Friday, July 21, 2006

Startups, Technology, and Inequality

There's been a lot of rather lively discussion on economics blogs lately about increasing income inequality. Don't think I can give a full roundup of the discussion—I probably haven't read even half of it—but Greg Mankiw kicks it off here arguing that the increase in inequality is caused by increasing returns to human capital (the differnece in pay between highly skilled workers and less skilled workers grows as technology improves). Brad DeLong responds here and here, arguing that inequality is increasing within classes of similarly-educated people; increasing returns to education don't explain why the top .1% have gained so much more than the rest of the top 1%. Tyler Cowen responds here, Alex Tabarrok here. There seems to be a consensus (among both bloggers and commenters) that a large part of the explanation in a winner-takes-all effect; most CEOs making $50 million a year have MBAs, for instance, but most people with MBAs aren't making $50 million a year. So the returns to human capital go disproportionately to a small fraction of those with the capital; in this sense human capital, like large stock market position, is a highly risky asset. Mankiw responds that given some reasonable assumptions, an increase in returns to education should usually increase income inequality. His commenters make a lot of good points (though the math-phobic in the audience should probably steer clear).

I wonder, however, if the proliferations of startups might not be the main cause. Startups are risky: ten equally educated and skilled entrepreneurs might found companies, but the odds are only one or two will survive. Those two founders cash out for millions of dollars; the others get good experience, but little cash. Moreover, the compensation to the successful founders would all occur in the one year they get acquired (or have an IPO); they'll skew the income distribution further by having an extremely high income in one year and low income in several others. Each year the guys who realize the value of their companies show up in the top .1% that disturbs DeLong; other years they show up somewhere much lower. This would suggest both that a lot of the inequality is a result of the beneficial tech-startup system that ultimately drives a lot of our economic growth, and that the inequality is partly a statistical artifact of compensation delivered as a lump sum rather than an annuity. Might be a lot less to worry about than it seems.