February 2013 • No.

383

Income Composition, Income Trends, and Income Shortfalls of Older Households By Sudipto Banerjee, Ph.D., Employee Benefit Research Institute A T

A

G L A N C E

 For all age groups above 65, Social Security remains the primary source of income. In 2009, households ages 65–74 and households with members age 85 or above received 54 percent and 66 percent of their total household incomes, respectively, from Social Security benefits.  The importance of Social Security income increases with age. For households that had members ages 65– 69 in 2001, the share of household income derived from Social Security rose from 47 percent in 2001 to almost 60 percent in 2009.  Income from pensions and annuities is the second-largest source of income for older households. In 2009, households ages 65–74 received 17.1 percent and households above age 85 received 15.3 percent of their incomes from pensions and annuities.  In 2009, two-fifths of households with members age 65 and above had incomes less than their expenditures—meaning they had deficits.  In 2009, 14.3 percent of households with members age 65 and above had spending that exceeded 175 percent of their household incomes.  Households that face income shortfalls not only have much lower levels of assets, they spend down their liquid assets at a faster rate than households with no income shortfalls.  The probability of running into an income shortfall is much higher for those with lower incomes. In 2009, 66.4 percent of households age 65 or older in the bottom-income quartile faced income deficits, while only 6.8 percent in the top-income quartile faced such shortfalls.  Median home- and health-related expenses, as well as median total expenses, are much higher for households that face income shortfalls, even if they have lower levels of income.  Singles, households with no pensions, African-Americans, and Hispanics have larger shares of households with income deficits.

A monthly research report from the EBRI Education and Research Fund © 2013 Employee Benefit Research Institute

Sudipto Banerjee is a research associate at the Employee Benefit Research Institute (EBRI). This Issue Brief was written with assistance from the Institute’s research and editorial staffs. Any views expressed in this report are those of the author and should not be ascribed to the officers, trustees, or other sponsors of EBRI, EBRI-ERF, or their staffs. Neither EBRI nor EBRI-ERF lobbies or takes positions on specific policy proposals. EBRI invites comment on this research.

Copyright Information: This report is copyrighted by the Employee Benefit Research Institute (EBRI). It may be used without permission but citation of the source is required.

Recommended Citation: Sudipto Banerjee, “Income Composition, Income Trends and Income Shortfalls of Older Households,” EBRI Issue Brief, no. 383 (Employee Benefit Research Institute, February 2013). Report availability: This report is available on the Internet at www.ebri.org

Table of Contents Introduction .......................................................................................................................................................... 4  Data ..................................................................................................................................................................... 4  Income Composition .............................................................................................................................................. 5  Income Comparison With CPS ................................................................................................................................ 7  Change in Household Income ................................................................................................................................. 7  Income Shortfalls .................................................................................................................................................. 9  Change in Wealth ................................................................................................................................................ 11  Income Shortfalls across Income Quartiles ............................................................................................................ 12  Expenditure Patterns of Households with and without Deficits................................................................................. 12  Does a Pension Make Any Difference? ................................................................................................................... 14  Demographic Differences ..................................................................................................................................... 14  Conclusion .......................................................................................................................................................... 14  References .......................................................................................................................................................... 17  Endnotes ............................................................................................................................................................ 17 

Figures Figure 1, Median Income, in 2010 $s, from Different Sources, and Average Percentage Share of Each Source in Total Income, by Age Group .......................................................................................................................................... 6 Figure 2, Comparison of Percentage of Different Income Sources in Total Household Income of 65+ Households between the Health and Retirement Study (HRS) and Current Population Survey (CPS) .................................... 8 Figure 3, Longitudinal Change in Total Household Income (Median), in 2010 $s, for Different Age Cohorts .................... 8 Figure 4, Longitudinal Change in Share of Income from Different Sources, for Those Ages 60-64 in 2001 ...................... 8 Figure 5, Longitudinal Change in Share of Income from Different Sources, for Those Ages 65-69 in 2001 .................... 10 Figure 6, Longitudinal Change in Share of Income from Different Sources, for Those Ages 70-74 in 2001 .................... 10 Figure 7, Percentage of Households With Household Incomes Less Than Household Expenditures, by Age Groups I‒III ............................................................................................................................................. 13

