It’s been widely reported that Romney’s effective tax rate is 13.9%. The low rate is largely because he receives most of his income from investments (stock portfolio appreciation and the like), which are taxed at 15% (a far lower rate than earned income).
Greg Mankiw (Harvard economist and Romney advisor) and others have argued that the tax rate on capital returns is actually much higher, since those returns come from corporate profits, which are subject to corporate income taxation (as high as 35%). Romney himself has said that this means his tax rate is “really closer to 45 or 50 percent.”
Economist Rajiv Sethi (Columbia) has argued that this reasoning is totally wrong. The existence of the capital gains tax, he says, was already priced into Romney’s shareholdings when he bought them. The idea is this: suppose you’re bidding on a $10,000 (pre-scratched) winning lottery ticket on eBay (probably not legal, but ignore that for now). Lottery winnings are heavily taxed — say at 50% — so that when the winning bidder turns in the ticket (s)he’ll have to send the IRS $5,000 in extra taxes on the lottery winnings. With this in mind, what would you expect the winning bid to be? Obviously not more than $5,000. Suppose you bid $4,900 and win the ticket. At the end of the year, should we say you paid an extra $5,000 in taxes because of this transaction? Of course not! It’s true that you cashed a lottery ticket worth $10,000 which was taxed at 50%. But that tax was already taken into account by bidders for the ticket, so the $5,000 tax payment didn’t cost you anything. Similarly (this argument goes) we shouldn’t take the corporate income tax into account when calculating Romney’s tax rate, because the prices of his shares were already reduced accordingly when he bought them.
These two stories seem glaringly inconsistent — according to Mankiw, Romney pays a higher tax rate than any of us; according to Sethi, Romney’s tax rate is very low. So what’s going on? I emailed Mankiw to ask for his take on Sethi’s post, and he responded with a one-liner:
If we levied a tax on land, to whom would you attribute the tax burden?
I thought about this a bit and wrote back:
Good point. I would expect land prices to fall immediately at the time the tax is announced, effectively making a transfer from the (then) landowner to the government. Thereafter, land prices would be such that cash flow from the land generates the risk free rate of return.
In the case of securities, I would expect share prices to be such that the post-tax cash flow (net of both corporate and capital gains taxes) equals the market interest rate. So it might not make sense to attribute a corporate tax burden (or a capital gains tax burden) to current capital holders at all, since the burden was born entirely by whoever held the capital when the tax was announced. Is that similar to how you think about this? (Obviously there are GE [general equilibrium] effects as well, since the higher tax means fewer projects will generate positive net returns, and the tax harms those marginal project-holders.)
To which Mankiw responded:
Yes, I think so.
In general, the CBO calculations on tax incidence are static. That is, they assume capital bears the burden of all capital taxes, and labor bears the burden of all labor taxes. That is right in one-period model if capital and labor are inelastically supplied.
But as you point out, there are complicated intertemporal issues, which surely complicate things. To the extent taxes are capitalized into prices, the original owners bear the entire burden.
I get the feeling that you won’t know the “right” answer until the question is more precisely posed. That is, what are these effective tax rates supposed to measure?
Upshot: it’s not really clear what we mean when we say “Jane’s tax rate was X.” Do we mean “The IRS received a check for X*(Jane’s income) from Jane”? The lottery ticket example demonstrates that this is a completely uninformative statement. Someone could have sent the IRS a very large check without incurring any costs. In fact, if the winning bidder had a low enough income, (s)he could have a tax rate over 100%! We probably mean something more like “Taxes imposed a cost of X on Jane”. That seems like a more interesting question — but it also means that we need to take into account price movements caused by taxes. These issues are covered in the tax incidence literature. The calculations depend on the elasticities of demand and supply in the relevant markets.
One final note: in these incidence calculations, both the corporate income tax and the capital gains tax enter in the same way — they create a wedge between the cash flow paid by a security and the net income received by the shareholder. Therefore it’s almost certainly not right to talk about Romney’s 14% tax rate — the range could be anywhere between 0 and 50% depending on how well you think prices incorporate future taxes, but there’s not a clear reason to count the capital gains tax while ignoring the corporate income tax.