Public pension systems should offer tiered benefits explicitly contingent upon the number of children a couple has had, but with the extra benefits not kicking-in until at least 20 years after a child was born. Lots of variations on this theme are possible (extra benefits to parents of the educationally successful, etc.), though there is a slippery slope to overintrusive social engineering.

This idea has the advantage that it offers an incentive to reproduce that is likely to disproportionally affect those well prepared to raise a child. An impulsive teen whose failure to use birth control leads to a pregnancy is unlikely to be swayed in her choice of whether to have the child by an incentive 40 years distant. But a distant incentive may well affect the decision of successful couples, already socking away funds in their IRAs an 401-Ks, and whose decision to have very few children is related to conscientiousness in managing their finances.

Aside from addressing broad perverse incentives against childrearing, fecundity-indexed public pensions would help counter the growing insolvency of social insurance programs. Any private pension manager knows that it is important to match the assets and liabilities of her fund, in terms of both quantity and timing of payoff. A private pension manager buys bonds as long-term, liability matching assets. But with a social insurance programs, owing to their scales, financial assets may become meaningless. If a society fails to produce real assets sufficent to support the aged, any bonds the social insurance program has become claims on stale air. The real asset that social insurance programs rely on is productive young workers. A social insurance program that wants to match assets to its liabilities ought to be encouraging the reproduction of successful wealth creators.