New questions to measure income in the Census Bureau’s Current Population Survey (CPS) revealed a significant amount of income from individual retirement accounts (IRAs) and 401(k)-type plans was being missed across all levels of income for Americans age 65 or older, according to EBRI.

However, even using new questions, Social Security remains by far the predominant source of income for Americans of retirement age (65 or older), especially for those with incomes in the bottom half. More than 60% of individuals in the lowest two income quartiles receive more than 90% of their total income from Social Security, even when accounting for the additional IRA and 401(k)-type plan income.

“The new measure of income in the CPS finds significantly more income that is from IRAs and 401(k)-type retirement plans than what has been reported under the old measure of income in CPS,” says Craig Copeland, senior research associate at EBRI. “But for those in the lower-income brackets—and even for those with higher incomes—Social Security income remains extremely important for older Americans.”

Specifically, the new EBRI analysis finds:

• The new measure of income in the CPS identifies significantly more income (and a much larger percentage of income coming from IRAs and 401(k)-type plans). Compared with the estimated amount under the traditional income questions for 2013, the redesigned questions resulted in an estimated total income 9.1% larger for those ages 65 or older, an aggregate amount of almost $133 billion more. Retirement income was 27.9% larger, or almost $71 billion in aggregate.
• Income from IRAs and 401(k)-type plans was an important component of this higher amount of retirement income found in the new questions, which overall was more than 250% higher than that found by the traditional measure. Phrased another way, the traditional CPS measure under-reported IRA and 401(k)-type income by more than 250%.
• Those with IRAs, 401(k)-type plan and defined benefit pension income from private- and public-sector employers are more likely to be in the upper-income quartiles because they are far more likely to have these sources. Consequently, those who both had access and took advantage of these plans are the ones with higher amounts of this income in retirement—and not necessarily those who had the highest overall incomes prior to retirement.

The Annual Social and Economic Supplement to the CPS (fielded in March of each year)—which EBRI says is a widely cited source of data on income for those old enough to be considered retired (age 65 or older)—has not kept up with the change in the private-sector retirement landscape from predominantly defined benefit (DB) plans to predominantly defined contribution (DC) plans. This survey has misclassified certain types of income and generally under-reported income— in particular, retirement income—because the questions did not properly ask about income from sources that did not provide an annuity payment like that has been traditionally paid from DB plans.

Last year the Census Bureau revised the income questions to in the hope of reducing the discrepancies from the prior design. The 2014 CPS conducted a test of the new set of income questions by doing a spilt-panel design, where 3/8 of the sample received the redesigned questions while the remaining 5/8 received the traditional questions. The new EBRI research compares results from the old and new methods.

The full report, “Individual Retirement Account Balances, Contributions, and Rollovers, 2013; With Longitudinal Results 2010–2013: The EBRI IRA Database,” is published in the May 2015 EBRI Issue Brief, online at www.ebri.org.