Skip to main content
HHS Author Manuscripts logoLink to HHS Author Manuscripts
. Author manuscript; available in PMC: 2024 Nov 13.
Published in final edited form as: Health Aff (Millwood). 2021 Dec;40(12):1943–1952. doi: 10.1377/hlthaff.2021.00841

Increasing Medicaid’s Stagnant Asset Test For People Eligible For Medicare And Medicaid Will Help Vulnerable Seniors

Noelle Cornelio 1, Melissa McInerney 2, Jennifer M Mellor 3, Eric T Roberts 4, Lindsay M Sabik 5
PMCID: PMC11558658  NIHMSID: NIHMS2024119  PMID: 34871073

Abstract

Low-income Medicare beneficiaries rely on Medicaid for supplemental coverage but must meet income and asset tests to qualify. We examined states’ asset and income tests for full-benefit Medicaid from 2006–2018 and examined how alternative asset tests would affect eligibility for community-dwelling seniors. Most states have not updated the dollar limit of Medicaid’s asset test since 1989, making the asset test increasingly restrictive in inflation-adjusted terms. We estimated that increasing Medicaid’s asset limit by the Consumer Price Index, to Medicare Savings Program levels, or to $10,000 for individuals and $20,000 for couples, would increase Medicaid eligibility by 1.7%, 4.4%, and 7.5%, respectively. Simplifying Medicaid’s asset test to focus only on certain high-value assets would increase eligibility by 20.5%. Increasing asset limits would lessen restrictions on Medicaid eligibility that arise from stagnant asset tests, broadening eligibility for certain low-income Medicare beneficiaries and allowing them to retain higher, yet still modest, savings.

Keywords: Medicaid, Medicare, eligibility, low-income, aged

Background

Approximately 8.7 million low-income elderly or disabled people with Medicare are dually enrolled in full-benefit Medicaid.1 For these beneficiaries, Medicaid pays for Medicare’s premiums and cost-sharing and services not covered by Medicare (e.g., long-term care, vision, and dental), exceeding the generosity of benefits provided through the Medicare Savings Programs (MSPs), which only cover Medicare’s premiums and, in some cases, cost sharing. Medicaid provides financial protection from health care costs that might otherwise pose a substantial financial burden on low-income Medicare beneficiaries, given their age, health status, and socioeconomic vulnerability.2

Most Medicare beneficiaries receiving full Medicaid qualify through a “categorically needy” pathway, which requires them to have income and assets no higher than state-specific eligibility thresholds. These income and asset limits were unaffected by the Affordable Care Act’s Medicaid expansion and differ from Medicaid eligibility criteria for non-elderly and non-disabled adults, which are based only on income.

Although the rationale for asset tests is to limit Medicaid coverage to people who lack resources to pay for care, asset tests can present challenges for both states and Medicaid applicants. For states, asset tests can be administratively complex and expensive to enforce, as they require detailed, typically annual, assessments of Medicaid applicants’ property (excluding a primary home and vehicle), investments, and other resources. For applicants, providing this information can be burdensome and may contribute to the low take-up of Medicaid benefits (only about half of Medicare beneficiaries who meet the eligibility requirements for Medicaid through the categorically needy pathway receive Medicaid).3, 4

Moreover, Medicaid’s asset test has grown more restrictive over time. In all but three states, the asset limit is a fixed dollar threshold, and in 2018 39 states used fixed dollar thresholds based on asset limits for the Supplemental Security Income (SSI) program. Asset limits for the SSI program were last updated in 1989, when they were increased to $2,000 for individuals and $3,000 for couples, and are not adjusted for inflation.5 As a result, inflation has made asset tests increasingly restrictive. In 1989, for example, an individual with $1,500 in savings would have met the asset test in most states. By 2021, someone with the equivalent inflation-adjusted level of savings (approximately $3,200) would be ineligible. In contrast to the asset tests, most states use a Medicaid income test that is based on federal poverty guidelines (FPG) or income limits for the SSI program, both of which are indexed for inflation.

The restrictiveness of Medicaid’s asset test for Medicare beneficiaries has elicited periodic scrutiny from policy makers, but federal reforms have not been implemented for full-benefit Medicaid.68 In contrast, asset tests for MSPs are adjusted for inflation. MSP asset tests have been updated annually since 2010 when the Medicare Improvements for Patients and Providers Act (MIPPA) aligned the MSP asset tests to Medicare Part D’s low income subsidy (LIS).9 However, changes to the asset limits for the MSPs did not apply to full Medicaid.

