Effect of Medicaid Expansion on Health Insurance for Low-Income Nursing Home Aides

Abstract

We examine how the Affordable Care Act Medicaid expansion affected the insurance coverage and the sources of coverage among low-income nursing home aides using the 2010–2019 American Community Survey data. Insurance coverage for low-income nursing home aides increased from about 60% to nearly 90% in expansion states but rose to only about 80% in nonexpansion states. Using a difference-in-differences regression design, we find that Medicaid expansion was associated with a 5.1 percentage-point increase in overall insurance coverage. Expansion states had a 12.2 percentage-point gain in Medicaid that was partially offset by a 6.4 percentage-point reduction in private insurance coverage. Our results show that ACA Medicaid expansion increased insurance coverage for low-income nursing home aides; however, there was substantial crowd-out of private insurance coverage in this population. Policymakers should consider expanding Medicaid while incentivizing affordable private health insurance options for low-income nursing home aides to improve insurance coverage.

Keywords

Affordable Care Act Medicaid nursing aides nursing homes health insurance

What this paper adds

• We find that ACA Medicaid expansion increased Medicaid coverage substantially among nursing home aides.

• However, our findings suggest that overall insurance coverage increase is much lower because of crowd-out of private insurance coverage.

Applications of study findings

• Since there is substantial crowd-out of private insurance, policymakers at the national and state level should ensure that private insurance is generous and affordable to nursing home aides.

• To increase insurance coverage among nursing home aides, states should consider expanding Medicaid.

Introduction

Nursing aides, also known as Certified Nursing Assistants or Aides (CNAs), make up over a third of the nursing home workforce (PHI, 2019) and take care of over 1.3 million vulnerable residents. Health insurance is essential to the wellbeing, work safety, and productivity of nursing home aides. Insurance also plays an important role in worker recruitment and retention. However, aides in nursing homes are less likely to have employer-sponsored insurance and the uninsured rate has been historically higher than the general workforce. For more than two decades, policymakers have been concerned about low insurance coverage among nursing home aides (United States General Accounting Office, 2001; US Department of Health and Human Services, 2004; Squillace et al., 2009).

We need policies to improve the welfare of nursing home workers if we want to improve the quality of care in nursing homes. Studies have suggested that high staffing level and low turnover rate of nursing home aides are key to better quality of care (Bostick et al., 2006; Castle & Engberg, 2007; Hyer et al., 2011; Lerner et al., 2014; Morley, 2014). During the COVID-19 pandemic, higher CNA staffing level were associated with a lower probability of an outbreak and fewer deaths (Gorges & Konetzka, 2020). However, the COVID-19 pandemic has highlighted and exacerbated the vulnerability of this essential workforce. Nursing home work was considered one of the most dangerous jobs during the pandemic (McGarry et al., 2020). As of June 5, 2022, more than 1.1 million nursing home staff have tested positive for COVID-19 and over 2400 staff have died from the virus (Center for Medicare & Medicaid Services (CMS), 2022). There was also a huge decline in the employment of CNAs during the pandemic (Buerhaus et al., 2022). The demand for nursing aides is projected to grow 8% by 2030 and the nursing home industry will face challenges to fill job openings as many workers may transfer to different occupations or exit the labor force (Bureau of Labor Statistics, 2022).

Nursing home aides perform a physically and emotionally demanding job and are at greater risk for work-related injuries. CNAs provide 80%–90% of the direct care to residents, including assistance with activities of daily living such as bathing, feeding, and dressing (Wunderlich et al., 1996). CNAs are over three times more likely to experience workplace injury involving days away from work than workers overall (284 per 10,000 workers vs. 90 per 10, 000 workers in 2019) (U.S. Bureau of Labor Statistics, 2020). They are also exposed to infectious diseases and hazardous drugs as well as psychosocial hazards such as burnout and violence/assault by patients or co-workers (Walton & Rogers, 2017). Compared to other healthcare and clerical workers, CNAs have poorer self-reported physical and mental health (Silver et al., 2020). Yet, about 17% of CNAs were uninsured between 2013 to 2016 and 21% reported forgone doctor visit in the past 12 months because of cost (Silver et al., 2020).

Nursing home aides have limited access to affordable health insurance. Compared to other health care workers such as hospital aides, nursing homes aides are financially more vulnerable (Muench et al., 2021). In 2019, nursing home aides earned an average of only $14 per hour (PHI, 2022). About 12% of CNAs lived in a household with total income below the federal poverty line and 13% did not have health insurance (PHI, 2022). Over 40% of the uninsured nursing home aides do not take up employer-sponsored insurance (ESI) because of high premiums (Squillace et al., 2009). Employers often require full-time employment and impose a waiting period for coverage eligibility, but many nursing home aides work part-time (PHI, 2019) and their turnover rate can exceed 100% (Gandhi et al., 2021), potentially making them ineligible for ESI.