ebri.org Issue Brief • February 2013 • No. 383

2

Figure 8, Percentage of Age 65+ Households With Different Degrees of Income Deficit ................................................. 13 Figure 9, Longitudinal Change in Median Total Household Wealth and Non-Housing Wealth (in 2010 $s) for those 65+ in 2001 ................................................................................................................................................................. 13 Figure 10, Percentage of 65+ Households With Household Incomes Less Than Household Expenditures, By Income Quartile ................................................................................................................................................................ 13 Figure 11, Median Expenditure (in 2010 $s) in Each Spending Category for 65+ Households, by Income Deficit .......... 16 Figure 12, Percentage of 65+ Households With Household Incomes Less Than Household Expenditures, by Pension Status .................................................................................................................................................................. 16 Figure 13, Percentage of 65+ Households With Household Incomes Less Than Household Expenditures, by Single/Couple Status ........................................................................................................................................... 16 Figure 14, Percentage of Age 65+ Households With Household Incomes Less Than Household Expenditures, By Race ............................................................................................................................................................... 16

ebri.org Issue Brief • February 2013 • No. 383

3

Income Composition, Income Trends, and Income Shortfalls of Older Households By Sudipto Banerjee, Ph.D., Employee Benefit Research Institute Introduction A steady, reliable source of retirement income is essential to financial security in retirement. Policymakers, employers, and financial-service providers have long tried to design policies and products that can help retirees acquire steady retirement income, but such policies and products must also respond to the constantly changing financial environment. Furthermore, in order to adapt these policies and products for future retirees, it is important to understand the current state of income in retirement. This Issue Brief examines the income patterns of older U.S. households. Using data from the Health and Retirement Study (HRS) and Consumption and Activities Mail Survey (CAMS), this study combines both income and expenditure data to gain a more complete picture of the retirement income adequacy of current retirees. It focuses on three different aspects of retirement income:  The sizes of the different components of income and their shares of total household income.  How the importance of each component of retirement income changes with age for different cohorts.  Finally, it estimates the percentage of older households with retirement incomes that are not sufficient to finance their spending needs. The study also identifies the expenditure components that drive households to income deficits and identifies the key demographic characteristics of these households.

Data As noted above, two different sources of data are used for this study. First is the Health and Retirement Study (HRS), a study of a nationally representative sample of U.S. households with individuals over age 50. It is the most comprehensive survey of older Americans in the nation, covering in detail topics such as health, assets, income, and labor-force status. It is a biennial, longitudinal survey with survey waves in even-numbered years beginning in 1992. The initial sample consisted of individuals born between 1931 and 1941 and their spouses, regardless of their birth years. Newer cohorts have been added in the following survey years. The study is sponsored by the National Institute on Aging (NIA) and the Social Security Administration (SSA) and administered by the Institute for Social Research (ISR) at the University of Michigan.1 The other data used in this study come from the Consumption and Activities Mail Survey (CAMS), which was started in 2001 as a supplement to the HRS. From the 2000 HRS, 5,000 households were selected at random and mailed the CAMS questionnaire. In couple households, the questionnaire was sent randomly to one of the two spouses. Since 2001, CAMS has been conducted every two years, with 2009 the latest round of available data. For those between ages 55–64, the aggregate expenditures in each of the 32 categories (six durable and 26 nondurable items) of CAMS are very close to the same categories in the Consumer Expenditure Survey (CEX) , the benchmark survey on household consumption in the United States. However, CAMS reports higher consumption expenditures for older households (Hurd and Rohwedder, 2011). The household income and wealth measures are taken from the RAND version of HRS data because it provides consistent measures of income and wealth across all waves.

