How government policy discourages people from having children
A debate broke out recently in the blogs about the ethics of having children. The occasion was the publication of a remarkably silly book arguing that reproduction is immoral. One blogger argued in response that people have an obligation to create new life as an expression of gratitude for the life they have been given. Another denied the existence of any such obligation but argued that having children is an important source of happiness for most people. In this fact he finds sufficient justification for having children, and for governments to help people afford to have them.
The discussion, while interesting, would have been unintelligible throughout most of human history. The absence of reliable means of contraception meant that having children was a less discrete decision than it is today. And while many people felt an obligation to bear children or wanted the emotional satisfactions they can bring, they also had an overwhelming practical reason for wanting them: They needed the help. They needed their offspring’s labor. They needed children, especially, to avoid hunger and privation in old age. The bargain was simple: Parents take care of their children until they are able-bodied, and in return get taken care of by their children when they no longer are.
We still need to have children so that we can enjoy a secure old age. Modern societies have disguised the old bargain by socializing it. They maintain expensive government programs to assist the elderly, financed by successive generations. The children still take care of the elderly when they grow up: But now it’s all the children providing for all the elderly, collectively.
In some ways this arrangement may represent an advance for civilization. Most people seem to think so. But it has a little-appreciated drawback: It imposes a heavy, if hidden, burden on parents, especially those with several children, and societies that adopt it therefore tend to have fewer children. For both moral and practical reasons it is time to revise the generational bargain again.
Incentives tend to change when activities are socialized, and provision for old age is no exception. Now it is possible to enjoy a free ride, as the economists say: Don’t raise children yourself, but benefit in old age from the fact that others have done so. Looking at it from the other direction: Parents contribute more to the programs than non-parents who pay the same amount of tax, but they get the same benefits. One ancient motivation for having children dramatically shrinks (although it does not vanish: many elderly people still get a lot of help from their kids). One might therefore expect that the introduction and expansion of old-age programs would lead people to have fewer children. One might further expect people to marry later in life, and for fewer people to marry at all, as they envision lives with fewer, or no, children.
The fact that children are not only future contributors to old-age programs but beneficiaries of them does not force any modification to this analysis. The childless still free-ride. Or think about it this way: Imagine a society where from time immemorial each woman has had two children. For one unusual generation, each woman has three children, and then the society reverts to the historical norm of two. The temporary increase in fertility would improve the finances of that society’s old-age programs, and this effect would never be undone. The ratio of contributors to beneficiaries, that is, would temporarily rise above what it had originally been and then fall back to its original level but not below it.
Nor does the fact that governments finance the education of children by taxing everyone, including the childless, affect the analysis. Educational expenses are only part of the economic cost of raising children, including the cost of forgone income. And everyone got an education paid for by someone else, whether his parents or taxpayers generally. Parents are not free-riding on the childless.
Even if entitlements reduce the number of children, it may still be the case that they improve social welfare. Hans-Werner Sinn, a German economist, has noted that old-age entitlements can be seen as a kind of insurance policy. They protect people against the risks that they will be unable to have children, or that their children will be unable to provide for them, or that they won’t want to. He suggests that the desire to enforce obligations toward parents was a major motive behind Bismarck’s creation of these programs. But this argument, he notes, can justify only a “moderately sized” set of entitlements. If the elderly often leave some of their pension funds to their children and grandchildren, the transfer programs are larger than optimal. In passing he suggests that the effects of entitlements on family size in his country have been anything but small: “In Germany, generations of households have learned that life in old age can be pleasant and economically sound even without children. The idea of marrying and having children in order to ensure satisfactory consumption in old age had been common before Bismarck’s reforms. A century later” — Sinn was writing in 2002 — “this idea has largely vanished, and a growing number of people prefer to stay single or at best form a ‘dink family’ — with double income and no kids.”
The American Social Security program is often said to contain a subsidy for “homemakers.” Both social-conservative activists who laud this “pro-family” feature and feminists and libertarians who consider it an illegitimate government favor for social conservatives say this. It is true that the system gives these women benefits as though they had contributed some taxes to the program. But what the government gives with one hand it takes away with the other: Take account of the anti-childrearing effect of entitlements, and only housewives with no kids — a rare social type — come out clearly ahead. A family in which the husband makes the income while the wife devotes herself full time to raising three children still loses.
In the U.S., the debate over entitlements has dwelt almost entirely on their effect on the government’s solvency, and a little bit on their effect on the capital stock. Martin Feldstein has argued, for example, that Social Security undermines the incentive to save. But research confirms that entitlements also reduce our stock of human capital by reducing the number of children we have.