Public health insurance is an important source of coverage for low-income nursing home aides and expanding public insurance can reduce uninsured rate among this workforce. Prior to the Affordable Care Act (ACA) Medicaid expansion, many CNAs were not eligible for Medicaid because of the low income thresholds. Under ACA, many states expanded Medicaid coverage to non-elderly adults who live in households with income below 138% of the federal poverty guidelines (FPG). Descriptive reports suggest that nationwide uninsured rate of nursing home aides decreased from 28% in 2010 to 11% in 2017 and Medicaid coverage increased from 16% in 2010 to 25% in 2017 (Campbell, 2017; PHI, 2019). Many studies have found substantial gain in insurance coverage among different populations in Medicaid expansion states (Guth et al., 2020). However, we do not know the extent to which Medicaid expansion helped increase insurance coverage among nursing home aides. Since ACA also established federally subsidized plans in health care marketplaces for individuals with household income between 100% to 400% of FPG, ACA can affect overall insurance as well as sources of coverage in expansion and nonexpansion states differently.

Although Medicaid expansion is likely to increase insurance coverage among nursing home aides, there are concerns about crowd-out of private insurance coverage (Gruber & Simon, 2008) as well as workforce shortages. Crowd-out occurs when previously insured individuals drop their private insurance and enroll in Medicaid. Medicaid can be an appealing option because it provides a rather comprehensive coverage with very low or zero cost. Moreover, public insurance coverage reduces job lock by separating insurance from employment. Individuals who were locked into their jobs to keep insurance might alter their labor supply behaviors after they became Medicaid eligible (Baker, 2015). For instance, individuals who experience disutility from work can reduce their labor supply by decreasing work hours or withdrawing from labor force. Increased job mobility also motivates people to re-sort to other jobs after obtaining Medicaid. While these outcomes may be good for workers, they create strain on an industry that is already short on workers. On the other hand, crowd-out may not happen because of the stigma associated with public programs, barriers to Medicaid enrollment, and/or concerns of access to providers (Cutler & Gruber, 1996). A descriptive report showed that among non-elderly adult workers with household income below 250% of FPG, ESI coverage fell from 40% in 2013 to 39% in 2014 in states that expanded Medicaid while ESI rose from 41% to 45% in states that did not expand Medicaid, suggesting that Medicaid expansion might have crowded out ESI in expansion states (Williamson et al., 2016). Finally, increased demand for services from millions of newly insured individuals as well as increased administrative burden due to ACA can lead to workforce shortages and adversely affect access to services (Anderson, 2014).

Several studies investigating the effect of Medicaid expansion have found little to no evidence of crowd-out of private insurance among the general population (Courtemanche et al., 2017; Kaestner et al., 2017). However, studies have found significant crowd-out effects in certain demographic groups such as low-educated young adults aged 19–26 years and minority women(Wehby & Lyu, 2018), or low-income unmarried older parents (Kaestner et al., 2017). These groups were more likely to be uninsured or underinsured prior to the Medicaid expansion because of the lower income threshold and categorical criteria (e.g., pregnant women, parents with dependent children). Some of previously uninsured individuals, in the absence of Medicaid expansion, would have obtained private coverage. Individuals who were underinsured by their previous private options might shift to Medicaid once they became eligible. Compared to part-time workers in nonexpansion states, part-time workers in Medicaid expansions states had substantial gain in Medicaid coverage that was offset by reduced take-up of ESI (Berdahl & Moriya, 2018). Among farm workers, researchers found no significant crowd-out of private health insurance following Medicaid expansion (Kandilov & Kandilov, 2022). Among low-income artists, crowd-out of private insurance was large enough to completely offset the gain in Medicaid (Woronkowicz et al., 2020). The unique demographic profile (e.g., predominantly female, nonwhite majority) and job characteristics (e.g., low-pay, limited benefits, part-time, and unappealing work environment) of nursing home aides are contributing factors for potentially large crowd-out of private insurance. Further, previous studies using earlier years of data are likely to underestimate the crowd-out effect because it may take time for individuals to shift from private coverage to Medicaid.

In this study, we seek to fill this knowledge gap by using more recent years of data and estimating the effect of Medicaid expansion on overall insurance coverage as well as the crowd-out of private insurance among low-income nursing homes aides.

Method

Data and Sample

We use annual ACS data from 2010 to 2019 accessed via Integrated Public Use Microdata Series (IPUMS) (Ruggles et al., 2021). ACS is a repeated cross-sectional survey conducted by the US Census Bureau. The average response rate is 94%, ranging from 86% to 97.5% over the 10-year period (ACS office, n.d.). ACS collects information on demographic and socioeconomic characteristics of individuals and households including the state of residence, health insurance coverage, occupation and industry, and employment status. With respect to insurance, ACS has information on not only whether an individual has insurance coverage but also the type of insurance coverage at the time of interview. Several ACA evaluation studies have used the ACS data to examine health insurance coverage (Courtemanche et al., 2017; Duggan et al., 2019; Kaestner et al., 2017; Leung & Mas, 2018; Stimpson et al., 2019; Wehby & Lyu, 2018), but only a few have taken advantage of the ACS data to examine the impact of ACA Medicaid expansion for a specific occupation group that is likely to have differential impact compared to the general population (DiNardi, 2021; Woronkowicz et al., 2020).

Consistent with the literature (Kelly et al., 2020), our primary sample includes currently employed nurse aides (occupation code: 3600) working in the nursing home industry (industry code: 8270). We restrict our sample to adult (19 years ≤ age ≤ 64 years) nursing home aides who live in a household with income below 138% of FPG since these individuals are likely to be affected by Medicaid expansion. We adopted the methods developed by the State Health Access Data Assistance Center (SHADAC) to determine the poverty levels (SHADAC, 2021). In sensitivity analyses, we investigated nursing home aides by income levels as well. We exclude a small portion of the sample (about 0.8%) that is covered by Medicare. Our final sample includes 17,111 nursing home aides from 50 states and District of Columbia.

