Tanya BykerAssistant Professor of Economics, Middlebury College
There is almost universal agreement in the United States that workers should have access to wage replacement during a work disruption due to their own or a family member’s serious illness or the arrival of a child. In the current patchwork of limited private provision and a handful of state mandates, however, access to this type of insurance against lost wages is uneven and inequitable. Given the positive externalities to health and well-being of families that have been extensively documented, particularly in the case of paid parental leave, there is a strong argument for government coordination to correct for the inefficient underprovision of paid leave. We propose a federal paid parental and medical leave program to be administered by the Social Security Administration and financed by payroll taxes paid equally by employees and their employers. Flexible eligibility criteria and benefit provision would extend access to groups most disadvantaged by today’s system: covering 75 percent of part-time workers, 83 percent of workers making under $15 per hour, and 87 percent of single-parents. Together, the components of this proposal lay out a path, towards establishing comprehensive paid leave coverage commensurate with that demanded by American workers.
Paid family and medical leave (PFML) refers to fully or partially compensated time away from work to care for a worker’s own or a family member’s medical needs or to care for the arrival of a new child by birth, adoption, or foster care placement (hereafter new child). In the current terminology, “medical leave” relates to a worker’s own medical care while “family leave” relates to both care of a new child and care of family members with medical needs. This designation is consistent with the history of paid leave’s evolution in the United States, starting with the introduction of short-term disability insurance (STDI) for a worker’s own medical-related work absences in the mid-19th century, followed by the much later introduction of pay during leaves for childbirth, and eventually to care of sick family members. We propose a new designation: “parental leave” referring to care of a new child, and “medical leave” covering a worker’s own and family members’ serious medical conditions. We will use current terminology when discussing existing policies (parental, family, and medical leave) to facilitate comparisons with this and other sources. We will transition to our proposed new terminology in the context of discussing our proposal for, what we call the Paid Parental and (Family and Own) Medical Leave (PPML) program.
In what follows we provide an overview of the evolution of protections and wage replacement for employment disruptions due to parental, medical, and family needs in the United States. These policies began with early, state-based protections for medical conditions, and, eventually, pregnancy care. Following on the heels of these policies, the federal government mandated job protection through the Family and Medical Leave Act (FMLA) of 1993. Several decades later, California became the leader of a wave of state-based paid parental and family leave programs. Finally, the most recent years have ushered in significant and varied federal policy proposals for paid parental, family, and medical leave.
Private STDI, which provides wage replacement to workers during leaves for their own serious medical conditions, has been available in the United States since the mid-19th century. In 2020, 40 percent of private-industry workers had access to employer-supported STDI (Bureau of Labor Statistics [BLS] 2020). In addition, STDI coverage is universally available in a handful of states that mandate coverage.1 In the 1970s, pregnancy was designated as a disability eligible for STDI, effectively creating a paid maternity leave benefit for all women covered by STDI (Timpe 2019).2 Following on the heels of these changes, 12 states and the District of Columbia mandated job protection, but not pay, for employment disruptions driven by medical, family, and parental needs.3 This effort culminated in 1993 with the federal enactment of the FMLA.
The FMLA provides 12 weeks per year of unpaid job-protected leave for mothers and fathers to care for a new child and for workers to care for their own or a family member’s serious medical condition. Leave under FMLA is available to eligible workers at covered firms. A covered firm is one with at least 50 employees across all worksites; an eligible worker (1) must be employed by a covered firm, (2) must work at a worksite at which there are at least 50 employees within 75 miles, and (3) must have worked for at least 1,250 hours within the past 12 months.4 The US Department of Labor estimates that, in 2018, 56 percent of US workers were both covered and eligible, and that 17 percent of those workers had taken an FMLA leave in that year (FMLA Surveys 2018). Figure 1 shows that higher wage, more educated workers are more likely to be covered and eligible for FMLA than low-wage, less-educated workers.
In 2004 California became the first state to offer a paid family leave (PFL) program providing partial wage replacement for employment disruptions caused by the arrival of a new child and to care for sick family members. New Jersey followed suit in 2009, and as of early 2021 PFL is also available in four other states—New York, Rhode Island, Massachusetts, and Washington—and the District of Columbia. Three other states—Connecticut, Colorado, and Oregon—have passed paid-leave legislation and are in the process of implementation. The earliest PFL programs were married to the state STDI system by expanding coverage and benefits beyond medical leave to include parental and family leave insurance. More recent legislation has moved away from this design by offering a maximum annual benefit to each worker that can be claimed among several needs categories, and, in some cases, by offering a cap on the maximum benefit that can be claimed for particular needs categories.
Although they are varied in specific implementation, these state-based paid family and medical leave programs share similar design strategies, summarized in table 1. The primary parameters of these programs are the wage replacement rate, the eligible needs categories, and the allowable duration of leave. In designing the wage replacement rate, policymakers face an additional design consideration: whether to incorporate a redistributive component by providing more generous benefits for low-income beneficiaries. With regard to eligibility categories, policymakers typically agree to cover parental, medical, and family care driven employment disruptions, but there is also an opportunity to subsidize additional categories of leave. Finally, policymakers must determine the permissible duration of leave, balancing the ex-ante expectation of the need for leave with the cost of increasingly long leave lengths, both from a budgetary standpoint and from the viewpoint of minimizing costs to the employer.
In addition to the benefits design, policymakers must determine the funding mechanism. Typically, these funds are raised by a small payroll tax on wages. The statutory incidence of the payroll tax can be imposed on employers, employees, or both parties, but the theory of tax invariance implies that this choice does not differentially influence any associated distortionary impact of the tax itself. Said differently, the incidence of the tax is not driven by who is statutorily required to remit the tax to the tax authority; instead, incidence is driven by market forces. The specific tax rate can be chosen to simply balance ex-ante expected claims and administrative costs, or, as with all taxes, the rate can be chosen to raise revenue above and beyond these needs. Finally, an important administrative parameter is the work history requirement. In this case, designers balance total expected claims with expected funds. Weak work history requirements might encourage some people to work only enough to become eligible. These workers, however, will not have substantially contributed to the funds given the short duration of their employment. This behavior complicates the ability of policymakers to balance claims and funds.
Table 1 summarizes the parameters of the state-sponsored PFML programs as of early 2021 (including states that have passed but not yet implemented legislation). All state-sponsored paid-leave programs include own-medical, parental, and family care among the allowable needs categories. A few state programs provide benefits for care of a military family member with serious illness or injury or for needs related to the military deployment of a close family member. Additionally, a few states provide benefits for employment disruption due to needs related to domestic or sexual violence that affects the employee or a close family member. In 2020 replacement rates ranged between 50 percent and 95 percent, and maximum weekly benefits ranged between $667 and $1,312. All state paid-leave programs are financed by a payroll tax ranging between 0.4 percent and 1.2 percent, but there is wide variation in whether the statutory incidence of the payroll tax falls on employees, employers, or both. Finally, all states have minimum earnings and employment requirements, but these range from California’s very minimal requirement of a minimum of $300 in wages over a worker’s base period to New York’s requirement that workers must have worked full time (at least 20 hours per week) for 26 consecutive weeks or must have 175 nonconsecutive days of part-time employment (see table 2).
