This chapter examines issues in defining resources for use in measuring medical care economic risk (MCER)—the prospective risk that an individual or family will be unable to afford needed medical care or will be at high financial risk.1 In the end, the choice of a measure of resources will be tightly constrained by the choice of a survey to serve as home to a measure of MCER, and in this decision the measurement of medical care risk is likely to dominate the measurement of resources. Nevertheless, it is important to understand the key issues that exist in defining resources and the potential implications of including or excluding particular types of resources. The official measure of poverty in the United States as well as the new Supplemental Poverty Measure (SPM) are income-based and therefore more appropriately described as measures of income poverty. The resources available to families in meeting their financial needs also include assets—the product of families’ saving and investment activities over the life course. In this chapter we consider what sources of income should be included in the definition of resources in measuring MCER and whether some portion of assets should be included in resources as well. In Chapter 5, we review the strengths and weaknesses of alternative data sources for measuring resources in addition to measuring premiums and other out-of-pocket medical care costs.
1 As we noted in Chapter 2, the proposed measure of medical care economic burden, which is derived from the SPM, will use the SPM definition of resources.
CONCEPTUAL ISSUES IN DEFINING RESOURCES
A fundamental question facing the panel is whether the definition of resources to be used in measuring MCER should be equated with either of the income concepts that the Census Bureau employs in producing the official estimates of poverty in the United States or the new SPM published in November 2011, or whether a different concept would be more appropriate. In this section we discuss the Census Bureau income concepts and some of their limitations, review two alternative income concepts (Haig-Simons and federal income tax) and a consumption-based concept, and discuss the role of assets in meeting financial needs.
Income Concepts in Poverty Measures
The Census Bureau uses a reasonably well-defined concept of money income to produce the official, annual estimates of household income and poverty for the United States. A family’s annual money income, as measured in the Current Population Survey Annual Social and Economic Supplement (CPS ASEC), is compared with a threshold value that varies by family size; the number of children under age 18; and for one- and two-person households, whether the family reference person is age 65 or older. For the SPM, the Census Bureau substitutes a measure of disposable income for money income and uses an alternative set of thresholds. The two sections below define these two income concepts, laying out what they include and what they do not include.
The Census Bureau’s concept of money income as applied in the CPS ASEC is defined as total pretax cash income excluding lump sum payments and capital gains (Ruser, Pilot, and Nelson, 2004). Common sources of income that may be received as lump sums and therefore excluded from money income include bequests, life insurance (both survivor benefits and withdrawals of accumulated cash value unless converted to an annuity), and cashouts or withdrawals of pension and retirement funds. In excluding lump sums, the Census Bureau distinguishes between lump sums and regular payments, implying that these are the only two ways that income from these sources can be received. With the growth of new types of retirement accounts, which we discuss below, people make periodic withdrawals that are neither regular payments nor lump sums as these terms are commonly understood. This ambiguity is one of the issues with the application of the concept of money income—particularly for the measurement of economic well-being.
Another issue is that, in being restricted to cash, money income excludes the value of noncash benefits, which have become increasingly important in sustaining a segment of the population. Benefits from the Supplemental Nutrition Assistance Program (SNAP, formerly the Food Stamp Program) have an explicit cash value, which recipients use to purchase food. The free and reduced-price meals that students receive through the National School Lunch and School Breakfast Programs have an explicit monetary value as well, although their value is more restricted in its use than SNAP benefits. Housing subsidies are another type of noncash assistance that can be assigned a value. For decades, researchers and the Census Bureau itself have used the reported value of SNAP benefits and assigned cash values to other noncash benefits in order to develop alternative measures of income for the purpose of measuring the contribution of federal and state programs to combating poverty (see, for example, DeNavas-Watt, Cleveland, and Webster, 2003; Smeeding, 1982).
As a general concept, disposable income subtracts taxes from a pretax measure of income. The Census Bureau’s concept of disposable income, as used in the SPM, adds the cash value of noncash benefits while subtracting not only taxes, but also work-related expenses (including child care), child support payments to another household, and medical care out-of-pocket expenses (including premiums).2 Disposable income is intended to reflect the income that is actually available to families to meet their economic needs for food, clothing, shelter, utilities, and other basic necessities.