ebri.org Issue Brief • February 2013 • No. 383

4

Income Composition Figure 1 shows the income composition of older households between the years 2001 and 2009, by age group. It is important to note the different components of income as shown below.  Labor income: sum of wage and salary income; bonuses, overtime pay, commissions, tips; second job; professional-practice or trade income.  Capital income: sum of business or farm income, self-employment earnings, business income, gross rent, dividend and interest income, trust fund or royalties, and other-asset income.  Pension/Annuity income: sum of all pension and annuity payments. This includes income from defined benefit pensions, annuities, as well as income from other retirement savings such as 401(k)-type plans and individual retirement accounts (IRAs).  Social Security income: includes Social Security retirement, spouse, and widow or widower benefits.  Other income: sum of Social Security disability benefits, unemployment and workers’ compensation, veterans’ benefits, food stamps, alimony, lump-sums from insurance, pensions or inheritance. All income figures are annual and reported at the household level, so the respondent’s and spouse’s incomes are added for a couple household. The figure reports both mean and median (mid-point) dollar amounts (in 2010 dollars) for each income component, as well as total household income. It also reports the mean percentage share of each income component in total household income. All numbers are reported for three different age groups: ages 65–74 (Age Group I), 75–84 (Age Group II), and ages 85 and above (Age Group III). Note that for all three age groups, Social Security remains the most important source of income, and its importance increases with age. For example, in 2001, households in Age Group I derived half of their income from Social Security while households in Age Group III derived almost 70 percent of their income from Social Security. These shares did not change much over the 9-year period studied. In 2009, the shares for Social Security income for Age Groups I and III were 54 percent and 66 percent, respectively. Also, due to the progressivity of the Social Security benefit formula, mean and median benefits from Social Security are very close. Note also that income from pensions and annuities is the second-most important source of income for most older households. The share of pension and annuity income increases from Age Group I to Age Group II, but then falls for Age Group III. For example, in 2009, the shares of pension and annuity income were 17.1 percent for Age Group I, 18.4 percent for Age Group II and 15.3 percent for Age Group III. There also exist large differences between mean and median incomes from pensions and annuities. In 2009, the mean of pension and annuity income were $11,612, $9,682 and $4,917 for Age Groups I, II, and III, respectively, while median pension and annuity incomes for all three age groups were zero, because only a share of the older population receives such income. Labor and capital are the other two important sources of income for older households. Labor income is a significant part of income for Age Group I, but, as expected, the share of labor income falls rapidly with age. In 2001, households in Age Group I derived 11.4 percent of their income from labor earnings, but for Age Groups II and III the shares were 2.6 percent and 0, respectively. These proportions did not change much across the years studied. On the other hand, capital income remains a steady source of income for all age groups; in 2001, capital income’s share remained almost unchanged at about 13 percent for all age groups, while in 2009, the share represented by capital income actually increased, from 9.9 percent for Age Group I to 13.8 percent for Age Group III. Income from other sources remains small and declines with age. Total household income expectedly falls with age, and there remain large differences between mean and median household incomes. The presence of a few large earners drives this difference between mean and median household incomes, but mean and median incomes of older households did not change much throughout the past decade. For Age Group I, the 2001 average and median household incomes were $50,975 and $34,192, respectively. For the same ebri.org Issue Brief • February 2013 • No. 383

5

ebri.org Issue Brief • February 2013 • No. 383

6

12,719 12,951 13,253 16,452 4,330 59,708 10,802 12,975 11,926 17,274 9,385 62,364 13,120 14,065 11,347 17,407 5,327 61,269 9,677 8,591 11,612 18,441 4,647 52,970

