The Effects of the Workforce Innovation and Opportunity Act of 2014 on Vocational Rehabilitation Engagement, Employment, and Work Incentive Use Among Supplemental Security Income Recipients Aged 14–24

Introduction

The Workforce Innovation and Opportunity Act (WIOA) of 2014 amended the Rehabilitation Act of 1973 and significantly shifted how state vocational rehabilitation (VR) agencies offered services to youths with disabilities, particularly to students. WIOA requires state VR agencies to offer preemployment transition services (pre-ETS) to students with disabilities and to reserve at least 15 percent of their federal program funds for that purpose (Department of Labor 2014). This article explores the extent to which WIOA and pre-ETS access affected employment-related outcomes for youths with disabilities who receive Supplemental Security Income (SSI) payments. Understanding whether pre-ETS help students transition from high school to better postsecondary education and employment opportunities is critical because a successful transition can improve a young person's future employment prospects and earnings, health-related quality of life, and well-being. Moreover, evidence on the effectiveness of transition support programs for this population is limited (Urdapilleta and others 2020).

The influence of WIOA on VR service applicants and participants has been previously documented. When youths approaching the transition to adulthood are exposed to services such as pre-ETS, they are more likely to sign an individualized plan for employment (IPE) and use VR services (Luecking and others 2018). Increased VR engagement may lead to better employment and earnings outcomes (Dean and others 2019; Yin, Siwach, and Lin 2023), including for young SSI recipients (Hoffman, Hemmeter, and Bailey 2018). Although youths constituted a larger proportion of VR service applicants after WIOA than before (Department of Education 2020), we know of no quantitative evidence showing how the employment outcomes of transition-age youths changed after WIOA was implemented and pre-ETS became available.

This article aims to measure how VR engagement, employment, and SSI work incentive use changed for SSI recipients aged 14–24 after the 2014 enactment of WIOA made pre-ETS available. Using rich administrative data from the Social Security Administration (SSA) and the Department of Education's Rehabilitation Services Administration (RSA), we construct two models and measure how youth outcomes changed from 2010 to 2021. In the first model, we adjust for state and individual characteristics to estimate the extent to which WIOA affected the percentage of young SSI recipients who applied for VR services, the percentage who signed IPEs, the percentage who had any annual earnings, the annual earnings amounts, and the percentage who used either of two SSI work incentives: the student earned income exclusion (SEIE) and a continuation of payment eligibility for recipients who use VR or similar services, named for its authorizing legislation, Section 301 of the Social Security Act Disability Amendments of 1980. In the second model, we explore the effects of different levels of pre-ETS access by state and year.¹ We then estimate the association between state-level pre-ETS access and changes in outcomes for young SSI recipients.

This article documents the influence of WIOA on youths who have disabilities that pose substantive employment barriers. More of these youths applied for VR services, signed an IPE that would allow them to access services beyond pre-ETS, and had higher annual earnings after WIOA than before. Moreover, those in states offering greater pre-ETS access also had higher annual employment rates and earnings amounts, as well as higher rates of SEIE use, after WIOA. For 2017 to 2021, we observe positive correlations between state-level pre-ETS access and the signing of IPEs, employment, earnings, and use of the Section 301 work incentive.

Background

In this section, we discuss how WIOA affected the provision of employment-support services for youths with disabilities. We then describe the challenges to employment for our study population and summarize the literature on their interactions with services like pre-ETS.

WIOA and Pre-ETS

WIOA instituted new requirements for state VR agencies that provide services for students with disabilities. For example, it requires state VR agencies to reserve at least 15 percent of their federal program funds to offer pre-ETS to students with disabilities (Department of Labor 2014). A state VR agency must make pre-ETS available to all students with disabilities, regardless of whether they apply for other VR services at the agency. In addition to preparing a student for employment, pre-ETS could lead some students to apply for additional VR services before entering employment.

The Department of Education (2020) defines “student with a disability” as “an individual who is in an educational program, meets certain age requirements, and is eligible for and receiving special education or related services under the Individuals with Disabilities Education Act [IDEA] or is an individual with a disability for purposes of Section 504 of the Rehabilitation Act (Section 7(37) of the Rehabilitation Act and 34 C.F.R. § 361.5(51)).” Students can be enrolled in high school, a recognized educational setting, or postsecondary education institutions. They are typically aged 16 to 21, although the specific age range varies with the state's age requirements for IDEA-mandated transition services and the minimum age agreed upon with the state VR agency (Carlson, Thompson, and Monahan 2020).

VR agencies and educators often collaborate to provide pre-ETS at schools (Fabian, Neubert, and Luecking 2018). The school setting is especially suitable because WIOA allows VR agency staff to work with students in groups rather than individually. WIOA defines five specific services that agencies must offer to students with disabilities: (1) job exploration counseling (such as career counseling or vocational interest inventories), (2) work-based learning experiences (WBLEs; examples include job shadowing or internships), (3) counseling on opportunities for enrollment in comprehensive transition or postsecondary educational programs at institutions of higher education, (4) workplace readiness training (such as life skills and financial literacy), and (5) instruction in self-advocacy (such as self-determination training or peer mentoring).