Limitations of CPS Income Concepts
Chapter 5 discusses a number of issues that affect the quality of income measured in household surveys. This chapter focuses on conceptual issues that contribute to the CPS ASEC underestimating income from two sources: retirement and self-employment.
As people approach age 65, they reduce their work hours at an increasing rate, and many move into formal retirement. As this process unfolds, earnings decline as a share of total family income and are replaced by a variety of types of retirement income. The vast majority of retirees receive
2 Work-related expenses are capped at the amount of the secondary earner’s earnings.
Social Security benefits, which many supplement with income from other sources. Not long ago, many received a traditional defined benefit (DB) pension, whereby employers paid retirees and their survivors a monthly benefit for life. In the private sector, DB pensions have been largely replaced by newer forms of retirement income in which employers and employees contribute funds to pension and profit-sharing accounts that the employees manage as financial investments until they need to withdraw funds to meet needs in retirement. DB pensions generate a regular income flow, very much like Social Security benefits, but the newer, defined contribution (DC) pension plans do not. Similarly, individual retirement accounts (IRAs) allow individuals to create their own tax-advantaged savings plans completely separate from employment. Here, too, the resources that people accumulate in such plans are later tapped by making withdrawals rather than receiving fixed payments. These new options for replacing the earnings that are forgone in retirement have given rise to both conceptual and measurement issues in determining how to value the resources that these plans generate.
Withdrawals from savings, in general, are not construed as income under the Census Bureau or alternative income concepts discussed below, but the tax-advantaged savings plans that are replacing DB pensions receive a different treatment. Thus, for purposes of federal income taxes, distributions from these plans are counted in adjusted gross income (AGI)—and taxed as ordinary income—except for returns of contributions made with after-tax dollars and rollovers to other tax-advantaged retirement vehicles. Similarly, CPS ASEC money income includes regular payments from an IRA, Keogh, 401(k), 403(b), or similar thrift plan. “Regular” is interpreted by the respondent, but comparisons with fund withdrawals that can be documented with administrative data indicate that very little of what is withdrawn from these funds is being reported as income in the CPS ASEC. For example, in 2004, Americans withdrew $139.9 billion from IRAs, excluding rollovers, according to data collected by the Internal Revenue Service (IRS) (Bryant, 2008). Of this total, $101.7 billion or 73 percent was taxable—that is, counted in AGI. For the same year, the CPS ASEC estimated only $6.8 billion in regular payments from not only IRAs but also Keogh and thrift plans (Czajka and Denmead, 2011). Another Census Bureau survey, the Survey of Income and Program Participation (SIPP), separates regular and lump sum withdrawals from IRAs, Keoghs, and thrift plans. The CPS ASEC estimate for 2009 was 23.4 percent of the SIPP estimate of regular withdrawals from these plans and 15.5 percent of the SIPP estimate of total withdrawals (Czajka and Denmead, 2011).
Withdrawals from the newer sources of retirement income are still dwarfed by payments from more traditional plans—namely, Social Security benefits and DB pension payments—and the CPS ASEC captures high percentages of these income sources (over 90 percent for Social Security; see
Czajka and Denmead, 2008). The mix is shifting, however, and the implication is that, without a new approach to defining and measuring retirement income from nontraditional sources, the CPS ASEC will understate the income of the elderly by an increasing amount in the years to come, which could introduce a trend toward overestimating medical care economic risk.
Theoretically, self-employment income is a net income: the revenue taken in by a business over a period of time minus the expenses incurred over the same period of time in order to generate that revenue. If a business generates inventory, then that would be factored in as well. If the business has a single owner (a sole proprietor), then all of the net income from the business should be assigned to the single owner. Alternatively, a business may have multiple partners, in which case the partnership’s net income is allocated among the partners. A given partner’s share of the net income is determined by that partner’s ownership share in the partnership.