The lack of updates to the asset limits for full Medicaid is somewhat surprising due to this population’s substantial health care needs and the importance of insurance coverage in facilitating use of care among low-income people.2, 10, 11 To inform policy, this study provides new information on the variation in full Medicaid asset and income tests for seniors (i.e., people age 65 and older) across all states and the District of Columbia over a 13-year period (2006–2018). We use detailed income and asset data from a national survey to estimate how four alternative asset tests would impact the number of community-dwelling aged Medicare beneficiaries who would qualify for full Medicaid, conditional on meeting existing income eligibility criteria. We also examine the characteristics of the potentially newly-eligible populations to inform policy makers considering changes to asset tests for full-benefit duals.

Methods

Data

We catalogued income and asset eligibility rules for full-benefit Medicaid in each state and the District of Columbia for the period 2006–2018, focusing on rules for categorically needy individuals aged 65 years and older living in the community (rather than a nursing home). We collected eligibility rules through interviews with state Medicaid program officials; from reports and databases published by the Kaiser Family Foundation (KFF)1214, American Association of Retired Persons (AARP)15, 16, Medicaid and CHIP Payment and Access Commission (MACPAC)1724, and Urban Institute25; and by searching state Medicaid websites (see the “Medicaid eligibility rules” section of the appendix for details).26 We used these data to characterize changes in states’ Medicaid eligibility rules for community-dwelling seniors from 2006–2018.

We linked state-specific eligibility data to the Health and Retirement Study (HRS), a longitudinal, nationally representative survey of approximately 20,000 adults aged 50 and older that collects extensive income and asset information in addition to demographic and health characteristics.27 We analyzed the publicly available RAND HRS, focusing on wave 13 of the survey (interviews conducted in 2016 and 2017), and used restricted state of residence data to link respondents to Medicaid eligibility rules for their state in 2016 or 2017.28 We limited analyses to respondents who were enrolled in Medicare and age 65 years or older at the time of the survey. We excluded respondents under age 65 because the HRS focuses on the elderly and near-elderly population. We excluded respondents who resided outside the 50 states and District of Columbia, and those who resided in or had a spouse residing in a nursing home. We retained observations in which a proxy responded to the survey on behalf of a subject, so as not to exclude subjects with physical or cognitive impairments that could have precluded survey participation, and conducted a sensitivity analysis that excluded proxy respondents.

Measures

We measured income following Medicaid’s income counting rules, which include separate disregards for earned and unearned income, using methods described in the “income and asset rules” section of the appendix.26 We calculated Medicaid-countable assets by summing all categories of assets the Medicaid program considers for the asset test, such as checking and savings accounts, retirement accounts, and business assets. We made several decisions about how to count assets or apply deductions, such as the deduction of burial funds from assets, and conducted sensitivity analyses to explore the impact of these assumptions on our estimates.

Analysis

We compared Medicaid-countable income and assets to income and asset eligibility limits for the respondent’s state and year to identify respondents who qualified for full Medicaid under existing income and asset tests. We examined the demographic and health characteristics of respondents who met both income and asset tests and of respondents who met the income test but not the asset test. For each group, we calculated the average value of the assets held by category, such as savings accounts or stocks, as well as the interquartile range (IQR) for each type of asset. We also examined the distribution of total assets among respondents who qualified for Medicaid under the income test but whose assets exceeded the asset test.

We simulated the number of senior community-dwelling Medicare beneficiaries who would be newly eligible for full Medicaid, based on data for 2016, under four alternative asset tests among beneficiaries who met Medicaid’s income test. These simulations reflect respondents’ current asset holdings and make no assumptions about behavioral responses to these alternative tests by older adults (e.g., changes in savings when asset limits are higher). However, our analyses do examine whether higher eligibility limits would continue to screen out individuals with substantial wealth in the absence of such behavioral responses.

In the first scenario, we simulated the number of low-income Medicare beneficiaries who would qualify for Medicaid if asset tests were inflated from the SSI program’s 1989 levels by the Consumer Price Index (CPI), a measure of price inflation that is used to update the FPG. In 2016, this would correspond to an asset limit of $3,913 for individuals and $5,869 for couples. We did not apply the inflated asset limit to the 8 states and DC that used asset limits above the inflated value (details in the “first policy simulation” section of the appendix).26

In the second scenario, we simulated the number of low-income Medicare beneficiaries who would qualify for Medicaid if the asset test were raised to match the 2016 federal MSP/LIS tests. This corresponds to an asset limit of $7,280 for individuals and $10,930 for couples.

In the third scenario, we simulated the number who would qualify for Medicaid if the asset test were raised to $10,000 for individuals and $20,000 for couples, as proposed in the SSI Restoration Act of 2019.29

In the fourth scenario, we simulated the number who would be eligible for Medicaid if asset limits were replaced by a hypothetical “high wealth category test” that expands eligibility to all persons who do not own investments (stocks, bonds, and certificates of deposit (CDs)), secondary residences, or businesses.