The successful implementation of pre-ETS depends on factors involving students, their families, VR providers, educators, and local area characteristics. Students and their families may not be adequately informed about the purpose of VR services and the availability of services in their areas (Schutz and others 2022; Awsumb, Balcazar, and Keel 2019). Students may also lack the resources and support required for their VR engagement because of their needs and disabilities (Fabian, Neubert, and Luecking 2018; Bromley and others 2022). VR counselors report challenges related to the increased caseload and paperwork involved with serving eligible and potentially eligible students after WIOA, in addition to insufficient time and financial resources to implement pre-ETS (Fabian, Neubert, and Luecking 2018; Awsumb and others 2020). Despite the overall collaborative relationships between VR counselors and local schools, some educators struggle to connect with students who could benefit from pre-ETS but are unfamiliar with the services (Carter and others 2021). Finally, successfully implementing some services depends on local area characteristics, such as the availability of employers interested in offering community-based WBLEs to students (Bromley and others 2022).

To date, VR agency implementation of pre-ETS has varied, which could lead to differences in the outcomes for students with disabilities across states. For example, in program year 2021 (July 2021–June 2022), the percentage of students with disabilities receiving VR services who used pre-ETS ranged from 100 percent in Illinois to 14 percent in Puerto Rico. Further, in that same year, the percentage of VR participants who were younger than 19 when they signed their IPE ranged from 63 percent in Illinois to 7 percent in Oregon (Department of Education 2022). The student participation patterns also differ: in Illinois, 94 percent of students with disabilities who used pre-ETS had applied for VR services, and the remainder were potentially eligible; in Oregon, the percentage of students with disabilities who used pre-ETS and had applied for VR services was only 1 percent.

Young SSI Recipients

This study focuses on SSI recipients aged 14–24. Most, but not all, in the younger part of that age range are probably students. SSI is a means-tested cash payment for individuals with significant disabilities.² Given their income, asset, and health situations, the post-school employment prospects for these youths may benefit from pre-ETS even more than those of other youths with disabilities.

Youths with disabilities, in general, might not be adequately prepared for employment because they lack career development, learning, and training opportunities. Despite the potential availability of public programs that offer these services, youths with disabilities might face challenges in using them because of complex eligibility rules, fragmented transition systems, and other barriers (Livermore and others 2019).

Such challenges in achieving employment are likely to be even more significant for SSI recipients because of their low household resources and limiting health conditions. Based on 2021 Current Population Survey data, the median household income for an SSI recipient aged 17 was $51,600 and for one aged 18 it was $60,500. For comparison, the median household income for youths with disabilities but not receiving SSI was $78,300 at age 17 and $93,300 at age 18 (Flood and others 2023). The employment rates of youths with a disability are 22 percent for those aged 16–19 and 46 percent for those aged 20–24. By contrast, youths in those age groups with no disability have employment rates of 33 percent and 68 percent, respectively (Bureau of Labor Statistics 2024). Moreover, young SSI recipients are less likely to use VR services after applying for them, and their VR cases are less likely to close with employment, than are nonrecipient youths with disabilities (Honeycutt, Martin, and Wittenburg 2017).

For these reasons, pre-ETS and other VR services can help young SSI recipients to transition from high school with better postsecondary education and employment opportunities. A successful transition can lead to upward mobility by improving future employment and earnings prospects, health-related quality of life, and well-being (Hartman and others 2019).

Employment-Related Outcomes of Young SSI Recipients

We are not aware of direct evidence on how pre-ETS affects employment outcomes for transition-age youths, but literature documents outcomes for youths who use services similar to pre-ETS. Although VR services can improve employment outcomes for transition-age youths, the findings are mixed. Correlational evidence shows that youths who use VR services, including SSI recipients, have better long-term employment outcomes than those who do not (Hoffman, Hemmeter, and Bailey 2018). Osmani and others (2022) showed that an immersive experience (Project SEARCH) was correlated with higher probability of successful VR case closure. Yin, Siwach, and Lin (2023) presented causal evidence that VR services increased youth employment rates and earnings for up to 2 years after case closure, with greater effects for those aged 14–18 than for those aged 19–24. Dean and others (2019) found that youths with disabilities participating in a transition program had increased employment and earnings for more than 2 years after the end of the program.

A series of recent studies measured the effect of offering WBLEs to high school students with disabilities in Maine, Maryland, Massachusetts, and Vermont. Despite the successful implementation of these programs, WBLEs were not consistently associated with improved employment outcomes up to 24 months after enrollment, although participants in Massachusetts had higher mean hourly wages (Foley and others 2022; Mann and others 2021; Sevak and others 2021; Siwach and others 2021). Finally, two demonstrations—the Youth Transition Demonstration (YTD) and Promoting Readiness of Minors in SSI (PROMISE)—offered employment and other services to young SSI recipients. In YTD, six independent projects tested a variety of service models, but all generally focused on providing employment services to youths aged 14 to 25. Although the projects had positive short-term effects on service receipt and other outcomes, the employment results were not sustained (Fraker and others 2014). PROMISE offered employment and other services through six school-to-work transition programs to SSI recipients aged 14–16. These programs used different service models, but all focused on state and local partnerships, case management, and employment, and all offered to connect youths with VR agencies. All programs affected service use and employment within 18 months of enrollment, a period that includes youths' direct involvement with the programs (Mamun and others 2019). Only two PROMISE programs had positive employment effects for youths 5 years after enrollment (that is, after they left the programs) (Patnaik and others 2022; Mamun and others 2019).

