Are They Better Off?

By Mary Williams Walsh.

The following essay was written by Mary Williams Walsh, managing editor of News Items, who has covered the pension story for numerous publications, most recently at The New York Times.

The most important news you didn’t hear about last week may have been a report by the Census Bureau, unenticingly titled, “National Experimental WellBeing Statistics.”  

On its face, the report described new ways to make the U.S. income and poverty statistics more accurate. Between the lines, it confirmed what some economists have been arguing for decades: that key Census data undercount the retirement income of older Americans, by hundreds of billions of dollars a year.

American seniors, as a group, turn out to be a lot better off than we thought.

That doesn’t mean there are no elderly living in poverty, of course; millions do. What it does mean is that millions of other seniors have significant resources that didn’t show up in the Current Population Survey. That’s the go-to data set for policy makers seeking to understand income distribution.

It’s a critical consideration now, as America casts around for ways to shore up its beloved but unsustainable Social Security program. Social Security’s long-term shortfall now measures $20 trillion, and if nothing is done by 2033 or so, the program is forecast to hit a wall, and everybody’s benefits will have to be permanently cut by 20 percent.

That disaster can be prevented, but it will mean either reducing retirees’ benefits, raising taxes, or some combination of the two. Getting it right will take intellectual honesty and a dedication to the public good seldom on display in Washington these days. A defensible package should be fair to all parties—rich and poor, young and old, working and retired—but that’s unlikely if you start from the false premise that retirees are all living on cat food.

Remember what happened when the Social Security 2100 Act was introduced in 2019? It didn’t just propose to shield the elderly from any benefit reductions. It would have increased all Social Security benefits, regardless of who was needy and who was set for life. It also offered all recipients better cost-of-living increases and bigger tax breaks.

The cost would have been borne entirely by working people. The bill called for phasing in a 14.8 percent payroll tax, up from the current 12.4 percent, and for phasing out the cap on how much earned income is taxed. Eventually all of it would be.

The bill did not exclude low earners from the tax increase, even though they have been falling behind the rest of the workforce for years and can’t typically save for their own retirement. Their taxes were to rise, implausibly, to support retirees whose incomes have been rising faster than those of the overall U.S. population.

And while removing the payroll tax cap put an appealing “tax the rich” face on the bill, even America’s highest earners haven’t kept pace with the elderly, according to Syl Schieber, an economist who has been questioning the Census numbers for more than 25 years. 

By his calculation, the income of the “5 percent” grew at 3.37 percent per year from 2012 to 2018, but in the median elderly household it grew faster, 3.83 percent. 

Schieber began his career at the Social Security Administration and later spent years consulting with companies on their retirement plans. His work put him into a long-running debate with other analysts, who argued that company retirement plans didn’t deliver much value to workers and weren’t worth the favorable tax treatment they got.

He disagreed, and in 1995 he set out to prove his point by comparing the retirement-income numbers in the Current Population Survey with those available from income-tax returns.

The I.R.S. numbers showed that retirees up and down the economic ladder were getting significant income from their retirement plans; they reported it to the I.R.S. and paid taxes on it. But the income didn’t show up in the Census statistics.

One reason: The Census back then counted as “income” only the money that people received at regular intervals, like weekly paychecks or monthly Social Security payments. If you dipped into your 401(k) account once in a while to pay your property taxes, that didn’t count as income. There were other survey problems, too.

Schieber reported his findings in the journal Benefits Quarterly. The response?

“The longest article ever written to date in the ‘Social Security Bulletin,’ lambasting me and my analysis,” he recalled in an email last week.   

Three years later he was appointed to the Social Security Advisory Board; he became its chairman in 2006. He continued questioning the retiree income numbers, reporting, for instance, that in 2000, retirees told the I.R.S. they withdrew $59 billion from their retirement accounts, when only $2 billion turned up in the Census numbers.

Defined-contribution retirement plans were catching on and growing, and that made the discrepancies grow, too. By 2007, retirees told the I.R.S. they had withdrawn $102 billion from their 401(k) plans, but only $6 billion of it showed up in the Census numbers.

Schieber said people still urged him to stop raising the issue, so as not to undermine the credibility of the Current Population Survey, or the Social Security Administration’s “Income of the Aged Population” charts that are based on it.

But others started questioning the numbers too. Two Census analysts, Adam Bee and Joshua Mitchell, confirmed that underreporting of retiree income was a growing problem. The Organization for Economic Cooperation and Development reported that the average incomes of America’s elderly were growing faster than those of the U.S. population at large. 

Andrew Biggs, a former principal deputy commissioner for Social Security now at the American Enterprise Institute, found a complete reversal of seniors’ fortunes over four decades. In 1979, retiree households were more than twice as likely to be in the poorest fifth of the U.S. population as in the richest fifth, he wrote last month in National Review. But in 2019 they were “nearly twice as likely to live in the richest income quintile as to live in the poorest quintile.”

In 2018, the Social Security Administration acknowledged the statistical problem and suspended publication of the “Income of the Aged Population” charts “as we evaluate the adequacy of the charts’ data source.”

Schieber called the charts “the bible on retiree income,” frequently consulted by policy staff on the House Ways and Means and Senate Finance Committees. They remain unavailable.

For now, it’s quiet in Washington. President Biden has said Social Security is off the table while the debt ceiling looms. Senator Rick Scott, the Florida Republican, last week removed Social Security and Medicare from his proposal to put all spending programs through a five-year review.

But Congress is going to have to take up Social Security at some point. Now, at long last, lawmakers may have the statistical tools to avoid “reforming” the program by showering dollars onto well-off retirees, while foisting the cost onto the young. If only they use them.