While the FMLA is the primary policy governing federally mandated employee protection for employment disruptions, including those due to family, medical, and parental leave, there have been three recent legislative changes that have expanded access to wage replacement for these leaves. While these policies stop short of a federally sponsored paid family and medical leave program for all employees, they nonetheless represent important steps toward this goal.
First, the 2017 Tax Cuts and Jobs Act introduced a temporary employer tax credit to employers that voluntarily provide PFML. This tax credit subsidizes wage replacement costs for qualifying low- and middle-income employees through the general business credit. The credit was made available for tax years beginning in 2018 and 2019, and a one-year extension of the credit to tax years beginning in 2020 was included as part of the Further Consolidated Appropriations Act of 2020.
Second, the National Defense Reauthorization Act of 2020 amended the FMLA, introducing paid parental leave benefits for most of the federal civil service. Although eligibility rules differ slightly for legislative branch employees and the broader federal civil service, employees are entitled to 12 weeks of paid parental leave for the arrival of a new child by birth, adoption, or foster care placement. This leave is in addition to accrued annual and sick leave, but the paid parental leave benefit must be coupled with the employee’s FMLA entitlement benefit. Unlike many of the state-sponsored paid parental leave programs, the federal civil service paid leave includes a return-to-work agreement whereby employees who do not return to work for 12 workweeks following the conclusion of the paid parent leave may be required to repay employer contributions toward employee health care premiums made during the leave period. The proposed Federal Employee Paid Leave Act would amend the FMLA to provide 12 weeks of paid leave for all FMLA-qualifying uses of leave and would remove the return-to-work provision.
Finally, the Families First Coronavirus Response Act of 2020 and the American Rescue Plan of 2021 temporarily amended the FMLA to expand access to paid medical and family leave in response to the unprecedented impact of the COVID-19 global health pandemic. For the first time, job-protected leave with wage replacement was offered to employees of small- and medium-sized businesses for child-care needs, in addition to medical- and family-care needs. Due to the exigent circumstances surrounding all coronavirus response legislation, there was no accompanying tax or other revenue raiser earmarked to finance the cost of the leave; instead, the federal government directly bore the cost. These benefits are set to expire on December 31, 2021.
Recently, there has been an active legislative agenda proposing several different solutions to the lack of federal PFML. These proposals have varied widely, some creating new social insurance programs (Family and Medical Insurance [FAMILY] Act), some modifying existing social insurance programs (New Parents Act of 2019, and the prospective CRADLE Act), and some introducing new tax credits (Advancing Support for Working Families Act, and Support Working Families Act) or tax-preferred savings accounts (Working Parents Flexibility Act of 2019, and the Freedom for Families Act). Former President Trump’s fiscal 2021 budget proposed six weeks of financial support for parental leave, to be administered through state unemployment compensation offices.
The vast majority of Americans—91 percent—from across the political spectrum (96 percent of Democrats and 86 percent of Republicans) believe that workers should receive pay when they experience employment disruptions for at least some exigent medical and family needs (Horowitz et al. 2017). While support is the strongest for paid medical leave (85 percent) and paid maternity leave (82 percent), there is also strong support for paid paternity and family medical leave. In spite of this consensus, there is a divide about how to implement and finance paid leave: 51 percent of Americans believe the government should mandate employers to provide paid leave and 48 percent believe that employers should be able to decide for themselves.
Voluntary, private provision is the current paid-leave regime in all but a handful of states in the U.S. Under this regime, the majority of workers do not have access to paid parental and family leave and for those that do receive any pay during work absences related to family or own medical needs, this pay is often short (fewer than 20 days) and incomplete (BLS 2020; FMLA Surveys 2018). This section describes the challenges surrounding PFML in light of the patchwork of policies in the United States and the divergent viewpoints on the optimal delivery mechanism. We discuss, in turn, (1) the economic rationale for the federal coordination of PFML, (2) evidence of inadequate and inequitable access to paid leave in the absence of a federal policy, (3) the link between paid leave and positive outcomes for health and well-being, (4) implications of paid leave for workers, (5) implications for businesses, and finally (6) the equity and efficiency challenges of a federal paid-leave program.
PFML provides insurance against lost wages due to short employment disruptions driven by exigent medical- and family-care needs. This insurance can be provided through one of several mechanisms. First, individuals can choose to self-insure over time. However, intertemporal budget and liquidity constraints make this option out of reach for many, especially for low- and middle-income workers. Second, employers can create small-group insurance pools across employees within a firm, funded by a combination of employer and employee contributions. However, this private provision of paid leave results in an inefficient underprovision of insurance due to coordination problems across firms, adverse selection in employees, and undervaluing of the positive externalities that paid leave can have on the health and well-being of employees and family members. Finally, state and federal governments can coordinate a social insurance program. This public option can improve upon several of the disadvantages of individual and private provision by pooling risk across all employees, making insurance more affordable and more widely available.
Figure 2 depicts variation in the share of workers in private industry with access to paid time off by wage level (figure 2a) and firm size (figure 2b). While the majority of workers at private firms have access to paid sick leave (75 percent) and to paid vacation (79 percent), which can be used during a health-related interruption, the duration of this wage replacement is relatively short: workers with one to five years of tenure with a firm have, on average, six paid sick and fifteen paid vacation days per year. Only a small fraction of workers have a benefit specifically designated as PFL—20 percent. Importantly, figure 2 reveals that access to paid leave of any form varies substantially across firm size and the income distribution, with fewer than half of workers in the lowest 10 percent of wages having access to any form of pay during work interruptions (BLS 2020).
Survey evidence, shown in figure 3, reveals how workers at different wage levels cope with the financial interruption of family and medical leaves. Figure 3a shows that, among workers earning less than $15 an hour, nearly 60 percent received no pay at all when they were away from work due to their own illness, the arrival of a new child, or a sick family member; fewer than a quarter received full pay. This stands in sharp contrast to the experience of workers with wages over $15 per hour: 57 percent received full pay and fewer than 16 percent of those workers received no pay. When looking at family and medical leaves between 11 and 60 days, even the higher-earning workers received full pay only 40 percent of the time. In light of the fact that the majority of workers receive, at most, partial pay during leaves, it is important to understand how they deal with the reduction in income. Figure 3b shows that both low- and high-wage workers are equally likely to limit spending and dip into savings to cover costs while on leave. However, low-wage workers are much more likely to need to borrow or go on public assistance.