Improved employment rates associated with the use of pre-ETS and VR services may lead more SSI recipients to use work incentives. Although SSI provides several work incentives, we focus on two that are especially relevant to transition-age youths: the SEIE and Section 301 payment continuations. For students who have an individualized education program (IEP), the SEIE allows SSA to exclude a portion of the SSI recipient's earnings in computing payment eligibility and amounts if the recipient is younger than 22 and regularly attending school, college, university, or a course of vocational or technical training. In 2023, an individual's maximum income exclusion was $2,220 per month, with the total annual amount not to exceed $8,950 (SSA 2023, 2024). Section 301 of the Social Security Act Disability Amendments of 1980 allows SSA to continue making monthly SSI payments to recipients who participate in VR or similar services, even if they no longer meet SSA's medical or work-related definition of having a qualifying disability.

The use of the SEIE and Section 301 payments historically has been low. From 2012 through 2015, less than 1.5 percent of SSI recipients aged 14–17 used an SEIE; and in 2015, about 1,200 recipients aged 18–19 used Section 301 continuations. Use of the SEIE may be low either because youths and their families have not heard of it or they fear that using it could negatively affect their payments. Potential reasons for low use of Section 301 continuations could include the limited number of individuals younger than 18 who used VR services and rules restricting eligibility for those aged 18–21 to those having an IEP (Government Accountability Office 2017, 2021). Additionally, the Section 301 incentive is available only if an individual has not requested continued payments while appealing a negative eligibility determination; because appealing and requesting payments is very common, even those otherwise eligible for Section 301 payments may not receive them.

Data and Methods

We used information from multiple administrative data sources. Our main source, SSA's 2021 Disability Analysis File (DAF), includes information on our study population—youths aged 14–24 who received SSI payments at any time from 2010 through 2021. The DAF combines (and links) extracts of administrative data files from SSA and other agencies. Our study uses DAF data drawn from SSA's Supplemental Security Record (the primary system for tracking SSI payments) and Master Earnings File; and from RSA's individual-level Case Service Reports (the RSA-911 file). We also used data from the Department of Education's Child Count and Educational Environments file, known as the IDEA Section 618 file after its authorizing legislation. We drew information on use of the Section 301 work incentive directly from a part of the Supplemental Security Record that is not available in the DAF. Similarly, RSA staff provided us with information on pre-ETS availability that is not included in the RSA-911 data in the DAF. We used the variables in these data to identify our analytical sample and define most of the outcomes we analyzed.

We acknowledge several limitations in the data:

• The records from the Master Earnings File include earnings as reported to the Internal Revenue Service, so they exclude informal earnings.
• Although WIOA established pre-ETS in 2014, the earliest RSA-911 data on pre-ETS use are for 2017, when RSA first required state VR agencies to report them.
• Besides having no pre-ETS information for 2014–2016, we cannot enumerate every person who used pre-ETS from 2017 onward. Because VR agencies must offer pre-ETS to youths regardless of their VR application status, a Social Security number is not required to access the services. Therefore, we cannot determine that young SSI recipients used pre-ETS if they did so before they applied for VR services. Moreover, even among youths who applied for VR services then used pre-ETS, not all RSA-911 records contain identifiers that allow a match to SSA data. For example, between 9 percent and 12 percent of nationwide RSA-911 records from program years 2019, 2020, and 2021 could not be matched to SSA data (Mathematica 2023). Further, record matching may vary by state. Our estimates of VR engagement therefore represent a lower bound.

Given the limitations on pre-ETS information in the individual-level RSA-911 and DAF data, we devised a way to estimate state-level pre-ETS use. RSA staff provided us with the number of students who used pre-ETS in each state and program year from 2017 through 2021. We complemented this state-level information with data from the IDEA Section 618 file. The latter file provided the total number of students aged 14–21 who had an IEP to use special education services each year from 2017 to 2021 (Dragoo 2024). When these data were missing for a state for a particular year, we imputed the missing value using data for that state from the previous year adjusted by the average national percentage change in the number of students in that year. Data were missing for three states in 2017 and for one each in 2018, 2019, and 2020; data for Wisconsin were missing for three of these years. Of note, the RSA-911 and IDEA Section 618 data are reported by program year, but DAF data are reported by calendar year.

Sample

The study sample is the universe of youths aged 14 to 24 who are eligible to receive SSI payments in December of each year. Because a youth may appear repeatedly across years in the sample, we view the sample as annual cross-sections of SSI-eligible youths from 2010 to 2021. The age distribution of the sample remains relatively unchanged over the years, although the total number of SSI recipients aged 14 to 24 in the sample decreased from 929,547 in 2010 to 820,650 in 2021 (Chart 1), reflecting the trend for the SSI program overall.

Empirical Strategy

We use two individual-level regression models to estimate how the WIOA, and specifically pre-ETS access, affects our outcomes of interest over time, nationally and across states. We estimate the marginal effects using linear models to avoid making assumptions about the true functional form of our models and the distribution of data. In all estimations, we cluster standard errors at the state level.

In Model 1, we estimate how outcomes for SSI recipients aged 14 to 24 changed nationally after the WIOA was enacted in 2014:

Y_ist represents each of the six outcomes for individual i living in state s in year t. AfterWIOA_t is a binary variable equal to zero for 2010 to 2013 and equal to one for 2014 to 2021. S_s represents fixed state effects, and X_i includes individual-level covariates (sex, age as of December 31 of each year, age at last SSI application, and impairment).