Measurement

Our key outcomes are health insurance coverage status and the sources of health insurance at the time of survey. We created three binary indicators for any insurance coverage, Medicaid, and private insurance (including ESI or self-purchased insurance).

The key independent variable is a binary indicator for the status of Medicaid expansion in each state by the end of 2019. We follow existing literature (Kaestner et al., 2017; Wehby & Lyu, 2018) and classify eligible nursing home aides from 20 states that expanded Medicaid in January 2014 and 9 states that expanded Medicaid between 2014 and 2019 in the treatment group. Nursing home aides from 17 states that did not expand Medicaid and 5 states that had full expansion prior to the ACA serve as controls (see Supplementary Table 1).

Consistent with existing literature (Courtemanche et al., 2017; Kaestner et al., 2017), we include individual-level characteristics (i.e., age, sex, race/ethnicity, citizenship, rural, marital status, number of children in the household, and education) and state unemployment rate as controls.

Statistical Analysis

We employ a difference-in-differences (DD) framework to examine the effect of the Medicaid expansion on insurance coverage among low-income nursing home aides. We use separate linear probability models to evaluate the changes in any insurance coverage, Medicaid insurance coverage, and any private insurance coverage (Equation (1))

Y_{i s t} = β (E x p a n s i o n_{s} \times P o s t_{t}) + λ X_{i s t} + θ U n e m p_{s t} + γ S t a t e_{s} + δ Y e a r_{t} + ϵ_{i s t}

(1)

Where

Y_{i s t}

is the outcome for individual

i

in state

s

in year.

E x p a n s i o n_{S}

is a binary indicator for states that expanded Medicaid (equals 1 for treatment group and 0 for control group).

P o s t_{t}

is a binary indicator for post-expansion time period (equals 1 for years with expansion and 0 for years without expansion). The coefficient of interest is

β

which captures the effect of Medicaid expansion.

X_{i s t}

is a vector of individual controls described under previous section.

U n e m p_{s t}

is state annual unemployment rate.

S t a t e_{s}

and

Y e a r_{t}

represent state and year fixed effects, respectively. Standard errors are clustered at the state level. All analyses account for the survey sampling weights.

To assess the validity of our findings, we conduct two additional analyses. First, to assess the parallel trends assumption of DD analysis and to explore how the effect of Medicaid expansion evolved over time, we estimate an event study model, interacting the expansion indicator with a set of relative year dummies (Equation (2))

Y_{i s t} = E x p a n s i o n_{s} \times \sum_{\begin{array}{c} y = - 5 \\ y \neq - 1 \end{array}}^{5} β_{y} I (t - t_{s}^{*} = y) + λ X_{i s t} + θ U n e m p_{s t} + γ S t a t e_{s} + δ Y e a r_{t} + ϵ_{i s t}

(2)

Where indicator variable

I (t - t_{s}^{*} = y)

measures the time relative to the year of Medicaid expansion (

t_{s}^{*}

). The omitted category is y = −1, the year prior to the expansion. Distant periods beyond

y = - 5

y = 5

are accumulated to the coefficients associated with

y = - 5

y = 5

. For states that did not expand Medicaid, the interaction terms are 0 in all periods. Each estimate of

β_{y}

provides the average treatment effect during period

y

relative to the reference period (1 year before the expansion). Specifically, coefficients

β_{0}

β_{5}

capture the immediate and lagged effects while coefficients

β_{- 5}

β_{- 2}

capture the lead effects prior to the expansion. A joint test of significance for

β_{- 5}

β_{- 2}

can show us whether insurance coverage trended in parallel between expansion and nonexpansion states.

Second, we conduct a placebo test among the subsample of nursing home aides with income at or above 251% of FPG. We chose a higher threshold of 251% of FPG instead of 139% of FPG to conduct the placebo test because there might be income related measurement errors for near-poor individuals whose income fell between 139% to 250% of the FPG. Nursing home aides whose income exceeds 250% of FPG were less likely to be affected by Medicaid expansion and we should observe no Medicaid expansion effect in this population.

We also conduct several robustness checks across different samples and model specifications. First, to investigate the source of changes in private insurance coverage, we examine ESI and self-purchased insurance separately. Second, time-varying covariates can be potentially endogenous to Medicaid expansion, that is, variables such as marital status, number of children, education, and unemployment might change following ACA, which leads to an over controlling issue (Courtemanche et al., 2017). So, in an additional sensitivity, we control for only age, sex, race/ethnicity, and citizenship. Third, we account for differential linear trends in states by including state-specific linear time trends. Fourth, we exclude prior expansion states because they might not be ideal control states. We also exclude both later expansion and prior full expansion states from our analysis to examine the effect of Medicaid expansion for states that expanded Medicaid in January 2014. Finally, the effect of Medicaid expansion can be different based on location and income levels. We also examine Medicaid expansion effect on health insurance coverage by rural/urban location and by income level.