Given the financial burden of lost wages during leave, it is not surprising that many workers do not take time away from work when facing their own or family health issues: 7 percent of workers in 2018 (9 percent of low-wage workers) report not taking a leave they needed. The most common reason for not taking a needed leave was the inability to afford losing wages; this was true for 80 percent of low-wage workers and 60 percent of workers earning more than $15 per hour (FMLA Surveys 2018). This finding is consistent with evidence that only college-educated mothers and mothers in married households increased parental leave-taking due to unpaid leave provisions in states and the introduction of FMLA (Han, Ruhm, and Waldfogel 2009). Meanwhile the introduction of paid leave in California was associated with increased leave-taking among less-educated, unmarried, and non-white mothers (Byker 2016; Rossin-Slater, Ruhm, and Waldfogel 2013).
The current US approach of primarily voluntary private provision of paid leave results in stark disparities in access. Figure 4 shows that, in states without paid leave, 45 percent of Black non-Hispanic workers received no pay while on leave compared to 28 percent of white non-Hispanic workers. While paid leave mandates are not the only difference between paid-leave and nonpaid-leave states, the proportion of Black workers in paid-leave states with no access to pay while on leave falls to 31 percent. In paid-leave states, the proportion of low-wage workers who report that the reason they did not take a needed leave because they could not afford it drops to 50 percent compared to 80 percent in nonpaid-leave states (see figure 5).
Taking a leave from work for one’s own or a family member’s medical needs or to care for a new child may have direct physical and mental health benefits for workers and their families. The undervaluing of these positive externalities could give rise to a market failure to provide adequate paid leave in the absence of a federal mandate (or coordination of paid-leave insurance). Given the disparities in access to the wage replacement that facilitates leave-taking described above, this market failure has important implications for those who are already the most disadvantaged in terms of health outcomes.
Much of the evidence for the health benefits of leave come from studies of parental leave. There is evidence of substantial declines in infant mortality for children born to mothers who have access to leave under FMLA (Rossin 2011). Importantly, however, these health benefits are experienced primarily by the children of college-educated mothers who are more able to afford to take unpaid leave. Given the parallel between the stark disparities in infant mortality by race—the infant mortality rate is currently twice as high for Black babies as for white babies (Matthew, Rodrigue, and Reeves 2016)5—and disparities in paid-leave access by race, this link between parental leave and infant health is of heightened importance.
The link between wage replacement and the ability to leverage these health benefits is reinforced by studies that explore the health impacts of states’ introduction of PFL mandates. Paid maternity leave offered through the early state disabilities insurance programs has been shown to reduce the incidence of low-birth-weight and preterm births, particularly among Black and unmarried mothers (Stearns 2015). Studies of California’s PFL find that women with access to paid parental leave breastfeed longer (Huang and Yang 2015; Pac et al. 2019), their children have higher immunization rates (Choudhury and Polachek 2019), as well as fewer infant hospitalizations (Pihl and Basso 2019). A study of longer-term outcomes found that children born to parents with access to paid parental leave had better behavioral and health outcomes in kindergarten (Lichtman-Sadot and Bell 2017). Access to California’s PFL is also linked to improved mental health for new mothers (Bullinger 2019). Much of the evidence on the impact of child and maternal health relates to mothers’ leave-taking, but new research from Sweden finds that the ability of fathers to take leave soon after birth and when the mother is also home, leads to better mental and physical health outcomes for mothers (Persson and Rossin-Slater 2019).
There are reasons to believe that leave for one’s own medical needs and to care for the medical needs of family member also has health benefits. However, there is very little evidence on the health impacts of medical and caregiving leaves, paid or unpaid (Jacobs 2018). One reason for the relative dearth of evidence on the impacts of leaves taken for own and family medical needs compared to leaves related to births is data constraints – most data sets either record or can be used to recover evidence of a birth, whereas own and (even less frequently family members’) medical events are rarely reported in nationally represented data sets or even administrative data sets.
There is a clear economic prediction that the provision of paid leave should increase leave-taking, and empirical evidence has borne this out (Bartel et al. 2018; Rossin-Slater, Ruhm, and Waldfogel 2013). With this in mind, the economic prediction of the impact of lengthier leaves on post-leave labor supply decisions is ambiguous. On the one hand, the ability to take longer leave might nudge some leave-takers—who would otherwise have exited the labor force—to remain attached to their employer. On the other hand, more leave may encourage some workers to opt out of the labor force, particularly in the context of extended parental leave that may directly alter new parents’ experience of child care and bonding (this theoretical framework is spelled out in Klerman and Leibowitz 1997 and Han, Ruhm, and Waldfogel 2009).
The labor-supply impact of paid leave to care for a new child has been studied most often in the context of California’s paid leave law, which was the first such law in the country. Despite the optimistic predictions of advocates that paid leave is a silver bullet to reduce the gender wage gap, the evidence is inconclusive. This literature is summarized in figure 6 (Bailey et al. 2019; Baum and Ruhm 2016; Campbell, Chyn, and Hastings 2018; Das and Polachek 2015; Rossin-Slater, Ruhm, and Waldfogel 2013). Recent work using administrative data to study the first decade of California’s law finds that paid parental leave has a negative impact on employment and wages for women experiencing first births and otherwise has no detectable impact on women’s (or men’s) employment, wages, or attachment to employers (Bailey et al. 2019). This conclusion is echoed in recent studies of paid parental leave policies in Europe (Kleven et al. 2020). In light of the theoretical ambiguity of potential labor supply impacts of paid parental leave, particularly for women, we consider negligible impacts on women’s work to be good news. While there are reasons to believe that family and own medical leave could impact employment outcomes, there is very little empirical evidence to support that belief (Jacobs 2018).
There are also concerns about potential unintended behavioral impacts of a paid-leave policy. For example, individuals may enter the labor force solely to gain access to leave if they know they will need to make a claim. Employment continuity requirements for benefits eligibility can help to counteract this concern about adverse selection while balancing the desire to maximize access to workers. A related concern is that employees may take leave with no intention of returning to work. In consideration of this concern, FMLA includes provisions for employers to recover health insurance premiums from an employee who does not return to work after a leave. However, the key provision of FMLA is job protection, while most iterations of PFML laws provide only wage replacement. Because workers (and sometimes employers) pay premiums in the form of taxes to fund this wage replacement, paid leave, as discussed above, can be thought of as an insurance benefit that has been paid for (assuming an appropriate tax rate), regardless of whether the worker returns to work at the end of the leave. To the extent that paid leave layers on top of FMLA job protection for eligible workers at covered firms, the (minor) penalties for labor-force exit remain in place.
