Did tax cuts in the '80's spur Internet entrepreneurship twenty years later? I can't prove it. But it seems very plausible, and no one's proven otherwise. After all, when marginal tax rates are 70%, the dream of striking it rich in business seems pretty hopeless, and you'd expect ambitious kids to look for other ways to excel.
I also don't know of any evidence for the impact of taxes on educational attainment. On the other hand, producing evidence would not be easy, as current taxes are less relevant for educational choice than expected future taxes. That is, in deciding how much to invest in skill acquisition, a young person would have to consider expected tax rates that would apply over the next several decades of working. Eddie's [Edward Lazear's] view that taxes matter for human capital accumulation seems like a plausible hypothesis. Reasonable people can disagree about the likely magnitude of the effect.
This situation exemplifies a common conundrum for policy advisers like Eddie. In the absence of hard evidence, should he act as if there is no effect, as Larry [Lawrence Katz] seems to be suggesting here? Or should Eddie rely on the general principle that people respond to incentives and make an educated guess about the magnitude?
Bottom line: If the best empirical methods probably aren't good enough to locate an effect if it exists, we're back to introspection and common sense. There's no escaping it.