Results

Our analytical sample includes 17,111 nursing home aides with income below 138% of FPG. Table 1 summarizes the characteristics of nursing home aides, overall and by Medicaid expansion status. The average age of nursing home aides was about 31.7, 92% were female, 38% were non-Hispanic black, and 11% were Hispanic. About 63% of nursing home aides were never married, and approximately 60% had high school education or less. Most of the characteristics were similar between low-income nursing home aides living in Medicaid expansion states and nonexpansion states. However, nursing home aides in expansion states were slightly younger (30.9 vs. 32.5), had a lower proportion of non-Hispanic black (33.0% vs. 43.6%), and were less likely to live in rural areas (40.3% vs. 47.8%).

Table 1.

Summary Characteristics of Low-income Nursing Assistants at or Below 138% Federal Poverty Guidelines in Nursing Home Industry by Medicaid Expansion Status.

SampleCharacteristics		Expansion Status
SampleCharacteristics	Full Sample	Expansion states	Nonexpansion states
Age	31.7	30.9	32.5
Female	91.9	91.2	92.8
Race/ethnicity, %
Non-hispanic white	45.1	48.8	40.9
Non-hispanic black	38.0	33.0	43.6
Non-hispanic other	5.9	7.2	4.5
Hispanic	11.0	11.0	11.0
Non-citizen, %	6.0	5.8	6.3
Rural, %	43.8	40.3	47.8
Marital status, %
Married	16.3	15.5	17.1
Separated, divorced, or widowed	20.8	18.8	23.1
Never married	62.9	65.7	59.8
Number of children	1.34	1.29	1.41
Education, %
Less than high school	9.2	8.0	10.6
High school	49.5	48.6	50.5
Some College	37.3	38.8	35.5
College	4.0	4.6	3.4
State annual unemployment rate, %	6.3	6.4	6.2
Number of observations	17,111	9022	8089

Notes. FPG indicates Federal Poverty Guideline. Analyses incorporated American Census Survey weights. Rural location was classified if the population density of the PUMA of residence was fewer than 500 persons per square mile.

Figure 1 shows the average unadjusted trends in any insurance coverage and different sources of coverage among low-income nursing home aides in states by Medicaid expansion status. Prior to 2014, states that eventually expanded Medicaid had higher rates of any insurance and Medicaid coverage compared to states that never expanded Medicaid. However, private insurance coverage was similar in both groups of states. Starting in 2014, states with Medicaid expansion had a substantial increase in any insurance coverage driven largely by an increase in Medicaid coverage.

Figure 1.

Unadjusted trends in insurance coverage among low-income nursing home aides below 138% federal poverty guidelines by Medicaid expansion status. Source: American Community Survey (2010–2019). Notes: Expansion states include 20 states that expanded Medicaid in January 2014 without prior full expansions and 9 later expansion states. Nonexpansion states include 17 states that did not expand Medicaid by the end of 2019 and 5 states that fully expanded Medicaid before January 2014 (details on state Medicaid expansion status can be found in Supplementary Table 1).

DD estimates show that any insurance coverage among low-income nursing home aides increased by 5.1 percentage points (p < .10) in states that expanded Medicaid compared to states that did not expand Medicaid (Table 2). This increase in any insurance coverage was largely driven by a 12.2 percentage-point gain (p < .01) in Medicaid coverage that was partially offset by a 6.4 percentage-point decrease (p < .01) in private insurance coverage. Our results appear valid based on our assessment of parallel trends assumption using event studies and placebo tests using a sample of nursing home aides with income above 251% of FPG. Event study results reveal no significant differences in trends in insurance coverage including different sources of coverage in the years leading to Medicaid expansion between states with and without Medicaid expansion (Table 2). The positive effect of Medicaid expansion on Medicaid coverage started in the year of implementation and remained significant over the years even though the magnitude decreased over time. When we examine the effect of Medicaid expansion on insurance coverage among nursing home aides with income above 251% FPG, we found no meaningful difference in Medicaid or other coverage types and this was expected because these individuals have income above Medicaid eligibility and should not be affected by Medicaid expansion (Table 3, Panel C).

Table 2.

Difference-in-Differences and Event Study Estimates of the Effect of Medicaid Expansion on Insurance Coverage among Low-income Nursing Assistants at or Below 138% Federal Poverty Guideline in Nursing Home Industry.

	Any insurance	Medicaid	Any private insurance
Average effect
Expansion × Post	0.051^* (0.027)	0.122^*** (0.025)	−0.064^*** (0.023)
Event study
Year -5 (and beyond)	−0.041 (0.050)	−0.02 (0.028)	−0.029 (0.060)
Year -4	−0.011 (0.028)	0.00 (0.035)	−0.014 (0.032)
Year -3	0.003 (0.023)	−0.024 (0.021)	0.025 (0.029)
Year -2	−0.001 (0.029)	−0.028 (0.022)	0.014 (0.030)
Year -1 (Reference)	0	0	0
Year 0 (Year of expansion)	0.02 (0.036)	0.050^* (0.029)	−0.027 (0.026)
Year 1	0.060^* (0.033)	0.150^*** (0.036)	−0.091^** (0.041)
Year 2	0.053 (0.044)	0.139^*** (0.041)	−0.068 (0.048)
Year 3	0.082^** (0.034)	0.145^*** (0.030)	−0.067 (0.044)
Year 4	0.053 (0.042)	0.107^*** (0.038)	−0.066 (0.049)
Year 5 (and beyond)	0.04 (0.034)	0.086^* (0.045)	−0.035 (0.049)
p-value from joint significance test for years prior to expansion	0.93	0.57	0.66
Number of observations	17,111	17,111	17,111

Notes. Robust standard errors clustered at the state level are presented in parentheses. All regressions controlled for age, sex, race/ethnicity, citizenship, rural, marital status, number of children, education, state annual employment rate, state and year fixed effects. Analyses accounted for American Census Survey sample weights. FPG indicates Federal Poverty Guideline.