Although FMLA and state paid leave policies are gender neutral, in practice paid parental and family leaves are taken predominantly by women. Figure 7 corroborates this, showing that, through the 15-year history of California’s PFL, women are much more likely than men to take leave and to take longer leaves to care for new children or for ill family members. Relating to the prior discussion about the potential labor supply impacts of extended leave-taking, this gender disparity in paid leave take-up has the potential to exacerbate rather than alleviate the gender gap. Paid-leave policy can be designed to better encourage participation in parental leave from both parents. For example, Scandinavian countries have designed their paid leave policies to explicitly encourage dual parent participation by allocating some use-it-or-lose-it leave to each parent. As described above, early research has found that these policies lead to improved physical and mental health for mothers during the immediate postpartum period. At the same time, legislating family behavior is neither possible nor desirable.
The primary objection to paid leave stems from the costs it imposes on employers. In some states, including California and Rhode Island, this objection was a stated rationale for imposing all of the tax incidence on employees. In other states, like Massachusetts, New Jersey, New York, and Washington, funding is split between employers and employees.6 Beyond the direct financial cost of wage replacement, there are concerns that firms will face an increased administrative burden from mandated paid leave. Additionally, because paid-leave policies increase leave-taking, firms could face costly disruptions to their business from workers’ absences. On the other hand, there are arguments that access to paid leave may improve morale or help employers retain valuable employees. As discussed above however, there is inconclusive evidence about the impact of paid leave on worker attachment to specific employers.7
Despite dire predictions of how employee leave-taking would impact firms in the wake of FMLA and state-sponsored paid leave, employer surveys consistently find that the overwhelming majority of firms have neutral and even very positive experiences complying with these mandates. Comprehensive evidence of this is available from the 2018 FMLA Survey, which is administered by the US Department of Labor and is nationally representative of worksites. When asked about the administrative burden of complying with FMLA, 70 to 80 percent of respondents found conducting these activities very or somewhat easy and fewer than 10 percent found them very difficult. In terms of the overall effect that complying with FMLA had on the business, 96 percent reported a neutral or very positive experience compared to just 4 percent reporting a somewhat or very negative experience. Figure 8 shows that employers’ perception was slightly more positive in paid-leave states. Regardless of whether the firms are covered by FMLA, the majority report that the most common way that they deal with worker absence is to temporarily assign work to other workers at the worksite (58 percent), with only a small proportion (6 percent) responding that hiring a temporary replacement is the most common way they cover work during absences.
The potential costs and benefits of paid-leave policies may play out differently for small businesses. On the one hand, there is concern that leaves are more disruptive to small businesses who have fewer employees to cover workload during employee absences. Additionally, small businesses are less likely to have human resources and accounting departments to administer leave policies, increasing the cost of implementing such policies. Finally, the private provision of paid leave by small firms is especially costly because they have a smaller pool of employees over which to spread the risk of employment disruption. On the other hand, because paid leave is a benefit valued by employees that potentially boosts morale and productivity at firms, mandated paid leave funded through a tax mechanism would allow small firms to be more competitive with large companies in terms of recruiting and retaining employees.
The existing survey evidence does not show that small businesses experience a disproportionate burden of compliance with FMLA or paid leave compared to larger firms. Figure 8 reveals very little variation in the reported experience of complying with FMLA for worksites with 50 to 99 employees compared to those with 100 or more employees. Because FMLA is applicable only to worksites with 50 or more employees, the FMLA survey is less informative about the experience of very small firms. To this end, Appelbaum and Milkman (2011) conducted a survey in California in 2010, six years after California’s PFL went into effect, that specifically addressed experiences with a paid-leave mandate and included small firms with fewer than 50 workers. Survey questions related to the impact of California’s PFL on productivity, performance, turnover, and morale. Employers reported positive or no noticeable effect 89 to 99 percent of the time and businesses with less than 50 employees were the least likely to report negative impacts. All told, the descriptive evidence does not corroborate the ex-ante concern of overwhelmingly burdensome costs to businesses of any size. It is worth noting, however, that this survey evidence is more than a decade old. As more state-based policies have come online, we look forward to more empirical evidence that speaks to the causal effect of paid leave on small businesses.
Publicly provided PFML programs are typically funded by raising taxes, which are generally inefficient and can cause both horizontal and vertical inequities. When the inequities are horizontal, economists worry about the extent to which behavior is distorted by the tax, and when the inequities are vertical economists compare relative tax burden and access to tax-funded programs among similar individuals (horizontally) and across the income distribution (vertically). To the extent that either the tax burden or access to the tax-funded program is disproportionate, this is labeled as either horizontal or vertical inequity. Next, we outline these concerns in the context of PFML.
In this specific case, state-sponsored PFML programs are typically funded via payroll taxes paid by some combination of employers and employees. The payroll tax, which is a flat tax on wages that phases out for wages beyond the maximum wage threshold, are regressive by nature. The Tax Policy Center, for example, estimates that the bottom fifth of earners will pay 6.9 percent of their income in payroll taxes compared to 5.9 percent for the top fifth, and just 2.3 percent for the top 1 percent of earners (Tax Policy Center 2020). When PFML programs are financed by additional payroll taxes, these tax burdens will likewise fall disproportionately on low-income earners, imposing a vertical inequity. Special attention should be paid to this vertical inequity for any proposed federal PFML program. For example, a policy where benefits are largest for the lowest earners provides some balance to the vertical inequity of the tax itself.
Aside from the regressive nature of the financing source, paid medical or family leave can cause a horizontal inequity both contemporaneously and over the life cycle of the worker. This is because these programs tend to be financed by all workers, but the expected beneficiaries can be highly targeted depending on the permissible leave. For example, consider a program that provides only maternity and paternity benefits. In this case, those working adults that do not have and do not intend to have children are simply subsidizing employment disruption insurance for those that will eventually become parents. This underscores the importance of designing a program that has some expected benefit for all workers, which supports the inclusion of medical and family leave as part of any paid-leave program.
Finally, the wage replacement schedule itself can induce inefficiencies in labor force participation, particularly if it is designed to have kinks and notches. For example, consider a wage-replacement schedule that decreases from 100 percent to 50 percent at a discrete earnings level. Economic theory predicts that individuals will respond to notches and kinks in tax and benefits schedules by inefficiently altering their behavior. Consider a worker who faces a notch in the wage-replacement schedule, where the next dollar of income earned discretely reduces paid-leave benefits. This worker is incentivized to reduce hours worked in order to avoid the benefits reduction. These stair-stepped schedules are common in federal policy design—for example, the Earned Income Tax Credit (EITC) schedule, the Premium Tax Credit, and even the individual tax schedule—and empirical evidence of this behavior abounds.8 In light of this empirical finding, policymakers might consider a smooth linear benefits schedule to avoid such distortions.
We propose the Paid Parental and (Family and Own) Medical Leave (PPML) program that would benefit wage employees and would be financed by a payroll tax paid by both employers and employees. The program would be administered by the Social Security Administration, leveraging the infrastructure already in place for supporting the Social Security Disability Insurance (SSDI) program. The program can be designed to be revenue neutral, with the payroll tax rate set to offset claims in expectation, in addition to administrative costs. All claims would be paid out of a newly established federal Paid Parental and Medical Leave Trust Fund.
