RD288 - Virginia Paid Family and Medical Leave Study – November 2021

Executive Summary:

This report presents results of an actuarial and policy analysis for a prospective Paid Family and Medical Leave (PFML) program for the Commonwealth of Virginia. Paid Family and Medical Leave provides temporary replacement income for workers with a serious health condition, who need to care for an ill family member, or who are welcoming a new child. Most developed countries have PFML programs and by 2021 nine U.S. states and the District of Columbia have enacted PFML. PFML has also been the feature of previous and current federal legislation such as the federal “Build Back Better" legislative package, which provides four weeks of paid leave. Several PFML program bills have also been introduced by the Virginia General Assembly in recent years, including HB2016 and SB1330, which would create a public PFML program of 12 weeks covering 80 percent replacement of wages up to 80 percent of the state average weekly wage.

Virginia workers are currently covered by a patchwork of federal programs and firm-based leave programs The federal Family and Medical Leave Act (FMLA) program has covered eligible workers up to 12 weeks of job-protected unpaid family and medical leave since 1994. However, eligibility conditions limit the protection to approximately 56 percent of the workforce, a percentage that has not improved in at least the last decade. Although Virginia-specific data on FMLA and private coverage is not available, private employers have increasingly covered short-term disability and paid family leave benefits to their workers. According to the Bureau of Labor Statistics National Compensation Survey, private employers nationwide covering short-term disability access rose from 37 percent in 2011 to 41 percent in 2021 and paid family leave access from 11 to 23 percent over the same period. However, many workers are less likely to be covered, particularly part-time, lower-wage, and small business employees. Increases in female labor force participation, the growth of single-parent families, population aging and some research suggesting beneficial economic, social, and health effects for participants and their families are reasons that the issue has received more attention from from policymakers.

This report examines the effects of a Virginia PFML program, with the focus being HB2016/SB1330 legislation introduced during the 2021 General Assembly Session. It examines the features of the Virginia legislation in comparison to other states that have adopted PFML and the potential effects of varying program design elements. It also reports on a professional actuarial analysis by Milliman, Inc. that projected costs needed for benefit payments and the direct and indirect costs of the operation and administration as well as to maintain a sufficient cash balance to ensure program solvency over the 2022 to 2033 period. The potential short-run and long-run economic, social, and demographic effects on Virginia residents are examined through the prism of recent scholarly research on U.S. state programs. Lastly, the study looks at the economic impacts of Virginia PFML legislation, considering expenditures, taxes and possible secondary economic and demographic effects, using a commercial economic impact model.

A variety of policy design features and parameters can affect the cost, utilization, distributional effects, and health, social, and economic impacts of PFML programs. They include the manner of funding, eligibility requirements, benefit structure, administration, and other characteristics. The table displayed on page 2 provides a summary of major features of the proposed PFML program.

Virginia’s proposed PFML program provides somewhat more generous eligibility requirements and benefits than older state PFML programs, some of which have short-term disability components that date back to the immediate post-war period. But it is fairly representative of PFML programs adopted in more recent years. The features of this proposed program in comparison to other states and their consequences are summarized below:

Funding Method. The Virginia program specifies a 50-50 split between employer and employee in payroll taxes. Payroll taxes to fund PFML are nearly universal among states. Employees pay the full tax in three states and employers do so in the District of Columbia. Most states cover employee/employer splits in the range of 40-60 (Massachusetts), 45-55 (Washington State), 50-50 (Colorado), and 60-40 (Oregon). New Jersey and New York allotments vary by type of leave; workers pay payroll taxes for Paid Family Leave (PFL) while employees and employers split the cost of short-term disability. Split allotments seem to be motivated by a combination of factors such as pragmatic political considerations, social equity, or the benefit principle. Regardless of the motivation, substantial empirical research suggest that the actual incidence of payroll taxes is roughly evenly split between employer (in the form of reduced profits) and employees (in the form of reduced earnings) in the short-run, while workers pay most of the tax in the long-run.

