Who Will Be Denied Medicare Prescription Drug Subsidies Because of the Asset Test?

Objective: To determine the number and characteristics of Medicare beneficiaries who will be excluded from low-income prescription drug subsidies because they do not qualify under an asset test.

Study Design: Cross-sectional, using the US Census Bureau's Survey of Income and Program Participation (SIPP); results were based on interviews occurring between October 2002 and January 2003. The sample included 9278 Medicare beneficiaries, 2929 with incomes below 150% of the federal poverty level (FPL).

Methods: Using SIPP, each sample member's income was compared to the FPL. Income was adjusted to include only liquid assets and primary residences. The number of individuals excluded by the asset test and their characteristics and types of assets responsible were calculated.

Results: Of 13.97 million noninstitutionalized Medicare beneficiaries, 2.37 million (17%) with low incomes would be excluded from subsidized drug coverage due to the asset test. Compared to higher-income beneficiaries, the excluded individuals tended to be older, female, widowed, and living alone. Almost half of their assets were checking and savings accounts. Half of the individuals failing the test had assets less than $35 000 above the allowing thresholds.

Conclusions: Widows are disproportionately affected by the asset test. When a husband dies, income plummets but accumulated assets often exceed those allowed under Medicare legislation. During their working years Americans are encouraged to save for retirement, but by accumulating modest amounts of assets, these same people often will then not qualify for low-income drug subsidies. Modifying or eliminating the asset test would help protect individuals disadvantaged by low incomes who have modest amounts of asset holdings.

(Am J Manag Care. 2005;12:46-54)

Amidst fanfare and controversy, the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 was signed into law by President George W. Bush in December of that year. The centerpiece of the legislation was coverage of outpatient prescription drugs, a benefit absent from Medicare during the program's first 40 years. This new drug coverage takes effect in January 2006.

The legislation provides voluntary, subsidized prescription drug coverage that can be obtained either from stand-alone insurance policies or through Medicare managed care plans. The specific style of benefit has been referred to as having a "doughnut hole" because during a given year, there may be a portion of expenditures for which no coverage is provided. This gap, combined with other cost-sharing features, means that many beneficiaries will still have to pay a sizeable portion of their prescription costs. An individual spending $5000 a year for covered drugs would pay a total of $3500 out-of-pocket, not including a premium averaging $420 per year in 2006.

A subsidy intended to provide assistance for low-income Medicare beneficiaries is the focus of this article. This subsidy is necessary because without it, drug coverage would not be affordable for such individuals and would be far more costly than is now the case for beneficiaries who are dually covered by Medicare and Medicaid. Once the prescription drug provisions are implemented in January 2006, individuals who are dually covered by Medicare and Medicaid will receive their drug benefits through the Medicare program rather than through Medicaid, as is currently the case.

To qualify for low-income subsidies, a beneficiary must meet specific income and asset thresholds. The low-income subsidies will offer substantial assistance in paying the Part D premium and cost sharing associated with drug coverage. The level of assistance will vary depending on an individual's income and assets. Individuals who meet the income threshold but whose assets exceed a specified limit will not qualify for low-income subsidies.

For example, a person who is not dually eligible for Medicare and Medicaid, and who has an income of less than 135% of the federal poverty level (FPL), will fail the test and thus will not be eligible for the low-income subsidies if he or she has countable assets exceeding $6000 (individual) or $9000 (couple). The definition of countable assets does not include the value of a house and automobiles, or household furnishings and possessions.

In this study, we addressed 3 key questions regarding the asset test:

• How many and what percentage of Medicare beneficiaries will be eligible for low-income prescription drug subsidies?
• What are the characteristics of individuals who are excluded from the subsidies because they do not qualify under the asset test? Do variations depend on age, sex, race, education, family composition, geographic location, supplemental health insurance status, or health status?