Table 3 describes key features of this proposal. The program would provide annual, partially paid leave shared across two categories: parental leave and medical leave. Parental leave is tied to arrival of a new child and provides partial wage replacement to working parents for work interruptions due to care of a new child. Medical leave provides partial wage replacement for work interruptions due to the employee’s own serious illness or care of a family member with a serious illness. In what follows, we provide details on the design, administration, implementation, potential revenue costs, and additional policy design considerations.
There are several core features in the design of the PPML program: (1) the amount of leave provided within each category of leave and annual cap in leave taken across both categories, (2) wage replacement rates, (3) eligibility requirements, (4) waiting period, (5) choice of payroll tax rate, and (6) payroll tax base. In what follows, we discuss the analysis that informed the specific policy parameters outlined in table 3.
Amount of Leave
The amount of leave provided for medical and parental leave must balance the needs of each exigent cause, the expected cost of extended work disruptions to employers, and the behavioral consequences of providing leave. Because the two leave categories are distinct, we describe each in turn.
We propose that up to 16 weeks of partially paid parental leave be allocated for each birth or adoption. Allocating the total parental leave amount to a birth or adoption rather than to the individual workers is a departure from leave policies in the United States. The goal of this design is to encourage more gender equitable use of paid leave.
Parental leave weeks are shared by both working parents according to a set formula. Ten of these weeks may be taken by either parent, and each parent is entitled to two additional weeks of leave. Parents can choose to take any of these leave weeks simultaneously or sequentially. In the case of two working parents, one bonus week is additionally available to both parents if it is taken simultaneously within two weeks of birth. Given this design there are, at most, 13 total weeks of leave available to any one worker. Parental leave must be used within one year of birth or arrival in the home.
While more complicated than existing structures, the proposed parental leave program is designed to be gender-neutral with regard to which parent provides early-life child care and to encourage participation by both parents by allocating some use-it-or-lose-it leave that must be taken by each parent. The bonus week of parental leave that is available to both parents and that must be used simultaneously and immediately postpartum encourages both parents to be home for a short period during a critical care period. This type of bonus week has been shown to improve postpartum maternal health by decreasing the likelihood of unexpected medical events and increasing mental health outcomes (Persson and Rossin-Slater 2019).
Some care must be taken in implementing this proposal so as to accommodate and address inequalities across households of different structures. These households include same-sex households and single-parent households where one parent is unavailable or unwilling to participate in child care but a domestic partner or spouse is willing and able to provide child care. Special consideration should also be given to situations where participation of a parent is unsafe for the other parent and child, and where encouraging involvement is undesirable.
In figure 9 we depict several examples of how parental leave might hypothetically be allocated across two parents for a given child. As previously described, there are 16 total weeks available to both parents for a given child: 10 weeks available to either parent (blue boxes), two weeks available to each parent (orange and red boxes), and two bonus weeks that must be taken simultaneously by both parents within two weeks after the child’s arrival (green). Example 1 demonstrates how two working parents might fully use all leave available in sequence, one after the other. Example 2 demonstrates how leave could be allocated if only one parent takes leave or if a household has a single working parent—in this case the total leave available per child is 12 weeks. Example 3 demonstrates how two working parents could take leave fully simultaneously—in this case, each parent would claim eight weeks of leave. Finally, example 4 demonstrates how both working parents might fully utilize all available leave in an interspersed manner. Each of these examples assumes that parental leave is fully extinguished in the weeks immediately after the child’s arrival. However, parental leave weeks are available for up to one year after the child’s arrival and can be used in a manner similar to these examples spread over a longer period of time.
We propose that up to 12 weeks of partially paid medical leave be available annually to eligible wage employees for self-care due to a serious illness, and six weeks of partially paid medical leave be available to care for a family member with a serious illness. Differences in available leave for own-medical and family medical care reflect the US experience with family care under the current FMLA unpaid leave system, wherein family care is the least likely reason cited for short-term employment disruptions, and these leave lengths are typically short.
For the purposes of eligible medical leave, we rely on the FMLA definition of serious illness, both for own-medical care and medical care of a family member. This medical leave category also covers employment disruptions due to pregnancy. These disruptions can be driven by medical complications, including incapacity due to morning sickness and medically required bed rest. Medical leave can also be used for prenatal birth preparation and postnatal recovery. In this case, we draw from the design of state-sponsored STDI programs in permitting up to two weeks of medical leave for prenatal birth preparation and up to four weeks of medical leave for postnatal recovery in the case of a normal pregnancy. Deviations from these prescribed pre- and postnatal leave amounts for a mother’s pregnancy complications would be at the discretion of a medical provider and the administrator of the program.
Finally, we propose a 16-week cap in total annual leave to be taken by any one claimant. This cap balances the possibility that a claimant may experience an adverse health outcome in the same year as a birth event with the cost of increasingly long leave lengths for employers. This situation is particularly likely for pregnant mothers, who may experience prenatal complications in the same calendar year as a birth. A pregnant mother would be entitled to wage replacement for up to 12 weeks through the medical leave benefit and could additionally receive benefits for up to 13 weeks of parental leave, subject to the 16-week total annual leave cap.
The combination of medical and parental leave creates an employment interruption insurance program similar to SSDI. Importantly, the program provides an expected benefit to all wage employees, whenever the program is enacted. We believe this element is important because the program is to be financed by a payroll tax levied against all wage employees. A counterfactual PFL program, for example, that is limited to wage replacement for family care and parental leave but that excludes own medical care imposes horizontal inequity across parents and childless adults over the life cycle in addition to intertemporal horizontal inequity for elderly workers at the time of enactment.
Wage Replacement Rates
Wage replacement rates for medical and parental leave are determined based on average weekly earnings during a base period. The base period is defined by the three most-recent quarters of employment. The quarter with the highest earnings in the base period is used to calculate the weekly benefit, where the total quarterly earnings are divided by 13 (the number of weeks in a quarter) to determine the average weekly earnings.
We propose a wage-replacement rate that is a nonlinear function of average weekly earnings, with low-income earners subject to a higher wage-replacement rate than higher-income earners. Specifically, those with average weekly earnings of up to 200 percent of the weekly single-family federal poverty level (FPL) are eligible for an 80 percent wage replacement rate.9 In 2020 these workers would be eligible for $393 in weekly benefits for up to 12 weeks of own medical leave, 6 weeks of family medical leave, or up to 13 weeks of parental leave. The wage-replacement rate will decrease linearly from 80 percent to 50 percent for average weekly earnings between 200 percent of FPL and the Social Security Payroll Tax Cap.10 In 2020 workers subject to the benefits cap would be eligible for a maximum of $1,324 in weekly benefits for up to 12 weeks of own medical leave, 6 weeks of family medical leave, or up to 13 weeks of parental leave. We expect that these benefits will be subject to federal income tax, just as unemployment benefits are federally taxable. Importantly, the maximum and minimum wage-replacement rates will be automatically adjusted for inflation because they are tied to the FPL and Social Security Payroll Tax Cap, each of which is independently adjusted annually for inflation.