We extend Model 1 by allowing the estimate of post-WIOA changes to vary by age group, both for the entire study period and for each year after 2014, to capture any shifting trends, such as increasing pre-ETS use over time, as WIOA provisions were implemented.

In Model 2, we allow the access to pre-ETS to vary by state and year. Instead of using a binary variable to capture the pre- and post-WIOA periods, we use a state-and-year-specific pre-ETS access ratio. Although we expect that pre-ETS availability began to increase once WIOA was enacted in 2014, there are no data on pre-ETS access before 2017. Therefore, this model estimates how outcomes for young SSI recipients changed with an increase in pre-ETS access from 2017 through 2021:

where PreETSRatio_st is the number of students using pre-ETS divided by the number of students receiving special education in state s and year t. Z_t represents fixed year effects, which capture variations across time common to all states—for example, improvements in data management systems that reflected a more accurate report of the number of students using pre-ETS. The other variables follow the Model 1 definitions. We also extend Model 2 to allow estimates to vary by age group.

To validate the findings, we conduct five sensitivity analyses:

• Excluding states with extremely high or low pre-ETS access ratios in 2017 (the pre-ETS access ratio for Iowa was 53 percent, whereas the next highest ratio was 39 percent; the ratios for California, New Jersey, and New York were all below 1 percent),
• Excluding the period during the COVID-19 pandemic (2020 and 2021) from the sample,
• Adding the state annual unemployment rate—calculated from the monthly rates extracted from the Bureau of Labor Statistics' Local Area Unemployment Statistics—as a control in Models 1 and 2,
• Testing whether the estimates for the states that implemented the PROMISE demonstration (Arkansas, California, Maryland, New York, Wisconsin, and states in the Achieving Success by Promoting Readiness for Education and Employment [ASPIRE] consortium—Arizona, Colorado, Montana, North Dakota, South Dakota, and Utah) differed from those for all other states,³ and
• Using binary indicators of pre-ETS access ratios instead of the continuous pre-ETS access ratio in Model 2.

Independent Variables

Three independent variables reflect pre-ETS availability: an indicator for the enactment of WIOA in 2014, an indicator for each year from 2014 to 2021, and a state-and-year-specific ratio that we use as a proxy for pre-ETS access. The pre-ETS access ratio captures young SSI recipients' access to and potential use of pre-ETS from 2017 to 2021.

Chart 2 shows the 5-year average annual ratio for each state. We need to use this proxy of pre-ETS access because, although we can determine pre-ETS use for many young SSI recipients, we cannot identify the use of any pre-ETS among young SSI recipients who used pre-ETS as potentially eligible students (that is, before they applied for VR services). The numerator consists of the number of unique students who used pre-ETS in a state and year. In most states, youths must be students aged 16–21 to use these services. The denominator is the number of students aged 14–21 using special education services under the IDEA—this population is more restrictive than the population in the numerator, as students using pre-ETS may also use educational support services under Section 504 of the Rehabilitation Act or be enrolled in postsecondary education. The pre-ETS access ratio varies by state and year. In addition to analyzing the ratios directly, we use them to split the sample into states with consistently low or high ratios, thus identifying states where students with disabilities had broader or more restricted access. The 15 states in the low group (Arizona, California, Colorado, Connecticut, Kansas, Maine, Massachusetts, New Jersey, New Mexico, New York, Ohio, Oklahoma, Rhode Island, Texas, and Washington) had pre-ETS access ratios below the median every year from 2017 to 2021, while the 14 states in the high group (Alabama, Hawaii, Indiana, Iowa, Kentucky, Michigan, Montana, Nebraska, North Dakota, South Carolina, Vermont, West Virginia, Wisconsin, and Wyoming) had ratios above the median for all years.

Each model includes four additional covariates that control for individual characteristics: sex, age (as of December 31 in each year), age at last SSI application, and impairment.

Outcome Variables

We explore six outcomes of interest, grouped into three domains, which could potentially be affected by WIOA. The first domain consists of two binary variables that indicate engagement with VR: whether the youth applied to a VR agency for services and whether the youth signed an IPE. The former captures a person's initial interest in VR services and the latter indicates that the VR agency found the applicant eligible for services and developed a plan with the person to identify an employment goal along with the services needed to achieve that goal. The second domain comprises two variables measuring employment outcomes—a binary variable indicating whether the youth reported any earnings in that year and a continuous variable indicating total annual earnings. We adjust earnings to 2021 dollars and cap them at the state's 99^th percentile. The average 99^th percentile state earnings level is $12,849.11 and truncating the top 1 percent of earnings in each state eliminates high values that could be a result of problems in the underlying data.⁴ The third domain addresses the use of two SSI work incentives—the SEIE and Section 301 payment continuations—both of which are binary variables. We identify all outcomes by calendar year of occurrence.

Results

Table 1 presents the descriptive statistics on the outcome averages, with detail by age group, for 2010–2021. In the subsections below, we discuss the results for each outcome domain.

VR Engagement

The percentage of young SSI recipients engaging in VR services increased rapidly after WIOA was enacted in 2014 (Chart 3). The path of VR engagement across the years is similar for all ages, but those aged 18–21 and 22–24 were most likely to apply for VR services and to sign IPEs.