Currently, Arizona is the only state that does not impose asset limits for full-benefit duals. Since the goal of this exercise is to examine policies that are more expansive than the status quo policies in all other U.S. states, all four scenarios assumed that Arizona continued to waive the asset test. All analyses used the survey weights provided in the HRS, producing estimates that are representative of the community-dwelling senior population in the US.28

Limitations

Our analyses had several limitations. First, all surveys, including the HRS, are subject to reporting bias and measurement error that may affect our results. However, analysis by RAND found that the distributions of income and assets reported by HRS respondents closely matched those of similarly-aged US adults in other data sources, and the HRS is widely used in analyses of household finances and Medicaid eligibility for older US adults.28 Second, we were unable to implement certain state-specific income and asset exclusions (e.g., those pertaining to welfare income and trusts) using HRS variables. Third, the HRS asks respondents to report their income in the year prior to the interview but to report the value of assets at the time of the survey. We assumed that the prior year’s income reflects the current year’s income, although this assumption may introduce some error in our eligibility calculations. Fourth, this analysis focuses on senior, community-dwelling Medicare beneficiaries. Our analysis does not provide estimates of changes in Medicaid eligibility among Medicare beneficiaries under age 65 or those in long-term nursing facilities.

Results

Variation in income and asset tests for full-benefit duals across states, 2006–2018

In 2018, 21 states used the SSI income and asset tests for determining full-benefit Medicaid eligibility for community-dwelling seniors enrolled in Medicare (Appendix Exhibit A2).26 The income test in these states was 74% of the FPG for individuals and 82% for couples. Across other states, income eligibility limits ranged from 50% FPG in some regions of Connecticut to 100% FPG in states that adopted an optional poverty pathway for seniors. In the same year, the asset test varied from $1,500 in New Hampshire to no limit in Arizona, with states that adopted SSI eligibility criteria using an asset test of $2,000 for individuals and $3,000 for couples (Exhibit 1).

EXHIBIT 1.

EXHIBIT 1

Medicaid individual asset tests for community-dwelling seniors enrolled in Medicare in 2018

SOURCE: Medicaid eligibility data collected from publicly available sources and state Medicaid departments.

NOTES: Maps present asset tests for individuals, not married couples. The asset test for couples is higher in most states and tends to be correlated with the individual asset test. Only Arkansas, Indiana, Missouri, Ohio, and South Carolina increased their asset limits between 2006 and 2018.

Between 2006 and 2018, three states (Connecticut, Indiana, and Ohio) increased income limits explicitly by raising the percent of FPG used to determine eligibility, while all other states implicitly adjusted income limits for inflation. However, only five states increased asset limits between 2006 and 2018 (Arkansas, Indiana, Missouri, Ohio, and South Carolina). All other states used the same asset dollar limit for the duration of the 13-year period.

Demographic and health characteristics

We compared the demographic and health characteristics of senior community-dwelling Medicare beneficiaries who met both the Medicaid income and asset tests, and those who met the income but not the asset test (Exhibit 2). Age, sex, marital status and health status did not differ significantly across the groups. In contrast, race and education differed across the groups. Nearly 80% of respondents who met the income test but not the asset test were white (compared to 32.8% among those who met both the income and asset tests) and 42.4% of respondents who met the income test but not the asset test had at least some college education compared with 24.2% of those who met both the income and asset tests (p<0.005 for both characteristics). In our sample, 53.1% of those who met the income and asset tests reported having Medicaid, consistent with other take-up estimates from the literature.3, 4

EXHIBIT 2:

Demographic and health characteristics of community-dwelling seniors enrolled in Medicare who met the Medicaid income and asset tests, and those who met the income but not the asset test in 2016