*p < .10, **p < .05, *** p< .01.

Table 3.

Difference-in-Differences Estimates of the Effect of Medicaid Expansion on Insurance Coverage among Nursing Home Aides: Alternative Sample Definitions.

	Any insurance	Medicaid	Any private insurance
Panel A:
Exclude prior full expansion
Expansion × Post	0.039 (0.027)	0.119^*** (0.026)	−0.076^*** (0.022)
Number of observations	15,857	15,857	15,857
Exclude later expansion and prior full expansion
Expansion × Post	0.036 (0.028)	0.139^*** (0.027)	−0.096^*** (0.026)
Number of observations	12,930	12,930	12,930
Panel B: By rural/urban
Rural
Expansion × Post	0.019 (0.028)	0.153^*** (0.031)	−0.125^*** (0.027)
Number of observations	8857	8857	8857
Urban
Expansion × Post	0.074^** (0.036)	0.101^*** (0.029)	−0.022 (0.028)
Number of observations	8254	8254	8254
Panel C: Nursing Home Aides with Different income levels
0–100% FPG
Expansion × Post	0.047 (0.031)	0.132^*** (0.033)	−0.087^*** (0.024)
Number of observations	10,273	10,273	10,273
251% FPG and above
Expansion × Post	0.006 (0.013)	0.018 (0.011)	−0.009 (0.019)
Number of observations	17,005	17,005	17,005
All income level
Expansion × Post	0.027 (0.018)	0.073^*** (0.016)	−0.041^** (0.018)
Number of observations	52,289	52,289	52,289

Notes. Panel A and B use sample of nursing home aides with income below 138% FPG. Robust standard errors clustered at the state level are presented in parentheses. All regressions controlled for age, sex, race/ethnicity, citizenship, rural, marital status, number of children, education, state annual employment rate, state and year fixed effects. All analyses accounted for American Census Survey sample weights. FPG indicates Federal Poverty Guideline.

*p < .10, **p < .05, ***p < .01.

Sensitivity analysis using the components of private insurance suggested that crowd-out of both ESI and self-purchased insurance were responsible for the reduction in private insurance in states with Medicaid expansion. Specifically, we observed 4.3 percentage points reduction in ESI (p < .10) and 2.9 percentage points reduction in self-purchased insurance (p < .01) among low-income nursing home aides in states with Medicaid expansion compared to those in states without Medicaid expansion (see Supplementary Table 2).

The results are robust to different model specifications. The effect of Medicaid expansion on insurance coverage is similar when we control for only time-invariant individual demographic characteristics or include state-specific linear time trends in the model (see Supplementary Table 3).

Although the effects are in the same direction, we observe some differences in the magnitude of the effect of Medicaid expansion on insurance coverage by rurality, income levels, and exclusion of later expansion or prior full expansion states (Table 3). First, when we drop prior expansion states from the control group, the effect of Medicaid expansion on Medicaid is close to the main model (11.9 percentage points). When we exclude later expansion states from the treatment group and prior expansion states from the control group, the effects of Medicaid expansion on Medicaid and private insurance were both greater than main estimates. The effect of Medicaid expansion on Medicaid coverage was greater among low-income nursing home aides in rural areas (15.3 percentage points, p < .01), but the crowd-out of private insurance coverage was also greater (12.5 percentage points, p < .01) which resulted in a small and insignificant increase in any coverage. By contrast, the crowd-out of private insurance in urban areas was small and insignificant (2.2 percentage points, p > .10) (Table 3, Panel B). Finally, when we examined all nursing home aides irrespective of income, the increase in Medicaid coverage was smaller and there was no significant difference in the overall insurance rate; the results still suggest a larger increase for Medicaid than the decrease for private insurance (Table 3, Panel C).

Discussion

Using ACS data from 2010 to 2019, we found that Medicaid coverage among low-income nursing home aides increased by 12.2 percentage points in states with Medicaid expansion compared to states without Medicaid expansion; however, there was a 6.4 percentage-point decrease in private insurance in the Medicaid expansion states. There was a net gain of 5.1 percentage points in overall health insurance coverage for low-income nursing home aides in Medicaid expansion states compared to states without Medicaid expansion. Our findings are robust to different samples and model specifications.

Our findings of substantial crowd-out of private insurance following Medicaid expansion among low-income nursing home aides are consistent with the existing literature. Several studies have found large, significant insurance crowd-out effects of Medicaid expansions among certain subpopulations such as young minority females (Wehby & Lyu, 2018), unmarried older parents (Kaestner et al., 2017), and occupations that lack affordable options for generous, private insurance (Woronkowicz et al., 2020). Even though some studies suggest that Medicaid mostly crowded out individually purchased insurance rather than ESI in certain demographic groups (Gruber & Simon, 2008; Leung & Mas, 2018; Wehby & Lyu, 2018), our study finds that Medicaid expansion crowded out ESI to a greater extent than individually purchased insurance among currently employed low-income nursing homes aides.