The linear benefit schedule that we propose is a departure from the piecewise replacement schedule in place in US states with paid-leave programs. While a piecewise replacement schedule is certainly easier to administer, it comes at an efficiency cost. In particular, there is a growing body of empirical evidence that individuals respond to piecewise benefits schedules by adjusting their behavior in a suboptimal way. For example, individual earners reduce employment in order to avoid the sharply decreasing wage subsidy schedule for the EITC (Mortenson and Whitten 2020). Similarly, we should expect that individuals that can predict a leave-taking event will reduce employment during the base period if they expect that increases in employment will sharply reduce their wage-replacement benefits. A linear wage-replacement schedule avoids this inefficiency.
Eligibility depends on labor force attachment and continuity in the period immediately before a claim is made. We propose that eligibility be determined based on a qualifying period, defined as the 52-week period immediately before the start date of the claim. During this qualifying period, individuals must have earned wages that are subject to program payroll taxes for at least 39 weeks (three quarters). This defines a minimum employment continuity. In addition, employees must be attached to the labor force for the five weeks immediately preceding the start of a claim. We follow the California Employment Development Department in defining labor market attachment as an employee who is employed, looking for or registered for work, or who has an active unemployment insurance claim in payment status within 90 calendar days from their last day of work.
Our eligibility requirements represent a departure from FMLA work history requirements, which include an element of specific employer attachment. In our PPML individuals can qualify for paid parental and medical leave provided they meet continuity and attachment requirements, which do not depend on the length of attachment to a particular employer. In this way, the proposed eligibility requirements are designed to be flexible with regard to spells of unemployment or labor force detachment that can be common among lower-income workers and in certain industries.
We propose a medical-only waiting period of one week in order to limit the use of paid medical leave for those conditions that require very short-term leave from work. A one-week waiting period is consistent with most state paid-leave policies. We do not impose a waiting period for parental leave claims.
Payroll Tax Rate
We propose that the PPML program be financed by payroll tax, to be shared equally by employers and employees. The tax would be imposed and administered in tandem with the Social Security payroll tax and would therefore phase out for earnings above the Social Security earnings cap. The choice of payroll tax can be designed to make the PPML program revenue neutral. In particular, the revenue raised by the paid leave tax could finance all expected claims in addition to covering expected program administrative costs. The revenue-neutral tax rate is explored further below.
By adopting a payroll tax as the financing mechanism, we have linked the program to wage employees. This introduces participation complexities for self-employed individuals, who are subject to different payroll taxes because an employer is not responsible for any portion of the tax. Some states have adopted mechanisms for including self-employed individuals in paid family and medical leave programs, just as the federal government has expanded some social safety net programs to self-employed individuals. While our proposal does not include a provision for self-employed individuals to be covered by paid parental and medical leave, we discuss this issue further in the section discussing further policy considerations.
Payroll Tax Base
We propose that the PPML payroll tax rate be imposed on the Social Security wage base for the Old-Age, Survivors, and Disability Insurance (OASDI) payroll tax, with one exception: we do not provide for an opt-out mechanism for certain employees of state and local government, such as police officers and public school teachers. These employees can be exempt from the OASDI payroll tax if they are beneficiaries of public pension programs because these public pensions provide retirement benefits in place of Social Security benefits. There is no such equivalent program that provides wage replacement for employment disruptions due to medical or parental leave for these employees. As a result, the PPML wage base is slightly more expansive than the Social Security wage base because we do not allow any employees to opt out. As with the Social Security wage base, the PPML payroll tax will phase out immediately for wages above the PPML wage cap, which is exactly coordinated with the OASDI wage cap.
We propose that the US Treasury Department collect the revenues from the PPML payroll tax and transfer these revenues to a newly established PPML Trust Fund. The PPML Trust Fund should be administered by a new program office established within the Social Security Administration. Benefits claims should be made through the Office of Paid Parental and Medical Leave, and benefits should be paid out from the PPML Trust Fund. By coordinating administration of PPML claims with the Social Security Administration, we hope to leverage the infrastructure that is already in place for receiving, evaluating, and paying SSDI benefits. At present, determination of eligibility for FMLA leave is left to the employer, possibly resulting in heterogeneous access for the same underlying claims across employers. A single federal administrator will harmonize eligibility rules and access.
We propose a phase-in of our proposal and a regular program review to ensure the solvency of the PPML Trust Fund. As earlier outlined, we suggest that the payroll tax be implemented one year before claims are permitted to be made in order to forward-fund the PPML Trust Fund. This will naturally roll in the qualification period of a 52-week lookback period with at least 39 weeks worked with eligible wages. This phase-in is consistent with the phase-in procedure implemented by Washington State, Massachusetts, and Washington, DC, where each state required premium contributions—paid by the payroll tax—for one to two years before benefits could be paid.
We propose that the Social Security Administration’s Office of the Chief Actuary provide a detailed annual review of the solvency of the PPMFL Trust Fund to the Treasury department, who administers the tax, and to Congress. These regular reports will allow Congress to make adjustments to the payroll tax rate should there be sufficient evidence that the Trust Fund will become insolvent. These adjustments are to be expected, since take-up of program benefits is likely to increase over time with increased awareness and experience.
The revenue costs of the PPML program are entirely at the discretion of policymakers through the choice of the payroll tax rate. In concept, the payroll tax rate can be set so that the PPML program has a net cost, provides net revenue to the federal government, or is roughly revenue neutral. While there are efficiency arguments in favor of all three outcomes, we propose that the PPML program be financed through a revenue-neutral payroll tax that covers both the expected benefits claims and the administration of the program.
To illustrate a rough approximation of a revenue-neutral payroll tax rate, we simulate the total PPML benefits that would have been claimed between 2012 and 2019, had such a program existed. Then we calculate the payroll tax that would have been required to exactly cover these benefits. This simulation abstracts away from administrative costs for the sake of simplicity. In addition, this simulation assumes that the PPML program is already in a long-run steady-state in terms of take-up, abstracting away from the transition to steady-state. This simulation requires assumptions about the eligible wage base, the expected take-up of benefits, the expected benefit level, and the expected duration of the typical claim for both medical and parental leave. We estimate the eligible wage base by adjusting historical OASDI wages to account for the expansion to include all public employees that are otherwise exempt from the OASDI tax.