Characteristics Met Medicaid income and asset tests Met Medicaid income but not asset test
Weighted, N 2,881,949 892,598
Age, years*
 ≥ 65 to ≤ 74 55.6% 45.7%
 ≥ 75 to ≤ 84 27.9% 29.1%
 ≥ 85 16.5% 25.2%
Sex, female 67.8% 67.8%
Race/ethnicity, % in category****
 White, non-Hispanic 32.8% 76.7%
 Black, non-Hispanic 30.9% 10.7%
 Other, non-Hispanic 4.3% 4.8%
 Hispanic 32.1% 7.8%
Education, % in category****
 Less than high school 50.0% 19.0%
 Finished high school 25.9% 38.6%
 Some college 17.6% 17.8%
 Finished college 6.6% 24.6%
Marital status, % in category
 Married/partnered 24.1% 33.0%
 Separated/divorced 25.9% 23.5%
 Widowed 31.1% 30.9%
 Never married 18.9% 12.6%
Region, % in category**
 Northeast 15.8% 15.9%
 Midwest 16.1% 27.3%
 South 40.4% 28.2%
 West 27.7% 28.6%
Share with at least 1 ADL 36.3% 31.0%
Fair/poor self-reported health status 54.2% 44.8%
Enrolled in Medicaid, %**** 53.1% -
Income, % FPG
 0 to ≤ 50% 40.7% 32.3%
 > 50 to ≤ 75% 34.1% 39.6%
 > 76 to ≤ 100% 25.2% 28.1%
Assets, 2016 dollars (% within range)****
 = $0 59.5% 0.0%
 > $0 and ≤ State limit 40.5% 0.0%
 > State limit and ≤ MSP/LIS limit 0.0% 14.3%
 > MSP/LIS limit and ≤ $50,000 0.0% 37.4%
 > $50,000 0.0% 48.3%

SOURCE Data from wave 13 of the Health and Retirement Study (HRS) and Medicaid eligibility data collected from publicly available sources and state Medicaid departments.

NOTES We used a Pearson chi-squared test to compute the p-value for differences in the distribution between those who met the income and asset test (column 1) and those who met the income but not the asset test (column 2). This table only includes income and assets that Medicaid would count in its eligibility determination. Some categories do not sum to one due to rounding. FPG: federal poverty guideline.

SIGNIFICANCE:

*

p<0.10

**

p<0.05

***

p<0.01

****

p<0.001

While 59.5% of the asset-eligible sample had no countable assets, 51.7% of those who did not meet the asset limit had assets less than $50,000. Appendix Exhibit A3 presents mean assets by category.26 Checking and savings accounts were the most commonly held asset and IRA accounts had the highest mean value. Most respondents did not have any assets in the other categories – stocks, bonds, CDs, secondary residences, other real estate, businesses, trusts, other assets.

We further examined the distribution of assets among persons in our sample who met the Medicaid income test but did not meet the asset test. Appendix Exhibit A4 reports the 25th, 50th, and 75th percentiles of total assets among individuals who have any versus no assets in the categories listed above.26 Respondents with high levels of Medicaid-countable assets held their wealth in different categories than respondents with lower assets. For example, respondents with checking or savings accounts or non-residential real estate had lower total assets and respondents with IRAs had higher total assets at each quartile than those who did not. Similarly, respondents who own investments, secondary residences, or businesses had considerably more total Medicaid-countable assets than those without these types of assets.

Alternative asset tests

Among all senior, community-dwelling Medicare beneficiaries in the weighted HRS sample (n=44,157,184 beneficiaries), 3,774,547 beneficiaries (8.5%) were eligible for Medicaid based on their incomes. Of those who met the income test, 2,881,949 (76.4%) also met the Medicaid asset tests in place in 2016 (Exhibit 3). If the 1989 SSI asset limits were adjusted by the CPI, the number of community-dwelling senior Medicare beneficiaries who were eligible for full-benefit Medicaid would increase by 1.7% relative to the number who were eligible in 2016 (an increase from 2,881,949 to 2,931,317 individuals). If states used the 2016 MSP/LIS asset limits, the number eligible for full-benefit Medicaid would increase by 4.4% relative to the number actually eligible in 2016 (to 3,009,522 individuals). If asset limits were increased to $10,000 for individuals and $20,000 for couples, the number eligible for full-benefit Medicaid would increase by 7.5% compared to the 2016 levels of eligibility (to 3,097,178 individuals). Finally, if a hypothetical “high wealth category test” deemed low-income persons eligible for Medicaid unless they held any investments, secondary residences, or businesses, the number eligible for full-benefit Medicaid would increase by 20.5% (to 3,472,422 individuals).

EXHIBIT 3.

EXHIBIT 3

Number of low-income, community-dwelling seniors enrolled in Medicare who would be eligible for Medicaid in 2016 if the asset test were increased

SOURCE: Data from wave 13 of the Health and Retirement Study (HRS) and Medicaid eligibility data collected from publicly available sources and state Medicaid departments.

NOTES: Dark gray bars show estimated 2016 eligibility numbers; light gray bars indicate projected eligibility numbers under different asset tests. CPI: consumer price index, MSP: Medicare savings program, LIS: low-income subsidy.