Determining the net income from a business can be exceedingly complex. The proprietor’s tax return provides one measure, but is a tax-based measure conceptually appropriate for determining the contribution of the business to the owner’s overall economic well-being? The designers of the SIPP thought otherwise. They recognized that a business owner may draw a salary from a business, which could provide a positive income flow even if the business lost money overall. Rather than measuring just the profit or loss from a business, then, SIPP included in self-employment income the salary that the owner drew from a business. The net profit or loss was added to the salary to produce a measure of total self-employment income from that business for a given reference period.
The Census Bureau has not adopted in the CPS ASEC the SIPP approach to defining and measuring self-employment income. For this and other reasons, the CPS ASEC identifies substantially less self-employment income than the SIPP. For 2009, the SIPP estimate of aggregate self-employment income was 80 percent higher than the CPS ASEC estimate (Czajka and Denmead, 2012). Most of the difference occurred among families above 400 percent of poverty, for whom SIPP self-employment income was double that of the CPS ASEC, but SIPP was higher at all ranges of relative income above 150 percent of poverty.
Other Income Concepts
Accrued Versus Realized Income
An issue that must be addressed in defining income is whether income is counted as it is accrued or when it is realized (Nelson, 1987). Consider, for
example, a DC pension plan. The employer makes periodic contributions to the employee’s account. Each year the account earns interest or dividends or both. After many years, the employee begins to withdraw funds from the account. If income is counted as it is accrued, then the employer’s contributions will be counted in the year that they are made, and the interest and dividends will be counted in the year that they are credited to the account. If income is counted when it is realized, neither the employer’s contributions nor the interest and dividends will be counted until they are withdrawn. The purpose for which income is being measured determines which of these approaches is more appropriate. For macroeconomic applications, counting income as it accrues is equivalent to counting income as it is generated by the economy, and that will generally be the preferred approach. For applications to measuring the adequacy of income to meet recipients’ needs, however, counting income as it is realized may be more appropriate. The distinction between accrual and realization of income will be important when we consider what to do about assets when defining resources for the purpose of measuring MCER.
A frequent starting point for discussions of alternative income concepts is the notion advanced by Haig (1921) and later Simons (1938) that economic income is consumption during a period plus the change in net worth. Stated somewhat differently, economic income is the amount that can be consumed (over a specified period of time) without changing net worth. Implicit in this notion is the idea that a net growth in assets—or a net reduction in debt—is as much a part of income as a salary or wages. Haig-Simons income provides a useful framework for thinking about the broad range of resources that might be included in a measure of MCER.
Providing a sharp contrast to the accrual focus of Haig-Simons income is the concept of income that is applied to individuals by the federal tax code. This tax-based concept, which recognizes income only when it is realized, for the most part, is important to the discussion in this chapter not only to highlight the differences that exist in how income is defined, but also because major household surveys—including some that we discuss in Chapter 5—sometimes refer their respondents to their tax returns when collecting data on income. Tax-based concepts of income have become more relevant to medical expenditures with the passage of the Affordable Care Act, which imposes a uniform income concept defined in the tax code for determining eligibility for Medicaid, the Children’s Health Insurance Pro-
gram, and the new health insurance premium tax credits and cost-sharing reductions created under the law.
AGI, which is calculated on the front page of IRS Form 1040, is the amount of the taxpayer’s income that is subject to tax. Exemptions and deductions are subtracted from AGI to determine taxable income. Although there is substantial overlap between Census money income and AGI, AGI excludes some sources that are included in Census money income, excludes portions of other sources that are in Census money income, and includes some additional sources that are not included in Census money income.
The following sources, which are included in Census money income (although not necessarily well reported by survey respondents), are not taxable and therefore are excluded from AGI (Henry and Day, 2005):
- Interest and dividends on funds held in tax-deferred retirement accounts—these become taxable only when funds are withdrawn;
- Tax-exempt interest from state and municipal bonds;
- Workers’ compensation;
- Veterans’ benefits;
- Benefits from private disability insurance if the premiums were paid by the taxpayer;
- Public assistance and Supplemental Security Income;
- Child support;
- Assistance from friends and relatives; and
- Educational assistance used for tuition and books (i.e., educational expenses).
