Large Business Vs. Small: The Ongoing Debate
Last week, Jared Bernstein wrote a good post on the differences in job gains between small and large firms. This is a perpetual debate that gets unfortunately (and inevitably) wrapped up in political persuasion, paeans to the virtues of small business, hostility toward big business, and arguments over "good" and "bad" jobs. All of this is further compounded by statistical confusion.
In the main, Bernstein is correct on three points. First, the continuing arguments about small and large firms is silly. In a series of research papers that Bernstein doesn't cite, John Haltiwanger and economists at the Census Bureau have found that firm size is basically irrelevant when it comes to net job creation. Instead, it is firm age that is the true determinant of net new jobs. Large companies (those with over 10,000 employees) account for a larger share of employment today than twenty years ago, but this is a stock versus flow distinction. On an annual basis, new and young companies account for a large share of net job creation.
It's true, of course, that new and young firms tend to be small, but Bernstein would probably agree that there is a big difference between a firm with 10 employees that is 20 years old and a firm with 10 employees that is two years old. There is also an important symbiosis between large and small companies that frequently gets lost in these sorts of arguments. For instance, big companies acquire smaller and younger companies. Looking at Bernstein's charts, it seems likely that the big job gains by large firms include such acquisitions, in which case we're conflating organic job creation by younger (and smaller) companies with acquisitive job creation by older and larger companies. For a certain subset of companies, for example, Jay Ritter and his colleagues have found substantial increases in acquisitions of newly public companies - or by newly public companies - to the exclusion of organic growth. They call this the dawn of an "eat or be eaten world."
This kind of symbiosis and churn lies underneath the aggregate data used by Bernstein and others (including ourselves at the Kauffman Foundation), but this also underscores the second point on which Bernstein is correct. Toward the end of his column, Bernstein notes the importance of "small startups that survive and grow." No matter how you slice the data-by size, age, sector, geography, astrological sign-you will never see a monolithic pattern from which you can conclude, "all jobs come from X." Large companies creates lots of jobs in terms of gross numbers, but they also shed lots of jobs; new and young firms create lots of jobs, but they take lots of jobs with them when they close or fail.
In fact, one persistent source of job creation that cuts across thresholds of size and age is high-growth companies. Haltiwanger and others have identified what they call an "up or out" dynamic among young firms, whereby surviving young firms tend to grow rapidly. Kauffman research, using Census data, has identified small slivers of firms as highly fertile sources of net new jobs, and forthcoming research from the Bureau of Labor Statistics will confirm this.
Importantly, high-growth firms can be found across all sizes and ages of companies, corroborating the relevance of something like the Inc. 500/5,000 list of fast-growing companies. The median age of these companies is six, and the median size is around 50 employees. In other words, these are young, smallish companies that grow rapidly and create lots of jobs. Forthcoming research, moreover, will show that a high fraction of them get acquired; which shows up, in the aggregate, as job gains by larger companies.
Finally, Bernstein is right to highlight the distinction between firms and establishments, as we have continuously done in our research. The only thing that needs to be added here is that we shouldn't overlook the entrepreneurial nature of franchising. Franchises-a new McDonald's, for example-would be included in the statistics as an establishment, rather than a firm, but that shouldn't diminish the importance of the jobs thereby created as well as the real risks borne by the franchisee. In many cases, these are independent operators bearing a good deal of cost and risk, similar to an entrepreneur who starts a brand new company.
Bernstein is absolutely correct that the obsession with firm size obscures a large amount of underlying variation. (We haven't even touched here upon the importance of also looking at revenues and innovation in addition to jobs. Judging the contribution of entrepreneurs solely on the basis of direct net job creation is incomplete.) It is probably a terrible idea to design and target policies based on firm size. Firm growth - small, big, old, young, high, low - is where we find the source of economic growth.