Table 2 shows the results of Model 1, estimating VR engagement before and after WIOA, and confirms the patterns observed in the descriptive statistics (Table 1). Model 1 estimates the differences in VR service application rates and signed IPE rates before and after 2014, adjusting for individual characteristics and fixed state effects. VR engagement increased after the WIOA was enacted. Among all SSI recipients aged 14–24, 2.79 percent applied for VR services during or after 2014, more than eight times higher than the rate (0.34 percent) for the years before WIOA. The post-WIOA increase in signed IPE rates was proportionally greater still, from 0.15 percent in 2010–2013 to 2.36 percent in 2014–2021.

Allowing the estimates to vary by age reveals that the increase in VR applications and signed IPEs was largest for youths aged 18–21, followed in rank order by those aged 22–24, 16–17, and 14–15 (Table 2). Although the population that applied for VR services and signed an IPE in each year is not necessarily the same, the estimates in Table 2 suggest that youths aged 18 or older experienced a larger post-WIOA increase in the likelihood of applying for VR services and signing an IPE than youths in other age groups.

Letting the estimates vary by age for each year 2014–2021 shows that the regression-adjusted differences in VR application rates increased for all age groups relative to the pre-WIOA mean in each year from 2014 through 2018 (Chart 4).⁵ The increase in VR application rates was sharpest for ages 18–21, followed by the 22–24 and 16–17 age groups; the rate increased slightly over this period for the youngest age group (14–15). For example, with all else equal, an SSI recipient aged 18–21 in 2019 was far more likely to apply for VR services than an SSI recipient aged 18–21 before 2014 (6.03 percent compared with 0.34 percent). Although changes in other factors may have affected the likelihood of applying for VR services, the increase in 2019 is potentially due to a combination of (1) 5 years of experience for VR agencies in implementing WIOA provisions, offering pre-ETS, and adjusting their service models toward youths with disabilities and (2) 5 years of a young person's potential access to pre-ETS and other changes related to WIOA. We observe similar patterns for the likelihood of signing an IPE.

VR engagement plateaued for all age groups from 2017 to 2019. The stability in the rates of VR applications and signed IPEs starting in 2017 could reflect a 3-year lag to implement WIOA policies or an improvement in their implementation once reporting of pre-ETS activities became mandatory in 2017.

The post-WIOA changes in VR engagement were similar both for the 14 states with consistently high pre-ETS access ratios and the 15 states with consistently low ratios (Table 3). The lack of a statistically significant difference between the states with consistently high and low pre-ETS access ratios suggests that VR engagement for young SSI recipients increased across all states after WIOA, regardless of ease of pre-ETS access.

In Model 2, we further examine the relationship between VR engagement and pre-ETS access by estimating the association between changes in the pre-ETS access ratio and changes in VR engagement from 2017 to 2021. As noted earlier, the ratio is a proxy for potential access to pre-ETS and VR agencies' experience with offering pre-ETS, which varies by state and year.

Pre-ETS access ratios have large, positive associations with VR engagement after controlling for individual characteristics and fixed state effects (Table 4). For each 10 percentage-point increase in the pre-ETS access ratio, the likelihood that a young person signed an IPE increased by 0.34 percentage points (p-value = 0.08). This finding implies a 13 percent increase from the baseline (pre-WIOA) scenario of no access to pre-ETS, in which 2.55 percent of youths signed an IPE. We observe a similar large, positive association between pre-ETS access ratios and VR applications, but that association is not statistically significant (p-value = 0.10).

The association between pre-ETS access ratios and VR engagement from 2017 to 2021 was statistically significant and positive for the youngest age groups. An increase in pre-ETS access of 10 percentage points correlated with significant increases in VR applications among youths aged 14–15 and 16–17 (0.24 percentage point and 0.87 percentage point, respectively). These changes represent an increase of about 50 percent from the pre-WIOA baseline scenario in which no youths used pre-ETS. Increases in the state's pre-ETS access ratios had no statistically significant associations with any changes in VR service application among youths aged 18–21 and 22–24. Changes in the rate of signed IPEs follow a similar pattern—the main difference is that youths aged 18–21 were also more likely to sign IPEs following increases in access to pre-ETS (p-value = 0.09).

Employment and Earnings

In this section, we evaluate how the increased use of pre-ETS and other WIOA effects corresponded with the employment of young SSI recipients. WIOA changes might either increase employment (as youths engage in pre-ETS activities and other VR services that help them transition to the labor force) or decrease employment in the short term (if greater access to pre-ETS leads them to seek further vocational training before pursuing work). The previous subsection showed the associations of WIOA and state-level pre-ETS access ratios with increased use of VR services. The evidence in the literature, as noted, also suggests VR service use improves employment outcomes.

The share of young SSI recipients who had any earnings increased during our analysis period (Chart 5). However, the increases were modest after WIOA was enacted in 2014. The findings of Model 1, which adjusts for individual characteristics and fixed state effects, suggest that employment outcomes improved after WIOA (Table 5). The likelihood that young SSI recipients overall had any paid earnings increased from 16.39 percent per year in the pre-WIOA period 2010–2013 to 19.99 percent after WIOA (2014–2021), a 22 percent increase. Average annual earnings increased more than 50 percent during the study period—from $457 to $696.

