September, 2010Startups and young companies dominate net job creation in the United States – and have done so for the last 30 years. This economic phenomenon occurs because of the patterns of firm formation and survival, according to a report released today by the Ewing Marion Kauffman Foundation.
While the extraordinary contribution of jobs from new and young firms often is treated as something unique to these firms, this study, Neutralism and Entrepreneurship: The Structural Dynamics of Startups, Young Firms and Job Creation, reveals some measure of structural order behind that reality. However, the study points out that forces creating this structure are subject to change. Recent shifts – including the declining cost of company creation in information technology and other sectors, and lower investment thresholds for seed-centric acceleration programs – indicate that we could very well be on the cusp of one such change: an increase in new company creation in certain sectors of the economy.
"This study reveals an important structural context in which firm formation and job creation occur that helps explain why new and young companies dominate net job creation," Robert E. Litan, vice president of Research and Policy at the Kauffman Foundation. "We need to understand the structural features of entrepreneurial capitalism—the why of firm formation and job creation—so we can take steps that support and encourage those features and not unknowingly undermine them."
Because startup entry and survival rates have proven to be relatively constant over time, the number of firms populating the American economy grows annually, with companies five years of age or younger comprising the largest demographic sector each year. In part, this bloc's size alone results in the most net new jobs contributed to the economy, according to the report.
The study bases its findings on an analysis of a Business Dynamics Statistics (BDS) dataset broken out by firm age to determine how total employment in startups changes as those companies age. The BDS, a U.S. government dataset compiled by the U.S. Census Bureau. tracks the annual number of new businesses (startups and new locations) from 1977 to 2005, giving data on firms and their establishments according to firm age for each of the first five years after the birth year and in five-year blocks thereafter.
"While startups and young companies' status as the largest demographic category accounts, at least in part, for the fact that they add more net new jobs to the economy each year than older companies do, a study of various datasets from the past several decades seems to show that U.S. firm formation has been remarkably consistent for the last 100 years," said Dane Stangler, research manager at the Kauffman Foundation and author of the study. "What this indicates is that the level of firm formation is largely neutral – that is, rather than being a process of constant turmoil, as often is assumed, it follows a natural structural order."
Research into the BDS dataset shows that, once the economy reaches a point at which it includes firms older than age six, new and young companies account for between 30 percent and 40 percent of all firms in the economy. Over time, as these companies age, they decrease in number. But, for about the past 20 years, net job creation from those that survive has been greater than that from businesses that open and close. Thus, the large amount of net job creation from continuing companies also appears to reflect the structural dynamics of firm accumulation.
Few companies survive past age 40. Although four-fifths of the companies on the Fortune 500 list were founded before 1970 – meaning that they have survived past age 40 – they represent only a tiny percentage of the 6 million-plus U.S. firms. Such long survival requires merging with and acquiring other companies – success factors for which new and young companies provide sustenance. In addition, firm entry likely stimulates older firms' demise, helping to account for the shrinking share of these companies.
The research also found: