National Academies Press: OpenBook
« Previous: 3 Growing Heterogeneity of the U.S. Population in Income and Life Expectancy
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

4

Implications of Growing Heterogeneity

How should society react to disproportionate historical and projected mortality gains among different segments of the population? Should groups that have larger gains in their life expectancy receive correspondingly larger gains in the present discounted values of their government benefits?

For most government programs, there is little concern for the question of the present value of benefits. For example, policy makers would not worry about people with a lower life expectancy receiving lower lifetime benefits from national defense or clean air because there is no obvious time dimension: in any given year, people who are alive pay taxes and receive benefits. But for programs with a strong or explicit time and age dimension, where the ages at which taxes are paid and benefits are received differ significantly, the principle of equal treatment requires consideration of such differences.

Since the inception of Social Security, for example, discussions of the philosophy of its benefit structure have revolved around two concepts. The first is the expected rates of return that individuals receive on the money that they and their employers pay in contributions during their working years. An “actuarially fair rate of return” would mean that an individual could expect to receive back from the system a stream of benefits with a present value equal to the present value of the contributions collected in his or her name. As is inherent in the start-up of a pay-as-you-go system, early cohorts of Social Security recipients received average rates of return on their contributions that were in excess of the actuarially fair rate of return. Given demographic trends, however, those retiring today and in future decades will receive average rates of return below the actuarially fair level, reflect-

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

ing the cost of transfers to earlier cohorts of retirees. Within a cohort (i.e., an age group at a point in time), one can compare expected rates of return for different segments of the population (e.g., income groups) and question the extent to which these are “equitable,” in the sense of different groups having similar expected rates of return.

The second central consideration in the philosophy underlying Social Security is the extent to which benefits provide an adequate safety net for individuals in the lower part of the income distribution. This safety net is achieved via a transfer from those with high lifetime income to those with low income.

The justification for considering equitable rates of return comes from several sources. The first is simple fairness: the notion that individuals should receive from the system a benefit commensurate with what they put in. This notion is particularly important if one views Social Security as a system to force people to save for their retirement. A second, related, consideration is based on political economy. If some elements of the population view Social Security as being unfair, then there would be support for dismantling the system. The third justification is efficiency: providing an actuarially fair rate of return minimizes the disincentive to work that would otherwise be associated with mandatory Social Security contributions, both during the working career and at the margin of the date of retirement.1 Many aspects of the Social Security system are designed to underline the equitable rates of return dimension. These include the designation of flows from workers as “contributions” rather than taxes, the fact that half of the contribution comes visibly from the worker’s paycheck, and the tracking of contributions over each individual’s working life.

The justification for the redistribution from high- to low-income individuals is a utilitarian concern for the poorest members of society. Unlike the return of one’s own contributions in the form of benefits, this redistribution is not as visible an element in the structure of Social Security. Rather, it is embodied in the benefit formula, which is opaque to most recipients. The tradeoff between equity in rates of return and redistribution from high- to low-income individuals is embodied most notably in the structure of the Social Security benefit formula that translates average indexed monthly earnings (AIME)2 into a primary insurance amount (PIA), as well as in provisions such as survivors’ benefits.

________________

1If workers recognize that their payroll taxes are associated with future benefits, then the effects of these taxes on labor supply should be muted relative to income taxes, which carry no marginal benefit. Whether workers actually respond differently to payroll taxes than to income taxes is unclear.

2The computation used by the Social Security Administration to define the AIME is described further below, in the section “Background on Social Security.”

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

IMPLICATIONS OF GROWING HETEROGENEITY FOR SOCIAL SECURITY

Because of the progressivity of the Social Security benefit formula, individuals with low lifetime earnings ceteris paribus on average receive higher expected rates of return on their contributions than those with higher lifetime earnings. Thus there is an inherent tradeoff between equity in terms of rates of return and the degree to which the system redistributes among income groups. In the current Social Security system, the gap in expected rates of return between low- and high-income individuals is neither as large as it would be under a flat old-age pension (in which all individuals received the same benefits regardless of the amount contributed) nor as small as it would be in a system of individual accounts in which there was no redistribution.

Another salient aspect of Social Security is that it is an annuity. Such a system necessarily entails redistribution from people who die young to those who die at older ages. This redistribution generates ex post inequity in terms of rates of return. Unlike the ex ante differences in returns that are generated by the benefit formula, this sort of ex post inequality is not necessarily perceived as a negative aspect of the system. There is a good reason for this: annuities are simply a form of insurance against living a long time, in which case there will be more years of consumption that have to be paid for. Similarly, people whose houses do not burn down earn a low rate of return on their fire insurance, while those whose houses do burn down earn a high rate of return; but the ex post inequality of rates of return does not seem problematic because those who earn the high rates of return need the money more and because, ex ante, one does not know which group one will be in.

The fact that Social Security benefits are paid in the form of an annuity may be taken to be a form of paternalism, in the sense that most well-informed and rational consumers would have chosen to sign up for an annuity anyway. Alternatively one may view Social Security as solving the problem of adverse selection in the annuity market.3

The distinction between ex post and ex ante inequity can be used to think about the effect of differences in life expectancy among groups on Social Security payouts. Much of the reason that ex post inequity may not offend notions of fairness is that it is not predictable. Among the population of living 60-year-olds, there are some who will receive high ex post rates of return because they live a long time, while others will receive low rates of return because they will die young, but mostly we do

________________

3Adverse selection in the annuity market refers to the observation that individuals who purchase annuities tend to live longer than people who do not buy such products. Longer-lived people are more costly for insurers, and their participation in the market raises overall prices (see Webb, 2006).

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

not know which people are which. Thus there is no perception of unfairness, and no distortion of decisions about labor supply.

However, when there are identifiable groups that vary in life expectancy, the inequity is more easily perceived. Furthermore, because the ex post inequity penalizes, on average, those with lower lifetime earnings, it undermines the progressivity of Social Security and has the potential to undo much of the redistribution embedded in the benefits formula.

To show the interaction of changing life expectancy with considerations of equitable rates of return and adequacy of benefits for lower-income individuals, the committee uses a very simple, stylized model. To match the analysis of data elsewhere in this report, we switch our focus from rates of return to the present value of net benefits received by different groups. However, these two concepts are closely related: for an individual with a given history of contributions, an increase in the present value of net benefits translates into a rise in the expected rate of return.

Consider a simplified scenario in which there are two equally sized groups: high income and low income. Within each group, all individuals have the same income. As a starting point, imagine a scenario in which the first group has higher lifetime wages (and thus higher contributions to Social Security) but in which the two groups have equal life expectancy. Also assume that the Social Security system is financially balanced, so that the present value of contributions from both income groups combined is equal to the net present value of benefits paid to both groups.

Figure 4-1 shows the relationship between the degree of redistribution incorporated into the benefit formula and the present value of benefits received by members of the two groups. Specifically, the horizontal axis represents the sensitivity of benefits to contributions. The left-most entry on this axis (“none”) indicates a system in which high- and low-income groups receive the same annual benefits. The right-most entry (“full”) indicates a system in which there is no redistribution embodied in the benefit formula. Each income group is represented by one curve on the graph.