BACKGROUND

Low-income Subsidies Under the Law

The new Medicare drug benefit will go into effect on January 1, 2006. Certain Medicare beneficiaries are eligible for substantial low-income subsidies (Figure 1). The first group includes individuals who qualify automatically because they are eligible for Medicaid, Supplemental Security Income (SSI), the Qualified Medicare Beneficiaries (QMB) program, the Specified Low-Income Medicare Beneficiary (SLMB) program, or the Qualifying Individuals (QI) program; such individuals are referred to as "dually eligible" for Medicare and Medicaid.² The second eligible group includes others who have incomes below 150% of the FPL and assets of less than $10 000 (individual) or $20 000 (couple). To provide context, the FPL in 2005 was $9570 for a single person, and $12 830 for a 2-person family.³

The subsidy a person qualifies for depends on his or her income and asset levels. Figure 1 and Table 1 provide labels for 3 different benefit levels (subsidy groups A, B, and C). The main data set used in this study, the Survey of Income and Program Participation (SIPP), is based on a sampling frame of the US civilian noninstitutionalized population. A fourth group–eligible individuals who are institutionalized–is not included because data limitations necessitated that this article focus on the noninstitutionalized population.

As shown in Figure 1, subsidy group A consists of dual eligibles with incomes less than 100% of the FPL. Subsidy group B comprises dually eligible individuals with incomes of more than 100% of the FPL and those not dually eligible but with incomes less than 135% of the FPL ($12 920/individual; $17 321/couple, in 2005) and assets of less than $6000 (individual) or $9000 (couple). Similarly, 2 subgroups of non-dually eligible individuals make up subsidy group C: those with incomes less than 135% of the FPL and assets between $6000 and $10 000 (individual) or $9000 and $20 000 (couple); and individuals with incomes between 135% and 150% of the FPL and assets less than $10 000 (individual) and $20 000 (couple). The following non-dually eligible individuals do not qualify for the subsidies: those with incomes of more than 150% of the FPL and those with assets exceeding $10 000 (individual) or $20 000 (couple).

Table 1 shows the drug benefit for each of the 3 subsidy groups in 2006. Persons in subsidy groups A and B will pay no premiums and will be responsible only for relatively small out-of-pocket copayments for each prescription they receive, varying from $1 to $2 for generic and $2 to $5 for brand name drugs. (The statute also requires that insurers apply the generic copayment levels to "preferred multiple-source drugs,"as specified but not defined in the statute.) Subsidy group C pays more, including a premium that is proportional to how close their income is to 135% of the FPL (vs 150%); 15% of drug spending between their $50 annual deductible and the $3600 out-of-pocket cost threshold (which could result in a maximum payment of about $530); and copayments of $2 (generic) or $5 (brand name) per prescription thereafter. No one receiving a low-income subsidy is subject to the doughnut hole of no coverage.

Impact of Eligibility on Potential Out-of-pocket Expenditures

Eligibility for low-income subsidies is likely to have a dramatic effect on out-of-pocket expenditures for prescription drugs. The Centers for Medicare & Medicaid Services (CMS) has estimated that on average, beneficiaries receiving the low-income subsidy would spend just $170 out-of-pocket in 2006, compared to $1122 without the subsidy–plus savings in premiums of up to $440/year.^4(p4468)

Similarly, another study using an actuarial projection model concluded that near-poor persons who receive the low-income subsidies would pay far less out-of-pocket than individuals who do not qualify. Among individuals with incomes between 100% and 134% of the FPL, annual out-of-pocket costs for prescription drugs are expected to average $149 for persons receiving the subsidies, compared to $1086 for those not receiving them. The figures for beneficiaries with incomes between 135% and 149% of the FPL are $283 and $979, respectively.⁵

Congressional Budget Office Estimates of the Impact of the Asset Test on Eligibility for Low-income Subsidies

The Congressional Budget Office (CBO) has estimated how many otherwise qualified low-income beneficiaries will not receive the low-income prescription drug subsidies because they will fail the asset test. The CBO's estimates were published before the final regulations were issued by CMS. This point is particularly important because the final regulations are, in some ways, more generous than the previous asset test requirements used by state Medicaid programs and the federal SSI program. Whereas the other programs include as countable assets the value of the first automobile exceeding $4500, and the total value of a second car, the regulations do not include any value from automobiles. As a result, fewer beneficiaries are likely to be excluded.