Labor Capital Pension/Annuity Social Security Others Total Income

Labor Capital Pension/Annuity Social Security Others Total Income

Labor Capital Pension/Annuity Social Security Others Total Income

Labor Capital Pension/Annuity Social Security Others Total Income

Labor Capital Pension/Annuity Social Security Others Total Income

0 73 0 17,666 0 37,201

0 354 0 17,164 0 41,001

0 401 1,205 16,780 0 39,537

0 245 910 15,499 0 37,024

Median $0 492 649 15,111 0 34,192

11.9 9.9 17.1 53.9 7.0

14.1 13.0 18.1 47.7 6.8

13.5 12.6 18.9 47.4 7.4

14.7 11.3 18.4 47.6 7.7

% of Total Income 11.4% 13.3 18.6 50.2 6.2

2,139 8,482 9,682 17,236 5,645 43,187

2,341 11,463 10,485 16,931 2,544 43,766

2,037 9,510 11,897 17,228 2,160 42,835

1,598 7,179 11,068 15,896 2,231 37,974

Mean $1,456 7,790 8,871 15,017 1,879 35,015 3.2 12.4 20.4 59.0 4.8

3.1 13.8 21.7 56.9 4.3

3.8 14.0 19.4 57.6 5.0

3.5 11.6 18.4 60.6 5.7

2005 0 497 3,562 16,712 0 30,846 2007 0 622 1,336 16,702 0 30,647 2009 0 100 0 17,068 0 29,626

% of Total Income 2.6% 13.1 19.2 60.0 4.9

2003 0 445 2,275 14,788 0 26,434

Median $0 615 1,418 14,225 0 25,112

2001

Ages 75–84

Source: Employee Benefit Research Institute estimates from Health and Retirement Study (HRS).

Mean $9,135 11,991 10,358 16,000 3,489 50,975

Ages 65–74

358 6,245 4,917 12,780 1,439 25,741

243 17,499 6938 13,854 2,615 41,152

173 4,744 5,856 13,793 1,706 26,273

188 5,009 5,460 13,166 1,458 25,284

Mean $18 5,431 4,454 12,832 1,402 24,199

Figure 1 Median Income, in 2010 $s, from Different Sources, and Average Percentage Share of Each Source in Total Income, by Age Groups

0 7 0 12,222 0 19,623

0 105 0 13,185 0 20,122

0 66 167 13,331 0 19,913

0 42 0 12,442 0 17,078

Median $0 147 0 12,763 0 16,761

Age 85+

0.5 13.8 15.3 65.7 4.5

0.5 14.4 16.7 62.6 5.5

0.3 11.6 17.1 66.2 4.6

0.3 9.7 15.5 69.5 4.8

% of Total Income 0.0% 13.0 12.9 69.1 4.7

age group in 2009, mean and median household incomes were $52,970 and $37,201, respectively, while for Age Group III mean household income increased from $24,199 in 2001 to $25,741 in 2009, and median household income increased from $16,761 to $19,623 during the same period.

Income Comparison With CPS Gustman, Steinmeier and Tabatabai (2012) compared pension-income data from the Current Population Survey (CPS) with pension-income data from the HRS and found that pension income data from the CPS understates its share of income among households ages 65–69. The CPS is one of the data sources often used for research on topics related to retirement income, so it is important to point out if there are other systemic differences in retirement income components between the CPS and other data sources. Figure 2 compares the shares of different components of household income between the CPS and the HRS for households age 65 and older. The CPS numbers are taken from the EBRI Databook on Employee Benefits (Table 7.5, http://ebri.org/pdf/publications/books/databook/DB.Chapter%2007.pdf). First, note that the share represented by Social Security income is much higher in the HRS than in the CPS. It should be noted that the CPS measure includes Social Security Disability Insurance (SSDI) payments, while the HRS measure does not. Still, the differences are very large. Secondly, the share represented by labor earnings is much higher in the CPS than in the HRS, though survey design differences can drive some of these differences. Thirdly, the differences between the two data sets are not large in pension and annuity income or in asset income, but in most years the share of income for both those categories in the CPS is slightly higher. Finally, income from other sources is higher in the HRS than in the CPS, although this could be due to the difference in components used to measure this income category.