We estimate take-up rates and average leave lengths from two data sources. For paid parental leave, we draw from state-specific reporting on take-up and duration of parental leave. Specifically, we extrapolate the 2018 experience in California to the full population by combining parental leave take-up estimates with annual natality counts from the National Vital Statistics System, relying more heavily on information that comes from recently reported years (Bedard and Rossin-Slater 2016; California Employment Development Department 2019). Because California’s PFL program has been in place for more than 15 years, these parameters are less likely to reflect transitional changes in take-up and roll out. Because state programs do not separately report leave length information for parental, medical, and family care, we draw from the 2018 FMLA Survey to estimate use of the medical leave benefits under the PPML program. We increase our take-up assumptions relative to the FMLA reporting under the assumption that PPML benefits will be more widely used than FMLA.
We estimate average claims for both parental and medical leave based on annual earnings and employment data reported in the Current Population Survey (CPS). These calculations require assumptions about expected weekly benefits, which are based on weekly earnings reported in the CPS monthly outgoing rotation group earner study for private and state/local government employees (excluding the self-employed, members of the military, and federal workers). We exploit the panel data structure of the CPS to estimate average earnings in the year before the arrival of a new child for the purposes of estimating average parental leave benefits. Finally, we make assumptions about the share of the labor force that would be eligible for PPML benefits based on the total number of weeks worked in the previous year, as reported in the Annual Social and Economic Supplement of the CPS. All calculations are gender-specific where possible.
The results of this simulation are shown in figure 10, which depicts annual benefits amounts in millions of dollars for each of the three PPML leave categories. Annual total benefits paid would have ranged between $50 billion and $65 billion during this period—roughly the same order of magnitude as the annual tax expenditure on the EITC (Office of Management and Budget 2011). Roughly 60 percent of these claims are related to own-medical leave, 30 percent are related to parental leave, and the remaining 10 percent of claims are driven by family medical leave.
The PPML program is financed by a payroll tax applied to the same wage base as the OASDI program. We calculate the total annual PPML payroll tax that would have been required to pay simulated benefits. This is shown in figure 11. This required revenue-neutral payroll tax rate ranges between 1.04 percent and 1.14 percent. If this payroll tax is shared by employers and employees, as the OASDI payroll tax is shared, then each would be subject to a 0.52–0.57 percent payroll tax.
This estimate excludes any administrative costs of administering the program. However, the payroll tax could be adjusted to recover these costs, which are assumed to be de minimus compared to total expected benefits. Moreover, these simulations reflect long-run assumptions about the expected take-up of benefits. In other words, we do not incorporate lower initial take-up followed by expected increases in program take-up that will likely occur during the first few years of implementation. We expect that there will be a considerable transition to full-equilibrium take-up. The roll-out required for this program includes educating employers and employees about benefits and an adjustment period while privately sponsored and state-sponsored leave programs adjusted to wrap around the federal benefits.
Interaction with Private and State-Sponsored Paid Leave
The creation of a federal PFML program will undoubtedly induce behavioral adjustments on the part of employers that already offer employer-sponsored paid leave and on the part of states with state-sponsored paid medical and family leave. Much like federal minimum wage laws, we imagine that the proposed PPML program will set a minimum standard for paid leave. This leaves room for employers and states to craft policies that expand upon the benefits provided by the PPML program. There are clues about how employers might respond to the PPML policy based on their behavior in states with paid-leave programs. For example, in California some employers and even some local governments have adjusted by “topping off” wage replacement rates to be more generous than what is offered by the state. Ultimately, it will be up to individual employers and state governments to determine this exact mechanism. In states with paid leave, rules exist to ensure that any additional employer-sponsored wage replacement does not push any single employee to receive more than 100 percent wage replacement. These rules are reasonable and should be incorporated into a federal PPML program.
An alternative policy design could follow the federal-state unemployment insurance system, which is run by the states and is overseen by the US Department of Labor. Following this model, the federal government could specify the basic design elements (minimum wage replacement rates, minimum leave durations, and minimum needs categories) and provide oversight but otherwise allow states to specify the full policy parameters according to their own local preferences and needs. However, there are important weaknesses of this design. The first was highlighted during the ongoing global health pandemic and ensuing macroeconomic crisis. In this case, state UI funds experienced an unexpected, large negative shock. While the federal government has some ability to provide short-term loans to these funds, and—as it has during the global pandemic—can provide temporary additional funding, this response can be inadequate in maintaining fund solvency, especially because all but two states are subject to constitutional requirements for a balanced budget. Second, there is important demographic heterogeneity across states that imposes design and funding consequences for state-based paid parental and medical leave insurance. For example, states with a relatively large aging or elderly population should expect a greater need for family care compared to states with a younger population and higher fertility rates. A federal policy can smooth the effects of this heterogeneity across the US population, providing the same benefits more efficiently.
As specified, the PPML program benefits are limited to wage employees by virtue of the funding mechanism. However, these benefits could be extended to self-employed persons by augmenting the Self-Employment Contributions Act tax. Most states with PFL programs have provided an opt-in option for self-employed individuals to participate in disability and paid family leave insurance programs. These opt-in clauses typically require a waiting period of two years before benefits claims can be made, and benefits are based on weekly earnings, comparable to weekly wages for wage employees. Giving the growing importance of alternative work arrangements for the US labor force, this expansion is worth serious consideration (Jackson, Looney, and Ramnath 2017).
One of the most-often-cited arguments against a federal paid medical and family paid-leave program is that such a program would be disproportionately costly for small businesses. For example, small businesses typically have fewer on-site employees and are therefore more likely to experience negative cost shocks when one or more employees takes a lengthy leave of absence. These cost shocks can come from the cost of temporarily hiring a worker to maintain productivity or negative productivity shocks induced by temporarily assigning more work to on-site employees, among other causes.
Although we have shown that empirical evidence does not corroborate these concerns, there are additional federal policy tools available to help alleviate some of the anticipated additional costs borne by small businesses. For example, the federal government could offer a general business credit to small businesses to subsidize the cost of temporarily hiring workers. General business credits are an often-used federal tool to provide targeted tax subsidies to specific activities, industries, or firm sizes. In fact, the use of tax credits to subsidize the cost of paid leave has become more popular in recent years. For example, the 2017 Tax Cuts and Jobs Act introduced a temporary general business credit to subsidize the cost to employers of providing PFML. Another approach, modeled after Washington State’s paid-leave program, would be to exempt small businesses from all or a portion of the payroll tax. Finally, the Families First Coronavirus Response Act of 2020 temporarily provided PFML to employees of small- and medium-sized businesses through a refundable payroll tax credit.
The FMLA, along with several state-sponsored PFML programs, covers leaves of absence related to extra family care induced by military deployment. The proposed PPML program could be expanded to include this leave category by adjusting the payroll tax to cover expected benefits claimed for this purpose. We expect that a very small adjustment to the payroll would be required – in 2018 less than 0.1 percent of employees took leave for a family military deployment. At the same time, it is not clear whether employment interruption insurance for this category of leave should be provided by the military in particular, the federal government more broadly, or some combination of the two. It is for this reason that we do not explicitly include this leave in the proposed PPML program.