The curve representing the relationship between the sensitivity of benefits to contributions and the present discounted value of benefits is downward sloping for the low income and upward sloping for the high income. Notably, in the case where there is no sensitivity of benefits to contributions (i.e., full redistribution) the two curves intersect, meaning that the two groups have the same present value of benefits despite their differences in earnings and Social Security contributions. At the right side of the graph, the gap in present value between high and low income is proportional to the gap in the present value of their Social Security contributions. Figure 4-1 also shows an initial level of sensitivity of benefits to contributions in the middle of the range, indicating a benefit formula with partial redistribution: the high-income group has higher present value of benefits than the

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-1 Present discounted value of benefits with equal mortality.

low-income group, but the gap in these benefits is smaller than the gap in the present value of contributions between the two groups. This “initial level” presumably represents some choice on the part of society regarding the proper balance of equity in rates of return versus adequacy of benefits for people with low lifetime income.

Now consider the effect of differential changes in mortality in this setting. In the simplest case, life expectancy rises for the high income but remains constant for the low income. The initial effect of this change, holding the benefits formula constant, would be to shift upward the curve representing the present value of benefits received by the high income without affecting the present value of benefits received by the low income (see Figure 4-2). For the same “initial level” in the sensitivity range, the gap in present value of benefits between the two groups would thus rise.4

________________

4Although this example focuses on present value of benefits, the same effect can be illustrated in terms of rates of return, as in the analysis by Goda and colleagues (2011). They look at the effect of differential mortality on the rate of return to Social Security contributions for stylized workers at different points in the earnings distribution. For example, considering the cohort born in 1938, males in the 25th percentile of the earnings distribution would receive an internal rate of return (IRR) of 1.51 percent versus an IRR of 0.75 percent for those at the 75th percentile, if the two groups experienced mortality at the average rate for males of their cohort. However, adjusting for the mortality rates actually experienced by these different parts of the earning distribution, the IRR for males at the 25th percentile decreases to 1.07 percent, while the IRR for males at the 75th percentile increases to 1.28 percent. In this particular case, the effect of differential mortality is sufficient to raise the IRR for high earners above that for low earners, but this is not always so; it is not true for some of the other male birth cohorts

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-2 The effect of decreased mortality among the high-income on relative present discounted value of Social Security benefits.

Results from the Future Elderly Model (FEM) can be used to illustrate the effect just described, anticipating more detailed results to be presented later in this chapter. Specifically, consider the change in the present discounted value of future benefits (taking all entitlement programs together) of 50-year-olds that results from projected changes in mortality between the cohort born in 1930 and the cohort born in 1960 (the latter based on the committee’s “best forecast”). For this exercise, the wage profiles of individuals are held constant, as is the policy environment regarding all government entitlement programs. For males in the bottom quintile of lifetime earnings, the present value of net benefits (i.e., benefits after age 50 minus taxes after age 50) would change very little, reflecting the small change in their life expectancy. Specifically, projected net benefits would fall from $319,000 to $310,000. By contrast, for males in the top quintile of lifetime earnings, the present value of projected net benefits would rise from $189,000 to $306,000.

The analysis of how differential mortality affects the present discounted value of net benefits is only a first step in the story. If life expectancy rises for high earners but is relatively constant for low earners, then that will raise the present value of net benefits of the former group and leave unchanged the present value of net benefits of the latter. However, if the Social Security system is to remain actuarially balanced, then something else has

________________

they examined or for any of the female cohorts they examined. Nonetheless, the accounting for differential mortality always made the system less progressive (Goda et al., 2011).

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

to change in response to this change in life expectancy. The easiest adjustment to think about is an increase in the normal retirement age (NRA), defined as the age at which beneficiaries receive “full” benefits under the Social Security benefit formula (described in more detail later in this report). Whatever its policy benefits and costs, an increase in the NRA has historically been part of the response of Social Security to rising life expectancies. An increase in the NRA will reduce the present value of benefits for both groups. In Figure 4-3, this is represented as a downward shift in the curves representing present discounted values for both high-income and low-income participants.

Given that the current system reflects a balancing of concern with equity in rates of return with adequacy of income for people with low income, one can see that the change in mortality in the absence of a change in the benefit formula moves the system in the direction of more equity and less adequacy—that is, in the direction of making it less redistributive. This can be seen in Figure 4-3, where the net result of changing longevity and the adjustment of Social Security NRA is that the present discounted value of benefits has risen for the high income and fallen for the low income and that the gap between these present values has increased. In the figure, these effects can be undone by shifting the vertical line representing the sensitivity of benefits to contributions to the left; in other words, making the formula that maps the AIME into a PIA have a larger redistributive component. This is shown in Figure 4-4.

image

FIGURE 4-3 An increase in normal retirement age (NRA) to offset lower mortality of the high income (red line). Arrows indicate the change in benefits curves, relative to Figure 4-2, due to the increase in NRA.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-4 A change in the benefit formula (shift of vertical line to the left) to counteract the combined effects of increased longevity for the high income and an increase in normal retirement age.

There are several subtleties that are ignored in this simplified presentation. Perhaps most significant, by focusing on the present discounted value of benefits rather than annual benefits themselves, the analysis above ignores the degree to which growth in life expectancy can strain the system. In the face of rising life expectancy, the present value of benefits is held constant by reducing annual benefits. But this constancy of expected return may be of little comfort to an elderly person who now has to get by with smaller annual benefits. In this simple example, it was possible to restore the present value of benefits of both high and low income to the same levels that existed before the change in longevity. Because the longevity of the low income did not change, this would require increasing their benefits to fully undo the increase in the NRA. As a result, the entire reduction in annual benefits would fall on the high income.

EFFECTS OF DIFFERENTIAL MORTALITY ON BENEFIT CLAIMING AND RETIREMENT INCENTIVES

The above discussion focused on the fairness aspects of mortality differentials. A second consideration is the effects of differential mortality on the incentives for Social Security benefit claiming. Social Security beneficiaries can choose to claim benefits earlier or later than the NRA, but the

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

benefits are adjusted in such a way as to provide the same present value of lifetime benefits, on average.5 Because those who claim early will, on average, receive more years of benefits than those who claim later, the average monthly benefit is adjusted down. Similarly, those who delay claiming benefits will receive fewer years of Social Security benefits, and their benefit is adjusted up.

The adjustment is intended to be actuarially fair so as to provide the same present value of benefits regardless of when someone claims benefits. But the benefits of early claiming depend on life expectancy. Consider those whose life expectancy after age 67 is 10 years; for them, claiming at 62 would increase their years of benefits by about 50 percent (15 years instead of 10 years).6 If, instead, their life expectancy at age 67 were 20 years, then early claiming would have a proportionally smaller effect on years of benefits (25 years instead of 20 years, or a 25% increase). Thus, in order for the system to be actuarially fair for high earners and low earners, the adjustment would have to depend on income in a way that changed over time in keeping with changes in life expectancy by income category.

Differential mortality in relation to lifetime earnings thus does two things. Relative to a situation in which the low and high income have the same life expectancy, differential mortality lowers lifetime benefits for lower earners. Second, it also raises the incentives for early claiming for the lower earners. A more efficient and arguably fairer system would have both the annual benefits and the early claiming adjustments indexed to life expectancy. This would mean that people with lower than average life expectancy (in this context, those with low lifetime income) would face a larger reduction in monthly benefits for early claiming, and similarly a larger increase in monthly benefits for late claiming, than would people with higher life expectancy. This sort of indexing could raise lifetime Social Security benefits for poor people and lower their incentives for early claiming relative to the current system. To the extent that retirement and claiming go hand in hand, this would also lower the incentive for early retirement.