That research acknowledges that large social trends other than entitlements contribute to the decline in fertility. People have fewer kids as infant-mortality rates drop. The shift away from farming has reduced the value of children as laborers. The development of financial markets has expanded the range of alternative investments. The growth of female participation in the market for paid labor has also reduced the fertility rate — although one has to be careful in analyzing this relationship, because causality runs in both directions. (A woman who expects to have one child is more likely to pursue a career than a woman who expects to have four.) A 2005 paper for the National Bureau of Economic Research by economists Michele Boldrin, Mariacristina De Nardi, and Larry E. Jones points out that “the size and timing of the growth in government pension systems” matches up nicely with fertility trends in the U.S. and Europe. They expanded on both sides of the Atlantic Ocean, and fertility fell on both sides, after World War II; and they expanded more in Europe, where fertility fell further. In their model, entitlements account for roughly half of the decline in fertility, and 60 percent of the difference between European and American fertility. When a pension system expands by 10 percent of GDP, the average number of children per woman drops by 0.7 to 1.6. “These findings are highly statistically significant and fairly robust to the inclusion of other possible explanatory variables.”
A 2007 paper by Isaac Ehrlich and Jinyoung Kim, also for the NBER, reached similar conclusions, finding that pension programs explained a little under half of the decline in fertility rates, and a little more than half of the decline in marriage rates, in developed countries between 1965 and 1989. One implication of this finding is that pension programs have contributed to their own financial woes by suppressing fertility.
In one of the last issues of the social-science quarterly The Public Interest, the sociologist Neil Gilbert looked at how the fertility decline played out in the lives of different groups of women. He constructed a useful, if rough, typology. He labeled women who reached age 40 without having children “postmoderns.” As a percentage of women of their age they had increased from 10 to 18 percent between 1976 and 2005. “Traditional” women, who had three or more children by that age and were usually “stay-at-home mothers,” had been 59 percent of the group and had fallen to 29 percent.
The percentage of “modern” and “neo-traditional” women, with one and two children respectively, had also jumped (by 90 and 75 percent). The moderns tended to be career-oriented, like the postmoderns, but the neo-traditionals tended to prefer part-time employment.
It is easy to surmise the differences between these groups that Gilbert does not emphasize (or sometimes even mention). Almost all of the traditional women (defined, remember, by number of children) are married; their average age at first marriage is surely lower than that of other groups; they are almost certainly more religious. They are also more conservative in politics: The list of states that went for each party’s presidential candidate in the close election of 2004 lines up pretty well, and in exactly the way you’d expect, with their rank in terms of average age of first marriage and white fertility rate. (The racial qualifier on the second correlation results from the overwhelming Democratic preference of blacks, and strong Democratic preference of Hispanics, which holds regardless of family type.) Republicans in presidential politics have illustrated the pattern almost too perfectly in recent years. The large families of John McCain, Sarah Palin, Mitt Romney, and Rick Santorum led a Washington Post reporter to comment on the party’s “smug fecundity.”
Gilbert’s essay was titled “What Do Women Really Want?” and his answer to Freud’s famous question is the obvious one: Different women want different things, and some of these differences form predictable patterns. His “traditionals” and “postmoderns” have different values and interests, on average. It is this basic fact that underlies the “mommy wars.” Hence our inability to wish those wars away.
Hilary Rosen, a Democratic lobbyist and talking head, set off a brief furor in April when she said on CNN that Ann Romney had no understanding of the economic circumstances of most American women because she had never worked a day in her life. Rosen is no more representative of “working moms” than Romney is of stay-at-homers: Each has far too much money for that. But the warring sentiments expressed during the brief controversy — Rosen, under pressure from the White House, apologized — reflected an enduring conflict. Many moderns regard traditionals as self-indulgent and retrograde. Many traditionals regard moderns and postmoderns as selfish and materialistic.
This division helps to account for the political weakness of “family-friendly” policies: They invariably help some families more than others. Moderns are the core constituency for subsidized day care. Traditionals and postmoderns often resent it as a tax on their life choices. These policies might be thought to counteract the negative effect of entitlements on fertility. But their actual effect is ambiguous because different women respond to them differently. The availability of subsidies might make it easier for women with no children to have one, or women with one child to have a second. They are much less likely to lead a woman with two children to have a third. They may even discourage her, precisely by making it easier to lead a life with one or two kids plus paid employment. If women considering having a third child are also considering scaling back their participation in the labor market, subsidized day care may be something they pay for in taxes more than it is something they receive as a benefit.
Which effect will predominate depends on, among other things, how many women of each type a given society has. In a society where full-time paid employment by women is nearly universal and almost nobody has three or four children, day-care subsidies might well increase fertility; not in a society with the opposite conditions. “Family-friendly workplaces,” Gilbert notes, are also friendlier to some family structures than to others — families with fewer children are more workplace-friendly, one might say — and their effects too are therefore ambiguous.