Maximum Taxable Wages. Virginia specifies a maximum taxable earning limit used for the Social Security payroll tax ($142,800 in 2021). Five states use this cutoff. Several other states have established lower taxable wage ceilings, with the lowest being $74,000. The District of Columbia is the only jurisdiction without a taxable wage ceiling. Lower taxable wage ceilings will narrow the tax base and increase the payroll tax rate for workers with wages below the ceiling.

Eligibility Requirements. State programs require evidence of some minimal level of labor force attachment to be eligible for PFML benefits. This ensures that workers have adequately paid into the system and that it does not become a general purpose entitlement program. This is typically measured by wages earned over some base period, usually four or five quarters immediately prior to taking leave. Tying PFML to Unemployment Insurance (UI) program eligibility as occurs with the Virginia legislation may help standardize eligibility and simplify administration. However, most states do not use the same eligibility standards as UI, perhaps because that would restrict eligibility more than desired.

Qualifying Family Members and Events. Family definition and qualifying events are important primarily in determining eligibility for caretaking leave. Virginia like most U.S. states specify that qualifying family members are immediate family, including spouses and common law partners; birth, adopted, and foster children; mothers and fathers; siblings; parent-in-laws; and grandparents and grandchildren. A few states expand qualifying members further to include brothers and sisters-in-law, spouses and domestic partners of siblings, any other person related to the worker by blood and individuals with close association equivalent to family relationships. Expansion of the family definition and qualifying family members will likely have only a very small impact on leave utilization since it affects only family caretaking leave, the smallest component of PFML. Moreover, the vast majority of caretaking leaves are for immediate family members.

Wage Replacement Rate. The wage replacement rate for Virginia is 80 percent. Most states provide replacement rates in the range of 60-90 percent. Higher replacement rates will increase the costs of PFML programs. Empirical research suggests that as the replacement rate increases, program utilization increases. Rates are generally less than 100 percent to minimize moral hazard, to reflect the fact that living expenses such as commuting will be lower when workers are on leave, and to allow private employers room to “top off" benefits as desired. Research suggests that lower replacement rates result in significantly disparities in program usage, with lower wage earners much less likely to utilize benefits for which they may be eligible because they may find it more challenging to live o the benefit. Thus, states with long-standing PFML programs such as California have increased their replacement rates in recent years. More recent adopters have tended to cover higher replacement and more progressive rates, replacing a larger share of wages for lower earners than many early adopters.

Maximum Leave. Virginia’s legislation affords up to 12 weeks of PFML during a 12-month period, which is the same allowance covered by federal unpaid leave through the Family and Medical Leave Act (FMLA). Expanding the available leave time also increases program costs by contributing to longer average leave durations. Total allowed leave varies from a low of 8 weeks to 52 weeks, with an average of 21 weeks. This is due largely to the fact that states with older short-term disability programs adopted maximum leave durations more typical of private plans, which average 26 weeks. Newer PFML programs have copied the FMLA model of covering 12 weeks of total annual leave regardless of leave type. Considerations in the development of maximum leave time are the policy objectives of promoting worker and family health and wellness and child development while assisting workers to transition from leave back to work. Leave needs to be long enough to facilitate recovery and bonding but not so long that it contributes to worker loss of human capital. International research suggests that negative labor market effects onset at much higher durations (one year duration or more) than allowed by state programs.

Minimum and Maximum Benefits. Virginia’s program benefits are capped at 80 percent of state average weekly wages (based on 2020 average weekly wages of $1,253 in Virginia, the maximum would be $1,002) while the minimum benefit is $100. For a small minority of part-time workers, this minimum benefit will slightly exceed 80 percent of wages earned. State programs set benefit caps, usually stated as a certain percentage of statewide average weekly wage (ranging from a low of 64 percent to a high of 120 percent), a fixed amount with and without annual cost of living adjustments, or a multiple of the minimum wage. Several state programs also specify minimum benefit amounts in constant dollar terms ranging from $20 to $50 or percentage of statewide weekly wages. Imposing benefit ceilings and floors decreases disparities in program benefit allowances and provides another mechanism to improve participation of extremely low earners.