The question of early retirement incentives is a difficult one, however. On the one hand, encouraging lower-income workers to delay retirement

________________

5Social Security benefit claiming and retirement may not occur simultaneously, because one can claim benefits but continue to work or retire but postpone claiming benefits. This discussion focuses on claiming because the benefit adjustment discussed depends on age of claiming, not age of retirement. For simplicity, our discussion assumes that the age of retirement is fixed (e.g., at the early retirement age), while the age of benefit claiming may vary. Of course, an incentive that encourages a worker to claim later may also lead him or her to retire later. If that occurs, then the worker’s monthly benefit amount would generally rise because of his or her longer work history (as well as because of the adjustment mechanism discussed here), although the worker would also be making additional payroll tax contributions.

6For simplicity, this calculation assumes that all retirees survive until at least age 67.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

would increase lifetime earned income as well as annual Social Security benefits. On the other hand, not penalizing the earlier retirement for lower-income workers might be a better social policy: these workers might be in poorer health, making work more difficult; they may have more physically taxing jobs; they may have worked for more years because they are less likely to have taken time out for education; and, because of their shorter life expectancy, they may want to retire early to ensure that they actually get to enjoy retirement for a few years in good health.

THE DISTRIBUTION OF SOCIAL SECURITY BENEFITS

The preceding discussion of conceptual issues surrounding the progressivity of Social Security under differential mortality abstracts from many details of how the program operates. In this section, the committee provides some additional background information on the Social Security program that is necessary for our subsequent discussion and reviews the empirical literature on the distributional effects of Social Security before turning to new estimates based on the FEM.

Background on Social Security

While the basic structure of Social Security is straightforward, there are many complexities that affect its distributional impact. Individuals are eligible to receive retired worker benefits if they have a minimum of 10 years (40 quarters) of covered earnings. To calculate the monthly benefit amount, past earnings are multiplied by a wage index to bring their value up to the present day. An average of the top 35 years of indexed earnings is calculated, which, converted to a monthly value, is the AIME. Next, a piecewise linear formula (straight lines connecting bend points) is applied to the AIME to create the PIA, which forms the basis for the monthly benefit amount. This formula introduces progressivity into the system because the rate at which the AIME is translated into PIA declines as AIME increases. In 2014, each dollar of average monthly earnings up to the first bend point of $816 is converted into 90 cents of PIA; the conversion factor is 32 percent of PIA until the next bend point of $4,917 and 15 percent for earnings beyond this value.

The monthly benefit amount also depends on the age at which benefits are first claimed. Workers may claim as early as age 62, the early entitlement age,7 and as late as age 70. Workers receive the PIA if they claim at the NRA, which has been rising over time from age 65 (for those born by

________________

7The term “early entitlement age” may also be referred to as “early eligibility age” in the research literature.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

1937) to 67 (for those born in or after 1960). Workers face an actuarial reduction (increase) for claiming before (after) the NRA, designed to ensure that the expected benefits received over a worker’s lifetime are roughly the same regardless of claiming age.8 A worker whose NRA is 67 receives a benefit equal to 70 percent of PIA by claiming at age 62 or equal to 124 percent of PIA by claiming at age 70.9

A few more relevant details pertain to other benefits and Social Security financing. Dependent and surviving spouses and children of insured workers are eligible for benefits, equal to 50 percent of the worker’s PIA for a dependent spouse and 100 percent for a surviving spouse. Many individuals are dually entitled as both a worker and a spouse but receive only the larger of the benefits to which they are entitled. The Social Security and Disability Insurance (DI) programs are integrated; the DI benefit calculation is largely similar to that for Social Security except that there is no reduction for early claiming, and DI eligibility requires passing a medical screening process as well as meeting recent work requirements. Finally, Social Security and DI benefits are funded by payroll taxes (or contributions) of 6.2 percent of earnings by both employers and employees (12.4 percent total) on workers’ earnings up to a taxable maximum amount: $117,000 in 2014.

Past Research on the Progressivity of Social Security

One way to estimate the progressivity of Social Security is to compare the replacement rate for workers at different points in the income distribution. The replacement rate is usually defined as the monthly benefit amount divided by pre-retirement average monthly career earnings.10 The Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds (2013) reported that the replacement rate for a worker who consistently earns the national average wage over his or

________________

8Whether the reduction factor is, in fact, actuarially fair for a typical worker is a matter of some dispute. Shoven and Slavov (2013) argue that the gains from delaying Social Security have increased dramatically since the 1990s because of a combination of low interest rates, increasing longevity, and legislated increases in the gain for claiming delays beyond the NRA (the Delayed Retirement Credit).

9A further complication in the benefit calculation is the Social Security earnings test. Before the NRA, workers face a reduction in benefits if they earn above an exempt amount ($15,480 in 2014). However, upon reaching the NRA, the worker is credited for any lost months of benefits through a recomputation of the actuarial adjustment. Although there is some evidence the earnings test may affect claiming behavior (Gruber and Orszag, 2003), it does not affect the (ex ante) progressivity of Social Security, and so the committee abstracts from it in our discussion.

10The replacement rate may also be calculated using final earnings or an average of earnings in the years just before retirement. Goss and colleagues (2014) compared replacement rates using alternative earnings measures.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

her career (a “medium-wage” worker) and retires at age 65 would be 41.7 percent. The replacement rate rises to 56.3 percent for a low- wage worker and to 77.4 percent for a very low-wage worker; it falls to 34.6 percent for a high-wage worker.11 The monthly benefit amount rises with past earnings, even though the replacement rate falls; the hypothetical high-wage earner would receive a benefit of $2,016 a month, versus $1,520, $923, and $705 for the medium-, low-, and very low-wage earners, respectively. Nonetheless, as measured by the replacement rate, the Social Security system is progressive in the sense that the system replaces a larger fraction of earnings for lower-income workers.

One clear drawback of the replacement rate measure is that it includes benefits but not contributions, yielding an incomplete picture of the program’s distributional impact.12 Several related measures of Social Security’s “money’s worth” address this shortcoming, as detailed by Geanakoplos and colleagues (1999). One such measure is the internal rate of return (IRR), the interest rate that a worker would have to receive on contributions to a (hypothetical) savings account so that the account balance at the time of the worker’s retirement would finance a stream of benefits equal to those promised by Social Security. A second measure is the benefit/tax ratio, which is the present value of lifetime benefits received divided by the present value of taxes paid (using an assumed rate of time preference, or discount rate). Another measure is the net transfer, which is the difference of the two present values rather than their ratio. As with the replacement rate, one might compare the money’s worth measures for people at different points in the income distribution to assess progressivity.13 An alternative approach, employed by Coronado and colleagues (2011), is to calculate the Gini coefficient, a measure of income inequality within a population, before and after Social Security benefits and taxes, to see if Social Security reduces (or increases) inequality.

It is well known that the money’s worth of Social Security has fallen over time, as the introduction of a pay-as-you-go system benefited early cohorts, whose benefits were quite generous in light of their modest contri-

________________

11Goss and colleagues (2014) compared these fictional workers to real workers from a large sample of 2011 claimants and found that the very low-, low-, medium-, and high-wage workers correspond to workers at the 12th, 25th, 56th, and 81st percentiles of the lifetime earnings distribution, respectively.