Subgroup analyses show that nursing home aides with income below 100% FPG and those in rural areas contributed to the crowd-out of private insurance associated with Medicaid expansion. In one of the studies that examined artists with income below 100% FPG, the magnitude of crowd-out of private insurance was large enough to fully offset the gain in Medicaid (Woronkowicz et al., 2020). Another study found a shift from individually purchased insurance to Medicaid coverage following Medicaid expansion among low-income rural populations resulting in no significant change in overall insurance (Soni et al., 2017). Medicaid represents a more generous coverage option and gives workers more flexibility in terms of job mobility (Kaestner et al., 2017).

Despite the methodological rigor employed in this study, there are several limitations. First, ACS is a census-based data and may not capture all occupation-specific information. However, ACS data has been extensively used to study nursing workforce as well as direct care workforce in different care settings including nursing homes (Buerhaus et al., 2017; DiNardi, 2021; Kelly et al., 2020). Second, ACS data is a repeated cross-sectional survey of individuals instead of a panel. Ideally, we would want to examine the effect of Medicaid expansion on individuals over time but we are limited by the data. Third, ACS may have fewer rural respondents because ACS 1-year data has information for areas with population of 65,000 or more only (US Census Bureau, n.d.). Depending on the insurance coverage change of low-income nursing assistants in small rural areas in both expansion and nonexpansion states, our estimates might have upward or downward bias. Finally, there can be reporting and measurement errors for outcomes in surveys and ACS is no exception.

Our findings of increased insurance coverage as well as substantial crowd-out of private insurance following Medicaid expansion have three key policy implications. First, despite the crowd-out, there is a net gain in insurance coverage among low-income nursing home aides. In nonexpansion states, the uninsured rate among currently employed, low-income nursing home aides is still considerably high (22% vs. 11% in expansion states as of 2019). Policymakers in nonexpansion states should either consider expanding Medicaid or provide incentives so that there are affordable private insurance options for low-income nursing home aides. Second, the crowd-out effect of Medicaid expansion is strongest among nursing home aides living under 100% FPG or living in rural areas, suggesting a lack of affordable private coverage options among nursing home aides living under poverty and those living in rural areas. To address the Medicaid crowd-out of private insurance, it is important to identify what drives insurance take-up decisions for low-income nursing aides. We need to understand whether private insurance plans are unavailable or unaffordable. Nursing homes that are heavily reliant on Medicaid reimbursement may opt to provide wages over insurance. Increasing Medicaid reimbursement rates with an employer mandate for affordable insurance offer is a potential policy to improve the availability of ESI. If the ESI is available but practically unaffordable for nursing assistants, lowering the federal definition of affordable ESI (currently set as 9.61% of the employee’s monthly household income) can increase the take-up of ESI. An additional benefit of ESI is its “job lock” effect, which potentially can reduce turnover rate (Baker, 2015). Similarly, increasing marketplace subsidies for nursing homes aides can improve the affordability of marketplace plans. As the nursing home industry is facing severe staffing shortage, policymakers can consider making permanent enhanced subsidies under the American Rescue Plan Act for nursing home aides to stabilize the workforce.

Third, before Medicaid expansion, private insurance coverage rates were about the same for low-income nursing home aides in expansion states and those in nonexpansion states. But after Medicaid expansion, low-income nursing home aides in expansion states are more likely to have Medicaid whereas those in nonexpansion states are more likely to have private insurance coverage. Different sources of coverage type have potential implication for labor supply behavior. For nursing home aides who experienced “job lock,” Medicaid potentially loosens the “job lock” and allows them to have more flexibility in work schedules (e.g., full-time to part-time) and job choices. However, as an income-based program, Medicaid can disincentivize work for those who would like to earn more income (Altig et al., 2020). To meet the income eligibility criteria of Medicaid, nursing home aides might work fewer hours and/or accept lower wages. Given the recent efforts to increase wages for nursing home aides, there are concerns that more nursing home aides, especially for those in the nonexpansion states, might fall into this coverage gap where they earn too much to qualify for Medicaid but they do not earn enough to purchase private insurance.

In conclusion, Medicaid expansion has been instrumental in increasing insurance coverage for low-income nursing home aides. While higher pay and affordable and adequate private insurance options for nursing home aides are necessary to strengthen the workforce, policymakers should also consider other reforms to Medicaid eligibility requirements to make transitions to private insurance smoother and to ensure that any efforts to increase wages for nursing home aides do not have unintended consequences.

Supplemental Material

Supplemental material - Effect of Medicaid Expansion on Health Insurance for Low-Income Nursing Home Aides

Supplemental material for Effect of Medicaid Expansion on Health Insurance for Low-Income Nursing Home Aides by Lili Xu and Hari Sharma in Journal of Applied Gerontology

Footnotes

Acknowledgments

The author(s) would like to express sincere thanks to the two anonymous reviewers whose comments/suggestions helped improve and clarify this manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Lili Xu

Hari Sharma

Supplemental Material

Supplemental material for this article is available online.

