Wednesday, December 10, 2014

Tabarrok on business dynamism

Last week's Cato growth forum included a session on declining business dynamism, which included presentations by John Haltiwanger, Amar Bhide, and Alex Tabarrok. Alex focused on the question of whether we should care about this. See his MR post here; some key points he makes:

1. More regulated industries are not less dynamic (he shows this by plotting a measure of industry-level dynamism against a measure of regulatory burden).

2. Entrepreneurship is not necessarily good. Poor countries are full of entrepreneurs of necessity. They have no better options.

3. Old firms are very entrepreneurial. The good ones are constantly reinventing themselves. Think of Ford, Zara, or Apple.

4. Increasingly, what matters is global dynamism. Looking at national statistics is not enough.

Point 1 has made everyone's job very difficult. It is very hard to find a smoking gun in terms of policy. That said, we should not overinterpret Alex's result. In most of the work on this topic, people look within industries because regulatory burden is not the only way industries differ. My prior is that, in many industries, other differences swamp regulatory differences. I would be more persuaded by time series evidence--for example, have the industries that have seen large changes in regulatory burden also seen large changes in dynamism (of the appropriate sign)? Davis and Haltiwanger describe evidence on the effect of certain labor market regulations, like occupational licensing and various worker protections, but as they say more work is needed.

Point 2 is a good one, and in my work with John Haltiwanger, Javier Miranda, and Ron Jarmin, we mention this frequently. This was also Noah Smith's first reaction--declining mom-n-pops doesn't hurt productivity. But as we show in more recent work, it's not just low-productivity entrepreneurs that are going away. We are seeing fewer high-growth firms.

I like Point 3. I know there is work being done on this avenue, including by us. What if creative destruction is being brought inside the boundaries of older firms? In terms of Alex's exposition, what we would want is some reason to think that the reinvention he mentions is going on more now than it was 30 years ago. A related hypothesis is the notion that startups get acquired soon, so their high growth happens inside other firms. I should also note that almost all of the good data on firm growth are based on employment, because that is easy to measure (at least for privately held firms). So that matters too--what if firms are more dynamic now but it shows up somewhere other than employment? At least among public firms, though, we do see these trends (speaking of gross job flows) in sales data as well as employment.

Point 4 is good too. I know people are researching this point as well. There are a few ways to think about it, so I think it's a question with an answer.


  1. Hi Ryan,

    What are the best measures of business dynamism? What are your thoughts on business starts as a measure of dynamism (I'm sceptical)?

    (I'm not familiar with this literature at all so sorry if this is a naive question.)


    1. Hi Tim,

      In this literature people typically use things like gross job flows (ie, jobs created at expanding establishments, jobs destroyed by shrinking establishments), various measures of young firm activity (so yes, business starts, but also the share of employment accounted for by young firms, the share of gross job creation accounted for by entrants, etc), worker flows, etc.

      What makes you skeptical of business starts? I guess it depends on how you want to define "dynamism". Certainly there are limits--as Alex notes, under some conditions a lot of entrepreneurship means that labor markets aren't functioning well, etc.

    2. That last point about lots of entrepreneurship meaning that labour markets aren't working well (through lack of trust, poor institutions etc.) is what I am on about.

      Is there any work discussing the 'optimal' rate of firm entry? Surely there's such a thing as too much churn just as there can be not enough?

      I found this article very interesting

    3. Yeah. In our JEP paper we say this explicitly. We don't know the optimal amount of entrepreneurship, and the literature on the subject--particularly in development--gives a fair amount of attention to entrepreneurs of necessity. More broadly, churn has costs--"Surely there's such a thing as too much churn." I think figuring out the optimal amount of that stuff is on the research frontier.