Job Protection. Virginia’s PFML program provides job protection for the full duration of 12 weeks for eligible workers. Although the federal FMLA program provides similar job protection, eligibility is much more restrictive, being limited to workers who accumulated at least 1,250 work hours of work over the previous year for a business that employs at least 50 workers within a 75 mile radius. Forty-four percent of the labor force, disproportionately lower earning, minority, part-time and small-business workers are not eligible for such protection according to 2018 survey data. Most states have extended job protection beyond FMLA, but these additional protections are sometimes restricted to particular categories of leave (e.g., parental bonding leave, maternity leave) or limited to workers who have demonstrated some minimum level of job continuity with a single employer. Only Massachusetts appears to cover job protection as broad as Virginia’s proposed legislation. Some studies indicate that such job protection is an important influence on worker decisions to take leave.

Exemptions for Businesses. Virginia’s PFML program does not cover special exemptions for certain categories of firms such as small businesses or firms covering competitive plans. However, it does allow self-employed individuals to opt in. The Chief Workforce Development Advisor and Secretary of Commerce and Trade Paid Family and Medical Leave Study does recommend that small businesses be permitted a exemption from paying their portion of a PFML payroll tax. This feature can be found in a handful of other state programs such as Colorado, Massachusetts, Oregon, and Washington State. These exemptions are sometimes covered because of concerns that smaller businesses face high costs or realize fewer benefits from implementing PFML. Moreover, most states allow firms to cover regulated private plans that provide similar or better benefits or coverage. However, various counterarguments against providing these exemptions are covered, including that permitting exceptions could create adverse selection problems, administrative complexity, or other problems.

Virginia legislation and other state PFML laws are silent on a few major policy issues that could affect workers and businesses. First, the Virginia legislation does not address the state taxability of benefits. Thus, one must assume that PFML benefits would be taxable since the state conforms to the federal definition of gross income, unless a specific exception is included in a future Virginia law. Some policy analysts recommends that PFML legislation explicitly state whether benefits are taxable income, which would also affect individuals options for having state income tax withheld along with any federal tax. Moreover, they recommend that states indicate if benefits can be used in ascertaining whether residents are eligible for means-tested public assistance and other benefits. Second, some policy analysts recommend that states adopt Return-to-Work (RTW) programs as part of their Medical Leave benefits. RTW programs provide financial incentives, therapeutical services, education, and employee workplace accommodations to transition workers back to work. They are sometimes covered as part of private short-term disability programs and can include “stepwise" payment models that provide a lower rate of wage replacement after a period of time and modified workplace duties to ease workers back into the workplace. RTW programs are not currently regular features of state Paid Medical Leave (PML) programs. However, studies of private RTW programs find that they reduce short-term disability leave lengths and costs.

Based on the features of the Virginia PFML program described above, Milliman, Inc. performed an actuarial analysis. The study assumed that the PFML program is established on July 1, 2022, initial staffing, procurement and education/outreach begins on January 1, 2023, implementation of the payroll tax on workers and businesses starts on January 1, 2023, and benefit payments are initiated on January 1, 2024. Since contributions begin one year before benefits are paid, a one-year period is used to build reserves for the PFML trust fund.

In addition to the HB2016/SB1330 legislation (termed the “Baseline" scenario), two additional scenarios were developed. The second scenario (Alternative 1) provides a combination of expanded eligibility, higher wage replacement, and longer leave duration than covered in the HB2016/SB1330 baseline scenario. Its policy parameters are based on programs in states that expand eligibility or benefits along these dimensions. The second scenario (Alternative 2) provides more restrictive eligibility, less generous benefits and a payroll contribution exemption for small businesses. This scenario is partially informed by the Offices of the Secretary of Commerce and Trade and the Chief Workforce Development Advisor (2020). This PFML study recommended that small businesses be exempt from contributing the employer payroll tax share to the program, though small business size is never defined. Other policy design parameters such as the program eligibility, wage replacement rate, maximum benefit, and benefit waiting period are based on the experiences of selected states with more restrictive eligibility and benefits along these dimensions.