12Economic theory suggests that the incidence of employer contributions to Social Security may fall on workers, in the form of reduced wages; evidence from Gruber (1997) supports this hypothesis, and virtually all analysts adopt this convention in their calculations.

13In theory, one might also wish to compare how Social Security affects the utility (happiness) of individuals at different points in the income distribution. However, comparing utility across individuals would require making additional assumptions for which there is relatively little guidance from economic theory. Therefore, discussions of Social Security progressivity generally rely on financial, rather than utility, measures.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

butions, at the expense of later cohorts, who necessarily fared less well as the system matured and the worker to beneficiary ratio fell. Leimer (1995) reported that the 1900 birth cohort received an IRR of 11.9 percent, as compared to 4.8 percent for the 1925 cohort, 2.2 percent for the 1950 cohort, and 1.9 percent for the 1975 cohort.

For our purposes, however, it is more relevant to look at money’s worth within a birth cohort. Liebman (2002) did so for a sample of individuals born in 1925-1929. Grouping individuals in quintiles according to the AIME (of the higher-earning spouse, for married individuals), he found that the system is progressive. The IRR was 2.70 percent for the lowest AIME quintile in his cohort, 1.32 for the middle quintile, and 0.85 for the top quintile. With a 3 percent discount rate, all quintiles experience a negative net transfer, but the lowest AIME quintile loses $22,103 or 3.3 percent of earnings, versus $196,230 or 7.9 percent of earnings for the top quintile. With a lower discount rate of 1.29 percent, the two lowest AIME quintiles experience positive net transfers, on average.

The results obtained in any analysis of money’s worth depend, to some extent, on decisions the researcher must make in order to carry out the calculations. Chief among these is the choice of earnings measure used to determine an individual’s place in the income distribution. Gustman and Steinmeier (2001) and Coronado and colleagues (2011) found that the estimated progressivity of Social Security may be reduced or even eliminated when using lifetime rather than annual earnings, household rather than individual earnings, and potential (with full-time work at the current hourly wage) rather than actual earnings. These changes reduce progressivity because there may be people who have low earnings by the initial earnings measure and receive high net transfers who would be reclassified as higher earners under the new definition, such as a part-time worker (higher potential than actual earnings) or nonworking spouse (higher household than individual earnings). The inclusion of earnings above the taxable maximum increases progressivity, because it boosts the earnings of the highest-income workers and thus lowers their replacement rate. Using a higher discount rate is another decision that tends to reduce progressivity, by reducing the value of benefits received at very old ages, which accrue disproportionately to higher-income workers (Fullerton and Mast, 2005).

Another key factor that may affect the progressivity of Social Security and is of particular interest here is differential mortality. As described earlier, there are large and growing differences in mortality by socioeconomic status (SES) that would be expected, by themselves, to reduce progressivity. Liebman (2002) explores this empirically, using education and race/ ethnicity as measures of SES. When money’s worth is calculated using mortality probabilities that vary only by age and sex, low-SES groups gain more from Social Security than do high-SES groups. The IRR is 0.60 per-

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

centage points higher for blacks than for whites and 0.55 points higher for high school dropouts than those with some college or more. When race- and education-specific mortality tables are used, the estimated black-white difference in IRR falls to 0.10 points and the education difference falls to 0.17 points. In essence, the progressive effect of the Social Security benefit formula is largely undone by the fact that low-SES groups have lower life expectancies and so receive fewer years of benefits, on average.

While the discussion to this point has focused on Social Security retired worker benefits, disability insurance benefits are also relevant. DI benefits are likely to be quite progressive for several reasons (Meyerson and Sabelhaus, 2006). First, DI benefits are calculated using the same progressive formula as retired worker benefits. Second, low-income workers are empirically more likely to enter the DI program and receive benefits. Third, workers who end up on DI have shortened careers, which may make them more likely to be classified in a low-income group, depending on the earnings measure used in the analysis. In the Meyerson and Sabelhaus analysis, the overall Old-Age, Survivors, and Disability Insurance system is strongly progressive, with workers in the lowest quintile of household lifetime earnings having a benefit-tax ratio of 1.65 versus 0.65 for workers in the top quintile (for a sample of workers born in the 1960s, using a 3 percent discount rate and incorporating differential mortality). The lion’s share of this progressivity is due to DI benefits, which account for approximately 0.60 of the ratio at the 10th percentile of income versus about 0.05 at the 90th percentile.

A final program worthy of mention is the Supplemental Security Income (SSI) program, which is included in the calculations below. The program provides cash benefits to low-income individuals who are aged 65 and older, blind, or disabled. In 2014, the maximum monthly benefit amounts were $721 for single individuals and $1,082 for couples, but these benefits are, with few exceptions, reduced dollar-for-dollar against other income, including Social Security or DI benefits. Therefore, the program is expected to provide benefits primarily to very low-income individuals and to add to the overall progressivity of old-age support programs.

FEM Results on the Distribution of Social Security Benefits

The committee’s discussion now turns to the new estimates of the distribution of Social Security benefits generated specifically for this study using the FEM (see Box 4-1). To put these results in context and facilitate comparison with the previous literature, it is worth highlighting several aspects of the committee’s approach. First, our estimates represent the projected present value of the stream of benefits that individuals can expect to receive from age 50 until death, using a real rate of 2.9 percent to discount future

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

BOX 4-1
Cohorts and Scenarios in the Future Elderly Model (FEM)

The committee’s analysis of public programs focuses on two hypothetical cohorts that have the health and mortality experience of people born in 1930 and 1960. The FEM takes a cohort of Americans at age 50—each of whom has a measure of lifetime income and an initial health status—and simulates their lifetime benefits in a baseline scenario, which is based on the 1930 cohort. It then modifies the health and mortality experience to mirror that of the 1960 cohort. The model starts in 2010 with the policy environment observed in that year and assumed to persist throughout the simulation. The model is run biennially until everyone in the cohort has died; lifetime benefits and other outcomes are tracked. This establishes a baseline scenario against which other scenarios can be compared, such as what would happen if mortality differences changed by income group or what would happen if program eligibility or benefits changed.

benefits back to age 50. These estimates are not directly comparable to the money’s worth measures discussed above because they do not incorporate Social Security taxes and they represent the value of benefits as of age 50. Second, to classify individuals into quintiles, we use the average of nonzero earnings between ages 41 and 50; for married individuals, we sum household earnings and divide by the square root of 2. This approach is likely to generate lower estimates of progressivity because it uses: (1) an average of earnings rather than a single year’s earnings; (2) household rather than individual earnings; and (3) by excluding zero-earnings years, something closer to potential rather than actual earnings for a worker with an intermittent work history. Although decisions such as the choice of earnings measure are always important, they are arguably less critical in this case, given that our goal is not to estimate the money’s worth of Social Security per se but rather to assess how the distribution of Social Security benefits is changing over time in light of unequal longevity increases.

The experiment that the committee performs is based on the program rules as of 2010 and compares outcomes for two hypothetical mortality and health regimes. The first is based on the experience of the 1930 birth cohort, with its initial health status distribution by income quintile and estimated mortality gradient. The second is based similarly on the experience of the (simulated/projected) 1960 birth cohort health status and gradient. Health status does not enter directly into mortality or medical spending, so those outcomes are driven entirely by the mortality gradient as described

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-5 Average lifetime Social Security benefits for males (in thousands of dollars).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

in Chapter 3.14 Health does influence some economic outcomes, so the differences in initial health prevalence and simulated health transitions for the two cohort regimes will lead to some cohort differences in trajectories in earnings, workforce participation, Social Security claiming, SSI claiming, and DI claiming. And again we emphasize that our estimates are dependent on projections of mortality after age 50, rather than observed levels, for the 1960 cohort.