In the Model 1 estimates by age group, the increase in the share of youths with any earnings is lowest for ages 14–15, and the average change in annual earnings increases with age. This pattern is not surprising, as labor force participation tends to be low at ages 14 and 15, and average earnings are expected to increase with age. In splitting the analysis by age group and year after 2014, Chart 4 shows that the any-earnings rate generally rose in relation to the 2010–2013 baseline period in the post-WIOA years before 2019. The likelihood of having any earnings and the mean annual earnings amounts increased each year after 2014 for all age groups until the COVID-19 public health emergency in 2020. Employment outcomes worsened for all age groups when the pandemic emerged in 2020 then started to recover in 2021.

After WIOA, young SSI recipients were more likely to have any earnings and had higher annual earnings amounts. The increases were largest in states with consistently high pre-ETS access ratios (Table 6). However, because we observed no differences in VR outcomes across these states, the difference in employment outcomes might be due to other WIOA features or state policy and economic environments that do not directly affect VR service use.

To examine the role of pre-ETS access in these changes, we use Model 2 to explore how employment outcomes varied across states by pre-ETS access ratios from 2017 to 2021 (Table 7). We find a positive association between pre-ETS access ratios and the likelihood of having any earnings after controlling for individual characteristics and fixed state effects. In a baseline (pre-WIOA) scenario with no access to pre-ETS, 18.89 percent of all-ages youths had any earnings. The model estimates that, for each 10 percentage-point increase in the pre-ETS access ratio, the likelihood that a youth had any earnings increased by 0.92 percentage points (p-value = 0.05), or nearly a 5 percent increase from the baseline estimate. We observe a positive and large but not statistically significant association between pre-ETS access ratios and total annual earnings.

The post-WIOA patterns of change in pre-ETS access and in earnings outcomes by age are similar. Table 7 shows that correlations between pre-ETS access and the percentage of youths with earnings were significant for those aged 16–17 and 18–21, but not for the oldest (22–24) and youngest (14–15) groups. The association between mean annual earnings and pre-ETS access ratios was positive and significant only for youths aged 16–17.

These results partially support the hypothesis of a positive association of WIOA and pre-ETS access with better employment outcomes. The association is stronger in the extensive margin—increasing the likelihood that young SSI recipients had any earnings—than in the intensive margin, where the associations with earnings amounts were less robust.

Use of Work Incentives

Two potential pathways could lead to a post-WIOA increase in work incentive use. First, greater VR service use (as evidenced by increased IPE signings) could result in more youths meeting the requirements for continued SSI payments through Section 301. Second, the post-WIOA increase in the use of pre-ETS and other VR services may have contributed to increased use of SEIEs if the positive association of post-WIOA earnings outcomes occurred for students.

Our descriptive statistics show that the association between WIOA and work incentive use is weak. The share of young SSI recipients who used the SEIE remained relatively stable from 2010 to 2019 and declined during the pandemic years, whereas the share using Section 301 continuations declined over time (Chart 6).

Adjusting for individual characteristics and fixed state effects, Model 1 shows that the likelihood of using Section 301 continuations decreased from 0.19 percent before WIOA to 0.13 percent afterward (Table 8) and the increase in SEIE use was negligible. However, these are percentages among all young SSI recipients, not among only those who are most likely to use the work incentives.

WIOA's effects on work incentive use differed by age group and incentive type. The decline in the use of Section 301 continuations was driven primarily by individuals aged 18–21—which is the group most likely to use that incentive.⁶ The point estimates by age also show that youths aged 16–17 had the largest increase in the likelihood of using the SEIE, from 2.19 percent for the period 2010–2013 to 2.54 percent for the period 2014–2021. Youths aged 22–24 were less than 0.03 percentage point more likely to use the SEIE after WIOA than before. Although the increase and the share of SSI recipients using this work incentive are small, this difference is precisely estimated.

Patterns in the use of the two work incentives also differed when we split the analysis in Model 1 by age group and year after 2014. Relative to the 2010–2013 baseline period, the likelihood of using Section 301 continuations fell in each post-WIOA year for the 18–21 and 22–24 age groups, for whom this policy is relevant. The decline continued until the pandemic (Chart 4). However, SEIE use increased for youths aged 16–17 and 18–21 after 2014, until the pandemic years likewise interrupted that trend. The use of both types of work incentives dropped for all age groups in 2020 and 2021. Although SEIE use might be expected to decline because of the decline in employment during the pandemic, the decrease in Section 301 continuation use is unexpected (because that incentive is tied to education and training).

The increase in SEIE use after WIOA was concentrated in states with consistently high pre-ETS access ratios (Table 9). This pattern corresponds with those of young SSI recipients in these states having (1) a higher likelihood of receiving any earnings and (2) higher annual earnings (Table 6).

When we consider access to pre-ETS from 2017 to 2021 in Model 2, we find a large, positive association between pre-ETS access ratios and the likelihood of using a Section 301 continuation but not the SEIE after controlling for individual characteristics and fixed state effects (Table 10). For each 10 percentage-point increase in the pre-ETS access ratio, use of Section 301 continuations increased by 0.05 percentage point (p-value = 0.06). This finding implies an increase of more than 80 percent from a baseline scenario of no pre-ETS access, in which 0.06 percent of youths used Section 301 continuations.

Letting the association between pre-ETS access and work incentive use vary by age in Model 2 reveals important heterogeneities, as expected, as work incentives are more relevant for recipients of certain ages than for others. Table 10 shows that higher pre-ETS access ratios were associated with increased Section 301 continuation use for all age groups. The point estimate is largest for ages 18–21, where an increase of 10 percentage points in pre-ETS access was associated with significant increases of 0.12 percentage point in the use of Section 301 continuations—corresponding to a 63 percent increase from a baseline scenario of no pre-ETS access.

