It’s a fact: low-wage workers don’t save much for retirement. States are aiming to fix that. But here’s a question: is it really a problem that needs to be fixed? How hard should we push the poor to save for retirement?
Bureau of Labor Statistics data show that 75% of workers in the upper half of the salary distribution — those who earn at least $36,000 on a full-time basis — participate in a retirement plan. Among workers in the bottom quarter of the wage distribution, with average salaries of about $24,000 per year, only 25% participate. That’s a big difference.
States like California, Oregon, Illinois and Connecticut are rolling out so-called “auto-IRA” programs, which would automatically enroll employees whose jobs don’t offer a retirement plan into a Roth Individual Retirement Account. Employees could choose not to participate, but otherwise they’ll have 3% to 5% of each paycheck automatically deposited in their IRA. Nobel-prize winning behavioral economist Richard Thaler, whose research helped inspire auto-IRAs, writes, “There is no coherent argument against these state plans.”
I agree that everyone should have the opportunity to save for retirement. But it is far from a no-brainer that lower-income Americans need to save more. In fact, there are a number of good arguments against it. And if we push low earners unnecessarily to boost their retirement savings, that leaves less money for everything else.
But first we need to get something straight: retirement saving isn’t about making yourself rich in retirement. In reality, retirement saving is about being able to maintain your pre-retirement standard of living once you stop working. And most Americans, including low earning workers, seem able to do that.
In fact, financial security increases in retirement. In the Federal Reserve’s Survey of Household Economics and Decisionmaking, at ages 57 to 61, 12.5% of individuals with a high school education or less and reported they were “finding it hard to get by,” versus only 7.25% of individuals aged 62 to 67. Likewise, Census Bureau research shows that the poverty rate also drops as Americans shift from work into retirement.
The reason is that nearly every U.S. worker actually does have a retirement plan: it’s called Social Security. And while everyone pays the same payroll tax rate, Social Security’s progressive benefits mean that low earners get a more generous payback than high earners. Congressional Budget Office figures show that a low-earning worker retiring today receives a Social Security benefit equal to about 84% of his career-average earnings, adjusted for inflation. Workers in the top half of the earnings distribution receive “replacement rates” of only 43%.
Differences in Social Security replacement rates explain why higher-income Americans save so much and lower-income Americans save so little. Recent research by economists at the Internal Revenue Serve and the Investment Company Institute found that the median low-income retiree has a retirement income equal to 103% of earnings just prior to retirement. Other research from Census Bureau economists found similar results: retirees at the 25th percentile of the income distribution had incomes equal to 93% of their average earnings in the 15 years prior to retirement. Given that living costs decline in retirement, low earning workers seem well prepared to maintain their pre-retirement standard of living.
That doesn’t mean there aren’t poor retirees. There are. But the poverty rate is lower among retirees than it is among working-age adults. Should a low-income worker reduce his standard of living today in order to raise it even further once he retires? It’s not at all clear that he should. Moreover, savings for retirement can displace savings to help finance education, a home down-payment or a business start, or simply build a rainy-day fund to protect against the stress of unexpected expenses. The thing about being low income is that you don’t have much income to go around and you have to choose. It’s not clear that low-income workers are choosing incorrectly.
Moreover, most low-income workers receive benefits from one or more government transfer programs. Because many government programs have means tests, low earners who build even modest savings could disqualify themselves for thousands of dollars in government benefits.
An Urban Institute analysis found that for a married couple with two children, increasing the household’s liquid assets from below $1,000 to between $1,000 and $2,000 would reduce annual benefits from means-tested transfer programs by almost $3,000. IRAs generally count as liquid assets because they can be readily converted into cash (with a penalty), though states have some leeway on how to apply means testing.
Once a worker with an auto-IRA account retired, the income pulled from the account could cause him or her to be ineligible for Medicaid, for which a retiree must generally have an income below around $1,500 per month. A 2017 AARP-funded analysis found that about one-fifth of the contributions that low earners could make to auto-IRA accounts would be offset by reduced means-tested benefits. That’s like having a 20% tax on low-income households’ retirement savings.
There’s another risk: once low earners see their take-home pay reduced via automatic enrollment, they may borrow to maintain their standard of living. Recent research found that when federal employees automatically enrolled in their 401(k)-type retirement plan, employee contributions indeed rose. But four years following auto-enrollment, there wasn’t any statistically significant increase in those employees’ net worth. While the researchers’ data were incomplete, the results hinted that less-educated workers may have borrowed more via higher credit cards, auto or mortgage loans. State auto-IRA plans will enroll many more low-wage workers than the federal government plans, so states should investigate how these employees’ total household finances react, not simply laud increases in retirement plan balances.
Does this mean states should abandon their auto-IRA plans? No. But they need to rethink the idea that everyone should be saving for retirement at all times of their lives, regardless of income. It just doesn’t work that way.
States could instead limit automatic enrollment to employees with earnings above some threshold. The United Kingdom’s nationwide auto-enrollment retirement plan automatically enrolls only workers with annual earnings over £10,000. This is about $12,325, very close to the federal poverty threshold which is often used in setting eligibility for government welfare programs. Employees with lower earnings may participate, but must affirmatively petition to join. This could more effectively target employees who truly need to save and who are less likely to rely on means-tested government benefits.
Limiting automatic enrollment won’t produce the flashy outcomes auto-IRA proponent say they want. It won’t close the gap in retirement plan participation between rich and poor, nor will it lead to low-income workers having substantial retirement savings. But it will realistically recognize the role that progressive Social Security benefits play in low-earners’ retirement planning, and it will prevent punitive cuts in means-tested benefits for low-income workers who may not need to be saving for retirement in the first place.
I’d call that a success.