The U.S. Startup Economy Is in Both Better and Worse Shape than We Thought17 March, 2016 / Articles
Startup activity can signal a city’s economic potential, but it’s actually the quality of the startups, not the quantity, that matters.
That’s just one of several important findings from a paper released this week by Jorge Guzman and Scott Stern, both of MIT. The paper surveys the landscape of American entrepreneurship, offering an optimistic picture of it and of the U.S. economy’s future prospects.
The study’s core point is simple. We’ve long known that new businesses matter to the economy and that it’s a small group of fast-growing firms that matter most, because of the jobs and innovation they bring. What Guzman and Stern add is a method for identifying the firms that are trying to grow. A new restaurant or dry cleaner probably won’t end up hiring thousands of employees or commercializing new technology. On the other hand, what the authors call “innovation-driven enterprises” — think Facebook or Google — do intend to grow and can have significant economic impact. Guzman and Stern’s quality measure seeks to separate the former from the latter.
With this measure of entrepreneurship “quality” in hand, the researchers can map the geography of startup potential, linking it to cities’ future growth.
“A doubling of entrepreneurial quality predicts an increase of 6.8% in GDP 11 years in the future,” the authors report. By contrast, startup quantity is less correlated and not significantly linked once the city’s current GDP is controlled for.
The authors say they are not implying that entrepreneurship, even the quality kind, necessarily causes growth. It could also be that “the reason people start firms is that they see opportunities,” said Guzman, which would mean measures of entrepreneurship could simply reflect a city’s economic strength rather than causing it.
To determine which new firms are likely to grow, Guzman and Stern developed an algorithm that predicts the chances of a startup going public or being acquired for a significant sum. Firms that register in Delaware are more likely to grow, for instance, as are firms that file as a corporation rather than a partnership or an LLC. Firms that apply for patents are also more likely to grow. Firms that are named after their founders are less so.
Amazingly, these and other factors can predict a startup’s prospects pretty well. Three-quarters of the startups that end up going public or being acquired score in the top 10% by this algorithm, based only on factors available at the time of their founding. The algorithm isn’t good enough to distinguish Facebook from MySpace, so it won’t help venture capitalists do their work. But by applying this algorithm to business registration records, the researchers can get a reasonable measure of the entrepreneurial quality of a city, state, or nation.
In addition to predicting growth, the researchers’ measure of startup quality challenges two common economic arguments. The first is that American entrepreneurship is in decline. That’s true as measured by the number of new businesses, but Guzman and Stern’s measure shows that America’s entrepreneurial potential has increased since the Great Recession and that in 2014 it was almost as high as its 2000 peak. In some places, including the Bay Area, it’s at record highs. (Guzman and Stern’s data is restricted to 15 states but includes the major U.S. startup hubs.)
The second bit of pushback concerns the argument, put forward by economists such as Larry Summers and Robert Gordon, that America is facing a period of slow economic growth. “Our index does say there’s been a steady growth of [quality] entrepreneurship,” Guzman said, and he sees this as reason for optimism. Whether or not startups directly cause economic growth, they seem to predict it, so the high number of growth-oriented startups in recent years may be cause for confidence in the future of the American economy.
The challenge, as FiveThirtyEight’s Ben Casselman explains in his coverage of the research, is that quality startups don’t seem to be as likely to grow as in the past. The number of startups that the algorithm scores as quality is high, but the likelihood of a successful IPO or acquisition hasn’t kept up. This could be a function of funding availability and the IPO market, or it could reflect something darker about the U.S. economy, perhaps that incumbents are more protected from competition than they used to be. In a companion policy paper, Guzman, Stern, Catherine Fazio, and Fiona Murray urge policy makers to take this drop in growth events seriously by focusing on the problems startups face when trying to scale.
The good news is that the potential is there. America is producing fewer new businesses than it once did but plenty of the ones that matter most. The trick is to make sure that they grow.