Sensitivity Analyses

Our main results remained largely unaffected under alternative specifications. First, because 2017 is the first year for which states reported pre-ETS statistics, we removed states with extremely high (Iowa) or low (California, New Jersey, and New York) pre-ETS access ratios in 2017. Although these states could be legitimate outliers in pre-ETS offered and provided, they also could have experienced data quality issues as they initiated reporting the numbers to RSA. To avoid capturing spurious correlations or hiding important patterns in the data caused by measurement errors, we reestimated our regressions dropping outlier states from the sample. Second, to test the sensitivity of the estimates to the COVID-19 pandemic years, we used the main models for the study period omitting 2020 and 2021. In these two tests, all estimates of WIOA effects on outcomes and the associations with pre-ETS access ratios still pointed in the same direction. Several estimates lost some statistical significance, which is not surprising given the reduced sample size. As expected, the association of pre-ETS access ratios with outcomes was attenuated for all outcomes when we removed the pre-ETS access outlier states from the sample. Third, to control for fluctuations in economic conditions, we added the state annual unemployment rate as a covariate in both models. The point estimates varied, but the coefficients pointed in the same direction and their significance remained largely unchanged from our main analyses. This result strengthens the confidence that our findings are not driven exclusively by the stronger economic environment after 2010.

We also tested how our estimates varied in the 11 states that participated in the PROMISE demonstration, which connected young SSI recipients with VR agencies, among other services. Because the PROMISE implementation period (2016 to 2019) overlaps our analysis period, the demonstration may have affected the dynamics of WIOA and pre-ETS in these states. After WIOA, changes to outcomes in PROMISE states were similar to those in other states. The only difference was that the post-WIOA increase in the rate of recipients with earnings was smaller in PROMISE states. The similarity in outcomes is not surprising, given that most young SSI recipients in PROMISE states were not PROMISE participants. However, the associations between pre-ETS access and VR engagement and employment were statistically larger in PROMISE states. These differences suggest that investments in pre-ETS by VR agencies may have had a cumulative effect in these states. Because PROMISE enrollment began in 2014 and services began in 2016, the period of our analysis—2017 to 2021—likely captures the additional experience these states had offering services connected to pre-ETS.

Our final sensitivity analysis uses alternative versions of Model 2 with binary indicators of pre-ETS access ratios to indicate states that exceed the median and the 75^th percentile ratios (instead of the continuous pre-ETS access ratio). The estimates have the same sign as those of our main model but are mostly not statistically significant. This finding supports the notion that associations between pre-ETS access and changes in outcomes were proportional to increases in pre-ETS access, as opposed to the availability of these services above a certain threshold.⁷

Conclusion

WIOA's enactment in 2014 represented a significant shift in how VR agencies offered services to youths with disabilities. RSA has previously documented how WIOA affected the characteristics of VR service applicants and participants, who trended considerably younger after WIOA than before (Department of Education 2020). This study is the first to measure quantitative patterns in the VR engagement, employment, and work incentive use outcomes for a population of youths with disabilities before and after WIOA. Using administrative data from SSA and RSA to track the experiences of young SSI recipients during this period offers a unique opportunity to describe WIOA effects.

After WIOA was enacted, young SSI recipients applied for VR services and signed IPEs at higher rates than before WIOA. The average VR application rate increased from 0.34 percent before WIOA to 2.79 percent afterward. Although the post-WIOA rate may seem small, it is not insubstantial: the number of VR applicants in 2010 was 1,181; in 2019, that number was 30,569. For added context, the number of young SSI recipients who applied for VR services in 2019 represents almost 7 percent of the 446,919 people of all ages who applied for VR services that year (Department of Education 2020). This number underscores the scope of the opportunity for VR agencies to offset these participants' service costs, as they can be reimbursed for those costs by SSA when adult SSI recipients have substantial earnings in 9 consecutive months (SSA n.d.). When we examine state-level service use rates, the story is similarly striking. In 2010, the highest state VR application rate for young SSI recipients was 0.4 percent. In 2019, the lowest such rate was 0.5 percent, or higher than the state with the best rate 9 years earlier. In the state with the highest rate in 2019, 7.9 percent (or about 1 of every 13 young SSI recipients) applied for VR services.

Changes brought by WIOA, including offering pre-ETS to students with disabilities, were associated with increased applications for VR services. Although this finding is not unexpected, it does quantitatively document the potential associations that correspond to a specific federal policy change. VR agencies offer pre-ETS to all students with a disability, without requiring them to apply for further services, and agencies must spend 15 percent of their federal funding on these services. Between 2017 and 2021, higher state-level pre-ETS access ratios were associated with larger increases in signed IPEs (but not VR service application rates) among young SSI recipients. The associations we observe during the later years could be due to a combination of two levers. First, by 2021, VR agencies had up to 7 years of experience in responding to WIOA requirements, offering pre-ETS to students with disabilities, and developing partnerships with state and local education agencies and other workforce partners. Hence, they could provide better-quality pre-ETS in later years than in the years just after WIOA enactment. Second, in those later years, youths with disabilities could have benefited from increased pre-ETS access and other changes brought by WIOA. For instance, a 21-year-old in 2021 (who was 14 in 2014) could have used pre-ETS for up to 5 years. Potentially, this long-term access could better prepare youths for educational and employment opportunities.

