christopher ruhm is a professor of economics and public policy at the University of Virginia.
Published January 21, 2019
The past half-century has witnessed dramatic changes in both labor markets and the composition of typical families. Most notable is the rapid rise in the employment of women — especially mothers with young children — and the increased likelihood that these children are being raised in single-parent households. Thus, as countless blogs, magazines and scholarly journals inform us, parents are subject to increasing stress in balancing the needs of work and family, particularly just before and after the birth or adoption of a child. And by no coincidence, the plight of these overburdened parents is no longer just the preoccupation of admirers of European welfare states.
Indeed, polls suggest there is widespread public support for paid leave, with majorities of both political parties in favor — as well as surprising numbers of owners of small businesses who are often assumed to oppose it.
The federal government does not, of course, currently mandate paid parental leave, but some states and a modest share of private employers do offer the benefit. Here, I take a close look at parental leave, explaining why voluntary provision by private employers is unlikely to be sufficient and — drawing on the experiences of other countries as well as the five pioneering U.S. states — how such programs might best be structured to contain costs and minimize disruptions to labor markets.
First, Some Terminology
Parental leave refers to the right to time away from the job around the birth of a child that is available to either or both parents. I include periods that are restricted to mothers only (often called maternity leave), those available exclusively to fathers (paternity leave), and, of course, rights that can be used by either parent. These are distinguished from family leave, which refers to work absences for other family reasons — typically for illness or care of a sick dependent — or medical leave that provides time off to deal with one’s own health problems. Existing U.S. programs often conflate coverage for parental leave with family and medical leave, whereas other countries generally make a distinction in their policies.
Who Gets What
The United States is one of just two countries — the other being Papua New Guinea, with a GDP per capita of less than one-tenth that of the U.S. in terms of purchasing power — that do not offer universal paid leave for new and expecting mothers. Most industrialized nations also offer paid time off to new fathers, although often giving them much less time than mothers.
The specifics vary, but paid leave typically involves several weeks of maternity leave, generally with high levels of wage replacement; an extended period of leave that can be taken by either mothers or fathers, often with lower levels of wage replacement; and, increasingly, a short period of leave available to fathers. The provision of paternity leave, sometimes referred to as “daddy days,” is a relatively recent phenomenon. It is motivated by a desire to reduce gender inequity and recognizes that mothers tend to take the vast majority of the time off available to either parent.
Parental leave policies in other countries are often closely integrated with early child care arrangements that extend through around age three, at which point formal early care and education programs are almost universal. The figure on page 57 shows the maximum amount of leave available to mothers in OECD countries in 2016. The average is just over one year, but there’s considerable variation. The longest entitlements — exceeding three years — are offered in Finland, Estonia, the Slovak Republic and Hungary, while the shortest are just 12 to 14 weeks, mandated in Mexico, Switzerland and Turkey.
This dimension of paid leave can be a misleading indicator of generosity, however, since wage-replacement rates often tail off sharply. For instance, the average wage-replacement rate in Finland is 74 percent during the first four months of maternity leave and 19 percent for the remaining 32 months, whereas Norway offers a considerably more generous replacement income for a shorter period. Leave durations were increasing in many countries until the Great Recession.
Parental leave refers to the right to time away from the job around the birth of a child that is available to either or both parents.
Prior to the 1993 passage of the Family and Medical Leave Act, the United States had no national entitlement to parental leave, let alone paid leave. The FMLA guarantees 12 weeks of unpaid time off work after the birth or adoption of a child, to care for sick relatives, or to recover from personal illness.
But it is far from universal: to qualify, the individual must have worked for the same employer for 1,250 hours during the 12 months prior to the leave. Moreover, enterprises with fewer than 50 employees working within a 75-mile radius need not offer the benefit. As a result, fewer than 60 percent of new parents qualify. It is noteworthy that the FMLA provides an individual right (covering mothers and fathers equally), while other nations often allow mothers and fathers to split the time as they wish.
Several states have made an effort to fill the breach with paid leave mandates. California was first, implementing its policy in 2004. New Jersey (2009), Rhode Island (2014) and New York (2018) followed. Washington State and the District of Columbia will join the club in 2020, while some 20 other states are mulling legislation. The four states currently operating paid leave programs layer these on top of their state temporary disability insurance (TDI) systems.