In estimates published on November 20, 2003, just a few days before Congressional passage, CBO estimated that 1.8 million of the 15.1 Medicare beneficiaries with incomes below 150% of the FPL (12%) would be ineligible for the low-income subsidies because their assets were too high. This number includes:

• 0.4 million of the 3.6 million beneficiaries (11%) with incomes between 101% and 120% of the FPL;
• 0.5 million of the 1.8 million beneficiaries (28%) with incomes between 136% and 150% of the FLP.⁶

DATA AND METHODS

Data

The data set used in this study was the SIPP, conducted by the US Census Bureau. The purpose of SIPP is "to provide accurate and comprehensive information about the income and program participation of individuals and households in the United States, and about the principal determinants of income and program participation."⁷ SIPP also collects information on individuals' assets, a necessity for the conduct of the present study. The sampling frame for SIPP includes only the noninstitutionalized population.

SIPP is a nationally representative panel survey (details available at www.bls.census.gov/sipp). To ensure adequate representation, it oversamples individuals with low incomes. We based our findings on data collected from the 2001 panel, whose members were interviewed 3 times annually over 3 years. Most of the data in this study were based on interviews that occurred between October 2002 and January 2003 and refer to the period September through December 2002.

The overall sample size in the SIPP file from which the data were extracted was 69 143 cases, of which 9278 were Medicare beneficiaries and 2929 had incomes below 150% of the FPL. The weighted counts were 37.6 million Medicare beneficiaries, 11.1 million of whom had incomes below 150% of the FPL.

Ascertaining the accuracy of the SIPP data is difficult. The SIPP does represent the US government's major effort to collect accurate data on income, assets, public program participation, and labor force issues (costing more than $30 million in 1998).⁸ Detailed comparisons have been conducted with respect to the accuracy of income data by comparing the SIPP with other data sources, notably the Current Population Survey.⁹ Although each survey has advantages, it is difficult for researchers to establish which is, in the aggregate, more accurate.

Methods

The current study was designed to address how many, and what percentage of, Medicare beneficiaries will be precluded from low-income prescription drug subsidies because their assets exceed the legislation's thresholds. All noninstitutionalized Medicare beneficiaries on SIPP were selected, including seniors and persons eligible due to disability (N = 9278). The first task was to estimate the number of low-income beneficiaries eligible for the subsidies if there were no asset test. Two groups are potentially eligible: those who receive Medicaid, SSI, QMB, SLMB, and QI; and those who do not receive any of these benefits but who have incomes below 150% of the FPL. The SIPP indicates if a person has Medicaid or SSI, but not QMB, SLMB, or QI. This lack of detail does not present a formidable problem, however, because an individual eligible for one of these programs must have an income below 135% of the FPL and therefore can be captured with SIPP through their income.

The estimated number of Medicare beneficiaries in this study was based on 2002 data from SIPP. These figures were then adjusted upward to provide the estimated number of beneficiaries in 2006. According to the CMS regulations, income was defined in accordance with SSI rules. Under those rules, certain exclusions are made from income, including the sum of each of the following amounts: the first $20 per month of any type of income, the first $65 per month of earned income, and half of the earnings above $65 per month. (In Section 1612 of the Social Security Act, earned income includes wages, net earnings from self-employment, and royalties.) Each sample member's income was compared to the FPL. According to the final CMS regulations, income includes that of both the individual and spouse (if any).^4(p4368) The FPL is based on family size, and naturally is higher for larger families.