The actuarial analysis of program costs and expenditures relies on several key inputs. They include data used to estimate eligible workers and taxable wages, incidence rates, leave durations, average benefits, and PFML administrative costs. The Weldon Cooper Center provided Milliman demographic, employment, wage escalation, and other data for use in determining the number of eligible employees and total taxable wages over the 2022-2033 period for each of the scenarios. Milliman developed claims incidence rates, duration rates, and benefit administrative costs for the Virginia PFML from public and private insurance PFL, PML, and short-term disability insurance data sources.

The baseline legislative scenario indicates that the number of eligible workers is projected to rise from 3.386 million in 2024 to 3.353 million in 2033. This represents an average of approximately 82 percent of all Virginia workers during the period. Total benefit payments increase from approximately $1.507 billion in 2024 to $2.142 billion in 2033. A payroll tax of 0.950 percent would need to be levied at the start of the program, dropping to 0.890 percent in 2026 and to 0.880 percent in 2029. This rate is similar to that levied by other states covering comparable PFML programs (i.e., benefits and taxable wage levels) such as California (1.20 percent) and Colorado (0.90 percent).

The other scenarios show the effect of easing and tightening eligibility guidelines and expanding or restricting benefits. The high benefit scenario, Alternative 1, results in an increase in the number of eligible employees with many lower wage part-time earners becoming eligible. Benefit payouts rise due to the increase in eligible employees, higher wage replacement, and longer leave allowance, which result in higher claims incidence rates and longer claim durations. The number of eligible workers is projected to rise from 3.579 million in 2024 to 3.702 million in 2033 with benefit payments increasing from $2.111 billion in 2024 to $3.011 billion in 2033. The percentage of workers covered by PFML in this scenario represents approximately 86 percent of all Virginia workers, an increase of 4 percentage points over the Baseline scenario. Higher contribution rates are needed to support the expanded eligibility and higher benefits, starting at 1.325 percent at the start of the program and decreasing to 1.250 percent in 2026 and to 1.235 percent in 2029.

The low benefit scenario, Alternative 2, results in a decrease in projected number of eligible workers. The higher wage threshold for eligibility removes many low-wage and part-time workers from the population of eligible workers. Benefit payouts decrease due to the drop in eligible workers and smaller wage replacement rates which result in lower claims incidence rates and shorter leave durations. The number of eligible workers is projected to grow from 3.076 million in 2024 to 3.182 million in 2033 with benefit payments increasing from $887 million in 2024 to $1.253 billion in 2033. The percentage of workers covered represents an estimated 74 percent of total Virginia workers, an 8 percentage point drop from the estimated baseline coverage. Contribution rates are, by far, the lowest of the three scenarios, starting at 0.625 percent at the beginning of the program, decreasing to 0.575 percent in 2026 and to 0.570 percent in 2029.

In addition to the costs and spending that result from the program, Virginia PFML can be expected to have other secondary economic, social, and demographic effects. Based on a review of literature restricted to peer-reviewed papers that used contemporary causal econometric methods and examined U.S. programs, the evidence is mixed. There is ample causal evidence that PFML increases leave utilization. Several studies also find that infant and toddler health and parental well-being improve along various dimensions. Studies of maternal labor outcomes find varied results; studies of short-term (1-2 years after childbirth) find generally positive outcomes while a few longer-term studies find no such effects. The effects of paid leave on businesses are likely fairly small, but they may be more problematic for small businesses. There is generally little evidence that adult caretaker or medical leave users realize improved labor outcomes. Finding in these areas are summarized here:

PFML Utilization and Duration. A substantial body of empirical research shows that utilization and length of parental child bonding leave increases following the introduction of both unpaid and paid family leave, but there is less causal empirical evidence about other forms of family leave taking and short-term disability. Moreover, while actuarial studies find that program design variables such as wage replacement rate and maximum leave allowance affect program utilization and length of leave, supportive causal empirical research is lacking in this area.

Labor Market Outcomes. At least eight studies have examined the effect of paid family leave on labor market outcomes such as mother’s labor force participation, employment, unemployment and wages. Most studies find that state PFL programs have positive labor market effects in the short term, including improved labor force participation and increased earnings. Two studies that focus on longer term labor market outcomes find that PFL has a negative effect on female employment.