Of these, only tax-exempt interest is even reported on the tax return, and it appears on a separate line rather than as part of a total interest amount. If a survey questionnaire follows the tax return, then it would have to include separate questions to capture these several sources.
Moreover, the two largest components of Census money income— wage and salary income and Social Security benefits—are not fully taxable for most people and therefore may not be fully included in AGI. Taxable wage and salary income excludes pretax deductions for a variety of special purposes, which have been growing in type and total value. These include
- Contributions to a 401(k) or similar plan, which can be as high as 12 percent of gross earnings for workers who are not nearing retirement and higher for older workers, who are allowed to make contributions above 12 percent if they set aside less than the maximum amount in earlier years;
- Funds set aside for health care flexible spending accounts (up to $5,000 annually through 2012 and dropping to $2,500 thereafter);
- Funds set aside for dependent care (up to $5,000 annually);
- Health insurance premiums paid by the employee (this can be well above $10,000 annually for family coverage); and
- Transportation expenses (up to $125 monthly for transit fares as of 2012 and even more for parking).
Amounts excluded as pretax deductions are not reported on the tax return, so a survey questionnaire that asks respondents to report amounts from their tax returns will exclude these amounts from wage and salary income unless they are collected separately. These exclusions can add up to a sizable fraction of gross income over much of the earnings distribution, although there is age variation in the use of these different deductions. Younger families are more likely to use the dependent care deduction, whereas older families and individuals are more likely to set aside large amounts for flexible spending.
Social Security benefits may be wholly or partially excluded from taxation, depending on the total amount of the benefits and the taxpayer’s other income.3 Unlike other nontaxable income or the nontaxable portion of wage and salary income, all Social Security benefits must be reported on the tax return so that the nontaxable portion can be calculated. A survey questionnaire that asks respondents to report their Social Security benefits could request either the total or taxable amount (or both). If the questionnaire is not explicit about which one should be reported, a respondent could report either the total or taxable benefits, and which one was reported might not be evident from the response.
AGI also includes sources that are not included in Census money income. AGI includes capital gains except for the one-time exclusion of gains from the sale of a principal residence and the exclusion of capital gains that occur within a tax-deferred retirement account until they are withdrawn from the account. AGI also includes state income tax refunds received in the prior year, gambling winnings, and all withdrawals from retirement plans—not just regular withdrawals—except when such withdrawals are rolled over into another tax-deferred plan.
Consumption-Based Resource Measures
A number of economists argue that for the purpose of measuring the adequacy of resources for people at low-income levels, a measure based on consumption is more appropriate than a measure based on income (see, for example, Meyer and Sullivan, 2003). Income, it is said, understates
3 Railroad Retirement benefits are treated the same way, but we focus on Social Security benefits because they cover far more people.
well-being to a greater extent than consumption. This may be more a measurement issue than a conceptual one. That is, consumption tends to be reported more accurately than income among those with low income. At higher income levels, the reverse may be true. In addition, at higher income levels, people consume less of their income, so consumption will tend to understate well-being.
For the measurement of MCER, consumption-based measures are problematic, as MCER represents the likelihood of incurring medical consumption beyond what a family or individual can afford. If one includes out-of-pocket medical expenditures in the measure of resources, then such expenditures become affordable by definition. Another, more general issue with consumption-based measures of resources became evident in hindsight in the lead-up to the global recession beginning in 2008. Spending beyond one’s apparent means (one’s income) may indicate a risk of defaulting on future obligations—creating exactly the situation that MCER is intended to quantify. But consumption-based measures do have merit in pointing out that families that are consistently able to spend more than they take in as measured income are tapping into additional resources that are readily available. At a minimum, this should lead us to consider more directly the role of assets as resources.
ROLE OF ASSETS IN MEETING FINANCIAL NEEDS
In the context of how people pay for extraordinary and, especially, unexpected expenses, the role of assets cannot be overlooked. Assets accumulate over a lifetime. Under models of life-cycle saving, people accumulate savings (including funds held in retirement accounts) during their working years and then draw on these savings in retirement. Savings, together with Social Security and pensions, replace the earnings forgone in retirement.
