Does the U.S. Have Higher Elderly Poverty Than Mexico?

Some would have you think so.

In a recent article, Teresa Ghilarducci, a prominent economist who focuses on retirement issues, highlights what she sees as the problem of elderly poverty in the United States:

“Already, America’s elderly suffer a far higher rate of poverty – defined as half the median income or lower – than their peers in other high-income countries. As I explained in recent testimony before the US Senate Committee on Health, Education, Labor, and Pensions, the old-age poverty rate in the US is 23%, compared to about 15% in the United Kingdom, 12% in Canada, 4.4% in France, and just 3.1% in the Netherlands.”

All of that sounds compelling. Clearly, something has gone very wrong with retirement in the U.S.

Until you learn that, according to the Organization for Economic Cooperation and Development (OECD), which is the source of Teresa’s figures, Mexico has a lower elderly poverty rate than the United States. Clearly, there’s more going on than either the OECD or Prof. Ghilarducci is telling you.

The OECD definition of poverty is doing two things. First, it’s comparing the incomes of retirees to the overall income level in a country, which is dominated by the more numerous working-age households. A country where seniors are rich relative to younger residents will have a lower elderly poverty rate, while one with relatively poor seniors will have a higher poverty rate.

And second, the OECD definition of poverty focuses on the distribution, not the level, of incomes in a country. If a country has a wider distribution of incomes, with more very rich and more very poor residents, then a larger percentage of those incomes will tend to be below half the median.

And there are significant problems with both aspects of the OECD’s poverty figures.

First, the OECD’s source data for its poverty figures – the U.S. Census Bureau’s Current Population Survey (CPS) – has been shown to dramatically underestimate the incomes of seniors. The Census Bureau itself calculates that in 2018, the CPS understated the true income of the median American age 65+ by 27%. The CPS does not, however, significantly understate the incomes of working-age Americans. Since the OECD’s measure of elderly poverty compares seniors’ incomes to those of the overall population, this data bias makes U.S. seniors appear “poorer” than they actually are. This is indisputable and skews the OECD’s elderly poverty figures upward.

What is the problem with the OECD’s data? It is that the CPS is particularly poor at capturing benefits paid from retirement accounts such as IRAs and 401(k)s. The reason is that the Census Bureau defines “income” as regular payments, such as from Social Security or a pension. Irregular payments, such as as-needed withdrawals from a retirement account, aren’t “income.” A 2017 Census Bureau study showed that the CPS captured only half of all the income seniors received from private retirement plans, including traditional pensions, which means the CPS likely catches well less than half of retirement account benefits.

If that sounds like a problem to you, it is. In 2018, the CPS claimed that the median senior had an income equal to 69% of the overall median income. But when the Census Bureau corrected for the CPS’s errors using administrative data from the IRS and other sources, the median 65+ income rose to 83% of the population-wide median. If those data were incorporated into the OECD’s calculations, elderly poverty in the U.S. would fall.

Second, the OECD’s poverty measure differs fundamentally from how poverty is measured in the United States. In the U.S., the federal poverty threshold is a fixed dollar amount. For a single senior in 2023, the poverty threshold was $14,614. If your income is below that amount you’re considered to be in poverty; if it’s above that level, you’re not. Maybe that dollar figure should be higher or lower, but the point is that it’s a fixed dollar figure. Other people’s incomes don’t affect whether you are considered to be poverty.

The OECD sets a poverty threshold equal to half the median income in a country. If your income is below half the median, then you’re considered to be in poverty. If this sounds like a measure of inequality rather than poverty, it is. And inequality, whether it’s good or bad, is an entirely different issue than whether a retirement system is working.

Most economists and financial planners believe that retirees wish to have enough income to maintain their pre-retirement standard of living. Now, imagine a retirement system that accomplished this perfectly: every retiree at every age has precisely what they need to maintain the standard of living they enjoyed prior to retirement. It’s perfect!

But what would the elderly poverty rate, as defined by the OECD, be with this perfect retirement system? Would it be zero?

No, not at all. Because what a retiree needs to maintain their pre-retirement standard of living depends on what their pre-retirement income was. If pre-retirement incomes are distributed unequally, as they are in the U.S., then you’re going to have an unequal distribution of retirement incomes as well, even if the retirement system is working perfectly.

In other words, the OECD’s measure of old-age poverty isn’t about the level of U.S. seniors’ incomes but about the distribution of pre-retirement incomes. Maybe that’s a problem, but it’s not a problem caused by the retirement system nor is it really one the retirement system can fix. And, of course, even the OECD’s distributional figures aren’t right because they don’t even have accurate data on U.S. seniors’ actual incomes.

All of this combines to give the nonsensical result that Mexico has a lower elderly poverty rate than the United States. Sure, Mexico’s seniors might have higher incomes relative to working-age Mexicans than U.S. seniors to do working-age Americans; or perhaps Mexico has lower income inequality than the U.S. But in 2018, according to the OECD, the disposable income of a U.S. at the 10^th percentile of the income distribution was $12,431 per year. In Mexico, it was $1,668. That’s a big difference.

And what’s actually happened to elderly poverty in the United States? The answer is harder to find than you would think, because the flawed dataset I discussed above – the Census Bureau’s Current Population Survey – also is the dataset used to measure poverty. However, the Census Bureau is aware of the CPS’s problems and is working around them. According to U.S. Census Bureau research, 9.7 percent of seniors in 1990 had incomes below the federal poverty threshold. By 2018, the Census Bureau found, the elderly poverty rate had fallen to just 6.4 percent – a one-third drop in the risk of old-age poverty in less than three decades. The Social Security Administration projects that seniors’ poverty rates will continue to fall in coming decades.

That doesn’t mean the U.S. retirement system shouldn’t do more to prevent poverty in old age. We’re a rich country and providing a true guarantee against poverty isn’t expensive, if we’re willing to tell middle and high-income Americans that they’re going to have to save more for retirement on their own. I argued for such a plan in greater detail in the 2023 AEI book American Renewal.

The point here, as always, is that so much of what you read about the U.S. retirement system is either incorrect or irrelevant. Readers need to be wary of it before they panic.