Change in Household Income How income sources change in retirement is a topic of interest not only to retirees, but to financial service providers and policy makers as well. Figure 3 shows how median household income changed for different age cohorts during the period of the study, 2001 to 2009. Three age cohorts were considered: households with at least one member ages 60– 64 in 2001 (Cohort I), ages 65–69 in 2001 (Cohort II), and ages 70–74 in 2001 (Cohort III). One of the advantages of using HRS data is that it tracks the same households over time. Fig. 3 exploits this panel aspect of the HRS to show the change in household income of the same set of households. First, looking only at calendar year 2001, the cross-sectional difference in income for households of different ages can be noted. Cohort I had a median household income of $47,100, while Cohort II had $39,419, and Cohort III had $36,043. Now, observing the change in the panel over time, this cross-sectional finding can be noted as well. For example, for Cohort I, median household income dropped from $47,100 in 2001 to $37,653 in 2009, close to the cross-sectional difference in 2001. For Cohort II, median household income dropped from $39,419 in 2001 to $32,284 in 2009, and for Cohort III it dropped from $36,043 to $30,199 during the same period. The steady drop in income of retired households coincides with a drop in their household expenditures (Banerjee, 2012). This might indicate that retired households adjust their expenditures to their declining income. The next three figures show how the composition of income changed for these cohorts during the period of 2001— 2009. Figure 4 shows the changes in income composition for Cohort I, where it can be noted that, in 2001, the largest share (40 percent) of income for this cohort came from labor earnings. Over the period of the study, this share declined steadily, slipping to 12.2 percent in 2009. The second-largest source of income (20.3 percent) for this cohort in 2001 was Social Security, with a share of income that increased steadily, becoming the largest source of income between 2003 and 2009, when it reached 56.2 percent. The third-largest component of income (14.7 percent) in 2001 for this cohort was capital income, but its share fell to 10.1 percent in 2009. On the other hand, the share of pension/annuity income increased slowly from 12.7 percent in 2001 to 15 percent in 2009, when it was the second-largest source of income. Finally, the share of income represented by other sources almost halved during the period of the study, falling from 12.2 percent in 2001 to 6.4 percent in 2009.

ebri.org Issue Brief • February 2013 • No. 383

7

Figure 2 Comparison of Percentage Share of Different Income Sources in Total Household Income of 65+ Households between the Health and Retirement Study (HRS) and Current Population Survey (CPS) OASDI (Social Security) CPS HRS* 42.0% 55.6% 41.9 53.7 40.1 52.2 38.6 52.5 41.5 57.6

2001 2003 2005 2007 2009

Pension & Annuities CPS HRS 19.5% 18.2% 20.6 18.7 19.3 19.6 18.6 18.4 19.2 17.4

Asset Income CPS HRS 17.0% 13.2% 13.9 11.5 13.6 12.9 15.6 13.5 11.3 10.9

Labor Earnings CPS HRS 19.4% 7.2% 21.7 9.4 24.8 9.1 25.3 9.2 25.7 7.7

Other CPS 2.1% 1.9 2.2 1.8 2.3

HRS 5.6% 6.5 6.1 6.1 6.2

Source: Employee Benefit Research Institute estimates from Health and Retirement Study (HRS). * HRS shares include only OASI (Social Security) benefits.

Figure 3 Longitudinal Change in Total Household Income (Median), in 2010 $s, for Different Age Cohorts Cohort I Cohort II Cohort III

2001 $47,100 39,419 36,043

2003 $45,171 38,737 34,475

2005 $43,298 36,578 33,283

2007 $42,036 35,313 31,382

2009 $37,653 32,284 30,199

Source: Employee Benefit Research Institute estimates from Health and Retirement Study (HRS). Cohort I : Ages 60–64 in 2001. Cohort II : Ages 65–69 in 2001. Cohort III : Ages 70–74 in 2001.