Given this setup, we calculate the present value of lifetime benefits received by each lifetime earnings quintile in each generation; for the 1960 cohort, this calculation is entirely based on projected net benefits. Later in this chapter, we also include taxes paid after age 50 to compute overall net benefit profiles from Social Security and other programs combined. (A description of the FEM’s estimation of taxes and net benefits is included in Chapter 2.) Because we do not attempt to allocate income taxes to each individual program, however, we use the net benefit concept only when examining the major entitlement programs in combination. We use a benefit-only approach when examining each program in isolation.

The baseline estimates of the present value of survival-weighted Social Security benefits for males by earnings quintile are displayed in Figure 4-5. For the 1930 cohort, benefits rise with earnings quintile. Workers in the

________________

14In the committee’s use of the FEM, health status does not directly influence mortality outcomes but the lower quintiles have both worse health status (e.g., more diabetes) and higher mortality.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

lowest quintile (quintile 1) can expect to receive, on average, $126,000 of benefits over the rest of their lives (discounted back to age 50), while workers in the top quintile (quintile 5) can expect to receive $229,000, a value which is $103,000 or 82 percent more than that for bottom quintile workers. The fact that higher earners receive higher benefits is not surprising, because the monthly benefit amount rises with the AIME, albeit in a nonlinear relation.

The real point of this calculation, however, is to see how the results change when one moves to the mortality and income experience based on the 1960 cohort. The committee’s results suggest that between the 1930 and 1960 cohorts, projected life expectancy at age 50 falls slightly for quintile 1 males (from 26.6 to 26.1 years), rises slightly for quintile 2 males (27.2 to 28.3 years), and rises more substantially for quintile 3 (28.1 to 33.4 years), quintile 4 (29.8 to 37.8 years), and quintile 5 males (31.7 to 38.8 years). The additional 6 to 8 years of life expectancy for the top three quintiles leads to large increases in their expected lifetime Social Security benefits, as seen in Figure 4-5, with projected benefits for the top quintile in 1960 reaching $295,000. For this cohort, the difference between the top and bottom quintiles is $173,000, or 142 percent of the bottom quintile’s benefit.

These results suggest that Social Security benefits are becoming more unequal over time because of gains in projected life expectancy that accrue disproportionately to those in the upper half of the income distribution. Under the mortality conditions of the 1960 cohort, the lifetime benefits advantage of the top quintile over the bottom quintile has grown by $70,000 ($173,000-$103,000). Although payroll tax contributions are not included in these calculations, it seems unlikely that their inclusion would change the key finding, given the magnitude of the benefit increases enjoyed by the top three quintiles.

The results for females, shown in Figure 4-6, also show benefits rising with earnings quintile in the 1930 cohort, with expected benefits of $112,000 for quintile 1 female and $208,000 for quintile 5 females. Benefits here are any received by the individual, including dependent spouse and survivor benefits derived from the earnings record of the spouse. Because of their lower career earnings and benefit entitlements, females’ total expected benefits are about 90 percent as large as those for males, even though they can expect to live 4 to 5 years longer. As for males, the gap between the top and bottom quintiles for female is large and widening over time, with values of 86 percent of bottom quintile benefits for the 1930 cohort versus 158 percent of bottom quintile benefits for the 1960 cohort. These percentage changes are a bit larger than those for males because the model predicts a decline in life expectancy for the four lower quintiles of females over time, so the expected benefits for bottom quintile females decline between the 1930 and 1960 cohorts. At the same time, the dollar gain by the

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-6 Average lifetime Social Security benefits for females (in thousands of dollars).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

top quintile relative to the bottom quintile of $48,000 is smaller than for males. But the overall message is the same for females as for males: diverging life expectancy is making Social Security benefits vary more by earnings quintile over time.

The results for DI benefits are shown in Figures 4-7 and 4-8 for males and females, respectively. Expected DI benefits are much smaller than expected Social Security benefits because the probability of ever receiving DI benefits is far lower. As discussed above, DI benefits are predicted to be distributed more toward lower career earnings, and the results bear this out. While Social Security benefits rise with earnings quintile, DI benefits decline sharply. In the 1930 cohort of males, for example, benefits are $25,000 for the lowest quintile, $14,000 for the second, and $4,000 for highest quintile. While a low-AIME worker on DI receives a smaller benefit than a high-AIME worker on DI, the low-AIME worker is so much more likely to receive DI that his expected DI benefit is larger. The pattern for females is the same, but the values are less than half as large, because of their lower career earnings and somewhat lower probability of ever going on DI, compared to males.

In this experiment, expected benefits for both males and females are quite stable across cohorts. This is perhaps unsurprising, given that the increases in life expectancy are concentrated in the third through fifth quintiles, which have relatively low probabilities of DI claiming. In results not shown here, the FEM predicts that the probability of claiming DI over a

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-7 Average lifetime Disability Insurance benefits for males (in thousands of dollars).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

image

FIGURE 4-8 Average lifetime Disability Insurance benefits for females (in thousands of dollars).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

2-year period for the 1930 cohort peaks around age 62 at nearly 20 percent for quintile 1 males versus roughly 10 percent for quintile 2 or 3 males and 5 percent or less for males in quintiles 4 or 5. The claiming behavior predicted by the FEM for later cohorts is similar. Thus, even though the 1960 cohort has a projected longer life expectancy than the 1930 cohort, this

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

does not necessarily translate into larger expected DI benefits because the types of people who are living longer are unlikely to claim DI. Furthermore, the increases in life expectancy are largely occurring after the NRA, when beneficiaries are no longer receiving DI benefits.

Finally, the results for SSI benefits are shown in Figures 4-9 and 4-10. As with DI, SSI benefits are larger for the lower quintiles because of their

image

FIGURE 4-9 Average lifetime Supplemental Security Income benefits for males (in thousands of dollars).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

image

image

FIGURE 4-10 Average lifetime Supplemental Security Income benefits for females (in thousands of dollars).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

higher probability of SSI receipt. For males in the 1930 cohort, expected benefits are $11,000 for the lowest quintile, $4,000 in the second, and $1,000 or less in quintiles 3 through 5. Values are about twice as large for females because of their longer life expectancy and higher probability of ending up with the very low income necessary to qualify for SSI. As with DI, changes across cohorts are relatively small.

