doesn't compare to, say, using lasers for microwave communications or DVD players.

In fact, e-commerce business models are almost always grounded in only the slightest bits of incremental innovation. For example, eBay wasn't the first Internet auction site—that honor probably goes to OnSale Inc., founded in 1994. And the most novel thing about Amazon.com, the other e-commerce elephant, is its recommendation software ("customers who bought this book also bought..."), which was invented by the Software Agents Group of the MIT Media Laboratory, also in the early 1990s.

After subtracting spinoffs (21 companies) and e-commerce (459 companies), we had 823 firms in our study. Even with all those companies subtracted, we found few truly innovative companies and a sharp decrease in their number over time. We used 1996 as a peak dividing two temporal watersheds. That year saw the seminal U.S. Telecommunications Act, the Lucent spinoff, and the start of the tech bubble.

In studying the four years from 1993 to 1996, we found 20 highly innovative companies, ones that fell into our two highest levels [see table below "10 Years, 20 Companies"].

That's about five per year, only 4.4 percent of the four-year total. For the next six years, though, that already low percentage plunged to 1.4 percent—only five such highly innovative companies in the entire period. That's not even one per year.

There was also a big drop in the middle range of innovation—companies that took some less-than-fundamental innovation and brought it to market in a clever way. Some of these companies, scoring a 3 on our scale, have been highly successful. The number of these midrange companies decreased substantially, from 29 per year during 1993 to 1996 to only 7 during 1997 to 2002.

By far the largest number of IPOs over the course of our 10-year period received a low rating of 4, which was essentially the bottom. (Only a few companies remained in the very lowest tier of 5 once we excluded Internet e-commerce plays.) For the first four of the 10 years studied, two-thirds of the 823 companies fell into this No. 4 tier. That proportion rose to 87 percent in the next six years—a further decrease in an already disappointing level of innovation.

The reasons for this failure are complicated and deeply entrenched in the VC way of doing business. But a common thread runs through many of them, and it has to do with risk. Based on our experience, we believe that VCs really aren't the risk takers they're often made out to be.