Figure 4 Longitudinal Change in Share of Income from Different Sources, for Those Ages 60‒64 (Cohort I) in 2001 60%

50%

Labor

Capital

Pension/Annuity

Social Security

Other

40%

30%

20%

10%

0% 2001

2003

2005

2007

2009

Source: Employee Benefit Research Institute estimates from Health and Retirement Study (HRS).

ebri.org Issue Brief • February 2013 • No. 383

8

Figure 5 shows similar changes in income shares for Cohort II between 2001 and 2009. The main difference between this cohort and Cohort I is that the relative importance of the top two components of income remained unchanged for Cohort II. Throughout the period, Social Security remained the primary source of income for this cohort, rising from 47 percent of household income in 2001 to almost 60 percent in 2009. Income from pensions and annuities remained a distant, second-most-important source of income, with its share remaining almost unchanged throughout the period. As expected, the share represented by labor income fell steadily, from 15.8 percent in 2001 to 5.7 percent in 2009. In fact, in 2009, labor income was the smallest share of household income for Cohort II. Capital income and income from other sources remained more or less stable, the former’s share dropping from 12.3 percent in 2001 to 10.5 percent in 2009, while other income’s share changed slightly from 7.3 percent to 6.5 percent during the same period. Similarly, Figure 6 shows the changes for Cohort III. The patterns are similar to Cohort II: Social Security remained the primary source of income throughout the period, but its share was higher than it was for Cohort II. In 2001, the households in Cohort III derived 53.8 percent of their income from Social Security, which increased to 63.5 percent in 2009. Pensions and annuities, the second-largest source of income, provided 20.8 percent of Cohort III’s income in 2001, which remained almost unchanged until 2007 before dropping to 17.7 percent in 2009. Capital income remained the third-largest source of income for this cohort throughout the period, but its share declined from 13.4 percent in 2001 to 10.4 percent in 2009. The share represented by labor income dropped from 7 percent in 2001 to 3.2 percent in 2009 (when it was the smallest source of income).

Income Shortfalls In retirement preparation studies much attention is given to income adequacy in retirement. Sophisticated models have been built to project future income and expenditures to determine appropriate levels of income adequacy. By examining the existing data on both income and expenditure for current retirees, it is possible to get a picture of income adequacy. As mentioned above, the HRS and its supplement, CAMS, can be used together for this purpose. Figure 7 shows the percentage of households with household incomes less than household expenditures. An important point to note is that the income used here, and throughout the study, is gross income, while expenditures are financed by net income. As a result, some households that have gross incomes higher than their expenditures, but with net incomes lower than their expenditures, may be wrongly classified as having incomes that exceed their expenditures. It is very difficult to estimate the tax bracket of each household with the survey information, and the analysis does not attempt to do so. Rather, it is recommended that the percentage of households with incomes less than expenditures shown in this study be treated as the lower bound, which means the actual percentage of such households could be higher. Also, it should be noted that merely having incomes less than expenditures does not necessarily mean that these households cannot afford their expenditures or that they have run out of money. Most households have positive asset holdings, which can be tapped to bridge the gap between income and expenditures. But this chart still provides a good sense of the percentage of households that do not have regular sources of income sufficient to finance their spending. Finally, the following analysis in Figure 7 is based on the yearly cross-sectional data and does not refer to changes in a panel. Figure 7 shows the percentage of inadequate-income households by age group and calendar year. As in Figure 1, three separate age groups are considered: households with at least one member age 65–74 (Age Group I), 75–84 (Age Group II) and 85 and above (Age Group III).The final column provides the numbers for all households age 65 and older. Some important observations can be noted. First, in all different age groups, the percentage of inadequate-income-households decreased during the period studied. In 2001, 53 percent of households in Age Group I did not have adequate income to finance their expenditures, but by 2009 that fell to 37.2 percent. For Age Group II, the drop in income-inadequate households was from 57.5 percent in 2001 to 43.9 percent in 2009. For the oldest group (Age Group III), the percentage of income-inadequate households dropped from 64.6 percent in 2001 to 46.3 percent in 2009. Overall, the percentage of households age 65 and older with inadequate income fell from 55.5 percent in 2001 to 40.5 percent in 2009. While the drop in income-

ebri.org Issue Brief • February 2013 • No. 383

9

Figure 5 Longitudinal Change in Share of Income from Different Sources, for Those Ages 65‒69 (Cohort II) in 2001 70%

60%

50%

40%

Labor

Capital

Pension/Annuity

Social Security

Other

30%

20%

10%

0%

2001

2003

2005

2007

2009

Source: Employee Benefit Research Institute estimates from Health and Retirement Study (HRS).