THE DISTRIBUTION OF MEDICARE BENEFITS

Background on Medicare

The Medicare program is composed of two parts: Medicare hospital insurance, also known as Medicare Part A, helps pay for inpatient care in hospitals and skilled nursing facilities, as well as home health and hospice services. Medicare supplementary medical insurance, which consists of Medicare Parts B, C, and D, helps pay for physician, outpatient, prescription drug, and other services. People who are 65 and older and who have a minimum of 10 years (40 quarters) of Social Security covered earnings (or whose spouse has that minimum) receive Medicare Part A without paying a premium; most enrollees must pay a premium to receive the other parts of Medicare, although the premium covers just a small portion of the costs.15

Medicare shares many characteristics with Social Security. Medicare Part A is financed by payroll taxes paid during the working years, and the Medicare benefit is limited (largely) to those aged 65 and older.16,17 However, the Medicare program has different distributional effects from Social Security for a number of reasons. First, the Medicare hospital insurance payroll tax, which finances Medicare Part A, is levied on all wages, rather than on wages up to a cap, as in Social Security. Second, Medicare supplementary medical insurance is financed by general revenues, which consist mostly of income taxes collected through a progressive income tax structure. Finally, whereas the Social Security benefit increases as lifetime earnings increase, Medicare offers essentially the same benefit package to

________________

15Other than for high-income enrollees, the premiums for Medicare Part B (which covers physician, outpatient hospital, and some home health services) and Medicare Part D (which covers prescription drugs) are set at about 25 percent of program expenditures (Cubanski et al., 2014). The premium for Medicare Part C, which allows Medicare beneficiaries to enroll in private health insurance plans as an alternative to traditional Part A and Part B coverage, varies based on the chosen plan.

16Medicare hospital insurance also receives funding from taxation of the Social Security benefits of high-income taxpayers.

17Two other major groups that are eligible for Medicare are those with end-stage renal disease and those who have received DI benefits for 2 years (Rupp and Riley, 2012).

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

all beneficiaries; the value of this benefit is arguably greater for those with lower income.18

Valuing the Medicare benefit requires addressing some conceptual issues. First, because Medicare is an in-kind benefit rather than a cash benefit, the value placed on it by its recipients might not be equivalent to the cost to the government of providing it. That is, if beneficiaries were given the cash value of Medicare, some might choose to spend that cash on items other than health care. So, to measure the utility effect of Medicare, one might want to make an adjustment to reflect the fact that not all beneficiaries would value Medicare at its cost. For example, the Census Bureau, when valuing Medicare benefits, chooses to value them at their “fungible” value, which is a measure of what beneficiaries might have spent on health insurance in the absence of Medicare.19 However, that method ignores the fact that even beneficiaries who couldn’t afford to purchase Medicare’s health benefits on their own still place some positive value on the benefits received.

Second, one might think that, because all Medicare beneficiaries receive the same medical insurance, one should value it the same for all—perhaps at the average per-beneficiary cost. However, the committee’s view is that people who expect to use Medicare benefits more—those in poorer health, for example—would place greater value on it.

For the purposes of this report, we take the simple approach and value Medicare expenditures by lifetime income at their actual cost. Thus, we are measuring the actual government transfers received by people of different lifetime income and not necessarily measuring the welfare effects of such transfers on those individuals.

________________

18One exception is that high-income beneficiaries face higher premiums for Medicare Parts B and D, further increasing the progressivity of Medicare. For a full description of the high-income premiums under Medicare, see http://kff.org/medicare/issue-brief/income-relating-medicare-part-b-and-part/.

19The Census Bureau explains this concept as follows: “The fungible approach for valuing medical coverage assigns income to the extent that having the insurance would free up resources that would have been spent on medical care. The estimated fungible value depends on family income, the cost of food and housing needs, and the market value of the medical benefits. If family income is not sufficient to cover the family’s basic food and housing requirements, the fungible value methodology treats Medicare and Medicaid as having no income value. If family income exceeds the cost of food and housing requirements, the fungible value of Medicare and Medicaid is equal to the amount which exceeds the value assigned for food and housing requirements (up to the amount of the market value of an equivalent insurance policy (total cost divided by the number of participants in each risk class).” See http://www.census.gov/hhes/www/income/data/historical/measures/redefs.html [July 2015].

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

Past Research

Previous analyses of Medicare progressivity have come to differing conclusions. Bhattacharya and Lakdawalla (2006), using years of education as a measure of SES, found that annual Medicare expenditures are much larger for the less well educated than for the better educated. They calculated the net present value of Medicare Part A, which is funded only by payroll taxes, and conclude that the net actuarial value for Medicare Part A is significantly larger for the less well educated, noting “While Medicare is actuarially unfair for college graduates, high school dropouts almost double their money” (Bhattacharya and Lakdawalla, 2006, p. 278).

In contrast, McClellan and Skinner (2006), using the income of a beneficiary’s zip code as an indicator of SES, found the picture less clear cut. For example, they showed that the annual distribution of health spending by zip code income decile changed over time. In the 1980s, beneficiaries living in lower-income neighborhoods had lower Medicare expenditures than those living in higher-income neighborhoods; by the late 1990s, this trend had reversed. They attribute much of this change to the growth in home health spending. Including both the distribution of annual benefits by zip code and differential mortality, they found the distributional consequences of the Medicare program to be roughly neutral in dollar terms. Thus, even though—unlike Social Security—Medicare provides a uniform health insurance benefit to all, McClellan and Skinner found that the higher expenditures of the rich combined with their longer life expectancy are enough to offset their higher tax payments.

Thus, the literature about the progressivity of Medicare is inconclusive. The difference between the Bhattacharya and Lakdawalla study and the research by McClellan and Skinner might be attributable to the differences in Medicare Part A (examined by Bhattacharya and Lakdawalla) versus overall Medicare expenditures (examined by McClellan and Skinner) or to differences between individual education and zip code income as measures of SES.

Results from the FEM

An advantage of the FEM for assessing the value of Medicare is that it is able to link lifetime income to actual medical expenses. Figures 4-11 and 4-12 show the distribution of annual Medicare expenditures at ages 67 and 77 for males and females born in 1930. The findings from the FEM are unambiguous: those with lower lifetime income have higher annual Medicare expenditures. For example, for 67-year-old males, the Medicare expenditures in the lowest income quintile are 48 percent higher than in the top quintile; for females at this age, the ratio is 69 percent. The ratio

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-11 Average annual Medicare spending for males born in 1930, by income quintile.

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

image

FIGURE 4-12 Average annual Medicare spending for females born in 1930, by income quintile.

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

attenuates somewhat with age, most likely reflecting the fact that the least healthy people in the bottom quintile die earlier (and are out of the sample). At age 77, for example, the Medicare expenditures for males and females in the bottom quintile are 32 percent and 47 percent higher, respectively, than expenditures for those in the top quintile.

The committee’s analysis now turns to the estimates of the lifetime distributional effect of the Medicare benefit generated using the FEM. These estimates reflect Medicare expenditures in the period 2002 to 2004—that is, they abstract from rising overall Medicare expenditures over time—adjusted so that they are in 2010 dollars. The result is that differences in benefit receipts across the two hypothetical cohorts arise only from changes in underlying health and life expectancy, not from the ongoing rise in cost per beneficiary across cohorts.

The baseline results showing average lifetime Medicare benefits for males by earnings quintile are displayed in Figure 4-13. For the 1930 cohort, lifetime Medicare benefits are relatively flat by earnings quintile: males in the lowest quintile can expect to receive, on average, $162,000 in lifetime Medicare benefits, only 6 percent more than those in the top quintile. Thus, for the 1930 cohort of males, the higher annual Medicare expenditure of those in the lower-income quintiles is roughly offset by their shorter life expectancy.

image

FIGURE 4-13 Average lifetime Medicare benefits for males (in thousands of dollars).
SOURCE: Committee generated using Health and Retirement Study data and cohort
assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-14 Average lifetime Medicare benefits for females (in thousands of dollars).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Looking forward, however, widening disparities in life expectancy and associated health status change this picture substantially. For the 1960 cohort of males, the projected mortality gradient produces an upward gradient in the distribution of lifetime Medicare benefits with income. For example, those in the bottom income quintile can expect to receive $158,000 in lifetime Medicare benefits, just 78 percent of the lifetime benefits for those in the top quintile.20

The results for females, shown in Figure 4-14, are somewhat different, reflecting both the distribution of annual Medicare benefits and the smaller disparities in life expectancy for females in the 1930 cohort. Females in the lowest quintile receive about 30 percent more in lifetime Medicare benefits than those in the top quintile. But, as with the males, the income gradient changes over time. For the 1960 cohort, for example, the lifetime Medicare benefit for females in the lowest income quintile is expected to be only 92 percent of the benefit in the top quintile.

