Monday, November 03, 2008

I am not an economist...

and Greg Mankiw is.

And yet, I'm pretty sure about this bit.

Mankiw writes:

On a regular basis, I am offered opportunities to make some extra money. It could be giving a talk, writing an article, editing a journal, and so on. What incentive is there to put forward that extra work effort?

To a large extent, the beneficiaries of that extra effort are my kids. My lifestyle is, as a first approximation, invariant to my income. But if I make an extra few dollars today, I will leave more to my kids when I move on. I won't leave them enough so they can lead lives of leisure, but perhaps I will leave them enough so they won't have to struggle too much to afford a downpayment on their houses or to send their own kids to college.[...]

Let's suppose Greg Mankiw takes on an incremental job today and earns a dollar. How much, as a result, will he leave his kids in T years?

The answer depends on four tax rates. First, I pay the combined income and payroll tax on the dollar earned. Second, I pay the corporate tax rate while the money is invested in a firm. Third, I pay the dividend and capital gains rate as I receive that return. And fourth, I pay the estate tax when I leave what has accumulated to my kids.

Let t1 be the combined income and payroll tax rate, t2 be the corporate tax rate, t3 be the dividend and capital gains tax rate, and t4 be the estate tax rate. And let r be the before-tax rate of return on corporate capital. Then one dollar I earn today will yield my kids:


Mankiw goes on to analyze the different long-term returns ostensibly generated by the tax plans of Obama and McCain.

Let's look at something again:

"On a regular basis, I am offered opportunities to make some extra money. It could be giving a talk, writing an article, editing a journal, and so on. [...] Let t1 be the combined income and payroll tax rate[...]"

Obama is proposing a significant hike in the payroll tax on people who earn more than $250,000, and it's fair to note that this would elevate top marginal tax rates considerably above the Clinton-era rates that people often refer to Obama as reinstating. It's also fair to note that it makes Social Security much more redistributive (it's always been somewhat redistributive) and much less like the insurance or pension plan the average voter perceives it to be. Indeed it makes Social Security so much more redistributive that we may see a test of a longstanding article of faith among left-leaning opponents of means-testing: that Social Security that taxed some in order to give to others would be politically vulnerable, and that keeping the program universal with payouts tied fairly closely to "contributions" was a necessary part of the political strategy of keeping the program untouchable and immortal.

But nothing in Obama's plan involves changing the fact that the payroll tax is a payroll tax, that is, a tax on wages and salaries. The honorarium Mankiw receives for giving a talk or writing a commissioned article, the royalties he receives for writing a book-- these things aren't subject to the payroll tax. The payroll tax is paid on the check he receives every month from Harvard, and not on those other bits and pieces.

Now I understand that he's trying to model the general case of "incentives to do a little more work for a little more compensation." But most people can't fine-tune their work-income tradeoffs the way a famous academic with lots of discretionary paid speaking engagements and writing assignments can. A regular salaried white-collar worker can work hard in the pursuit of whatever merit raises are on offer next year, without knowing precisely what they'll be. A working-class worker is going to be far below the threshold that's needed to make Mankiw's model work (top income tax brackets as well as maxed out on current consumption, so the marginal dollar is entirely invested and subject to compounded taxes on investment as well as inheritance taxes later). So the choice of these little lumps of extra work-for-income was important for the simulation to proceed. And it proceeds on a false basis: that extra lump of work is not taxed at the (payroll+ top income tax) rate, but only at the top income tax rate.

It also seems to me that one has to be pursuing a pretty odd investment strategy if one is paying capital gains taxes on the dollar investment every year. It may be worth noting the distorting effect created by the incentive not to realize your gains every year-- but given the existence of the tax, behavior adjusts accordingly, and people don't realize their gains every year and subject themselves to the tax. Maybe that means they forgo the chance to maximize each year's return by switching around-- but you only maximize that way if you're an omniscient forecaster. Mortals maximize expected value by buying and holding index funds. The omniscient forecaster who would otherwise be capable of getting a 10% return won't continue to chase it at the price of paying capital gains taxes every year, unless he's awfully dumb for an omniscient guy; he'll do what us mortals do and live with the market-tracking return that isn't taxed every year.

The capital gains oddity might be a tolerable approximation; the misrepresentation of marginal lumps of income and the payroll tax seems to me entirely out of bounds.