Figure 6 Longitudinal Change in Share of Income from Different Sources, for Those Ages 70‒74 (Cohort III) in 2001 70%

60%

50%

40%

Labor

Capital

Pension/Annuity

Social Security

Other 30%

20%

10%

0%

2001

2003

2005

2007

2009

Source: Employee Benefit Research Institute estimates from Health and Retirement Study (HRS).

ebri.org Issue Brief • February 2013 • No. 383

10

inadequate households is good news, serious concerns remain since more than two-fifths of households age 65 or older outspent their income. It is important to know what percentage of households lack adequate income to finance their expenditures, but a single number may not be very informative, because some households might have incomes that fall just short of their expenditures while others might have far deeper shortfalls. It is important to know how many households have income shortfalls that are acute, as they are the ones that might have to draw down their accumulated savings faster than anticipated. Figure 8 shows the percentage of households age 65 or older with different degrees of income shortfalls between the years 2001 and 2009. The income shortfall, or deficit, is calculated as the difference between income and expenditure, measured as a percentage of income. For this analysis, households are divided into six different deficit categories: Households with no deficit (Category I); households with deficits of less than 10 percent (Category II); households with deficits between 10 percent and 25 percent (Category III), households with deficits between 25 percent and 50 percent (Category IV), households with deficits between 50 percent and 75 percent (Category V), and households with deficits of more than 75 percent (Category VI). Note that among the categories with deficits (Categories II through VI), the highest percentage of households belong to Category VI—meaning the highest level of deficit. While this is alarming, the percentage of households belonging to this category has dropped steadily over time: in 2001, 22.7 percent of households age 65 or older had income deficits of 75 percent or greater, which fell to 14.3 percent in 2009 (phrased another way, they spent 175 percent or more of their income in 2009). As with the drop in the overall percentage of inadequate-income households, the percentage of households in every deficit category (except Category III) dropped between 2001 and 2009. The percentage of households in Category III increased slightly, from 7.1 percent in 2001 to 7.5 percent in 2009, while in Category II the drop between 2001 and 2009 was very small: 6.4 percent to 5.5 percent. In other categories, the drops were larger— for example, in Category IV the percentage of households with income deficits fell from 10.8 per-cent in 2001 to 7.8 percent in 2009.

Change in Wealth As noted earlier, an income shortfall does not necessarily mean that a household cannot sustain its expenditures. It might have enough savings to cover the income shortfall, for instance. Consequently, it is also important to check the level of savings that these households have, and the rate at which they are drawing down their savings. Figure 9 tracks the change in total household wealth and non-housing wealth of a group of households that had at least one member age 65 or older in 2001. Comparing their 2001 incomes and expenditures, these households are divided into two groups: households with deficits and those with no deficits. Notice that the levels of wealth (both total wealth and nonhousing wealth) are much lower for households with 2001 deficits than in households with no 2001 deficits. For example, in 2002, households with no 2001 deficits had a median total wealth of $304,939, compared with $152,061 for households with 2001 income deficits. The difference (in percentage terms) was even higher for nonhousing wealth. This suggests that households with larger accumulated savings can keep expenditures within their incomes, while households with smaller savings cannot. Of course, this renders the latter group more vulnerable to running out of money in retirement. A better picture of this can be formed by looking at the rates at which both types of households liquidate, or decumulate, their wealth. Households with no deficits in 2001 experienced a 15 percent decline in their median total wealth between 2002 and 2010, from $304,939 to $260,000. On the other hand, households with deficits in 2001 had a slightly lower (10.4 percent) decline in their median wealth, from $152,061 to $136,200, during the same period. It is important to note that total wealth includes housing wealth, and that part of the change in total wealth could be due to the change in the value of houses, rather than active spending down by the households. A better understanding of spending down can be obtained by examining the change in non-housing assets. Between 2002 and 2010, non-housing assets for households with no 2001 deficits dropped 21 percent, compared with a drop of ebri.org Issue Brief • February 2013 • No. 383

11

31 percent for households with 2001 deficits. So households with deficits not only have much lower levels of liquid assets but they also draw down their assets at a faster rate than households without deficits. The rest of this article identifies how households with and without deficits are divided across demographic lines and the factors contributing to income shortfalls.

