Though the U.S. government is touting claims that it has created an economic engine better than at any time in the 20th century, the reality is much different and a large part of the reason why Congress has lurched to the opposition party of U.S. President Barack Obama.

If people have been unemployed for so long they can no longer get unemployment benefits, the government frames it as that they went back to work. And we have been engaged in a culture war on business, everything from nuclear energy and natural gas to manufacturing of every kind.

Silicon Valley works because it came into existence before modern political 'you didn't build that' belief, and the notion that government is responsible for all success. If America is going to get back to economic prowess, we need to make it easier to create Silicon Valleys rather than picking winners and losers with subsidies, mandates and penalties for corporations that the government happens to like or dislike. No one is going to lure Silicon Valley or Wall Street to their town after the fact.

Such clusters for almost any industry helps a region's economy grow and a new study shows empirically that clusters of almost all kinds help drive overall economic growth in multiple ways, from job creation and development of intellectual property to the formation of new industries. 

The study, based on a unique long-term project collecting data on regional economies, has important implications for policymakers: It suggests that a region can improve its economic performance by improving its existing assets, rather than attempting a transformation by chasing industries situated elsewhere. 

"Very often, regions are given advice that they should become the next Silicon Valley, or be like some other region," says Scott Stern, the David Sarnoff Professor of Management at the MIT Sloan School of Management, and co-author of a new paper detailing the study. "They're told that they're in a global war for talent, or they should try to put out very expensive incentives to attract a single plant. What our research suggests is that regions succeed by investing in and extending their comparative advantage." 

The study examined 41 industrial clusters, 589 different subfields of industry, and 177 regions in the U.S. Each cluster is composed of multiple related industries that, as the paper states, offer 'complementary activities that give rise to increasing returns.' Quantifying the cluster effect in detail, Stern hopes, will help local leaders make sound economic decisions.

"Policymakers can use analytics to understand what their sources of relative advantage are," Stern says. "And while they, of course, want to avoid picking winners and we want to let a lot of experimentation flourish, we can prioritize those activities that leverage the things about our regions that are unique, distinctive, and meaningful. That leads to a smarter type of economic development than simply chasing the next big thing."

The paper, titled 'Clusters, convergence, and economic performance,' appears in the December issue of the journal Research Policy.  Co-authors are Mercedes Delgado, a professor at Temple University's Fox School of Business, and Michael Porter, a professor at Harvard Business School. The research uses data from the U.S. Cluster Mapping Project, a partnership between MIT, Harvard University, Temple, and the U.S. Economic Development Administration in the Department of Commerce.

An economic cluster, as defined by the study, has about 15 distinct types of industries within it, and multiple regions can host a given industry. For instance, while the Detroit and Cleveland areas feature well-known automotive clusters, so do a pair of economic regions in Kentucky: Lexington and Louisville.

The study spans the years 1990 to 2005. Among other results, it finds that a one-standard-deviation increase above the mean in industry specialization, within a region, leads to an expected 1.3 percentage point annual increase in employment growth in that area -- something Stern calls a 'first-order' result of the research.

"To me, jobs are a really important outcome, and we've tied that back to economic structure," Stern says.

However, jobs growth is also part of a virtuous cycle of innovation, as measured by patenting activity, that occurs in successful clusters. A cluster that is one standard deviation above the mean according to some of the study's criteria will also have a 1.2 percentage point annual increase in its growth of patents.

"We see more innovation in strong clusters, and strong innovation clusters are also associated with stronger employment,' Stern explains. "So there's a duality between the innovation and the jobs."

In other words, the numerous claims that 'government investment in research yields 1.3X' are misguided if the work has no applied benefit that leads to jobs, it is instead robbing the taxpayer Peter to hire Paul.