Articles

In Silicon Valley, Job Hopping Contributes to Innovation

The New York Times, "Economic Scene" , December 01, 2005

For four decades, through booms, busts and bubbles, Silicon Valley has maintained an amazingly innovative business environment.

Companies and technologies rise and fall. Hot start-ups morph into giant corporations. Cutting-edge products become mature commodities. Business models change. Through it all, the area remains creative and resilient -- and more successful than other technology centers, notably the Route 128 area around Boston.

What makes Silicon Valley special? Thanks to some new data, economists have finally been able to test statistically some popular explanations.

In her influential 1994 book Regional Advantage: Culture and Competition in Silicon Valley and Route 128 (Harvard University Press), AnnaLee Saxenian, an economic development scholar at the University of California, Berkeley, argued that Silicon Valley's innovative edge comes from two unusual characteristics. First, talented employees move easily and often to new employers, far more so than people elsewhere. ''The joke is that you can change jobs and not change parking lots,'' one of her interview subjects said.

Second, instead of vertically integrating, Silicon Valley computer makers rely on networks of suppliers. They also design open systems that can flexibly accommodate all sorts of new components.

''The system's decentralization encourages the pursuit of multiple technical opportunities through spontaneous regroupings of skill, technology and capital,'' she wrote.

Many people, especially in Silicon Valley, found Professor Saxenian's argument convincing. But while her research was careful, it depended on interviews and had no large-scale statistical backing. Perhaps her subjects' impressions were unreliable.

After all, the argument that Silicon Valley's job hopping fosters innovation contradicts economists' common assumptions. ''It didn't feel right to me,'' James B. Rebitzer, an economist at Case Western Reserve University, said in an interview.

When employees jump from company to company, they take their knowledge with them. ''The innovation from one firm will tend to bleed over into other firms,'' Professor Rebitzer explained. For a given company, ''it's hard to capture the returns on your innovation,'' he went on. ''From an economics perspective, that should hamper innovation.''

He found a possible answer to the puzzle in the work of two management scholars, Carliss Y. Baldwin and Kim B. Clark. In their book Design Rules (MIT Press, 2000), they argued that when there is a lot of technological uncertainty, the fastest way to find the best solution is to permit lots of independent experiments. That requires modular designs rather than tightly integrated systems.

''By having a lot of modular experimenters, you can take the best, which will be a lot better than the average,'' Professor Rebitzer said. Employee mobility may encourage productive innovation, as people quickly move to whichever company comes up with the best new technology.

But you would not expect to find people moving around all the time in every industry, only those where technical uncertainty justifies spending lots of resources on experiments -- including many that will not pan out. ''In most other settings,'' said Professor Rebitzer, ''it's going to be easier simply to design things with special purpose parts that fit in.''

In a forthcoming article in The Review of Economics and Statistics, he and two economists at the Federal Reserve Board, Bruce C. Fallick and Charles A. Fleischman, empirically test the claim that Silicon Valley employees move more often than computer industry employees in other places. (The article, ''Job Hopping in Silicon Valley,'' is available here.)

The two Fed economists, who are interested in the macroeconomic patterns created when people move from employer to employer, use data from the Current Population Survey. Until recently, economists almost entirely ignored such voluntary job hopping because they did not have good data to track it.

Since 1994, however, the Current Population Survey has asked respondents whether they have changed jobs in the last month. This question was added simply to cut down on repetition. Respondents who are still at the same workplace do not fill out questions about their employer. But the change created the first large database that tracks job changes. Since the survey is geographically based, it is ideal for examining job changes within the same area.

To Professor Rebitzer's surprise (though not his co-authors'), it turns out that Silicon Valley employees really do move around more often than other people. The researchers looked at job changes by male college graduates from 1994 to 2001. During that period, an average of 2.41 percent of respondents changed jobs in any given month.

But, they write, ''living in Silicon Valley increases the rate of employer-to-employer job change by 0.8 percentage point.''

''This effect is both statistically and behaviorally significant -- suggesting employer-to-employer mobility rates are 40 percent higher than the sample average.''

Computer industry employees in other California technology clusters also seem to switch jobs more often than those in other states. This result supports an argument made by Ronald J. Gilson, a law professor at Stanford and Columbia. In a 1999 article, he suggested that a 19th-century California law helped create Silicon Valley's hypermobility by prohibiting the enforcement of noncompete agreements.

In other states, businesses use these agreements to keep employees from easily hopping to other companies in the same industry. (That article is available at papers.ssrn.com/sol3/papers.cfm?abstract--id=124508.)

Finally, the economists test whether computer industry employees are more likely to move than employees in other industries, as the modularity hypothesis would predict. Again, statistical tests suggest that the theory is right.

Looking at cities within California, they write: ''We find no evidence that outside the computer industry, job changes are more likely within Silicon Valley. Indeed, rates of job hopping appear to be lower in Los Angeles and San Diego than elsewhere in the nation.''