Notably, we find that VR engagement, employment, and use of SSI work incentives before 2017 (as well as before WIOA) were higher among young SSI recipients in states with the highest pre-ETS access ratios in the 2017–2021 period than among those in the states with the lowest pre-ETS access ratios. Thus, pre-WIOA state policy environments oriented toward the success of youths with disabilities in general and to young SSI recipients specifically may have influenced both WIOA implementation and other youth services and outcomes.

Although earnings for young SSI recipients increased after WIOA, and though it would be consistent with a view that increased pre-ETS access could encourage youths with disabilities to enter the labor market directly, the observed change may relate not exclusively to WIOA but also to the stronger postrecession economic environment after 2010. Earnings for young SSI recipients in most age groups increased after WIOA, but the change was strongest for ages 16–17, and the association with pre-ETS access ratios was positive for the presence of any earnings (though not statistically significant for annual earnings amounts) among all youths in the sample from 2017 to 2021, suggesting potential WIOA influence on employment.

We expected increased use of SSI work incentives among young SSI recipients, given increased VR engagement and earnings, but the evidence is mixed. About 20 percent of this population had any earnings in a given post-WIOA year; however, annual SEIE use rates were around 2 percent. Only students can use SEIE; and although not all young SSI recipients were students, we can assume that a good portion were. We have noted the lack of association between the pre-ETS access ratios and SEIE use for the broader population of young SSI recipients. In 2019, almost 23,000 16- and 17-year-old SSI recipients had earnings; comparatively few, just over 4,500 of them, used the SEIE work incentive. Similarly, despite the rise in VR applications, use of Section 301 continuations among youths aged 18–21 (the age group most likely to use them) declined after WIOA in aggregate, although youths in states with higher pre-ETS access ratios during 2017–2021 were more likely to use them. Youths can use Section 301 continuations to restore SSI eligibility after payments cease because of an age-18 redetermination, as long as they use VR or similar services.⁸ VR agency staff could direct young clients with SSI payment cessations to work-incentives counseling to increase use of Section 301 continuations.

A final point to consider is that states varied widely toward the latter part of our observation period in their pre-ETS access ratios and their VR application, SEIE use, and employment rates. Applications for VR services among young SSI recipients in 2019, for example, ranged across states from 0.5 percent (New Hampshire) to 7.9 percent (North Dakota). In 2019, if all states had VR application rates similar to the state at the 90^th percentile, an additional 22,000 young SSI recipients would have applied for VR services. In other words, VR agencies would have received 72 percent more applications from this group (and additional potential SSA reimbursement for their associated costs).

Although access to pre-ETS emerged and VR service use increased after WIOA, did they lead to better employment outcomes? Our descriptive findings cannot answer that question, but hint at a positive relationship. First, increases in pre-ETS access ratios (that is, more access to pre-ETS) were associated with higher percentages of recipients with any earnings (but not with higher average annual earnings). Second, increases in the rate of recipients with any earnings, along with the amount of annual earnings, were greater for youths aged 16–17 than for other age groups. Third, after WIOA, the increases in employment rates, annual earnings amounts, and SEIE use were larger among states with high pre-ETS access ratios than in states with low pre-ETS access ratios.

Our findings should be interpreted in light of three limitations. First, our analyses are descriptive. We cannot attribute a causal connection between WIOA or pre-ETS access ratios and the outcomes for the study population.

Second, we cannot directly observe student use of pre-ETS. The RSA-911 Case Service Reports include data on pre-ETS use beginning in 2017, with linkable records only for those who applied for VR services. Our pre-ETS access ratio, which compares the number of students using pre-ETS with the number of high school students receiving special education services, approximates a young person's access to services. Both numbers have potential biases. VR agencies may have underreported the number of students using pre-ETS, particularly in the first reporting years (2017 and 2018) as they adjusted their data management systems to accommodate changes in their reporting to RSA; during this time, reporting bias may have been more prominent for some agencies than others. High school students receiving special education services represent a subset of the population affected by WIOA, excluding college students with disabilities as well as students with disabilities who do not use special education services or who have Rehabilitation Act Section 504 special education plans. The latter group is important for this study, as around one-quarter of young SSI recipients do not use special education services (Rupp and others 2005/2006; Wittenburg and Loprest 2007).

Third, the COVID-19 pandemic confounded the final 2 years of our observation period. In addition to its broad effects on public health and economic outcomes, the pandemic suppressed VR use, earnings, and SSI work incentive use for our sample in 2020 and 2021.

This study is the first to document the potential influence of WIOA on a group of youths with disabilities who have substantial employment barriers. After WIOA, more of these youths applied for VR services, signed an IPE that would allow broader access to services beyond pre-ETS, and had earnings. Greater pre-ETS availability in a state (as evidenced by its pre-ETS access ratio) was associated with higher rates of signed IPEs, employment, and use of Section 301 continuations. These outcomes are expected, as they are the law's intended effects. Two research avenues using these data could explore other aspects of WIOA's influence. First, the data could be used to identify the specific connections between young SSI recipients, pre-ETS and VR service use, and employment outcomes, to understand the effectiveness of pre-ETS and whether its use results in decreased reliance on SSI. Second, analyses could consider differential access to VR services and outcomes by young people's characteristics, particularly their disability type and race and ethnicity, and the influence of social determinants of health on these relationships.

Appendix A