Regulations Aren't What They Used to Be

As the angry debate over Ryan's budget plan wears on, one suggestion I'm seeing over and over is that we should just raise personal and corporate income tax rates back to where they were in the 1950s, when marginal tax rates were 50% on corporate income and up to 90% on personal income, while the effective rates were arguably nearly as high.

(Why don't we know exactly how high they were?  Because the very high taxes of the 1950s encouraged people to consume things in tax-free forms--shareholders were happy to used retained earnings for conglomerated empire building rather than having them distributed as taxable dividends, while managers were happy to take their income in the form of company cars and generous expense accounts.  These things showed up as actual expenses on the corporate income statement, lowering reported personal and corporate income.)

But we can't just go back there.  For one thing, while the GOP is wrong about getting Laffer Curve effects when marginal income tax rates are at 30% of income, such effects would virtually certainly apply at 90%: we'd slow capital formation and GDP growth while encouraging tax refugees and loads of cheating.

Why, then, was growth so high back then, with tax revenues flowing like wine?  Well, nothing's saying growth couldn't have been even higher.  But I think the more important problem is that lots of regulation is a one shot deal: you can get certain effects temporarily, but not permanently.

Think about the tax environment of the  early 1930s.  Top marginal income tax rates are 25% on people, 12% on corporations.  No doubt you have a fellow who does your taxes, and looks for ways to minimize them, but it's not particularly urgent--the top bracket kicks in at the equivalent of about $1.3 million a year.

Then FDR ratchets the top bracket on very high incomes into the 80s, and takes the corporate income tax into the 40s.  Suddenly, minimizing your income taxes becomes a much more lucrative endeavor.  You can afford to spend quite a lot on tax accountants and lobbyists, and still save money as long as they deliver tax savings.

Over time, an industry grows up dedicated to minimizing taxes.  The important thing to remember about trying to "reinstate" the conditions of the 1950s is that this time, the industry already exists: the army of professionals with long experience at parsing IRS rules and developing innovative strategies to avoid taxes. So even if we enacted exactly the same tax rules as we had in the 1950s, we wouldn't raise as much money this time around.

This is especially true because capital is more mobile than it was in the 1950s.  That magnifies both the income-shifting effects, and the negative economic effects, of high taxes on capital and talent.  This time around, if you bring income taxes back to 90%, the wealthy have a variety of very attractive places where they could move--and there are lots of countries that would be happy to accept their investments without demanding a 9/10s cut of the proceeds.  So we lose not only the taxes, but the benefit of the capital.  In the early 1950s, a combination of capital controls and war devastation made this sort of thing very difficult.

This is broadly true of regulation in general: if you think some tax or regulation worked in the past, you need to consider the possibility that this was simply a temporary effect before companies gained the inevitable talent at gaming your regulation.  You cannot simply assume away the gaming on the ground that companies oughtn't to do that sort of thing--and you should be wary of assuming that you can get back to the halcyon days of yore simply by "eliminating the loopholes".  Heavy regulation encourages people to get good at finding loopholes--and since "eliminating the loopholes" usually takes the form of making the law more complicated, it is at least as likely to create new loopholes as to get rid of them.

Over the long run, it is better to seek a simple set of rules than a perfect set of rules.  The perfect set of rules may get closer to your ideal--but it is also likely to be easier to game.