The second task was to estimate how many otherwise eligible individuals will be excluded from subsidized prescription drug benefits due to the asset test. The final regulations defined eligible assets as follows:

[W]e intended to only consider liquid resources (that is, those that could be converted to cash within twenty days) and real estate that is not an applicant's primary residence as resources that are available to the applicant to pay for the Part D premiums, deductibles, and copayments. Thus, we would not consider their nonliquid resources (for example, a second car) to be available to the applicant for this purpose.^4(p4369)

As a result, only the following assets from SIPP were counted for both the individual and his or her spouse: bank accounts; stocks; bonds; mutual funds; retirement accounts such as IRAs, Keoghs, and 401(k)s; rental and vacation property; and other investments. To convert the asset amounts from 2002 dollars to 2006 dollars (to compare to 2006 asset thresholds), they were multiplied by the predicted rate of growth in consumer prices over this 4-year period (1.092). (Predicted inflation factors were obtained from the CBO: http://www.cbo.gov/ftpdocs/18xx/doc1824/EconProjectionsTables.pdf.)

The final task was to determine the assets that are primarily responsible for precluding eligibility and the characteristics of individuals who are precluded from the subsidies because they fail the asset test. We also examined sociodemographic characteristics (age, sex, race/ethnicity, marital status, family composition, education, geographic location); health status/usage (self-reported health status, hospitalizations, physician visits, prescription drug use); and possession of supplemental health insurance.

FINDINGS

Overall Impact of the Asset Test on Eligibility for Low-income Drug Subsidies

An estimated 13.97 million noninstitutionalized Medicare beneficiaries would qualify for low-income subsidized prescription drug benefits in 2006 based on income alone. Among this group, an estimated 2.37 million are expected to be ineligible for low-income subsidies due to the asset test. This represents 17% of otherwise eligible individuals. Approximately 70% of the 2.37 million who are ineligible because of the asset test have incomes less than 135% of the FPL, with the remaining 30% having incomes between 135% and 150% of the FPL.

This figure of approximately 2.4 million individuals failing the asset test in 2006 is somewhat higher than the 1.8 million calculated by CBO in its July 2004 report. Discussions with CBO staff revealed that a major reason for this difference was related to different methods of calculating income. As noted earlier, our estimates were calculated according to CMS regulations that had not been released at the time of the CBO report and which indicated that in determining eligibility for subsidized drug benefits, income is to be defined in accordance with SSI regulations. Under SSI regulations, the first $20 per month of any type of income, the first $65 per month of earned income, and half of earnings above $65 per month are excluded. Using these exclusions, more low-income beneficiaries fall below 150% of the FPL, and therefore could be excluded from the drug subsidies by the asset test. Because they tend to have higher incomes relative to other near-poor individuals and families, they are in fact more likely to have assets in excess of the thresholds. Had these exclusions not been made, an estimated 2.1 million beneficiaries would have failed the asset test. Thus, this difference in income calculation appears to explain half of the difference between the 2 sets of estimates.

Several other reasons may explain the different estimates. The CBO estimates included the institutionalized population whereas our study did not. Because low-income institutionalized seniors tend to have very low asset levels, their inclusion in the total population resulted in CBO estimating a lower percentage of people failing the asset test. The 2 studies used different data sets that were based on different sampling frames. The CBO used multiple data sources including the Medicare Current Beneficiary Survey, modified by CBO projections of population growth; the Current Population Survey, to adjust income estimates; and SIPP, for asset measurement. Our data source was SIPP alone. In addition, we used more recent data on assets than did CBO, and differences are likely to have arisen from CBO's estimates of asset growth.

Characteristics of Beneficiaries Excluded From Drug Subsidies Due to the Asset Test

Table 2 provides the characteristics of individuals who are ineligible for the low-income prescription drug subsidies due to the asset test (group IV), as well as 3 other groups of beneficiaries: individuals who are dually eligible for Medicare and Medicaid and who therefore will automatically receive the low-income subsidies (group II); those who are not dually eligible but who qualify to receive the low-income subsidies due to low incomes and assets (group III); and those whose incomes exceed 150% of the FPL and therefore do not qualify to receive the subsidies because of their incomes (group V). The first 2 columns (group I) provide figures for the total Medicare noninstitutionalized population, that is, the sum of the 4 subgroups (groups II through V). Groups were compared using Chi-square tests that corrected for the SIPP sampling design.

