There’s No Evidence of a Retirement Crisis

The Claim

Last month, The New York Times printed an article about BlackRock BLK CEO Larry Fink, bearing the headline “He Wants to Address the Retirement Crisis Looming in the U.S.” (The online version carried a different title.)

The argument that the United States faces an impending retirement crisis has been struck many times, over many years. “Unless we as a society do a better job of investing,” said Eli Broad of SunAmerica in a 1994 Associated Press article, “future retirees face a drastically reduced standard of living.” Later in that article, Merrill Lynch’s Daniel Tully fretted that “members of the baby boom generation” could endure lower “living standards” than the retirees of the 1990s.

A decade later, The Coming Retirement Crisis, published in 2004 by Forbes, anticipated a “slow-motion crisis as baby boomers head into retirement.” Ten years after that, in 2014, came the video Broken Eggs: The Looming Retirement Crisis in America. There is no shortage of such examples. Until now, though, I had not seen a major newspaper cross that line.

The Research

In 1994, the nation’s retirement structure was changing rapidly. Over the previous eight years, the number of defined-benefit plans had fallen by more than 50%, from 172,642 to 74,422. Meanwhile, the number of 401(k) accounts was soaring. Baby boomers were the first defined-contribution generation. Unlike their predecessors, they would largely be planning for retirement on their own.

As baby boomers are currently between the ages of 60 and 78, enough time has passed to reveal the early results of this experiment. Fortunately for our purpose, the US Census Bureau publishes a history of median national incomes, sorted by sex and age. (The data used in this column can be found in Table P-8 within this link.)

Unfortunately for our purposes, although the Census Bureau’s research is regarded as the gold standard for US income data, it is nevertheless suspect. For example, the Bureau reports that in 2019, the median real income for female workers from ages 25 through 34 ballooned by 9%. Meanwhile, income for male workers of that age was slightly lower. The next year, it then states, the female workers’ income was flat, while the group’s males increased by 3%.

I’m not buying. Young American working women did not suddenly receive a decade’s worth of real income growth in a single year, even as their male counterparts went nowhere. The Census Bureau’s measurements contain the amount of fluctuation associated with small sample sizes, but for cohorts of 20 million.

The Age Groups’ Results

Thus, I would strongly caution against attempting to decipher the wiggles of the following illustration, which shows the results since 1994 for each of the Census Bureau’s seven age groups. (With these calculations, I combined the female and male results within each age group, to simplify the presentation. Each cohort is indexed, with an initial value of 100 assigned to its 1994 total. Thus, the chart shows changes in relative income, rather than the levels of absolute income—which, for reference, are highest for the 45 to 54 group and by far the lowest for the 15 to 24 group.)

That caveat aside, the overall trend is instructive.

Over the past 30 years, the median real income for every cohort has improved—for retirees as well as workers. Indeed, retirees have fared somewhat better than the norm. The 65 to 74 group, charted in purple, has enjoyed the second-highest overall gain among the seven cohorts, while the 75-plus crowd places fourth.

One potential concern is the purple line’s recent downturn. Does this indicate that, at long last, retirees’ problems have surfaced? To answer my rhetorical question, I think not. For one thing, as previously written, the data are lumpy, which makes interpretation hazardous. For another, the age group suffered two unusual shocks, just as the research period ended. First, covid-19′s arrival sent many older workers into early retirement and/or cost them their part-time jobs. Second, 2022′s high inflation eroded their purchasing power that year.

However, in 2023—which does not appear in the illustration, as those figures have not yet been released—the Social Security Administration gave retirees an 8.7% cost-of-living raise, while the Consumer Price Index rose by only 3.4%. That makes for a one-year after-inflation raise of 5% for Social Security benefits, which will almost have certainly reversed that near-term trend.

Where’s the Beef?

But I digress. The point isn’t that the evidence has been tried and found lacking. It is instead that it really hasn’t been offered at all. For several decades, people worried about future retirees have argued their case based on broad trends, such as the waning of defined-benefit plans, surveys suggesting that workers don’t own enough investment assets, and the ever-increasing length of retirement as Americans live longer. Those are all worthwhile topics, but they represent questions to be addressed, not answers to be delivered.

To put the matter another way, unless they are accompanied by specific supporting evidence, demographic arguments are insufficient. I should know. I spent much of my investment youth hearing that stock market returns would be strong until about 2010 but would then collapse as baby boomers became more conservative and sold their equity shares. Of course, the opposite occurred. When they were supposed to be declining, stocks instead soared.

Conclusion

None of this should be mistaken for complacency. My feelings about the transition from defined-benefit plans to defined-contribution plans are decidedly mixed. The current system possesses weaknesses as well as strengths. Nor do I discount the demographic concerns. The world is becoming older rather than younger. That shift will have profound implications, including with retirement funding.

But there is a difference between speculation and certainty, particularly when the speculation has existed for several decades. With its headline, The New York Times presented an assertion as a fact. It should not have done so.