So Which Is It: Do Startups Create Jobs or Not?
In 1998, Google was still a dream in the eyes of Larry and Sergey. Today, it employs 31,000 people and made about $35 billion in the past year. In 2004, Facebook was still a figment of Mark Zuckerberg’s imagination. Today? It’s expected to generate $4 billion this year and has 3,000 employees. In those two stats, you can see the question and answer to the debate: Do startups create jobs, or more accurately, do they create jobs the way they used to?
The issue is bothering Congress, which is focused on the economy and the nation’s continued high unemployment rate. In a hearing Wednesday, Brink Lindsey, a senior scholar in research and policy at the Ewing Marion Kauffman Foundation, told the House Subcommittee on Technology and Innovation that the number of startups, long a source of new jobs, has been declining since 2006, and when startups are created, they tend to generate fewer jobs. For those in Silicon Valley, the hub of American innovation, such a conclusion might be shocking, given the startups proliferating like weeds and the shortage of talent.
The Kauffman Foundation may be onto something that even Silicon Valley, which accounts for only a small portion of overall American job creation, should pay attention to. Startups, tracked from their inception through two-, five-, and 10-year periods, do seem to be producing a smaller number of jobs. So which vision of startups is true: the Silicon Valley impression of a startup bubble or the worrying stats that the Kauffman Foundation has collected for a 2011 report?
FEWER STARTUPS, FEWER EMPLOYEES
The statistics cited by Lindsey are disturbing. He said the number of new employer businesses created annually began falling after 2006, dropping 27 percent by 2009. Additionally, the average number of employees per new company has been trending gradually downward since 1998, while the pace of job growth at new businesses during their first five years has been slowing since 1994. So we have fewer startups hiring fewer people and tending to grow more slowly during their first five years.
Those stats are from Kauffman-supported research by John Haltiwanger, a professor of economics at the University of Maryland College Park. Citing the same study, Lindsey said average annual gross job creation by startups has fallen 25 percent since the 1980s, resulting in an overall reduction in both gross and net job creation for the U.S. economy. Lindsey suggests that the problem is structural and offers some suggestions that could help fix the process by making it easier for startups to build their businesses.
He reiterated the call for the Startup Act laid out in detail here, which has suggestions that range from relaxing Sarbanes-Oxley regulations to decreasing the backlog for patents and making it easier for foreign entrepreneurs to get visas. His point is that making it easier to start a new business will help create new businesses.
But what if it’s too easy to start a new business today? For example, in the technology and media sector, the Internet, cloud computing, and open-source software have broken down the distribution barriers that once had to be overcome, while also lowering the cost of starting and conducting business. Sure, if you’re going to be manufacturing physical goods or dealing with physical inventory, the story is very different, but technology is changing the very nature of startups and business creation.
Clearly, Lindsey’s data are measured over decades as opposed to just the past few years, but he appears to be associating productivity gains created by startups with job gains. From his testimony:
“Generally speaking, exiting firms are less productive than existing firms, which in turn are less productive than surviving new firms. According to a recent paper written by economist John Haltiwanger and supported by the Kauffman Foundation, net entry of firms has contributed about 30 percent of total productivity growth in the manufacturing sector and virtually all productivity growth in the retail sector. New firms are thus the lifeblood of rising productivity, and, consequently, rising living standards. And, when it comes to promoting prosperity through job creation, the role of new enterprises can hardly be overstated.”
ALTOGETHER NEW INDUSTRIES
But what if new enterprises are boosting productivity without boosting jobs? My hunch is that job gains may be stalled because technology has enabled existing employees to be more productive and has eliminated some jobs altogether. So what’s needed is the next step, in which startups and new businesses move beyond existing businesses and figure out the way to build something truly different that would take advantage of the Web and lower capital costs to create new industries instead of just remaking the old.
That’s the promise of advances in data analytics and a Web that talks back. It’s also the promise behind adding sensors to more and more physical elements to create a connected web of people and devices. At that point, we’re not just remaking an existing business, but starting a fresh cycle of innovation with the tools the tech industry has perfected over the years. So maybe this is a cyclical change—one that’s measured in 50-year increments as opposed to a single decade.
Also from GigaOM:
Flash Analysis: The Tech Startup Investment Environment, Q3 2011 (subscription required)