References

ACS Office . (n.d.). American Community Survey Response Rates. Retrieved January 4, 2022, from https://www.census.gov/acs/www/methodology/sample-size-and-data-quality/response-rates/

Altig

Auerbach

A. J.

Kotlikoff

L. J.

Ilin

(2020). Marginal net taxation of americans’ labor supply. National Bureau of Economic Research.

Anderson

(2014). The impact of the Affordable Care Act on the health care workforce. The Heritage Foundation.

Baker

(2015). Job lock and employer-provided health insurance: Evidence from the literature. AARP Real Possibilities Public Policy Institution.

Berdahl

T. A.

Moriya

A. S.

(2018). Difference in uninsurance rates between full- and part-time workers declined in 2014. Health Affairs, 37(10), 1669–1672. https://doi.org/10.1377/hlthaff.2018.0345

Bostick

J. E.

Rantz

M. J.

Flesner

M. K.

Riggs

C. J.

(2006). Systematic review of studies of staffing and quality in nursing homes. Journal of the American Medical Directors Association, 7(6), 366–376. https://doi.org/10.1016/j.jamda.2006.01.024

Buerhaus

P. I.

Skinner

L. E.

Auerbach

D. I.

Staiger

D. O.

(2017). State of the registered nurse workforce as a new era of health reform emerges. Nursing Economic, 35(5), 229–237.

Buerhaus

P. I.

Staiger

D. O.

Auerbach

D. I.

Yates

M. C.

Donelan

(2022). Nurse Employment during the first fifteen months of the COVID-19 pandemic: Study examines nurse employment trends during first fifteen months of the COVID-19 pandemic. Health Affairs, 41(1), 79–85. https://doi.org/10.1377/hlthaff.2021.01289

Bureau of Labor Statistics . (2022). Occupational outlook handbook: Nursing assistants and orderlies. BLS. https://www.bls.gov/ooh/healthcare/nursing-assistants.htm

10.

Campbell

(2017). The impact of the affordable care Act on health coverage for direct care workers (16). Phinational.

11.

Castle

N. G.

Engberg

(2007). The influence of staffing characteristics on quality of care in nursing homes. Health Services Research, 42(5), 1822–1847. https://doi.org/10.1111/j.1475-6773.2007.00704.x

12.

Center for Medicare & Medicaid Services . (2022). COVID-19 nursing home Dataâ€”Centers for Medicare & Medicaid services data. CMS. https://data.cms.gov/covid-19/covid-19-nursing-home-data

13.

Courtemanche

Marton

Ukert

Yelowitz

Zapata

(2017). Early impacts of the affordable care act on health insurance coverage in Medicaid expansion and non-expansion states. Journal of Policy Analysis and Management, 36(1), 178–210. https://doi.org/10.1002/pam.21961

14.

Cutler

D. M.

Gruber

(1996). Does public insurance crowd out private insurance. The Quarterly Journal of Economics, 111(2), 391–430. https://doi.org/10.2307/2946683

15.

DiNardi

(2021). Affordable Care Act Medicaid expansions and the nurse labor market. Southern Economic Journal, 88(1), 367–398. https://doi.org/10.1002/soej.12499

16.

Duggan

Goda

G. S.

Jackson

(2019). The effects of the affordable care act on health insurance coverage and labor market outcomes. National Tax Journal, 72(2), 261–322. https://doi.org/10.17310/ntj.2019.2.01

17.

Gandhi

Grabowski

D. C.

(2021). High nursing staff turnover in nursing homes offers important quality information. Health Affairs, 40(3), 384–391. https://doi.org/10.1377/hlthaff.2020.00957

18.

Gorges

R. J.

Konetzka

R. T.

(2020). Staffing levels and COVID-19 cases and outbreaks in US nursing homes. Journal of the American Geriatrics Society, 68(11), 2462–2466, https://doi.org/10.1111/jgs.16787

19.

Gruber

Simon

(2008). Crowd-out 10 years later: Have recent public insurance expansions crowded out private health insurance? Journal of Health Economics, 27(2), 201–217. https://doi.org/10.1016/j.jhealeco.2007.11.004

20.

Guth

Garfield

Rudowitz

(2020). The effects of Medicaid expansion under the ACA: Updated findings from a literature review (100). KFF.

21.

Hyer

Thomas

K. S.

Branch

L. G.

Harman

J. S.

Johnson

C. E.

Weech-Maldonado

(2011). The influence of nurse staffing levels on quality of care in nursing homes. The Gerontologist, 51(5), 610–616, https://doi.org/10.1093/geront/gnr050

22.

Kaestner

Garrett

Chen

Gangopadhyaya

Fleming

(2017). Effects of ACA Medicaid expansions on health insurance coverage and labor supply. Journal of Policy Analysis and Management, 36(3), 608–642. https://doi.org/10.1002/pam.21993

23.

Kandilov

A. M. G.

Kandilov

I. T.

(2022). The impact of the affordable care act medicaid expansions on agricultural workers’ health insurance coverage, medical care utilization, and labor supply. American Journal of Agricultural Economics, 104(3), 1026–1049. https://doi.org/10.1111/ajae.12263

24.

Kelly

Craft Morgan

Kemp

C. L.

Deichert

(2020). A profile of the assisted living direct care workforce in the United States. Journal of Applied Gerontology, 39(1), 16–27. https://doi.org/10.1177/0733464818757000

25.

