When Startups Don't Create Jobs
Myth: The best way to create companies that provide jobs and wealth is to encourage business creation of any kind. Reality: On Dec. 3, President Barack Obama opened his jobs summit. I wasn't invited, but thanks to the power of the Internet, I can contribute anyway. The goal of the summit is to create jobs. Policymakers in the U.S. and around the world know that high- growth startups create a lot of jobs. Therefore, one of the goals of President Obama's team and policymakers everywhere is to figure out how to create more high-growth companies. Unfortunately the basic approach of many policymakers to stimulating such companies is flawed. They think that simply creating more companies regardless of business type creates more high-growth businesses. Some compelling data indicate that it doesn't. Places with the most high-growth companies tend not to be the places with the most new businesses. Consider the example of Silicon Valley, the part of the U.S. thought of as a locus of many high-growth companies. When measured on the rate of per-capita new business formation, the poles of Silicon Valley—the metropolitan areas of San Jose and San Francisco—do not score high. Negative CorrelationExamined more formally, the data indicate that the correlation between the share of high-growth companies and the rate of company formation is actually negative. The places with higher company birth rates have a lower share of high-growth companies. In a study for the U.S. Small Business Administration, academics Zoltan Acs, William Parsons, and Spencer Tracy used data from the American Corporate Statistical Library to measure the Dun & Bradstreet (DNB) company birth rate for all of the U.S. metropolitan statistical areas. They also measured the number of "high-impact" firms, companies that at least had doubled their revenues over a four-year period and had what they called an "employment growth quantifier" of more than two. When I compared the different MSAs on the share of companies that met their criteria for high-impact firms with the birth rate of firms, I got a significant negative correlation. In addition, Acs and his colleagues found that most high growth companies are not young companies. They explain in their study: "On average, they are 25 years old…[and] less than 3% of high-impact firms were born in the previous four-year period." If high-growth companies tend to be more mature ventures, then efforts to create more new outfits is unlikely to generate a lot of high-growth companies. Cross-national comparisons also don't show a strong positive relationship between the firm birth rate and the share of high-growth startups. The Organization for Economic Cooperation & Development (OECD) measures firm birth rates, share of high-growth companies, and share of high-growth young companies in its member nations. It calls businesses five years old or younger, with at least 10 employees, and with annualized employment growth in excess of 20% over the three previous years "gazelles." U.S. Gazelle PopulationThe countries that have the highest firm birth rates tend not to have the highest share of gazelle companies and vice-versa. For instance, among the 13 OECD countries where the share of gazelle companies in manufacturing was measured, the U.S. ranked last. It had a small proportion of high-growth young companies in manufacturing. But it ranked a respectable 12th (out of 22 countries) in terms of manufacturing firm birth rate. In contrast, Bulgaria ranked No. 1 (out of 14) in terms of the share of gazelles in services, but it was only 16th out of 22 in firm birth rate in services. Similarly, Canada was ranked third from the lowest in gazelles in services, but sixth in terms of birth rate in services. The ranking of countries on the share of high-growth new firms is also not similar to the ranking of the share of high-growth firms overall, indicating that some nations may have a big share of high-growth companies, but a small share of high-growth new businesses. For instance, Italy ranks No. 2 in terms of the share of high-growth companies in manufacturing, but ranks 10th (out of 13) in terms of the share of high-growth new manufacturing firms. In short, two empirical patterns measured both across countries and within the U.S. indicate that the assumption that more company formation will result in a higher rate of creation of high-growth outfits is flawed. There is little relationship between places with high rates of firm formation and high shares of high-growth companies. Moreover, a large portion of high-growth companies are older ventures.