What Is Going on with the Income of the Elderly?: Friday Focus (January 24, 2014)

Dean Baker reads Sylvester Schieber and Andrew Biggs:

Dean Baker: The WSJ Says the Elderly Are Rich But We Didn’t Know It: “Sylvester Scheiber and Andrew Biggs have good news for us in a Wall Street Journal column….

Scheiber… and Biggs… tell readers that the standard numbers on income for the elderly are way off…. The Census Bureau’s Current Population Survey CPS)… misses a large portion of the income of retirees….

For 2008, the CPS reported $5.6 billion in individual IRA income. Retirees themselves reported $111 billion in IRA income to the Internal Revenue Service. The CPS… 2008 households receiving Social Security benefits collected $222 billion in pensions or annuit[ies]…. Federal tax filings… show… $457 billion…. The Current Population Survey… ignores at least 60% of the income being delivered to retirees….

This is an interesting argument. It is certainly newsworthy…. There is one major problem… the CPS is not the only data set… the Survey on Income and Program Participation (SIPP) that yields largely similar numbers to the CPS…. The Federal Reserve Board’s Survey of Consumer Finance (SCF) also shows numbers consistent with the CPS… for people between the ages of 65-74 of $42,700 and $29,100 for people 75 and over… the CPS showed median incomes of $39,800 for households between the ages of 65-74 and  $25,400 for households over the age of 75….

The SCF also has useful data on assets…. The 2010 SCF showed that 49.0 percent of families between the ages of 65-74 hold retirement accounts and 32.8 percent of families over age 75 have retirement accounts… [with] median holdings for those who have retirement accounts [of] $100,000 for those between the ages of 65-74 and $54,000 for those over the age of 75…

When I take Scheiber and Biggs $335 billion of unrecorded CPS income–or, rather, of inflows to elderly households’ bank accounts that the CPS thinks are not properly income but rather asset drawdowns–and divide it by the 43 million Americans over 65, I get that the CPS income concept understates actual elderly life-cycle resources by $7,000 per elderly–which means, given the skew in the distribution and the number of single individuals in the elderly population, an underestimate of the median of $3,500 or so.

That seems to closely fit the difference we see between the CPS on the one hand and the SCF (and SIPP) on the other. So one way to read Scheiber and Biggs is as saying: “CPS income is not a good measure of the economic resources available to the elderly–they have assets, and can and should be spending their assets down.”

And that is a very true and important point.

For the top quarter tile of the elderly, at least.

My go-to source on this set of issues has been Bosworth and Burke (2012):

Barry P. Bosworth and Kathleen Burke: Changing Sources of Income Among the Aged Population: “We begin by comparing the estimates of the income of aged households in three surveys: the Annual Social and Economic Supplement (ASEC) to the Current Population Survey, the Health and Retirement Study (HRS); and the Survey of Consumer Finances (SCF)….

Prior work (Czajka and Denmead, 2008) suggests that the ASEC fails to capture large portions of the withdrawal of funds from defined-contribution pension plans by limiting the income measure to regular payments. The HRS also inquires in more detail about pension income. The most puzzling discrepancy is for self-employment income. The differences are large and they are concentrated at the top of the income distribution…. When incomes in the two surveys are arrayed using the quintile breaks of the ASEC, the dollar magnitudes of the discrepancies are highly concentrated in the top fifth of the income distribution….

In the lower half of the income distribution, the aged are almost fully dependent upon Social Security because most have little or no income from earnings, private pensions or own saving. The diversity of income sources is important only for those in the top portions of the income distribution….

Approximately two-thirds of the capital income recorded as household income in the national accounts is excluded from the household surveys because households have no knowledge of the income earned in fiduciary accounts, such as pension funds and life insurance…. Furthermore, many individuals… make phased withdrawals from their savings…. Large portions of these withdrawals appear to not be reported as income…. Finally, one source of capital income missed in most household surveys is the flow of services from owner-occupied homes. Homeownership is higher among the aged than among non-aged adults, so the omission of this income flow is likely to cause a significant understatement….

We have experimented with replacing the standard measure of asset income from dividends and interest with the predicted annuity value of the family’s net worth…. The quality of the wealth data can be evaluated by equating it with estimates of wealth from the SCF…. When the comparison is restricted to the households below the 95th percentile of the SCF, the two estimates are virtually identical. It is somewhat surprising that the HRS performs so well with such a small number of questions about wealth holdings…

The takeaway? The average elderly household born in the 1930s not in the top 5% has $250,000 in wealth to spend on retirement in addition to their Social Security, defined-benefit pension, and other income. That means that the median elderly household born in the 1930s has about $150,000.