Doing it this way allows paid parental leave to be integrated with the existing administrative structure for TDI, which already entitles new mothers (but not fathers) to six to eight weeks of paid leave during the period surrounding birth. However, since only one other state has a TDI system of any sort in place (Hawaii), this administrative shortcut is not an option for most states.
Is Government Really Needed Here?
One argument often advanced by conservatives is that parental leave is not qualitatively different than other employment benefits, and thus should be left to private negotiations between employees and employers. Allowing for such flexible, market-based solutions does have some advantages. Employees are in the best position to assess the value of parental leave to their own families, while employers are best able to calculate the costs to the enterprise and thus how much they would need in cuts to wages or other benefits to break even. If both sides can see a way to be better off with parent leave on the books, there is room to make a deal.
However, there are many reasons private markets can fail to achieve the social optimum in this sort of calculation. Parental leave of modest length seems to lead to higher rates of future employment and wages down the road. Although these gains primarily accrue to the employee, there are also spillovers (“externalities” in econ-speak) to society in the form of increased tax revenues, reductions in unemployment compensation and other outlays for safety net programs. By the same token, parental leave has been linked to better health for both mothers and their children, resulting in reductions in a variety of public expenditures. Then, too, some analysts believe that, without encouragement, women will choose to have too few children to sustain a stable population, exposing the broader society to the consequences of rapid aging. Such concerns go a long way toward explaining why Japan and much of Europe are now heavily subsidizing motherhood in a host of ways.
More subtly, market solutions are likely to provide less-than-optimal access to parental leave even in the absence of externalities. One reason is that there are likely to be rigidities in the labor market (such as minimum wages) that prevent employers from reducing outlays for labor sufficiently to compensate for the added costs of parental leave. Even more important, efficient market solutions generally require both employers and employees to have similar (and substantial) information about the probability that workers would make use of the leave benefit.
Of course, employees are more knowledgeable than employers about their own plans to procreate, and the result of this “asymmetric” information is that an employer providing paid leave is likely to attract a disproportionate share of employees with a propensity to use it. Actually, there are wheels within wheels here: employers, fearing that offering the benefit will attract potential parents, may overshoot and demand an excessive cut in other labor costs as compensation. (This problem, by the way, is elegantly analyzed by Nobel Prize winner George Akerlof in his 1970 paper, “The Market for Lemons.”)
The number of companies offering paid leave has increased in recent decades, but coverage remains thin; only around 15 percent of employees are eligible for explicit paid parental leave benefits. Leave rights, moreover, are concentrated among high-wage workers, with low-wage and part-time employees getting the short end of the stick. Even where rights to leave are provided, they are often of short duration. What’s more, employees may be (rightfully) concerned about suffering negative consequences, such as relegation to the “mommy track,” if they use the benefit. It’s thus hard to escape the conclusion that, if we aspire to anything close to universal paid parental leave rights, some kind of government intervention will almost certainly be needed.
It’s hard to escape the conclusion that, if we aspire to anything close to universal paid parental leave rights, some kind of government intervention will almost certainly be needed.
One intervention seems to ruffle the fewest political feathers in the United States: tax breaks to firms that offer leave. Indeed, the sweeping tax changes in 2017 took a step in this direction, including a tax credit of 12.5 to 25 percent to employers for providing up to 12 weeks of leave paid at 50 percent or more of the employee’s normal wage rate. Note, however, that this approach garners relatively little bang for a taxpayer’s buck because the bulk of the benefit will almost certainly go to employers already offering paid time off.
The polar opposite approach is to require employers to provide leave as a cost of doing business. Yet, along with the political opposition that such unfunded mandates generate, there are strong reasons to be concerned about the unintended consequences. In particular, employers facing a mandate will have financial incentives to avoid hiring workers they suspect will make use of the benefit. As a result, women of child-bearing age may be less likely to be hired or promoted.
It thus makes sense for the cost of parental leave to be covered by taxpayers rather than employers — as is done in all states (as well as other countries) that provide benefits. Interestingly, the City of San Francisco’s Paid Parental Leave Ordinance requires employers to top-up the parental leave benefits mandated by the California Paid Family Leave program to 100 percent wage replacement, which may lead to the unintended consequences just discussed.