Two sets of comparisons are relevant: (1) low-income beneficiaries who fail the asset test (group IV) versus other non-dually eligible, low-income beneficiaries who pass the test (group III); and (2) low-income beneficiaries who fail the asset test (group IV) versus those who do not qualify for it because their income exceeds 150% of the FPL (group V). Most of the comparisons at the characteristic level were statistically significant at the 5% level. The only exceptions were for comparison 1 for metropolitan statistical area (MSA), hospitalized in the past year, number of physician visits in the past year, and daily use of prescription drugs; and for comparison 2 for race (significant at 10% level), MSA, region, health status, hospitalized in the past year, number of physician visits in the past year, and daily use of prescription drugs.

Comparison With Other Low-income Beneficiaries

Those with low incomes who are expected not to meet the asset test (group IV) have somewhat different characteristics than other low-income beneficiaries who are eligible to receive the low-income drug subsidies and who are not dually eligible for Medicaid (group III). Those failing the asset tests are more likely to be older, female, unmarried, and living alone. In these respects one might view them as more vulnerable, but in other respects they tend to be better off. They have higher education levels, are in better health, and are more likely to have private insurance. They are also more likely to be white. To illustrate, 55% of those failing the asset test are aged 75 or older, compared to 48% of the low-income group who do not fail the test; 73% graduated from high school, far higher than the 55% figure for the other group.

Comparison With Beneficiaries With Incomes Above 150% of the Federal Poverty Level

This comparison (group IV versus group V) is important because low-income beneficiaries who do not pass the asset test will receive the same coverage (and lack of subsidies) as those who do not have low incomes. In nearly all ways, individuals failing the asset test are much more vulnerable than beneficiaries with higher incomes. They are far more likely to be older, female, widowed, and living alone. They also have lower education levels and are less likely to have private health insurance.

Some of the key differences are illustrated in Figure 2. Fifty-five percent of those failing the asset test are aged 75 and older, compared to 41% of beneficiaries with higher incomes. Seventy one percent failing the test are female, compared to 52% of non-low-income beneficiaries. The most dramatic difference concerns marital status. Nearly half (46%) of those failing the asset test are widowed, twice the share (23%) for those with higher incomes. Forty-three percent of those who would fail the asset test are female widows. A related finding not shown in Figure 2 is the fact that those failing the asset test are far more likely to live alone (46%) than are higher-income beneficiaries (24%). Finally, Figure 2 shows that 22% of people failing the test have no private insurance, compared to 14% of higher income beneficiaries.

A clear pattern thus emerges. Persons who fail the asset test are disproportionately older widows who live alone. The most likely scenario is that when a husband dies, income plummets, making the widow potentially eligible for the low-income prescription drug subsidies. However, her accumulated assets exceed those allowed under the legislation. Aggravating the situation is that asset thresholds are lower for individuals than for couples (Figure 1). These people are vulnerable to financial catastrophe but, because they have some accumulated savings, they are ineligible for the subsidized prescription drug benefits.

Types of Assets That Preclude Eligibility for Low-income Drug Subsidies

Of the beneficiaries failing the asset test, on average, 44% of their total assets is with financial institutions; that is, checking and savings bank accounts and the like. Most of the remainder comprises stocks and mutual funds (18%) and retirement accounts such as IRAs, Keoghs, and 401(k) accounts (13%). On average, only 16% of the assets is equity in real estate (other than one's own house, which is not counted) and 3% is in ownership of a business.

In some instances, one type of asset alone puts a beneficiary over the threshold. Almost half (49%) of those who fail the asset test would fail it based solely on their assets in financial institutions. No other type of asset would, alone, disqualify more than 25% of individuals.

