Mitt Romney has proposed a child allowance of $4200/year for children under age 6 and $3000/year for children age 6 to 17, which is gradually phased out for people making over $200,000 (depending on the child’s age.) It is to be paid for without boosting the budget deficit, by reducing certain other poverty programs and also eliminating certain tax deductions, such as what’s left of the SALT deduction. (This last element is one of my favorite parts of the plan.)

I don’t know enough about the plan to have a firm opinion, but from a utilitarian perspective it seems to have some positive features:

Equity: The net effect is to shift money from the affluent to the poor, which probably results in a significant gain in aggregate utility.  (Yes, we can’t measure utility, but it seems likely that this factor is a net plus.)

Efficiency: It’s hard to say whether Romney’s plan improves or reduces efficiency, and that’s where I’ll focus the rest of the post.  But the mere fact that “it’s hard to say” is a sort of plus for the plan, because the equity considerations seem to be pretty clearly utility improving. With most welfare proposals, greater equity comes at a cost of lower efficiency.  I think it’s fair to say that either Mitt Romney is a very clever guy, or he has smart advisors, or both. At the end I’ll suggest a modification that would boost the equity of the plan, without any clear loss in efficiency.

1. Some conservatives like the fact that these child benefits would boost the birth rate, pointing to the fact that people say they want more children than they actually have.  I don’t share their worry that the birth rate is too low, and I don’t trust polls.  Some conservatives worry that paying poor people to have kids would cause so-called “inferior” people to reproduce.  I also don’t share this worry.  For me, the effect on births is a non-factor.

2.  Work disincentives can come from either the income or the substitution effects.  The substitution effect in Romney’s proposal is small, as parents don’t lose the child allowance until their income rises to well above $200,000.  So on that basis it won’t discourage poor people from getting a working class job.  There is a very mild work disincentive for upper middle class people experiencing the phase-out of the benefit.  The income effect refers to the fact that poor people might no longer work because they feel they can live on the child allowance without working (perhaps combined with other programs like food stamps.)  It seems to me that this disincentive would be quite modest for the size of benefits proposed by Romney.  Still, in net terms there’s probably a mild work disincentive from the issues I’ve discussed thus far.

3.  Many of the other provisions actually boost efficiency.  Several other (inefficient) poverty programs are either reduced or eliminated.  Furthermore, there’s a substantial gain from reducing the complexity of both the welfare system and the income tax system.  Eliminating the SALT deduction also discourages wasteful state and local spending.  So the various provisions that pay for the benefit have a significant positive impact on economic efficiency.

Combining points #2 and #3, I see no clear evidence of either an overall gain or loss of efficiency.  And again, the equity benefits seem pretty clear to me.

One final comment.  Why not make the child allowance fully universal, and then slightly boost the payroll tax (on wage income only) on people making over $200,000 a year to pay for it?  On an equity basis, that would redistribute money from the very rich down to the upper middle class, as people with very high wage income would pay more extra tax than they’d gain from the child allowance, while the opposite is true for the upper middle class—those making modestly above $200,000.

On efficiency grounds, my proposed modification would make the income tax system much simpler, so that’s a net gain.  The increase in the payroll tax rate would be smaller than the implicit marginal income tax during the phaseout range of Romney’s proposal (which mostly applies to people in the $200,000s), so extremely affluent people would face slightly higher MTRs while modestly affluent people would face significantly lower MTRs.  Overall, I doubt there’d be much change in economic efficiency, maybe even an increase.