Sylvester Scheiber and Andrew Biggs have good news for us in a Wall Street Journal column, apparently the elderly are much better off than we realized.  Scheiber, a pension consultant and former chairman of the Social Security Advisory Board, and Biggs, an economist at the American Enterprise Institute and former Deputy Commissioner of the Social Security Administration, tell readers that the standard numbers on income for the elderly are way off.

The most commonly used measures of income are from the Census Bureau's Current Population Survey CPS). Scheiber and Biggs say that this survey misses a large portion of the income of retirees. For example, they tell readers:

"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 suggests that in 2008 households receiving Social Security benefits collected $222 billion in pensions or annuity income. But federal tax filings for 2008 show that these same households received $457 billion of pension or annuity income.

"In combined terms, the Current Population Survey that ostensibly documents how poorly pensions and individual retirement plans provide retirement income ignores at least 60% of the income being delivered to retirees. Even that is not the whole story—because tax filings do not include distributions from Roth plans, since those distributions are not taxable."

Scheiber and Biggs go on to complain about the use of the CPS to assess retiree income and suggest alternative sources which they say would be more accurate.

This is an interesting argument. It is certainly newsworthy when someone finds major flaws in the most widely used survey for measuring income. However before we join Scheiber and Biggs in demanding that the Social Security Administration and other official bodies discard the data from the CPS, we may want to think this one over a bit.

There is one major problem with the Scheiber-Biggs story: the CPS is not the only data set that gives us these sorts of numbers about the income of the elderly. The Census Bureau has a totally separate survey, the Survey on Income and Program Participation (SIPP) that yields largely similar numbers to the CPS.

For example, an analysis of SIPP data by Pew showed a median income for families over age 65 in 2010 of $43,400. While this is considerably higher than the $33,100 median shown in the CPS data, the Pew number is adjusted as though it were the income for a family of three. If we used the CPS figure and assumed that it was for a family of 2, the equivalent Pew figure would be $40,500. If the CPS number were for a single person household, then it would be reported as $57,300 in the Pew analysis. In other words, the Pew calculation based on SIPP data would almost certainly imply a lower income figure for median household income that the CPS.

The Federal Reserve Board's Survey of Consumer Finance (SCF) also shows numbers consistent with the CPS. The 2010 SCF shows a median household income for people between the ages of 65-74 of $42,700 and $29,100 for people 75 and over (Table 1). By comparison, 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. While these numbers are somewhat higher than the data shown in the CPS, it does not qualitatively change our picture of the elderly population. If the median household over the age of 75 has an income of $29,100 no one will think that they are living well. (There are differences between the definition of "family" in the SCF and household in the CPS. I have not looked at their implications for measuring the income of seniors, but it is possible they explain some of the gap between the surveys.)

The SCF also has useful data on assets, exactly what Sheiber and Biggs claim the CPS is missing in the income data. (The SCF is designed first and foremost to measure wealth, so it would be surprising if its numbers on assets were far from the mark.) 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 (Table 6). The median holdings for those who have retirement accounts was $100,000 for those between the ages of 65-74 and $54,000 for those over the age of 75. This means the SCF data imply that 75.5 percent of families between the ages of 65-74 have less than $100,000 in a retirement account. The data also imply that 83.6 percent of the families over age 75 have less than $54,000 in retirement accounts.

In short, the SCF data support the notion that, contrary to Scheiber and Biggs, the vast majority of seniors are not getting substantial income from retirement accounts. We should always be willing to scrutinize the data we use in assessing income and other measures of well-being, but this means bringing to bear all the data sources available. The picture of the elderly given by the CPS, and being used by most people in policy debates, is consistent with the picture given by other commonly used data sets, even if it doesn't fit well with the story told by Scheiber and Biggs.



Andrew Biggs has called my attention to the fact that the income data in the Pew study were in fact taken from the Current Population Survey. While the wealth data come from the SIPP, they did a separate analysis of the CPS for income data.