________________

20As noted above, these calculations do not account for growth in overall Medicare expenditures; the average Medicare benefit received by those in the 1960 cohort is likely to be many times greater than the average benefit of the 1930 cohort.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

THE DISTRIBUTION OF MEDICAID BENEFITS

Background on Medicaid

Medicaid is a program that provides health insurance to those with low income and low assets. It is administered by the states within broad federal guidelines, and the eligibility requirements vary widely across the states. For example, in many states, nondisabled childless adults are ineligible for Medicaid, regardless of income. But all states cover the low-income disabled and low-income elderly, and it is these two groups who account for most of the Medicaid expenditures in the population aged 50 and older that this study addresses.

For elderly and Medicare-eligible disabled Medicaid beneficiaries, most acute health expenditures are financed by Medicare, although Medicaid helps with Medicare premiums and coinsurance. But Medicaid is the primary payer of long-term care services, particularly nursing homes, which are generally not covered by Medicare.21

Although Medicaid as a whole is undoubtedly progressive—it is financed by general revenues and provides health care for the poor—there is some question as to the progressivity of the Medicaid nursing home benefit. In most states, Medicaid can be a payer of last resort for nursing homes because people can become Medicaid eligible by spending down their assets.22 Thus, even people with relatively high lifetime incomes may end up on Medicaid if they require nursing home care for a lengthy period of time. For example, De Nardi and colleagues (2013) found that, for those retirees in the top two quintiles of the income distribution, Medicaid recipiency increases with age, rising from around 4 percent at age 89 to more than 20 percent at age 96. Given the sharp increase in nursing home use with age (Brown and Finkelstein [2008] report that the median age of first entry into a nursing home is about 83 years old) and the longer life expectancy of those with higher lifetime income, the lifetime impact of the Medicaid benefit is worth examining.

________________

21Medicare pays for short-term nursing home stays following hospitalizations but not for long-term nursing home use. The Medicare home health benefit has increased greatly over time and has now become an important source of financing for this form of long-term care (Brown and Finkelstein, 2008).

22Medicaid eligibility for the disabled and non-elderly requires that assets and income both fall below certain thresholds. But in many states, that income level is fairly high—up to $2,130 per month in 2013; see http://longtermcare.gov/medicare-medicaid-more/medicaid/medicaid-eligibility/share-of-cost/ [July 20115]. In these states, Medicaid-eligible individuals are required to spend most of their income on nursing home care, and Medicaid will cover the remainder of the costs.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

Results from the FEM

Figures 4-15 and 4-16 display the present value of Medicaid benefits at age 50 by lifetime earnings quintile. Despite the fact that some Americans with high lifetime income do rely on Medicaid for financing long-term care, the Medicaid benefit is much larger for lower earners than higher ones. For example, for males of the 1930 birth cohort, the present value of Medicaid from age 50 on is $77,000 for those in the lowest earning quintile, $35,000

image

FIGURE 4-15 Average lifetime Medicaid benefits for males (in thousands of dollars).
SOURCE: Committee generated using Health and Retirement Study data and cohort
assumptions.

image

FIGURE 4-16 Average lifetime Medicaid benefits for females (in thousands of dollars).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

in the second quintile, and just $16,000 for those in the highest quintile. For females—who are much more likely to use nursing homes—the disparities are even larger: the average lifetime Medicaid benefit from age 50 is about $164,000 for females in the lowest earnings quintile but only $21,000 for females in the highest quintile.

Widening disparities in life expectancy over time diminish the extent to which Medicaid benefits decline from lower to higher income quintiles. For example, moving from the 1930 to the 1960 birth cohort reduces the ratio of benefits of those in the bottom quintile to those in the top quintile from about 500 percent to 350 percent for males and from 800 percent to 500 percent for females.

THE DISTRIBUTION OF TOTAL NET BENEFITS FROM MEDICARE, MEDICAID, SOCIAL SECURITY, AND SUPPLEMENTAL SECURITY INCOME

In this section, the committee combines the present value of total benefits, by lifetime earnings quintile, from Medicare, Medicaid, Social Security (including retirement and disability), and Supplemental Security Income and then subtracts taxes paid after age 50 (including personal income taxes, personal payroll taxes on wages, and employer payroll taxes) into a “total net benefit” in present value. As Figure 4-17) and decline less rapidly as earnings rise for females than in the 1930 cohort (Figure 4-18).

For males, the impact of moving from the 1930 cohort to the 1960 cohort is to reduce total net lifetime benefits by 3 percent for those in the bottom quintile and to raise such net benefits by 62 percent for those in the top quintile. For females, that shift reduces net benefits by 17 percent for the bottom quintile and raises them by 28 percent for the top quintile. To examine what is driving these effects, one can analyze the impact on benefits and taxes separately. For males in the 1930 cohort, as shown in Figure 4-19, the present value of total benefits is estimated at about $400,000 in both the bottom and top quintiles. For females, as shown in Figure 4-20, the top quintile has lower average lifetime benefit levels than those at the bottom, largely because Medicaid benefits, which deliver larger benefits to those toward the bottom of the earnings distribution, are a larger factor in the total for females than for males.

The growing gap in life expectancy and associated health conditions, however, is projected to change these patterns significantly, as the figures illustrate for the 1960 cohort compared with the 1930 cohort. Whereas the gap in present value benefits between the highest quintile and the low-

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-17 Average lifetime total net benefits for males (present value in thousands of dollars), by lifetime earnings quintile. Net benefits equal total benefits minus taxes, all from age 50 onward. Total benefits include those from Medicare, Medicaid, Social Security (including retirement and disability), and Supplemental Security Income. Taxes include personal income taxes and payroll taxes (both employer and employee).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

image

FIGURE 4-18 Average lifetime total net benefits for females (present value in thousands of dollars), by lifetime earnings quintile. Net benefits equal total benefits minus taxes, all from age 50 onward. Total benefits include those from Medicare, Medicaid, Social Security (including retirement and disability), and Supplemental Security Income. Taxes include personal income taxes and payroll taxes (both employer and employee).

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-19 Average lifetime total benefits for males (present value in thousands of dollars), by lifetime earnings quintile. Total benefits include those from Medicare, Medicaid, Social Security (including retirement and disability), and Supplemental Security Income.

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

image

FIGURE 4-20 Average lifetime total benefits for females (present value in thousands of dollars), by lifetime earnings quintile. Total benefits include those from Medicare, Medicaid, Social Security (including retirement and disability), and Supplemental Security Income.

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

est in the 1930 cohort was zero for males and −$129,000 for females, for the 1960 cohort the gap is projected to become $132,000 for males and $28,000 for females. Thus the advantage in lifetime total benefits for the top quintile grew by $132,000 for males and by $157,000 for females.

