Reihan pointed me to Stephen J. Entin’s opinion piece in the Wall Street Journal today attacking Ramesh’s proposal for a family tax credit. As Reihan says, Ramesh is (more than) capable of defending himself, but there were a few statements in the piece that pushed my hot buttons about sloppy argumentation.
Entin says that:
We know, from decades of economic modeling, and observing the revenue feedback when taxes are changed, that each added dollar Washington collects, then spends or gives away (for goods and services, entitlements or tax-transfers), costs the public about $2.50.
Later in the piece, speaking of the 2001 tax cuts, he says that:
The apparent cost of the package led Congress and the administration to phase in the tax reductions over six years. One result was the anemic recovery from the 2000-2001 recession. The delay in bringing down unemployment cost millions of lower-income households more than they got back from the first small hikes in the child credit.
This is exactly the kind of false precision and excessive assertion of knowledge that gives econometrics a bad name. Entin knows what the economic recovery would have looked like over the past several years had various aspects of that tax package relating to changes in marriage penalty, child tax credit and so forth been designed differently? That’s a very impressive ability to predict the performance of the economy in the counterfactual world with different taxes. Naturally, such a “prediction” is completely non-falsifiable. Perhaps he can let us in on the secret of what next year’s GDP and jobs growth will be.
He knows that each dollar of tax costs us $2.50? Really – that would be tough to model with certainty without knowing, to choose one example, what the odds and costs of encouraging an adversary to start a war would have been in the counterfactual world with lower US defense expenditures. There are a bunch of guys running hedge funds right now who are feeling the effects of over-estimation of their ability to predict just such risk.
Further on, he says that:
The bottom half of the income distribution pays barely 3% of the income tax. According to the Tax Foundation, over 40% of the population owes no federal income tax…
True, but of course Ramesh has proposed that this benefit be credited chiefly against payroll taxes, not the income tax. Federal income tax receipts were a substantial minority of all taxes collected in the US last year. When you consider the total tax burden by income quintile, the picture looks a bit different, and a responsible argument would confront this.
Look, I think that in a perfect world the federal government would be much smaller, and federal income tax rates would be much lower, flatter and simpler. I think that we can get to such a world over time. I think that the central focus of conservative tax policy needs to be on growing the whole pie of the economy, rather than distribution of the pie. But we don’t live in a perfect world, and it seems to me that consideration of this proposal in the year 2008 needs to be seen in the light of achievable tax code changes, impacts of this proposed change on coalitional politics and so forth. There is a respectable pro-growth argument rooted in practical reality that the right thing to do is always to lean into the wind to make each incremental change to tax codes in the direction of lower, flatter and/or simpler. It’s too bad that Entin doesn’t make it.
( cross-posted at The Corner )