Finally, we examined selected demographic and health characteristics of community-dwelling seniors enrolled in Medicare who would be newly eligible for Medicaid under three alternative asset tests (Exhibit 4). We excluded the alternative CPI test due to small sample size and similarity with the MSP/LIS results. Individuals who would qualify under the MSP/LIS alternative asset test had lower levels of education (p=0.049), were more likely to have at least one ADL (p=0.113), and had worse self-reported health (p=0.022) compared to the complementary group who met the Medicaid income test but not the MSP/LIS asset test. The share of nonwhites who would newly qualify for Medicaid under the $10,000/$20,000 asset test (40.4%) exceeds the share of nonwhites who met the Medicaid income criteria but not the $10,000/$20,000 asset test (p=0.019). Those newly eligible under the “high wealth category” asset test were younger (p=0.023) and more likely to have at least one ADL (p=0.069) compared with those who meet Medicaid’s income test but would not meet the “high wealth category” asset test. Notably, under this alternative asset test, 67.3% of those newly eligible had assets below $50,000 (p=0.000).

EXHIBIT 4:

Demographic characteristics of community-dwelling seniors enrolled in Medicare who would be newly eligible for Medicaid under different asset tests in 2016

Beneficiaries who met Medicaid’s income test but not asset test If the asset limit was increased to MSP/LIS limit If the asset limit was increased to $10,000 for singles and $20,000 for couples If the asset test excluded those with investments, secondary residences, or businesses
Weighted, N 892,598 127,573 215,229 595,239

Age, years **

 ≥ 65 to ≤ 74 45.7% 43.2% 43.2% 53.2%

 ≥ 75 to ≤ 84 29.1% 30.8% 31.1% 24.7%

 ≥ 85 25.2% 26.0% 25.7% 22.1%

Race/ethnicity **

 White, non-Hispanic 76.7% 69.2% 59.6% 73.1%

 Black, Hispanic, and Other 23.3% 30.8% 40.4% 26.9%

Education **

 Less than high school 19.0% 26.6% 25.2% 18.7%

 Finished high school 38.6% 50.2% 44.4% 40.8%

 Some college or more 42.4% 19.6% 30.4% 40.5%

Share with at least 1 ADL 31.0% 46.5% 37.8% 36%*

Fair/poor self-reported health status 44.8% 68.6%** 54.6% 46.9%

Income, % FPG

 0 to ≤ 50% 32.3% 17.7% 24% 30.3%

 > 50 to ≤ 75% 39.6% 57.9% 54.8% 41.3%

 > 76 to ≤ 100% 28.1% 24.4% 21.2% 28.4%

Assets, 2018 dollars (% within range) **** **** ****

 > State limit and ≤ MSP/LIS limit 14.3% 100% 59.3% 18.8%

 > MSP/LIS limit and ≤ $50,000 37.4% 0.0% 40.7% 48.5%

 > $50,000 48.3% 0.0% 0.0% 32.7%

SOURCE Data from wave 13 of the Health and Retirement Study (HRS) and Medicaid eligibility data collected from publicly available sources and state Medicaid departments.

NOTES We used a Pearson chi-squared test to compare each newly eligible sample (last 3 columns) to its respective complement from the full sample of those who are currently income eligible but do not meet the asset test (1st column of data). Asterisks are used above the categorical variable values within asset test type to denote a statistically significant difference. FPG: federal poverty guideline, MSP: Medicare savings program, LIS: low-income subsidy, ADL: activities of daily living.

SIGNIFICANCE:

*

p<0.10

**

p<0.05

***

p<0.01

****

p<0.001

Sensitivity analysis

We conducted sensitivity analyses to assess whether methodological assumptions impacted our findings. Following a prior report,30 we subtracted whole life insurance policies up to $1,500, debt, and the burial fund allowance of $1,500 for individuals and $3,000 for couples from the measure of assets in our main analysis. The number of low-income seniors who were eligible for full-benefit Medicaid increased from the estimated 2016 eligibility by 1.3%, 0.9%, and 1.7% respectively. When all three exclusions were applied to the measure of assets, eligibility increased by 4.8% relative to the estimated 2016 full-benefit Medicaid eligibility levels. Though the number estimated to be eligible was higher when incorporating these exclusions, the relative changes to eligibility when inflating asset limits by the CPI or increasing asset limits to the MSP/LIS limits were similar to results reported in Exhibit 3. Our main findings were also similar when we excluded proxy respondents from the sample (see Appendix Exhibits A5 and A6).26

Discussion

For low-income Medicare beneficiaries in all but three states, asset limits in the Medicaid program are not adjusted for inflation and therefore have been stagnant for over 30 years. This has made Medicaid’s asset test for senior, community-dwelling, Medicare beneficiaries increasingly restrictive over time. The lack of updates to Medicaid’s asset limit are notable because low-income Medicare beneficiaries rely on Medicaid for financial protection and coverage of services that are not included in the Medicare program.1, 2 We examined options for policy changes by comparing how four alternative asset tests would affect Medicaid eligibility for low-income, community-dwelling Medicare beneficiaries who meet Medicaid’s current income tests.