The Congressional Budget Office (CBO) estimates that the FAMILY Act will increase direct spending by $547 billion over the 2020–30 budget window, $521 billion of which will come from paid benefits and $27 billion of which will come from administrative costs. The Joint Committee on Taxation estimates that the proposed 0.4 percent payroll tax will raise $319 billion over the window, covering only 65 percent of the program costs. In other words, the FAMILY Act is not revenue neutral. By comparison, we estimate that the PPML program would have resulted in $50 billion to $60 billion in claims annually had the program been in place from 2012 to 2019, and a 1 percent to 1.1 percent payroll tax would have been sufficient to ensure revenue neutrality with regard to paid benefits.
A direct comparison of these two revenue exercises is complicated by the fact that ours is backwards looking whereas the CBO score is a forward-looking projection. Moreover, our estimate abstracts away from transition costs by assuming that the PPML program is in a long-run steady state throughout the period, whereas the CBO score accounts for a lengthy transition period at the start of the budget window. To better facilitate a comparison, we highlight some major differences in program design across the two proposals, and we compare the 2030 CBO projection, which is likely closer to steady-state, to a 2030 projection of our 2019 projection.
Overall, the design of the PPML provides similar benefits to a larger share of workers than does the FAMILY Act. Specifically, the PPML program offers average wage replacement rate of 71 percent compared with 66 percent under the FAMILY Act, similar weekly maximum benefits ($1,324 vs. $1,000), and less-restrictive eligibility requirements (39 of the past 52 weeks vs. work history requirements as established under the SSDI program). With this in mind, the CBO estimates that there will be $84.3 billion in benefits payments in 2030 under the FAMILY Act and that the 0.4 percent payroll tax will generate $42.3 billion in tax revenues, covering just 50 percent of the expected claims. While the assumptions across the two exercises are certainly different, a projection of our 2019 estimate to 2030 based on the CBO’s projection of the Consumer Price Index for All Urban Consumers (CPI-U) suggests that PPML benefits paid would be $71.3 billion in the same period, and the program overall would be completely revenue neutral with a 0.98 percent payroll tax. In light of the considerable uncertainty underlying these projections, we find the two estimates to be largely consistent with each other.
No, the current proposal does not include job protection, although nonretaliation protections for workers who take leave should be part of the legislation. We intend for eligible workers at covered firms to be covered by the job protection provisions of the FMLA under current law. Because job protection relates to retaining a position at a particular firm, it makes sense that continuity of employment with a specific employer is stipulated in the FMLA. Firm size exemptions in FMLA respond to concerns that small firms’ business may be endangered by worker absences.
Employment disruptions taken under PPML are used simultaneously with FMLA leave. This is similar to layered use of paid leave and FMLA in the states with paid-leave mandates, which (mostly) do not provide job protection.
No, states and large employers cannot opt-out of the federal PPML program. We view the PPML program as defining a minimum set of benefits that are provided by the federal government. We anticipate that some states or employers may wish to offer “wrap” insurance plans that expand the benefits of the PPML program either through higher wage replacement rates (capped at 100 percent), or higher weekly benefits caps. In addition, these policies may wish to provide additional wage replacement for a longer duration than what is available through PPML. However, opt-out provisions of public insurance programs can induce administrative complexity for workers, adverse selection among the pool of beneficiaries covered by the PPML program, and a suboptimal separating equilibrium among those employees covered by the PPML versus state- or employer-sponsored programs. We believe that a less costly approach would be the development of wrap plans that provide additional benefits.
There is a well-developed, long-standing market that provides a private short- and long-term medical disability insurance product to those employees that choose to insure against these kinds of employment disruptions. As with state- and employer-sponsored plans, we imagine that this market will adjust its product to provide wrap plans that expand the benefits of PPML to those workers that choose to purchase additional employment disruption insurance.
There is almost universal agreement in the United States that workers should have access to wage replacement during a work disruption due to their own or a family member’s serious illness or the arrival of a child. In the current patchwork of limited private provision and a handful of state mandates, however, access to this type of insurance against lost wages is uneven and inequitable. Given the positive externalities to health and well-being of families that have been extensively documented, particularly in the case of paid parental leave, there is a strong argument for government coordination to correct for the inefficient underprovision of paid leave. We propose a federal paid parental and medical leave program to be administered by the Social Security Administration and financed by payroll taxes paid equally by employees and their employers. Under our proposal, all workers and employers would pay into the program and based on the range of leave-eligible events, workers at all stages of the life cycle would have some expected potential benefit.
In The Challenge section, we highlight disproportionately low access to paid leave among single parents, Black workers, and low-wage workers. Under our proposal all workers who have worked at least 39 weeks in the past year would be eligible for paid leave—the proposal does not impose restrictions based on hours worked, earnings, or firm size. Based on these eligibility criteria, we use work history information in the Annual Social and Economic Supplement of the Current Population Survey to estimate eligibility for the workers who currently have the least access to paid leave. Figure 12 shows that 88 percent of Black workers, 87 percent of single parents, and 83 percent of low-wage workers (those earning $15 or less per hour) would have access to paid leave under the proposal. While these rates of access are still lower than for white workers, dual-parent households, and high-wage workers, our proposal narrows this gap dramatically.
Given that benefits are based on average earnings for each worker, we find it unnecessary to impose any hours or wage restrictions. This means that 75 percent of part-time workers, who often lack access to these and other benefits, would also be eligible. It is important to note that adding exemptions for firms based on numbers of employees (even for firms with fewer than 25 employees) would considerably reduce access, particularly for Hispanic workers, according to our estimates as shown in figure 12. For this reason, we would resist carve-outs based on firm size.
There is evidence that, in paid-leave states, those not eligible for FMLA do not take the paid leave they are entitled to due to the lack of job protection, and that this is particularly true among non-white workers (Goodman, Elser, and Dow 2020). While not specifically in the scope of this proposal, a relaxation of FMLA attachment and firm size exemptions would further improve the socioeconomic disparities in access to paid leave. Furthermore, since both FMLA and the proposed PPML policy are federal mandates, we could imagine both of these being combined into a single policy, resulting in gains in administrative efficiency. A single piece of legislation covering leave could still have different eligibility requirements for wage replacement and job protection. The advantage of folding these two policies together would be administration simplicity, where one agency serves as a single source for information about rules for employers and employees. Given the reported confusion among both constituencies about both FMLA and state paid-leave policies, improvements in access and use of a single comprehensive leave policy could be substantial.
Acknowledgments: We would like to thank Kristen Broady, Wendy Edelberg, Vicki Shabo, Adam Looney, Peter Matthews, and participants in the Hamilton Project’s authors’ conference for their thoughtful comments and suggestions. We also thank Amy Cyr for exceptional research assistance.