Do I expect too much from journalists? When I see an article titled “Why America Stopped Being a Startup Nation” and I read it in the expectation that it will actually have a plausible explanation for why so many fewer companies are being started than used to be the case, am I wrong to be disappointed. That was the case with this post from Noah Smith at Bloomberg. Fortunately, it still had something of value—dashing cold water on the idea that it can be explained by over-regulation:
Goldschlag and Tabarrok find that regulation has increased quite a lot in most industries, especially in manufacturing…
But correlation doesn’t equal causation. If the increase in regulation is causing the decrease in dynamism, we would expect that industries where the former has been greater to have experienced more of the latter.
When Goldschlag and Tabarrok run the numbers, however, they find no measurable relationship between rising regulation and declining dynamism at the industry level. Sectors where rules have become increasingly stringent have seen declines in startup rates that are no greater than sectors in which the level of regulation has been little changed.
One of the co-authors of the paper in question is Alex Tabarrok, notably libertarian in his views, so the findings are sufficiently a testimony against interest that they’re worthy of extra credence. If he was grinding an ax, it would be the other way.
That doesn’t mean that regulation is not one of the factors in the reduction in the rate of new business formation but it does strongly suggest that regulation is not dispositive. Off-hand I’d say that there were multiple factors involved including regulation, the number and size of the established companies, the highly-evolved character of product marketing today, and generational differences in attitude towards risk and reward.
As long as new company formation is slow and the large, established companies are more predisposed to reducing the number of employees they have on their payrolls than increasing them, it’s going to be tough to put all of the people who should be working and would be working if there were jobs for them back to work.
Why are they only measuring regulatory impact for the Bush II administration, plus a year or two? How about this as a thesis: regulatory impacts did not appreciably increase during the Bush II administration?
The main issue though is that the stringency index is based upon the number of appearances of the words “shall,†“must,†“may not,†“prohibited,†and “required.†(1) This is nonsense because regulations often place restrictions on the government. (2) A single “shall” can impose one requirement or twenty-seven or more through compound setences. (Could I write the entire tax code with only one mandatory word? I think I could) (3) Clarifying regulations may appear to introduce a mandatory word, though in reality they are trying to reduce uncertainty and are not imposing any new requirement at all. (4) Some mandatory words impose more economic costs than others.
The number of lawyers employed in the District of Columbia metro might be the basis for a better index.
Suppose you want to start a company these days. You’ve an idea for a piece of software which does something potentially useful. It takes a few programmers, a few marketeers, some publicists, and so on — a manageable number of people, And some money, but there are venture capitalists who will take a risk with a couple million dollars. It’s doable. And if the idea doesn’t pan out … maybe next year you’ll have a better idea.
Now suppose you want to beat Elon Musk to Mars. You’re going to build really neat rockets and be one of the heroic 21st century entrepreneurs libertarians love to think about. You need maybe ten thousand people. You need facilities for them to work in, facilities to assemble your rockets, relationships with long term suppliers and customers — i.e., business types and lawyers and accountants. You need lots and lots of money, and you’re going to need for a long time to come. None of this telling your developers to sleep under their desks for three months while you’re approaching your release date, and paying them all with stock options. And you’re likely to fail, for half a dozen reasons, most of which aren’t under your control. It’s all a really bid deal, far beyond the scope of Kickstarter.
So what do you want to do? What’s your wife and family want you to do? What’s the friendly angel investor you’re talking to want you to do? And so on.
What I’m thinking is that modern American society and business and banking and so on is set up to encourage small scale software companies and to discourage large scale manufacturing. The payoff can be very big for software, very quick, and a handful of people there at the start can reap big profits. With manufacturing … not so much. So we’ve got a lot of three or four person video game companies, and not a whole lot of factories filled with engineers and people on assembly lines, not a lot of hole-in-the-wall machine shops making parts to sell to those factories.