Discouraging middle-class adults from having children is one of the federal government’s most important social policies, even if its existence is not widely recognized. It is hard to justify it in the absence of a domestic overpopulation crisis. We would never have adopted an explicit policy to this effect democratically. Neutrality on family size seems a much better policy for a limited government in a free society. We ought to end the federal government’s bias against having children.
The conceptually simplest way to eliminate the negative effects of entitlements on fertility would be to eliminate the entitlements. No way that’s happening. Some proposed reforms to entitlement programs would reduce the effect — but not all proposals would. Raising payroll taxes to finance future benefits would not help, and could hurt. Partially converting Social Security into a system of private savings accounts, whatever the other merits of the idea, would not reduce the program’s implicit tax on childrearing and could, again, increase it.
Reducing the size of entitlements would reduce their effects on family structure. Altering Social Security to slow the growth of benefits would be one such reform. But even a reined-in program would still entail a large, forced transfer of wealth from larger to smaller families. To prevent this transfer would require either paying parents more than non-parents in retirement or taxing them less beforehand. The rationale in either case would be that raising children is a contribution to the old-age programs just as taxes are, and the government should recognize it. The tax-cut approach seems preferable: Just let families have the money now instead of taking it from them to return later. Robert Stein, an economist at First Trust Advisors who served in the Treasury Department during the George W. Bush administration, has calculated that a $5,000 tax credit per child would fully offset entitlements. (Stein, I should note, has exerted a large influence on my thinking on the issues considered in this essay, and he pointed me toward some of the research it draws on.) The logic of the tax credit would require that it be applied against payroll taxes as well as income taxes. Conservatives sometimes resist payroll-tax cuts on the theory that payroll taxes fund entitlements, and tax credits that reduce people’s contributions give people something for nothing. Obviously that argument, whatever its force generally, would have none in this case, since the premise of the policy is that children and payroll taxes both finance old-age programs. If a childless couple making $100,000 has a total federal tax bill of $30,000, a similarly situated couple with two kids should pay $20,000. A couple that has no tax liability, on the other hand, shouldn’t get an annual $5,000 check for each child they have. That arrangement would enable them to start getting their own free ride: receiving pension benefits without having contributed through either children or taxes.
Unlike subsidizing day care or forcing companies to offer generous parental leave, an enlarged child credit would have an unequivocally positive effect on fertility. Families of three might often use the money for day care; of four, to move one parent from full-time to part-time employment; of five, to get a slightly bigger house; and of all sizes, to bank for future educational expenses. The choice would be theirs.
The social-science literature on the effects of the tax treatment of parents on fertility finds mixed results. Papers have found that tax benefits for children have raised fertility significantly in Quebec, in France, and in Israel. Research on the U.S. has tended (though not uniformly) to find small effects. The effects could be non-linear: Quintupling the existing $1,000 child credit could have an effect more than proportionally larger than the modest policies so far studied. The goal of the credit, it should be remembered, is not to bribe Americans to have more children than they want. Rather it is to rectify the government’s bias against children, which leaves families with children bearing an unjustifiably large share of the tax burden. Reducing that share would surely help some people who want more children to have them — and surveys suggest that in the U.S. and the West generally, desired family sizes are larger on average than actual family sizes. The credit would not make much difference to the very rich, or for those who have little in the way of federal tax liability to begin with. (Single parents would rarely get much benefit from it.) The biggest impact would be on middle-class families: exactly the people on whom one would expect old-age entitlements to have the largest effect.
Many Americans, especially conservatives, find the idea of flattening taxes appealing. They want a tax code that doesn’t discriminate between homeowners and renters, between people who buy “green” consumer goods and everyone else, and so on. In their pursuit of this goal, conservative politicians have sometimes proposed to eliminate the paltry child credit in today’s tax code. Their mistake is to consider the income tax in isolation from the payroll tax and what federal taxes pay for. Getting rid of the child credit would make the federal government less neutral with respect to family size, not more; and expanding it would make it more neutral, not less.
Readers may well wonder whether a large tax cut would be wise at a time of large deficits. But the appropriate tax structure is a separate issue from the appropriate tax level. Whether the tax code is designed to extract 15, 19, or 23 percent of the nation’s economic output for the federal government’s use, parents ought to pay a lower portion of that burden than they do now. To make room for a large child credit, my preferences would be, in order, to cut spending, to end or reduce truly discriminatory tax breaks, and to expand the top tax brackets so that a higher proportion of the income of high earners is taxed at the top rates. What’s important is that budgetary decisionmakers include the fair treatment of parents among their goals.
Polls show strong public support for a bigger child credit, especially among middle-income voters. Governor Romney was recently overheard telling donors that he would be campaigning on two things: “jobs and kids.” A presidential race is not the right forum for a discussion of trends in Western fertility rates. But there is more the governor could usefully say than he has so far, and he could say it in public.