Employer Outcomes. Studies generally suggest that employer impacts are relatively small. One possible reason for such findings is that employers do not bear the full direct costs of funding the PFML programs; statutory rates are usually split between workers and employers and the actual tax incidence is likely mostly borne by workers. Another explanation is that many businesses experience some productivity or retention improvements that offset other higher costs that some firms may experience. Employers may incur several costs from introduction of the program, including both administrative costs and costs resulting from worker absences. Evidence suggests that firms adjust to worker absences by: (1) shifting work to other workers without overtime, (2) shifting to other workers with overtime, (3) putting work on hold until an employee return, (4) hiring temporary workers, and (5) hiring permanent replacements. These adjustments may be more costly for smaller than larger firms.

Child Health Outcomes. Evidence suggests that PFML affects infant and children outcomes through intermediate improvements such as better feeding practices, improved vaccination, and reductions in low weight births. There is no evidence that PFML decreases overall infant mortality, perhaps because it does not improve outcomes for infants who are at greatest risk. Some studies find that PFML has positive effects on other child health outcomes, including infant hospitalization, parental assessments of infant and toddler health, and pediatric head trauma.

Parental Health and Wellbeing. Three studies find that PFL is associated with improved parental mental or physical health. Parents report better mental health status, lower psychological distress, and lower likelihood of being overweight and using alcohol.

The final piece of this study provides an ex-ante state economic and tax revenue impact analysis of Virginia PFML using REMI PI+ (Regional Economic Models Inc. Policy Insight Plus) software. REMI PI+ is a dynamic, multi-sector regional economic simulation model used for economic forecasting and measuring the economic impact of public policy changes on state and regional economies. Nine PFML scenarios in total were modelled. They included the HB2016/SB1330 legislative scenario ((Baseline), the more generous benefit scenario (Alternative 1) and the more restricted benefit scenario (Alternative 2). In addition, two scenarios were examined that vary the tax burdens for individuals and businesses, with the 50-50 percent split of payroll taxes between worker and firm specified in the baseline scenario changed to one scenario where 100 percent of the payroll burden is assumed by the worker (Employee Payroll Tax) and another where the total payroll burden is borne by the firm (Employer Payroll Tax). The final four scenarios explore the economic impacts of potential PFML secondary economic and demographic outcomes. These scenarios are much more speculative; they are based on program effects suggested by specific empirical studies of PFML or other information. The first scenario boosts maternal labor force participation (Labor Force Attachment). This scenario is based on substantial empirical evidence that PFML improves female labor force attachment. The second scenario considers the effect of reduced labor productivity (Labor Productivity). While the empirical evidence of PFML effects on worker productivity is mixed, most survey data suggest proportionally more firms report negative productivity effects than positive effects. The third scenario considers the effect of infant population growth due to either reduced infant mortality and/or increased fertility rates (Infant Population). The evidence for this outcome is limited; while PFL appears to improve parenting practices and child health, only a few studies show effects on infant mortality and U.S. empirical evidence of fertility effects is even more limited. The final scenario (Federal Tax Credit) considers the effect that the loss of firms’ continued eligibility to receive a federal tax credit for company-provided PFML benefits might have if Virginia adopted a PFML program.

The results indicate that the baseline scenario initially has a small positive economic impact. This occurs because administrative expenditures are needed one year before the onset of payroll taxes and two years before benefits are received by eligible employees to build the infrastructure and staffing for the program. This economic impact becomes negative in 2023 as the payroll taxes equal to .95 percent of payroll are levied to build the trust fund without a concomitant increase in benefit spending. The impacts for GDP and state tax revenue parallel those of employment. Although the economic impacts are large in absolute size, they are generally negligible relative to the size of the Virginia economy. The average employment and real GDP impacts over the 2022-2032 period represents less than 0.1 percent of average Virginia REMI PI+ forecasted total employment and real GDP over the period. The estimated total state tax revenue impacts of -$114.5 million over the period represent just 0.5 percent of the total $21.180 billion in tax revenue collected from PFML payroll taxes over the period. The lone exception is the 2023 employment impact of -16,349. Although still representing just 0.3 percent of total forecasted Virginia employment in 2023, the employment impact represents 35 percent of the REMI PI+ employment forecasted increase of 46,358 that would occur in the absence of a new PFML program. To avoid this disruption to employment growth, the General Assembly may want to consider issuing a revenue bond to smooth startup program costs over time.