Among our four alternative asset tests we found the smallest increases in Medicaid eligibility when Medicaid asset limits were indexed by the Consumer Price Index, increased to MSP/LIS asset limits, or increased to $10,000 for individuals and $20,000 for couples. We found the largest increase in eligibility under the high wealth category test (i.e., percentage increases in the number eligible were 1.7, 4.4, and 7.5 versus 20.5). These results are similar to those from prior research on SSI eligibility, which estimated that a 50% increase in the asset test would increase SSI participation by 1.3 percentage points among community-dwelling seniors.31 We also found that the asset limit in 2016 excluded around 23.6% of seniors who met Medicaid income limits. This suggests that eliminating asset tests entirely has the potential to substantially increase full-benefit Medicaid eligibility.

Our findings indicate that incrementally increasing asset limits may not greatly increase the number of newly-eligible full-benefit Medicaid beneficiaries relative to the overall population of low-income beneficiaries, but would target specific sub-populations of low-income people who could benefit from Medicaid’s supplemental coverage. For example, increasing the full-benefit Medicaid asset test to the MSP/LIS limit or to the $10,000/$20,000 asset limit minimally expanded the number of low-income Medicare beneficiaries who are newly eligible, but targeted a higher proportion of people who are non-white and have fair or poor self-reported health status. Thus, this relatively small eligibility increase could reach people more likely to benefit from full Medicaid coverage. Alternatively, using our “high wealth category test” would expand Medicaid eligibility to a larger low-income group with relatively modest assets. Under this policy, the newly-eligible population is demographically similar to those who meet the income test but not the “high wealth category” asset test.

Increasing the current asset limits for full-benefit Medicaid may benefit both low-income seniors and states. The current asset limits create unintended consequences, such as discouraging savings.32 Additionally, restrictive asset limits may perpetuate disparities by favoring better-resourced individuals who can access financial advisors, better educated people who can navigate the complex eligibility system, and people who own rather than rent their residence. If asset limits were increased, many low-income people could potentially acquire modest assets that would improve overall financially security. Simplifying Medicaid’s asset tests may also reduce state administrative costs arising from eligibility determinations.

Policy implications

The changes to Medicaid asset tests that we modeled would have important implications for beneficiaries and Medicaid programs. Broadening Medicaid eligibility will enable more low-income seniors to enroll in Medicaid, which protects them from high out-of-pocket health care costs and could help them retain more, though still modest, savings. These changes may disproportionately benefit specific subgroups of Medicare beneficiaries, including minority populations. In deciding whether to implement such changes to asset tests, policy makers also need to consider how these changes affect Medicaid enrollment and costs.

Increasing the Medicaid asset limit could have complex effects on Medicaid dual enrollment, and precisely estimating changes in enrollment is challenging for several reasons. Existing estimates of Medicaid take-up rates among low-income Medicare beneficiaries are imprecise, ranging from 50–73%.3, 4 Take-up could differ among the newly-eligible population; similarly, take-up among previously eligible persons could increase from its present level if streamlined asset rules make it easier to apply for and retain Medicaid. The size of this “welcome-mat effect” could also vary based on the level of outreach to Medicaid-eligible populations that accompanies policy changes.33 Finally, changes to asset limits may affect who qualifies for Medicaid through alternative coverage pathways, such as spend-downs for people with high medical costs and for nursing home-dwelling populations. Thus, Medicaid enrollment is expected to increase when the program’s asset limits are raised, but it is difficult to quantify by how much.

Similarly, the costs of these policy changes to the Medicaid program and states depend on complex, and potentially offsetting, factors. While covering more beneficiaries with Medicaid is likely to increase spending, these costs might be mitigated if new enrollees are healthier than individuals covered under current rules. Over the long term, improved access to care facilitated by Medicaid coverage might lead to offsetting reductions in other areas where Medicaid pays for care (e.g., nursing home care). Simplifying asset tests might also lessen the administrative costs states incur in processing Medicaid applications and renewals. Policy makers will need to weigh the effects of these different factors on net costs, and consider the implications of these costs on state budgets and programmatic features (e.g., benefit generosity). Both sets of considerations may have different short-run versus long-term implications.

A change in policy could also permit the elderly to have both modest savings and Medicaid coverage. This is especially important given that over half of older adults who live alone and nearly one-fourth of older couples do not have the income needed to pay for health care and other essential expenses such as food, housing, and transportation.34 There is a tradeoff between targeted and broad increases in asset tests for full-benefit Medicaid coverage, but either would benefit low-income seniors who have faced stagnant Medicaid asset tests for decades.