A number of researchers have used data from the Health and Retirement Study to explore the relationship between health and the accumulation of assets. Several studies focused on expenditures in the last year of life, most recently Marshall, McGarry, and Skinner (2010). Others have looked at a broader span of years. Coile and Milligan (2009) examined the response of asset holdings to acute health events and new diagnoses. De Nardi, French, and Jones (2010) investigated savings behavior as a response to potential medical costs. More recently, Poterba, Venti, and Wise (2010) assessed the relationship between health and asset accumulation among the elderly and near-elderly. Using an index of health status constructed from a combination of self-reports, diagnoses, and activities of daily living, they found positive relationships between health and asset accumulation, which imply that poor health reduces asset accumulation. These last findings are of particular interest because they suggest that individuals with poor health
not only face greater prospects of high medical expenditures in the future, but also will be less well prepared to finance such costs.
Using data from the Medical Expenditure Panel Survey (MEPS), Banthin and Bernard (in Part III of this volume) compared the distribution of net assets by relative income of the elderly and the nonelderly in 2006, 2007, and 2008 (pooling the 3 years to increase sample sizes for key subgroups). The poor and low-income elderly had substantially more assets than the nonelderly, and in the upper deciles of the asset distribution, these assets became substantial. MEPS understates net assets relative to the Survey of Consumer Finances—a survey focused on the measurement of wealth—so it is possible that sizable assets extend even lower in the wealth distribution than these findings suggest.
To exclude assets entirely from the resources used to measure MCER, and in so doing make this a measure of income-related economic risk, ignores accumulating evidence on how families prepare for potentially high medical expenditures and how well they are able to absorb them. Unlike the measurement of income poverty, which compares a family’s income with a poverty threshold representing minimally sufficient expenses defined over a broad class of families, our proposed measure of MCER is intended to reflect the risk of incurring not only ordinary or expected expenses but also extraordinary expenses that are specific to each family. To meet these expenses without being pushed (further) into poverty, a family with sufficient assets could elect to draw on these additional resources. If the goal of the measure of MCER is to assess a family’s ability to pay for both expected and unexpected medical care costs, then the resources component of that measure must take account of at least a share of the assets that a family could readily convert into income.
In expanding the definition of resources in this way, our goal is not to provide an alternative measure of poverty, but to assess how large a future medical expense (over the next calendar year) a family could absorb without falling (deeper) into poverty. Resources counted under a measure of MCER but not the SPM will not move a family out of poverty but will reduce the risk that a family’s medical expenses could send the family deeper into poverty. We note that with the SPM definition of income, funds withdrawn from a retirement account to help cover medical or other expenditures in the prior year could in fact show up as income (depending on how the respondent interprets the questions on income from retirement accounts). Our recommendation to include a portion of assets in the resources used to measure MCER is loosely equivalent to applying this notion prospectively. Instead of counting only assets that were in fact converted to income (and from limited types of assets), however, our prospective measure of MCER would count assets that could be converted to income, and from a potentially broader array of sources.
What Assets Should Be Counted?
If assets are to be included as potential resources for assessing MCER, then we must consider what assets should be counted and how these assets should be included. In determining what assets should be counted in resources, the panel considered the types of assets that families hold, the access that they have to these assets, and whether assets should be included in resources for the entire population or just a portion of the population.
Types of Assets
Assets are commonly divided into financial assets and property assets. Financial assets include checking and savings accounts, certificates of deposit, stocks and mutual funds, and a variety of retirement accounts, which are primarily tax-advantaged. Property assets include homes, vehicles, rental and other real property, and businesses. The net value of property assets is the difference between what they would command if sold and the amount of debt that is held (through mortgages and loans). Families may have other liabilities in addition to those related to the purchase of property. Such liabilities must be considered in assessing the amount of assets available for the consumption of medical care. These include the tax liabilities that would be incurred in withdrawing funds from tax-advantaged retirement accounts and from any other accounts for which withdrawals would generate capital gains.