The economic impacts are negative over the 2024-2032 period for essentially two reasons. First, program operation requires that reserves be maintained at a level of at least 40 percent of program expenditures over time. Thus, program expenditures during the first two years of benefits are approximately 86 percent of contributions and never exceed 98.2 percent over the entire period. The bulk of these funds are removed from the Virginia economy as Trust Fund savings that are invested in national capital markets. Second, business taxes charged to fund half of program expenses have slightly more deleterious effects on employment than personal taxes because of capital substitution for labor and effects on state business competitiveness.

The other two benefit scenarios either amplify or diminish the magnitude of economic impacts. Alternative 1, which enhances worker benefits by introducing a progressive wage replacement structure and expanding the number of weeks of eligible leave, results in higher payroll taxes and program expenditures which have more negative economic impacts throughout the period. Compared to the average annual baseline scenario impacts of -3,311 jobs, -$442.2 million in real GDP, and -$5.7 million in state tax revenue over the 2024-2032 period, this scenario results in an average annual impact of -4,644 jobs, -$620.8 million in real GDP, and -$8.1 million in state tax revenue. Alternative 2, which decreases the replacement rate and makes eligibility more difficult, results in a significantly lower payroll tax, smaller program-related expenditures, and smaller negative economic impacts. The average annual employment impact is -1,832, real GDP impact -$242.9 million and state tax revenue impact -$4.0 million over the 2024-2032 period.

Results from the tax burden scenarios suggest that shifting the payroll tax from employers to employees reduces the magnitude of the negative employment and real GDP impacts (an annual average employment impact of -1,086 and GDP impact of -$128.9 million) while shifting it to employers increases the magnitude of the negative impacts (an annual average employment impact of -5,533 and GDP impact of -$754.8 million). On the other hand, an employer payroll tax has a positive effect on state tax revenue (annual average state revenue impact of $5.6 million), while an employee tax has a negative impact (-$17.1 million). This result is obtained because payroll taxes raised on workers reduces consumer disposable incomes and consumer expenditures on goods that have a disproportionate impact on sales tax collections.

The remaining economic and demographic scenarios show varied economic and tax revenue impact. The first scenario (Labor Force Attachment) shows the effect of increasing the labor force participation rate of childbearing age females by 1.37 percentage points. This scenario results in an average annual employment impact of 7,726, real GDP impact of $840.0 million and state revenue impact of $28.1 million over the 2024-2032 period. The hypothetical scenario more than offsets the negative employment, real GDP and state tax revenue impacts of the baseline PFML operational scenario.

The second scenario shows the effect of a loss in worker productivity due to PFML (Labor Productivity). Firms respond to the loss in labor productivity (and thereby comparatively higher expense of labor) by substituting capital for labor. This scenario reinforces the negative economic impacts of the baseline scenario, resulting in an average annual impact of -4,158 jobs, -$385.3 million in real GDP, and -$18.4 in state tax revenue.

The third scenario shows the effects of an increased birth rate for PFML eligible childbearing age females (Infant Population). One major effect of the population growth is increased consumer spending, which contributes to an average annual impact of 7,960 jobs, $714.9 million in real GDP, and $66.2 million in state tax revenue over the 2024-2032 period. The economic impacts gradually grow over time as the additional births add to the Virginia population starting from a base of zero in 2023 and gradually growing to approximately 72,000 additional people in 2032.

The final scenario shows the economic impact of the loss of a federal PFML tax credit because of federal restrictions in using the credit in states with mandatory PFML programs (Federal Tax Credit). The scenario shows by far the smallest economic impacts. Results indicate that loss of the federal credit would have an average economic impact of -747 jobs, -$73.7 million in real GDP, and -$3.3 million in state tax revenues.