Conclusion

Medicaid asset limits for the low-income senior population are low and have been stagnant since 1989 in most states. Incrementally increasing the asset limit may target specific low-income sub-populations, while broader increases based on categories of assets would allow a substantial proportion of low-income seniors with modest assets above the current limits to qualify for coverage. Increasing the Medicaid asset limit will allow seniors to retain modest savings and better prepare for unexpected or essential expenses while qualifying for Medicaid coverage that can increase access to care and improve health outcomes.

Supplementary Material

Supplementary Material

Acknowledgments

This project was supported by grant number R01HS025422 from the Agency for Healthcare Research and Quality. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality.

Contributor Information

Noelle Cornelio, Department of Health Policy and Management, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania.

Melissa McInerney, Department of Economics, Tufts University, Medford, Massachusetts.

Jennifer M. Mellor, Department of Economics, William & Mary, Williamsburg, Virginia.

Eric T. Roberts, Department of Health Policy and Management, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania.

Lindsay M. Sabik, Department of Health Policy and Management, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania.

Endnotes

  • 1.CMS Medicare-Medicaid Coordination Office. Data Analysis Brief: Medicare-Medicaid Dual Enrollment 2006 through 2018 - Twelve-year Enrollment Trends: 2006–2018 2019. [Google Scholar]
  • 2.Medicare Payment Advisory Commission (MedPAC). A Data Book: Health Care Spending and the Medicare Program, Section 4: Dual-eligible Beneficiaries. [Internet]. Washington, DC: MedPAC; July 2020. Available from: http://medpac.gov/-documents-/data-book [Google Scholar]
  • 3.Pezzin LE, Kasper JD. Medicaid Enrollment among Elderly Medicare Beneficiaries: Individual Determinants, Effects of State Policy, and Impact on Service Use. Health Services Research. 2002;37(4):827–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Ungaro R, Federman AD. Restrictiveness of eligibility determination and Medicaid enrollment by low-income seniors. J Aging Soc Policy. 2009;21(4):338–51. [DOI] [PubMed] [Google Scholar]
  • 5.Scott CG. Resources of supplemental security income recipients. Soc Secur Bull. 1989;52(8):2–9. [PubMed] [Google Scholar]
  • 6.Dorn S, Shang B. Spurring enrollment in Medicare savings programs through a substitute for the asset test focused on investment income. Health Aff (Millwood). 2012;31(2):367–75. [DOI] [PubMed] [Google Scholar]
  • 7.Summer L How asset tests block low-income Medicare beneficiaries from needed benefits. Issue brief (Commonwealth Fund). 2004(727):1–12. [PubMed] [Google Scholar]
  • 8.Medicare Payment Advisory Commission (MedPAC). March 2008 Report to the Congress - Medicare Payment Policy: Chapter 5. [Internet]. Washington, DC: MedPAC; March 2008. Available from: http://medpac.gov/docs/default-source/reports/Mar08_Ch05.pdf [Google Scholar]
  • 9.GAO. Medicare Savings Program: Implementation of Requirements Aimed at Increasing Enrollment. Washington, DC; 2012. [Google Scholar]
  • 10.Brook RH, Ware JE, Rogers WH, Keeler EB, Davies AR, Sherbourne CD, et al. The Effect of Coinsurance on the Health of Adults: Results from the RAND Health Insurance Experiment. Santa Monica, CA: RAND Corporation; 1984. [Google Scholar]
  • 11.Ku L, Wachino V. The Effect of Increased Cost-Sharing In Medicaid: A Summary of Research Findings. Washington, DC: Center on Budget and Policy Priorities; 2005. [Google Scholar]
  • 12.KFF. Medicaid Financial Eligibility: Primary Pathways for the Elderly and People with Disabilities. Washington DC: Kaiser Family Foundation; 2010. [Google Scholar]
  • 13.O’Malley Watts M, Cornachione E, Musumeci M. Medicaid Financial Eligibility for Seniors and People with Disabilities in 2015. Washington DC: Kaiser Family Foundation; 2016. [Google Scholar]
  • 14.Musumeci M, Chidambaram P, O’Malley Watts M. Medicaid Financial Eligibility for Seniors and People with Disabilities: Findings from a 50-State Survey. Washington DC: Kaiser Family Foundation; 2019. [Google Scholar]