Access to Assets
A critical consideration in determining what types of assets to include in resources and how much value to assign these assets is the extent to which families have ready access to these assets. Financial assets are more accessible, clearly, than property assets, but through home equity loans and reverse mortgages, the elderly (and even younger persons) are able to extract fungible resources from the equity that they have built up in their homes. Assets held in tax-advantaged retirement accounts present unique issues for access. First, there are age restrictions on who can withdraw funds from these accounts without incurring significant financial penalties. For example, 59-and-a-half is a critical age for withdrawing funds from retirement accounts. Second, accounts held by employers may require substantial lead time to withdraw funds, making them effectively inaccessible for meeting short-term needs. This is particularly true prior to age 59-and-a-half, although much less of an issue after that age. Third, the taxes that must be paid on funds withdrawn from tax-advantaged accounts of any type (which are generally taxed as ordinary income) may not present
an immediate issue for withdrawing funds, but they lower their effective value.
Many of the surveys that could potentially support a measure of MCER collect no asset data or limited types of asset data (see Chapter 5), so fine distinctions among asset types may be of little use in the implementation of a measure of MCER. With that in mind, the panel recommends that a share of the value of financial assets held outside retirement accounts along with the posttax value of assets held in retirement accounts be taken into consideration as resources in measuring MCER (see “Conclusions and Recommendations” section). In principle, we would also include the amount received from a reverse mortgage, and we would include it as income rather than an asset, but we recognize the limitations of existing data.
For those who are still dependent on earnings to meet their basic needs, having to dip into financial assets to cover a needed medical expenditure may create a financial hardship down the road, due to a reduction in the savings needed later in life. The panel considered whether to include assets in the resources of persons at all ages or to do so only for persons who were no longer working or who had reached an age at which most people were not working. The availability of Medicare to those who reach age 65 changes the calculus for expected out-of-pocket expenditures for medical care, which affects prospective MCER when sick. For this reason, the panel thinks that MCER needs to be estimated separately for persons over and under age 65.
People retire across a wide range of ages, however, and they accumulate assets at widely varying rates. If assets were counted for all persons, regardless of employment or age, then the amount of assets accumulated—not an arbitrary cutoff—would determine the amount of assets included in resources. Other things being equal, retired persons and elderly persons would tend to have more liquid assets than those who are still working or who are not elderly, but the assets of younger persons or working persons who have accumulated more than their peers would not be excluded. Hence, although MCER needs to be estimated separately by age, the panel thinks it is a more acceptable approach to take account of assets for people of all ages and employment statuses in the estimation than counting or not counting assets depending on age or employment. However, the restrictions on access to funds held in retirement accounts prior to age 59-and-a-half would have to be recognized—by either excluding such funds from liquid assets below this age or discounting their value.
How Much Asset Value Should Be Counted in Resources?
Treating a family’s entire pool of liquid assets (as the panel defines them) as a resource that is available to offset a medical need is not a viable option, because this approach does not address the long-term financial hardship that would be created if a family consumed too large a share of its assets on medical expenditures—or any other purpose—in a single year. But if only a portion of a family’s accessible liquid assets can be counted toward offsetting MCER, how should that portion be determined? We consider two general approaches and their pros and cons.
If a family’s liquid assets grow at a rate of x percent per annum, on average, a family can withdraw that fraction of its liquid assets each year— minus any taxes owed on distributions from tax-advantaged accounts— without consuming principal. Adding a uniform fraction of liquid assets to income would provide a measure of resources that is consistent with a family’s ability to spend without drawing down its assets at an unacceptable rate.
Banthin and Bernard (in Part III) added 5 percent of net assets to annual income in order to assess the relative burden posed by family medical expenses. Families were identified as having high burdens if their expenditures exceeded a specified percentage (for example, 5, 10, or 20 percent) of this adjusted family income. They performed the same calculation without adding assets to income in order to assess how much the inclusion of assets reduced the burden of medical expenses at different income levels. The figure 5 percent was chosen as being “very close to what financial planners advise” as a draw-down rate for families in retirement.