As noted, five states plus the District of Columbia have passed paid leave laws, and many more are considering them. However, the state route (as opposed to a national policy of paid leave) is bound to be problematic. States dominated by conservatives are unlikely to follow suit. Indeed, the divergence of state policies may lead to a “race-to-the-bottom” in which states otherwise inclined to mandate paid leave decide to pass on it in order to attract or retain employers. That’s why a federal approach appeals to many (including me) — though its political viability remains in doubt.
The Nuts and Bolts
Parental paid leave programs can vary in (at least) five dimensions: duration, wage replacement, eligibility criteria, job protection and financing. Although many of these may be intertwined (e.g., the cost of longer durations may be offset by lower wage-replacement rates), they’re worth individual consideration. But before doing so, it should be emphasized that leave is equally available to mothers and fathers in the states that offer it. Even though women currently use the benefit far more frequently than men, this availability does increase the number of fathers opting for parental leave.
Other countries frequently provide a year or more of paid parental leave, and always at least several months. By contrast, no serious proposals in the United States would cover more than 12 weeks. This makes some sense, at least initially, even though the evidence from other countries provides little indication of serious negative effects to economic efficiency until leaves extend beyond six or nine months. The United States has sufficiently different institutional arrangements and attitudes about workplace relationships that cross-national comparisons may be misleading.
Rights to around 12 weeks of paid leave are also in line with current entitlements to unpaid leaves under the FMLA and with the duration of paid leave provided to mothers in existing state programs. Washington State’s program, which takes effect in 2020, will also offer 12 weeks of leave (without an accompanying TDI program), and by 2021, New York will provide for 12 weeks of paid parental leave plus TDI leave benefits for mothers. Finally, 12 weeks of paid leave are also proposed in the Family and Medical Leave Insurance (Family) Act, which has been sponsored in Congress by Representative Rosa DeLauro of Connecticut and Senator Kirsten Gillibrand of New York during the last several sessions.
However, somewhat shorter-duration leaves may be more viable in political terms. The bipartisan AEI-Brookings Working Group on Paid Family Leave (of which I am a member) recently developed a proposal for eight weeks of federally funded paid leave. This represents a compromise between those on the left (including me) who would prefer at least 12 weeks and others on the right who would prefer durations of around six weeks.
Six weeks, by the way, is also the leave offered under President Trump’s most recent budget proposal. The District of Columbia’s program, to be implemented in 2020, will offer eight weeks — less than other states, but still above the very low six-week threshold.
Balancing the needs of work and family yields benefits for society as well as households. This externality, I believe, makes it desirable to provide significant assistance to parents of infants, but it also makes sense for families to have some skin in the game. It is probably also desirable to offer higher replacement rates for low-end earners, since those with limited savings are likely to find it difficult to afford time off work if their incomes are even modestly reduced. On the other hand, there may be less political support for programs when benefits are not shared by the middle class.
The four states with paid parental leave currently in place offer wage replacement ranging from 50 to 70 percent of prior earnings, with the lowest of these rates (in New York) scheduled to increase to 67 percent by 2021. Replacement rates somewhere in this range seem reasonable for a national program. However, two other nuances are worth noting. First, it is not only the replacement percentage that matters, but the maximum benefit level. These vary fairly dramatically. For instance, the ceiling is $1,011 per week in California in 2018 versus $677 in New Jersey. Second, the programs planned for Washington State and the District of Columbia will replace 90 percent of wages up to a modest threshold and 50 percent for additional earnings up to a ceiling. Replacing more of the wages of low earners than of high earners adds progressivity to the system — a welcome outcome in light of growing income inequality.
As mentioned, only around 60 percent of new parents are eligible for unpaid FMLA leaves because of the relatively stringent work-history requirements and because small employers are exempt. Both restrictions are probably undesirable. Current state programs offer broader eligibility: California requires only $300 of earnings in a qualifying quarter and covers those working for small private employers as well as large ones.
