Wednesday, July 4, 2012

Has David Slain Goliath?

The future of the working American lies not with the conglomerate Goliath but with the start-up David.

Using the firm-size breakdowns of ADP employment report data on FRED, I found that:

(1) The contraction of employment in the United States was driven primarily by large firms, and so has the recovery of employment been by small and medium-size firms.

(2) The American economy is undergoing a long-term structural shift of the distribution of employment away from large firms to small firms.

First, a few definitions: I define firm size here as number of employees. Firms with 500 employees or more are "large," firms with 50 employees or less are "small," and firms with between 50 and 500 employees are "medium" sized.During the recession of 2008, employment in large firms declined far more than did employment in small and medium-size firms. At the trough of the recession, the number of employees at large firms had declined 8.6 percent relative to December 2007. That is a substantially larger drop than the 7.9 percent of all employees shed by medium-size firms, or the 5.6 percent shed by small firms.When nonfarm payroll employment began to recover in January 2010, hiring at small and medium-size firms drove the employment recovery. Employment at small firms has grown 3.5 percent, and at medium firms by 4.1 percent, since the start of 2010 -- whereas at large firms, employment has increased by only 1.1 percent over the same period of time.

Yet the data do not suggest that the recession functioned as a single shock which asymmetrically affected firms by their size, causing large firms to contract employment or fold more frequently than small and medium-size firms.

No. What's going on is bigger than the most recent recession. It appears to be a long-term structural shift in which larger firms are losing ground to small firms. Medium-size firms are holding even. The most compelling explanation I can think of for this phenomenon is that firm size is becoming less determinative of a firm's ability to achieve economies of scale, and to complement this, that macroeconomic and microeconomic change implies increasing returns to "nimbleness," innovation, low fixed costs, and other traits which characterize small firms.Consider this. In December 2000, 41.7 percent of Americans on nonfarm private payrolls worked for firms who employed less than 50 people; in May 2012, 44.9 percent of Americans work for such firms. That increase of 3.2 percentage points mirrors the decrease seen at large firms. In December 2000, 18.7 percent of Americans on nonfarm private payrolls worked for firms who employed over 500 people; in May 2012, 16.3 percent of Americans now work for such firms -- a decrease of 2.4 percent.

Another way to think about the secular shift from large firms to small ones is by using Riemann sums to approximate the average size of a nonfarm private firm in the United States. In December 2000, the average firm employed 213 people; in May 2012, the average firm employed 199 people. That amounts to a decrease in average firm size of 6.6 percent.Notably, the recession of 2008 is barely visible in this graph, which confirms the point that this isn't about cyclical movements, but rather a structural change to market structure and the distribution of employment.

7 comments:

  1. Perhaps one could make the argument that Keynes and Schumpeter were correct on the entrepreneur, judging from this data. Wouldn't you agree, Mr. Soltas?

    P.S. Even though I'm not an American, Happy Fourth of July!

    ReplyDelete
    Replies
    1. You'll have to tell me what Keynes and Schumpeter said about entrepreneurs. I know the both chiefly as a macroeconomists, although I understand they have contributions in other fields, I just don't know them well enough for informed comment.

      And thank you.

      Delete
  2. Really terrific post. You hypothesize: "that firm size is becoming less determinative of a firm's ability to achieve economies of scale, and to complement this, that macroeconomic and microeconomic change implies increasing returns to "nimbleness," innovation, low fixed costs, and other traits which characterize small firms." I think that's exactly right.

    It makes me think of Vizio. It's 10 years old, it has 417 employees (over half of those are tech support in a call center in South Dakota, with the other half being design, engineering, and sales), making it a medium-sized company in your analysis. It's the number one HDTV vendor in America, and it's expanding into lots of new markets. As far as I can tell, their main expertise is in design and supply chain management. The manufacturing is completely outsourced. As you point out, they don't need to develop economies of scale manufacturing consumer electronics, because Chinese companies are already really good at that.

    On a smaller scale, a friend of mine runs his own business selling accessories for consumer electronics (iPod earbud holders, mounting devices for Roku, stuff like that). He designs the products, has them built to spec by chinese manufacturers, and has a sales guy that helps him get into good channels. His wife handles the accounting. He's very, very successful.

    That's the future. There's never been a better time to be an entrepreneur.

    ReplyDelete
    Replies
    1. I hear anecdotes like yours every time I talk about small firms -- I didn't know about Vizio's story, but I do know them as a company -- and although anecdotes can get you in trouble if you don't look at data, I think that sometimes you need a context for the data, and an economy of Vizio-type firms will be profoundly different in a qualitative sense.

      It makes me deeply hopeful about the long run, due to the implied productivity gains. I think the short run is manageable with good macro policy, i.e. NGDP targeting. But there's this question mark for me in the medium run which is to some extent troubling. I don't know where the millions of jobs are going to come from.

      Delete
  3. great data I would love to see this data going farther back, is this a trend of the last decade or does it go farther back

    Ideally it would be great to see it going back to the industrial revolution

    ReplyDelete
    Replies
    1. I wish. Unfortunately, what I graphed is ADP's full public data set. Aargh.

      Delete
  4. Interesting post. There have been prior studies finding that large employers are more cyclically sensitive than small employers. See here: http://www.nber.org/papers/w14740.pdf


    Might be interested in the following analysis regarding gross versus net job creation: http://www.bls.gov/opub/mlr/2004/07/art1full.pdf

    Also, look into Steven Davis' work regarding this issue.

    ReplyDelete