Income Shortfalls across Income Quartiles To identify which households are more likely to run into deficits, it is important to understand their income and expenditure patterns. A key question: How are the households with income shortfalls spread across the income spectrum? Are they concentrated on the lower end of the income scale or are they spread evenly across it? Figure 10 sheds some light on this, showing the percentage of households age 65 or older in each income quartile with household incomes less than household expenditures. Not surprisingly, the share of households with income deficits is highest in the lowest-income quartile. However, the second- and third-income quartiles also contain significant shares of such households. In 2001, 86.5 percent of households in the bottom quartile had income shortfalls, while almost 20 percent of households in the top-income quartile had expenditures that exceeded their incomes. However, during the period of 2001 to 2009, the share of households with income deficits decreased in every income quartile. The top quartile had only 6.8 percent of households with income deficits in 2009, but the bottom three quartiles still had large shares of households with income deficits: 66.4 percent in the bottom quartile, 50.9 percent in the second quartile, and 29.2 percent in the third income quartile.

Expenditure Patterns of Households with and without Deficits To identify what factors contribute to those income deficits, it is important to know both the income and expenditure patterns of households. Figure 10 shows that these households are more likely to have lower incomes. But why can’t these households restrict their spending to avoid income deficits? This cannot be answered without knowing on what items they are spending their money. Figure 11 shows the expenditure patterns of households age 65 or older with or without income deficits. Median expenditures in the following categories, as well as median total expenditures, are shown for both types of households:  Home-related expenses include mortgage; property taxes; homeowner’s or renter’s insurance; rent; utilities; home repairs; home furnishings; housecleaning supplies; housekeeping and laundry services; gardening and yard supplies; and gardening and yard services.  Food expenses include food and drink, including alcoholic beverages that are bought in grocery and other stores. Dining out is not included here.  Health expenses include out-of-pocket (uninsured) health insurance costs, including Medicare supplemental insurance; out-of-pocket costs on prescription and nonprescription drugs; out-of-pocket costs of hospital care, doctor services, lab tests, eye, dental, and nursing home care; and out-of-pocket costs for medical supplies.  Transportation expenses include car payments (principal and interest), vehicle insurance, vehicle maintenance, and gas.  Clothing expenses include clothing and apparel (including jewelry) as well as personal-care products and services.  Entertainment expenses include trips and vacations; tickets to movies, sporting, or performing-arts events; hobbies and leisure equipment (photography, reading, camping, etc.); dining out in restaurants, cafes, and diners; and take-out food.  Other expenses include contributions to religious, educational, charitable, or political organizations, and cash and gifts to family and friends outside the household (including alimony and child-support payments).

ebri.org Issue Brief • February 2013 • No. 383

12

Figure 7 Percentage of Households With Household Incomes Less Than Household Expenditure, by Age Groups I-III Age Group I (65–74) 53.0% 50.4 43.3 38.1 37.2

2001 2003 2005 2007 2009

Age Group II (75–84) 57.5% 60.5 52.4 43.5 43.9

Age Group III (85+) 64.6% 53.8 52.3 53.0 46.3

All 65+ 55.5% 53.8 46.8 41.5 40.5

Source: Employee Benefit Research Institute estimates from Health and Retirement Study (HRS).

Figure 8 Percentage of Age 65+ Households With Different Degrees of Income Deficit*

2001 2003 2005 2007 2009

Category I (No Deficit) 44.6% 46.4 53.3 58.7 60.0

Category II Category III (Deficit (Deficit =/>10%