Figure 3 illustrates by how much individuals fail the asset test. Each bar represents $5000 in assets. Thus, the first bar indicates that of the people who fail the asset test, about 13% exceed it by $5000 or less; the second bar shows that another 9% exceed it by $5000 to $10 000, etc. The noteworthy pattern that emerges is that a large proportion of the 2.37 million people who are excluded from the low-income drug subsidies have assets that are not excessively high by most definitions. In fact, half of those who fail the asset test have excess assets of $35 000 or less. These savings would not pay for a year of nursing home care in most areas of the country. The amounts by which the unmarried fail the asset test are even more modest: 41% exceed the threshold by $25 000 or less.

Suppose that a woman (the typical case) exceeds the asset test by $35 000, which is the median amount. Liquidating these assets to pay for the prescription drug spending in the doughnut hole would reduce her current income through forgone interest; moreover, it would leave her with a very small financial buffer should she become ill and fall subject to large out-of-pocket costs, or suffer any other financial reversal. She would eventually be eligible for the subsidized prescription drug benefits, but at a cost of having spent down nearly all of her life savings.

CONCLUSIONS

This study estimates that in 2006, when the new Medicare prescription drug benefit goes into effect, 2.37 million low-income Medicare beneficiaries will not qualify for subsidized coverage because they fail the asset test. As a result, these individuals will face the same "doughnut hole" cost-sharing requirements as wealthier beneficiaries.

We further examined the types of beneficiaries who will be excluded by the asset test, as well as the types of assets responsible. Perhaps the most noteworthy finding was that the asset test will fall most heavily on widowed individuals. Whereas only 29% of Medicare beneficiaries are widowed, nearly half–46%–of those failing the asset test are widowed and nearly all of these (43% of the 46%) are women.

It is hardly surprising that most individuals who do not meet the asset test have relatively modest assets, which tend to be bank accounts rather than stocks, mutual funds, and bonds. They have little in the way of private retirement accounts such as IRAs and 401(k)s, real estate beyond their own home, and almost no equity in businesses. This financial state would be expected among a population of low-income individuals.

not

The study's findings raise serious questions about the equity of the asset test. During their working years, Americans are encouraged to save for retirement and the possibility that they will face sizable long-term care expenses. Yet many will have little or no income beyond what they receive from Social Security. By accumulating modest amounts of assets, either through bank accounts or retirement-savings vehicles, these same people have guaranteed that they will qualify for the low-income Medicare drug subsidies, but the vast majority use prescription drugs every day. This burden tends to fall on the most vulnerable of seniors: older, low-income widows living alone. Thus, modifying or eliminating the asset test would help protect individuals disadvantaged by low incomes who would be excluded from subsidized prescription drug benefits due to the asset test.

Acknowledgments

We express our appreciation to Tricia Neuman and Michelle Kitchman of the Kaiser Family Foundation for their helpful advice and support. We also thank Patricia Nemore of the Center for Medicare Advocacy for helping us understand the application of the asset test to low-income Medicare beneficiaries, and Jeanne De Sa and Eric Rollins of the Congressional Budget Office (CBO) for discussing with us the methodology CBO used in its study of the asset test. The conclusions do not represent those of the Kaiser Family Foundation or any reviewers; all errors are the responsibility of the authors.

From the Department of Health Services, UCLA School of Public Health, Los Angeles, Calif (TR) and Culver City, Calif (KD).

This study was supported by funding from the Henry J. Kaiser Family Foundation. A longer version of this report was published in April 2005 by the Henry J. Kaiser Family Foundation and is available on its Web site: http://www.kff.org/medicare/loader.cfm?url=/commonspot/security/getfile.cfm&PageID=52270.

Address correspondence to: Thomas Rice, PhD, Professor, Department of Health Services, UCLA School of Public Health, 650 S. Young Drive, Los Angeles, CA 90095-1772. E-mail: trice@ucla.edu.

Download PDF: Who Will Be Denied Medicare Prescription Drug Subsidies Because of the Asset Test?