Why yes, Warren, there is a Laffer curve

Warren Buffett wrote a column the other day. (Why he can't stick to being a fantastic fund manager, I'll never understand.) He made two rather curious claims. Well, curious when taken together. Watch.

Back in the 1980s and 1990s, tax rates for the rich were far higher, and my percentage rate was in the middle of the pack. According to a theory I sometimes hear, I should have thrown a fit and refused to invest because of the elevated tax rates on capital gains and dividends. I didn’t refuse, nor did others. …

Since 1992, the I.R.S. has compiled data from the returns of the 400 Americans reporting the largest income. In 1992, the top 400 had aggregate taxable income of $16.9 billion and paid federal taxes of 29.2 percent on that sum. In 2008, the aggregate income of the highest 400 had soared to $90.9 billion — a staggering $227.4 million on average — but the rate paid had fallen to 21.5 percent. (src, excerpted and re-paragraphed)

So in the first paragraph, he claims that the Laffer curve doesn't exist. But, um, what about the second?

Gee, even I can tell, without the aid of an electronic calculating machine, that 21% of $91bn is going to be rather more than 29% of $17bn. So they cut the rate and the receipts went up.

Of course, this needs caveating. The biggie, as usual, is that you can't say anything about causality. Moreover, there is an issue to do with netting off background growth over the period. But still, it is fairly clear: Warren, you just found yourself some prima facie evidence of a Laffer curve, and tried to use it to claim there is no such thing. Well done. Now, go back to identifying undervalued stocks. You're good at that; not so good at logic or economics.