  • 15.Bruen BK, Wiener JM, Thomas S. Medicaid Eligibility Policy for Aged, Blind, and Disabled Beneficiaries Washington, DC: AARP Public Policy Institute and The Urban Institute; 2003. [Google Scholar]
  • 16.Walker L, Accius J. Access to Long-Term Services and Supports: A 50-State Survey of Medicaid Financial Eligibility Standards Washington, DC: AARP Public Policy Institute; 2010. [Google Scholar]
  • 17.Medicaid and CHIP Payment and Access Commission (MACPAC). Medicaid and CHIP Program Statistics: March 2012 MACStats, table 11. [Internet]. Washington, DC: MACPAC; March 2012. Available from: https://www.macpac.gov/publication/macstats-archive/ [Google Scholar]
  • 18.Medicaid and CHIP Payment and Access Commission (MACPAC). MACStats: Medicaid and CHIP Program Statistics, table 11. [Internet]. Washington, DC: MACPAC; March 2013. Available from: https://www.macpac.gov/publication/macstats-archive/ [Google Scholar]
  • 19.Medicaid and CHIP Payment and Access Commission (MACPAC). MACStats: Medicaid and CHIP Program Statistics, table 11. [Internet]. Washington, DC: MACPAC; March 2014. Available from: https://www.macpac.gov/publication/macstats-archive/ [Google Scholar]
  • 20.Medicaid and CHIP Payment and Access Commission (MACPAC). MACStats: Medicaid and CHIP Data Book, exhibit 36. [Internet]. Washington, DC: MACPAC; December 2015. Available from: https://www.macpac.gov/publication/macstats-archive/ [Google Scholar]
  • 21.Medicaid and CHIP Payment and Access Commission (MACPAC). MACStats: Medicaid and CHIP Data Book, exhibit 36. [Internet]. Washington, DC: MACPAC; December 2016. Available from: : https://www.macpac.gov/publication/macstats-archive/ [Google Scholar]
  • 22.Medicaid and CHIP Payment and Access Commission (MACPAC). MACStats: Medicaid and CHIP Data Book, exhibit 37. [Internet]. Washington, DC: MACPAC; December 2017. Available from: https://www.macpac.gov/publication/macstats-archive/ [Google Scholar]
  • 23.Medicaid and CHIP Payment and Access Commission (MACPAC). MACStats: Medicaid and CHIP Data Book, exhibit 37. [Internet]. Washington, DC: MACPAC; December 2018. Available from: https://www.macpac.gov/publication/macstats-archive/ [Google Scholar]
  • 24.Medicaid and CHIP Payment and Access Commission (MACPAC). MACStats: Medicaid and CHIP Data Book, exhibit 37. [Internet] Washington, DC: MACPAC; December 2019. Available from: https://www.macpac.gov/publication/macstats-archive/ [Google Scholar]
  • 25.TRIM3 (Transfer Income Model, Version 3) project website. [Internet]. Urban Institute; 2012. Available from: trim3.urban.org [Google Scholar]
  • 26.To access the Appendix, click on the Appendix link in the box to the right of the article online.
  • 27.Health and Retirement Study, (RAND HRS Longitudinal File 2016 (V1)) public use dataset. Produced and distributed by the University of Michigan with funding from the National Institute on Aging (grant number NIA U01AG009740). Ann Arbor, MI, (2019). [Google Scholar]
  • 28.RAND HRS Longitudinal File 2016 (V1). Produced by the RAND Center for the Study of Aging, with funding from the National Institute on Aging and the Social Security Administration. Santa Monica, CA: (2019). [Google Scholar]
  • 29.Supplemental Security Income Restoration Act of 2019, S. H.R.4280, House of Representatives, 116th Congress (2019–2020) Sess. (2019).
  • 30.Meijer E, Karoly LA, Michaud P-C. Estimates of Potential Eligibility for Low-Income Subsidies Under Medicare Part D. Santa Monica, CA: RAND Corporation; 2009. [Google Scholar]
  • 31.Davies PS, Huynh M, Newcomb C, O’Leary P, Rupp K, Sears J. Modeling SSI financial eligibility and simulating the effect of policy options. Soc Secur Bull. 2001;64(2):16–45. [PubMed] [Google Scholar]
  • 32.Hubbard RG, Skinner J, Zeldes SP. Precautionary Saving and Social Insurance. J Political Econ. 1995;103(2):360–99. [Google Scholar]
  • 33.McInerney M, Mellor JM, Sabik LM. Welcome Mats and On-Ramps for Older Adults: The Impact of the Affordable Care Act’s Medicaid Expansions on Dual Enrollment in Medicare and Medicaid. J Policy Anal Manage. 2021;40(1):12–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Mutchler J, Li Y, Roldán NV. Living Below the Line: Economic Insecurity and Older Americans, Insecurity in the States 2019 Center for Social and Demographic Research on Aging Publications. 2019;40. https://scholarworks.umb.edu/demographyofaging/40. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material

RESOURCES