Lerner

N. B.

Johantgen

Trinkoff

A. M.

Storr

C. L.

Han

(2014). Are nursing home survey deficiencies higher in facilities with greater staff turnover. Journal of the American Medical Directors Association, 15(2), 102–107. https://doi.org/10.1016/j.jamda.2013.09.003

26.

Leung

Mas

(2018). Employment effects of the affordable care act Medicaid expansions. Industrial Relations: A Journal of Economy and Society, 57(2), 206–234. https://doi.org/10.1111/irel.12207

27.

McGarry

B. E.

Porter

Grabowski

D. C.

(2020). Opinion | Nursing home workers now have the most dangerous jobs in America. They deserve better. Washington Post. https://www.washingtonpost.com/opinions/2020/07/28/nursing-home-workers-now-have-most-dangerous-jobs-america-they-deserve-better/

28.

Morley

J. E.

(2014). Certified nursing assistants: A key to resident quality of life. Journal of the American Medical Directors Association, 15(9), 610–612. https://doi.org/10.1016/j.jamda.2014.06.016

29.

Muench

Jura

Spetz

Mathison

Herrington

(2021). Financial vulnerability and worker well-being: A comparison of long-term services and supports workers with other health workers. Medical Care Research and Review: MCRR, 78(5), 607–615. https://doi.org/10.1177/1077558720930131

30.

PHI . (2019). U.S. Nursing assistants employed in nursing homes. PHI. https://phinational.org/resource/u-s-nursing-assistants-employed-in-nursing-homes-key-facts-2019/

31.

PHI . (2022). PHI’s workforce data center. PHI. https://www.phinational.org/policy-research/workforce-data-center/

32.

Ruggles

Flood

Foster

Goeken

Pacas

Schouweiler

Sobek

(2021). Ipums USA: Version 11.0 [dataset]. IPUMS. https://doi.org/10.18128/D010.V11.0

33.

Silver

Boiano

(2020). Patient care aides: Differences in healthcare coverage, health-related behaviors, and health outcomes in a low-wage workforce by healthcare setting. American Journal of Industrial Medicine, 63(1), 60–73. https://doi.org/10.1002/ajim.23053

34.

Soni

Hendryx

Simon

(2017). Medicaid expansion under the affordable care act and insurance coverage in rural and urban areas: Medicaid expansion and insurance coverage. The Journal of Rural Health, 33(2), 217–226. https://doi.org/10.1111/jrh.12234

35.

Squillace

M. R.

Remsburg

R. E.

Harris-Kojetin

L. D.

Bercovitz

Rosenoff

Han

(2009). The national nursing assistant survey: Improving the evidence base for policy initiatives to strengthen the certified nursing assistant workforce. The Gerontologist, 49(2), 185–197. https://doi.org/10.1093/geront/gnp024

36.

State Health Access Data Assistance Center . (2021). Defining “family” for studies of health insurance coverage . University of Minnesota Minneapolis. https://www.shadac.org/publications/2020-update-shadacs-health-insurance-unit-and-defining-families-issue-brief

37.

Stimpson

J. P.

Kemmick Pintor

McKenna

R. M.

Park

Wilson

F. A.

(2019). Association of Medicaid expansion with health insurance coverage among persons with a disability. JAMA Network Open, 2(7), Article e197136. https://doi.org/10.1001/jamanetworkopen.2019.7136

38.

United States General Accounting Office . (2001). Recruitment and retention of nurses and nurse aides is a growing concern. In Before the committee on health, education, labor and pensions, US senate, United State General Accounting Office.

39.

US Census Bureau . (n.d.). When to use 1-year or 5-year estimates. Census.Gov. Retrieved June 22, 2022, from https://www.census.gov/programs-surveys/acs/guidance/estimates.html

40.

U.S. Bureau of Labor Statistics . (2020). Employer-reported workplace injuries and illnesses – 2019 (9). U.S. Bureau of Labor Statistics.

41.

US Department of Health and Human Services . (2004). Nursing aides, home health aides, and related health care occupations: National and local workforce shortages and associated data needs. Available at Bhpr. Hrsa. Gov/Healthworkforce.

42.

Walton

A. L.

Rogers

(2017). Workplace hazards faced by nursing assistants in the United States: A focused literature review. International Journal of Environmental Research and Public Health, 14(5). Article 544, https://doi.org/10.3390/ijerph14050544

43.

Wehby

G. L.

Lyu

(2018). The impact of the ACA Medicaid expansions on health insurance coverage through 2015 and coverage disparities by age, race/ethnicity, and gender. Health Services Research, 53(2), 1248–1272. https://doi.org/10.1111/1475-6773.12711

44.

Williamson

Antonisse

Tolbert

Garfield

, Jun 10, A. D. P., & 2016. (2016). ACA coverage expansions and low-income Workersâ€”Issue brief. KFF. https://www.kff.org/report-section/aca-coverage-expansions-and-low-income-workers-issue-brief/

45.

Woronkowicz

Soni

Freedman

Simon

(2020). How have recent health insurance expansions affected coverage among artist occupations in the USA? Journal of Cultural Economics, 44(1), 117–154. https://doi.org/10.1007/s10824-019-09352-5

46.

Wunderlich

G. S.

Sloan

Davis

C. K.

(1996). Nursing staff in hospitals and nursing homes: Is it adequate? NCBI.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.23 MB