Pension actuaries have a widely accepted approach to converting a lump sum amount into an annual payment. Commonly, the annual payment grows at an assumed rate of inflation. Life expectancy, as reflected in age and sex at a minimum, is the critical variable in determining the amount of the annual payment, with an interest rate and an inflation rate being included in the calculations as well. The actuarial approach is especially well suited to assigning an income value to assets, and the key assumptions can be based on those that are used in the annual reports of the Social Security and Medicare trustees or those that are used by the IRS to calculate minimum required distributions from IRAs (see, for example, Internal Revenue Service, 2011). A critical difference between the actuarial approach and the
first approach is that it allows the share of assets that would be added to resources to grow with age (declining life expectancy). More importantly, the methodology is well grounded in theory and well established in practice.
Pros and Cons
Defining the asset contribution to resources as a fixed percentage of asset value, rather than amount earned on assets during the most recent year, would prevent a large decline in the value of assets from producing negative family income. Likewise, calculating an annuitized value from the balance of liquid assets would also avoid generating a negative contribution during a year in which the value of asset holdings declined broadly. Furthermore, the actuarial approach is consistent with established methods of converting asset balances into income streams, which is exactly the need that we are addressing, and the assumptions that it requires could be obtained from those that are published each year by the Social Security and Medicare trustees. The drawback of the second approach is its complexity, given the limitations of the asset data to which it would be applied.
CONCLUSIONS AND RECOMMENDATIONS
The official poverty measure uses a concept of Census money income in conjunction with a set of thresholds, originally developed as the cost of a minimum diet times three for all other needs. A new supplemental poverty measure uses a different concept of income that includes tax credits, expenditures, and certain cash-equivalent benefits in conjunction with different thresholds. We have recommended (see Chapter 2) continuing the Census approach for purposes of defining income and resources for the SPM poverty measure, including recommended additional analysis related to medical care economic burden.
A fundamental question for the panel is whether the resources used to assess prospective MCER should be equated with either of these two income concepts or whether a different concept would be more appropriate.
We find, first, that there is a growing deficiency in both income concepts with respect to the treatment of retirement income, which is critical to the measurement of resources for the elderly, which is the age group with the greatest medical care needs. If resources are to be equated with income, then at a minimum this deficiency must be addressed—down the road if not in the near term. A deficiency with respect to the measurement of self-employment income is also notable, and this affects the nonelderly population.
We find, second, that for those who have access to them, assets are a potentially important resource for meeting unexpected medical needs—
particularly among the elderly, who may be depending in part on accumulated assets to offset the loss of earnings in retirement. We recommend that a portion of liquid assets be included in the resources of all persons, regardless of age or employment status except where restrictions on access may apply (as in retirement accounts). Although the panel finds the calculation of an annuitized value from the family’s liquid assets a compelling approach, we have not examined all the operational ramifications of adopting it, and thus we defer to those who are charged to implement a medical care economic risk measure. As to the choice between Census money income and disposable income, we recommend the use of disposable income, augmented by a portion of liquid assets, to facilitate comparisons with the SPM.
Recommendation 3-1: The panel recommends that the U.S. Census Bureau modify its concepts and measurement of money income and disposable income to better account for income flows from self-employment and from new forms of retirement income for use in measures of poverty and medical care economic risk and burden that are derived from its household surveys.
Recommendation 3-2: The panel recommends that, for measuring medical care economic risk, a portion of liquid assets be included in the resources of all persons, regardless of age or employment status. Only assets that the family or individual can access relatively quickly should be considered in determining the amount to be included—namely, financial assets held outside retirement accounts, the posttax value of assets held in retirement accounts, and, in principle, the amount potentially received from a reverse mortgage (treating it as income rather than as an asset), acknowledging the limitations of existing data.
Recommendation 3-3: The panel recommends that the method for calculating the share of liquid asset contribution to resources for measuring medical care economic risk be determined by the federal agency charged with producing the measures and that the methodology be based on one of two options—either a fixed share of assets or an annuitized value. The share of liquid asset contribution derived in this manner should be added to disposable income to provide the measure of resources for evaluating medical care economic risk.