Well, Ramesh, if you object to my characterizing the introduction of new distortionary policies to offset the effects of old distortionary policies as “social engineering,” perhaps we could compromise on “offering top recruiters a rebate on the operating costs associated with participating in a pyramid scheme.” In any case, I think that to argue that a tax incentive isn’t a tax incentive because it is not intended to be a strong enough incentive to change child-bearing behaviors is cutting it pretty fine. Presumably, manufacturing tax credits also are not intended to be so strong as to get people who have no interest or expertise related to manufacturing to open factories. The circle of “social engineering” grows ever narrower.
But in any case, the implicit arithmetic—a growing population of contributor-beneficiaries enabling the easier payment of benefits—only really works with the assumption of infinite population growth. Reconsider your “temporary spike” contributor-beneficiary-ratio scenario:
The ratio of contributors to beneficiaries temporarily rises and then falls, but not to below where it had been originally. Taxpayers with more kids contribute more to the entitlements than taxpayers with fewer kids.
That’s single-entry accounting. In a situation in which per capita benefits exceed per capita contributions — that being our situation — then those taxpayers “contribute” more in the form of providing taxpayers, and take more in the form of providing beneficiaries. The opposite would be the case if per capita contributions exceeded per capita benefits. But Social Security recipients are net tax-takers vis-à-vis that program, not net tax-payers. As long as that is the case, there is no way to arrange things so that additional contributor-beneficiaries are net contributors rather than net beneficiaries.
In that sense, the more straightforward social-engineering argument—that we want to encourage, or at least reward, people for having children—is in fact more persuasive. But I do not think it is persuasive enough to overcome the case for a simpler tax code oriented toward raising revenue rather than a more complex one oriented toward other ends.