Oddly enough, this statement, which is the basis of his argument, is also why his conclusion is wrong.
If by productivity he means revenue per employee, then he may be correct even of technology startups. Any successful startup's revenue per employee is likely to be higher in year 5 than in year 1, when they probably had no revenue. But what matters more is the derivative: successful startups have the fastest increase in revenue per employee. In other words, the reason startups are good is they're the companies in which "firm productivity increases with firm age" the most.
Except that he's talking about "startups" not "successful startups." As you said, productivity is essentially revenue per employee, and there's lots of startups out there where that is a small number (and not growing fast).
Productivity is probably something closer to 'value per employee', i.e. the increase in acquisition price due to an employee's actions. I do think that, even for unsuccessful startups, employee #3 is able to do more to proportionally increase their market value than employee #30.
Sure, but failing startups don't grow as big as successful ones. The median startup employee (which is what you care about when deciding whether startups are good for a national economy) is doing way better than the median startup.
But he doesn't define what a startup is, and his data is for all new businesses, not just tech startups, which have a completely different set of data than local barber shops.
The WSJ says that Obama hasn't defined it either. It also isn't clear what the cited study in the blog post says, though the title ("Productivity Differences across Employers: The Roles of Employer Size, Age, and Human Capital") of the paper does suggest it is for all startups.
Link in case someone is an AEA member (to get through the paywall):
Encouraging startups is a good direction to look for improving society; it plays to our strengths and encourages value creation. Unlike the current administration, I think Obama's administration would take an open mind about how best to aid startups.
I'd be most interested in a low-doc corporation. For example, don't make me pay estimated taxes, payroll, etc, except once a year. Give me a simplified tax setup that's brainless. Don't make me file a pointless annual report to please the gov't.
I'd be interested in a complete overhaul of the tax system. Unfortunately, real tax reform would require
A) all of the lawyers and accountants to learn a new system
B) If the system is simpler, lawyers and accountants will become less valuable, because people can figure out tax law themselves
I don't think we're very likely to see real tax reform in the US unless something truly and obviously catastrophic happens, because there will be a significant and powerful group of people opposing it.
A few comments (if you RTA, you'll know to what I refer):
1) WTF? His underlying presumption seems off base. He defines cutting taxes as "government intervention" and keeping taxes as "letting the market work." That's news to me.
2) He assumes that the average new business owner will create a "startup." First, the policy mentioned in the WSJ doesn't give a definition of startup, and neither does he (although he mentions the problem). And second, if a small business isn't a "startup," what is it?
3) Of course economic growth is inversly correlated with self-employment. It's because most people suck at owning a business. But to leave in economic barriers to reduce failure is a dumb idea.
4) He claims that "government intervention" by lowering cap gains taxes would encourage competition in areas that have low barriers to entry. I thought that was the point.
5) The first article he cites showing the negative correlation between firm age and productivity comes from a paper in 1999.(http://www.jstor.org/pss/117087). Bad data.
6) The second piece of evidence he cites pulls data from 1975 - 1996. Once again, bad data. And if he's talking about tech startups, then he should isolate that data to draw the proper correlation
This isn't about bias or politics. This is about an article full of bullshit. Either it was poorly written or he tried to take a conclusion and started selectively looking for evidence to support it.
My first thought upon reading about a correlation between declining GDP and increasing entrepreneurial activity was exactly the opposite of what this article states. If the GDP is dropping, my current employer is displaying inefficiencies and lack of product. I'll just start my own company and overcome all of that!
What's interesting about the article is not the article but the quote in its first sentence: Barack Obama "proposes eliminating capital gains taxes on start-up companies". Is that not astonishing? I hadn't heard it before. It's indicative of how deep Obama's connections are in the Valley.
As for what the article calls the "obvious problem" with the idea ("the tax lawyers will have every company in America defined as a start-up"), what a lame point. The same objection could be made to any policy or any legislation. I think Obama's dismissive answer to this is right (he basically said, duh, we'll write the definitions competently).
Having worked in both, I can't possibly agree that large companies are more productive than startups - at least not in IT. That might be true in other industries though.
Oddly enough, this statement, which is the basis of his argument, is also why his conclusion is wrong.
If by productivity he means revenue per employee, then he may be correct even of technology startups. Any successful startup's revenue per employee is likely to be higher in year 5 than in year 1, when they probably had no revenue. But what matters more is the derivative: successful startups have the fastest increase in revenue per employee. In other words, the reason startups are good is they're the companies in which "firm productivity increases with firm age" the most.