Figures 4-21 and 4-22 illustrate the tax side, at least for taxes paid at age 50 and above. As seen, higher earners pay more in taxes than lower earners. However, the pattern is not markedly different between the 1930 and the 1960 cohort. In other words, the changes in mortality do not generate substantial changes in taxes paid. The implication is that almost all of the change in the pattern of lifetime net benefits shown in Figures 4-17 and 4-18 is due to the impact of mortality on benefits and not taxes. (Because the committee had anticipated this result, which makes intuitive sense, and because the focus here is on the impact of a steeper mortality gradient rather than the level of net benefits, the committee’s initial work had excluded taxes altogether.)

We now turn to measures of progressivity. Figures 4-23 and 4-24 show how total net benefits change as a share of the committee’s inclusive wealth measure because of the change in the mortality gradient. For both males and females, three features of the progressivity measures are noteworthy. First, these government programs represent a substantial share of inclusive wealth at age 50; net benefits amount to greater than half of inclusive

image

FIGURE 4-22 Average lifetime total taxes paid for females (present value in thousands of dollars), by lifetime earnings quintile, from age 50 onward. Taxes include personal income taxes and payroll taxes (both employer and employee).
SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-23 Total net benefits as share of inclusive wealth as of age 50 for males.
SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

image

FIGURE 4-24 Total net benefits as share of inclusive wealth as of age 50 for females.

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-21 Average lifetime total taxes paid for males (present value in thousands of dollars), by lifetime earnings quintile, from age 50 onward. Taxes include personal income taxes and payroll taxes (both employer and employee).
SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

wealth for females in the lowest quintile and almost half for males in the lowest quintile. Even for the middle quintile of earnings, net benefits represent one-quarter to one-half of wealth. Second, for both the 1930 and the 1960 cohorts, net benefits are a larger share of inclusive wealth for lower earners than higher earners, suggesting that at least on this measure, the government programs as a whole are progressive for both cohorts. Third, our focus is mostly on the change in progressivity, not its level. And on that score, the change in mortality has made these government programs less progressive; the difference between the highest quintile and the lowest quintile has fallen by 7 percentage points (from 31 to 24) for males and by 9 percentage points (from 44 to 35) for females because of the more rapid rise in life expectancy for higher earners than lower earners.

SENSITIVITY OF RESULTS TO MORTALITY CHANGE

As discussed in Chapter 3, there is considerable uncertainty about whether the differences in life expectancy by midcareer earnings will continue to widen. The committee’s baseline projection or simulation, referred to in this report as the mortality regime of the 1960 birth cohort, is based

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

on the estimated mortality model. The actual mortality differences, though, could be larger or smaller than these estimates. How would that affect the results? Suppose, for example, that the actual differences for this cohort turn out to be substantially smaller than in the estimated model. To assess this possibility, we constructed a scenario, described in Chapter 3, in which the widening of the life expectancy differences between the 1930 and 1960 birth cohorts is only half as great as our fitted model would imply. We have called this the “half dispersion” regime. We have investigated how this alternate mortality outcome would affect our results for the differences in the present value of benefits, taxes, and their difference, net benefits.

Figures 4-21 and 4-22 show that the present values of taxes paid after age 50 under the mortality regimes of 1930 and 1960 barely differ at all. Therefore it will come as no surprise that the same is true for the half dispersion regime, which lies between the other two. The largest percentage difference between the 1960 and half dispersion mortality regimes is 1.3 percent, and all other differences are less than 1 percent. The real question is how the half dispersion regime affects the present value of benefits and net total benefits.

Figure 4-25 plots the difference between the present value of total benefits for the top income quintile and the bottom quintile, for the mor-

image

FIGURE 4-25 Difference in present value (in thousands of dollars) of total lifetime benefits between top and bottom income quintiles, for three mortality regimes: 1930 cohort, half dispersion, and 1960 cohort.

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×

image

FIGURE 4-26 Difference in present value (in thousands of dollars) of total lifetime benefits net of taxes between top and bottom income quintiles, for three mortality regimes: 1930 cohort, half dispersion, and 1960 cohort.

SOURCE: Committee generated using Health and Retirement Study data and cohort assumptions.

tality regimes of 1930, half dispersion, and 1960. The lines are very nearly straight, which means that the effect of mortality dispersion on the gap in the present value of benefits is approximately linear. The gap is about half as great in the half dispersion regime as it is for the 1960 cohort. Because mortality and survival rates enter into the calculation of the present value of benefits in a nonlinear way, this result was not obvious before the half dispersion scenario was run. It is convenient, however, because it means that one can evaluate the outcomes for any degree of change in the mortality dispersion that one believes to be appropriate.

Similarly, Figure 4-26 plots the gap between top and bottom income quintiles for total benefits net of taxes for the three mortality regimes. The result is the same. The relationship is approximately linear. When the mortality dispersion increases only half as much, the increase in the gap in present value of net total benefits is only half as great.

The committee is not able to provide a probability distribution for the size of the increase in mortality-income dispersion between the 1930 and 1960 birth cohorts, but we note that all the evidence reviewed in Chapter 3 indicates that the mortality-income differential has continued to widen during the past two decades.

Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 65
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 66
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 67
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 68
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 69
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 70
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 71
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 72
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 73
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 74
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 75
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 76
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 77
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 78
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 79
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 80
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 81
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 82
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 83
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 84
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 85
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 86
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 87
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 88
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 89
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 90
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 91
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 92
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 93
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 94
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 95
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 96
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 97
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 98
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 99
Suggested Citation:"4 Implications of Growing Heterogeneity." National Academies of Sciences, Engineering, and Medicine. 2015. The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses. Washington, DC: The National Academies Press. doi: 10.17226/19015.
×
Page 100
Next: 5 Policy Responses to an Aging Population »
The Growing Gap in Life Expectancy by Income: Implications for Federal Programs and Policy Responses Get This Book
×
Buy Paperback | $64.00 Buy Ebook | $49.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The U.S. population is aging. Social Security projections suggest that between 2013 and 2050, the population aged 65 and over will almost double, from 45 million to 86 million. One key driver of population aging is ongoing increases in life expectancy. Average U.S. life expectancy was 67 years for males and 73 years for females five decades ago; the averages are now 76 and 81, respectively. It has long been the case that better-educated, higher-income people enjoy longer life expectancies than less-educated, lower-income people. The causes include early life conditions, behavioral factors (such as nutrition, exercise, and smoking behaviors), stress, and access to health care services, all of which can vary across education and income.

Our major entitlement programs – Medicare, Medicaid, Social Security, and Supplemental Security Income – have come to deliver disproportionately larger lifetime benefits to higher-income people because, on average, they are increasingly collecting those benefits over more years than others. This report studies the impact the growing gap in life expectancy has on the present value of lifetime benefits that people with higher or lower earnings will receive from major entitlement programs. The analysis presented in The Growing Gap in Life Expectancy by Income goes beyond an examination of the existing literature by providing the first comprehensive estimates of how lifetime benefits are affected by the changing distribution of life expectancy. The report also explores, from a lifetime benefit perspective, how the growing gap in longevity affects traditional policy analyses of reforms to the nation’s leading entitlement programs. This in-depth analysis of the economic impacts of the longevity gap will inform debate and assist decision makers, economists, and researchers.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!