A useful way of avoiding a work-history requirement would be to base wage replacement on average weekly earnings over, say, a one-year period ending several months prior to the birth. That way, virtually all workers having children would be eligible for some paid leave, but only a small benefit would go to those with limited recent work experience. For example, using a 75 percent wage-replacement rate for average weekly earnings, an individual working 40 hours per week and earning $10 an hour over the full 12-month period would receive a weekly leave benefit of $300. Meanwhile, their counterpart averaging 10 hours per week at the same wage would get $75 per week.
Employees in small establishments have been excluded from the FMLA under the assumption that such leave is more disruptive to their employers. However, the evidence provides little support for the assumption, and surveys of small employers (with 10-99 employees) in Rhode Island, New York and New Jersey indicate that a substantial majority support their states’ paid leave programs. The exemption, then, hardly seems justified.
Giving leave recipients the right to return to the same or an equivalent position at the end of the leave is important because, in its absence, many individuals may be hesitant to take time off work after a birth or adoption. The unpaid leaves mandated by the FMLA are fully job-protected, as are parental leaves in most other industrialized countries. Surprisingly, job protection is not provided in several state leave programs, although persons also qualifying for FMLA leave are covered through the latter. A potential concern is that it would be burdensome to require employers to hold jobs for workers who started work for them shortly before taking leave. Some kind of exemption may be appropriate in these cases.
Family and Medical Leave
This article is about parental leave. However, the national FMLA and all the existing state paid leave programs (but not the soon-to-be-implemented ones in Washington State and the District of Columbia) provide equal coverage for medical and family leaves. Whether such combined coverage is desirable is a vexing issue, since the issues surrounding these other sorts of leave are somewhat different. Employers are particularly concerned about their ability to verify eligibility for medical leaves and may have greater difficulties in covering job responsibilities during them.
If family and medical leave are to be combined with parental leave into a single program, it may be useful to consider mandating shorter durations or lower wage-replacement rates, or including waiting periods. To deter abuse and limit difficulties for employers, it may also be desirable to cap maximum leave durations from the combination of all three sources at less than the sum of the maximums for each component.
Any paid leave program needs to be paid for. The existing state programs are financed through modest increases in payroll taxes that have usually been levied exclusively on employees, although employer payroll tax contributions are also included in the upcoming programs in Washington and the District of Columbia.
This history makes payroll taxes an obvious source of financing for broader programs. However, from the perspective of economic efficiency, increasing the size of the wedge between what employers pay and what employees receive is problematic. Using general tax revenues to cover the cost seems more sensible, since it would insulate employment decisions and broaden the tax base. On the other hand, the use of general revenues may not be viable because of political opposition that would make the program hostage to annual appropriations — unless paid parental leave were given the preferred status of a budget entitlement, like Medicare.
Recently, the suggestion has been floated by conservatives to finance parental leave by allowing employees to borrow against earned Social Security benefits. Specifically, receiving paid leave benefits would delay eligibility for Social Security retirement.
This approach appeals because it avoids the need to develop a new funding source — although some transitional Social Security funding would still probably be required and it may prove difficult to calculate a budget-neutral delay in Social Security. The more serious downside: this financing mechanism would impose the greatest costs on low-earners who currently have the least access to privately provided parental leave benefits.
Or, to put it another way, unlike financing from general revenues, the net impact would not reduce income inequality and might even increase it.
Well-designed paid parental leave programs would make a real difference in helping Americans balance the competing needs of work and family. But they would certainly not be a panacea. Many other related challenges would remain, among them:
• The difficulty of obtaining high-quality affordable child care.
• The pressing need to reformulate disability insurance to support needy individuals without undermining incentives to work.
• Helping prime-age workers stay on the job without letting down aging parents or ailing relatives.
Finally, any parental leave program enacted in the United States — either at the national or state level — is bound to be modest compared to those in most other industrialized countries, where generous family benefits are typically supported by both the left and right, albeit for different reasons. Arguably, this is as it should be: paid parental leave is just one of many social programs competing for public funds in a political culture that abhors taxation and is suspicious of using government to redistribute income. But it also highlights the long-term importance of challenging dogma about the virtues of a very limited role of government in a society in which the burdens faced by ordinary households seem to be multiplying.