In 2005, President George W. Bush attempted to partially privatize Social Security. He centered his argument for this change on the claim that people would fare better investing in asset markets than contributing to Social Security. The privatization push proved highly unpopular, as research from EPI and others highlighted the high transition costs and investment risks.

Nevertheless, the belief that Social Security amounts to a low-risk but low-return investment persists, hampering proposals to expand the popular program. This is unfortunate, as Social Security looks better than ever in comparison to low-performing 401(k)s and IRAs.

As shown below, a young worker today with average career earnings will receive Social Security retirement benefits equivalent to total employer and employee retirement contributions plus a 5.7 percent annual rate of return. This “internal rate of return” is not much lower than the 7.0 percent net return for 401(k)-style defined contribution plans between 1990 and 2012, and it’s higher than more recent returns for these plans and IRAs (3.1 percent and 2.2 percent, respectively, over the 2000–2012 period).

This calculation doesn’t take Social Security’s projected long-term shortfall into account. But even if we closed the shortfall by raising the contribution rate from 10.0 percent to 12.6 percent (excluding contributions going toward disability benefits), the internal rate of return for a medium earner would be 5.0 percent.

Though Social Security is primarily funded through worker contributions, a small share of the cost is paid for by taxes on the benefits of better-off retirees that revert to the program. If these taxes on high earners were eliminated so that the entire cost of retirement benefits were funded by worker contributions, the internal rate of return for a medium earner would be a healthy 4.5 percent, still an excellent return for such a low-risk investment.

Rates of return on 401(k)-style plans vary widely and are subject to market downturns. To reduce the risk of worse outcomes, most investors, especially retirement savers, would choose a secure 5 percent return over a volatile return averaging 7 percent, since, contrary to popular belief, investment risk doesn’t disappear over long time horizons.

The returns discussed above are the internal rates of return for a prototypical medium earner who lives long enough to receive individual retirement benefits. Actual returns for any given individual will vary depending on several factors, including longevity, lifetime earnings, spousal earnings, and family size and structure. For example, a one-earner married couple will receive higher returns due to spousal benefits. Conversely, an unmarried worker without dependents who dies before retiring will receive no cash benefits, though there’s value in being insured even if you never collect.

An interesting question is how internal rates of return vary by income. If everyone lived to the same age, internal rates of return would be higher for low earners and lower for high earners due to Social Security’s progressive benefit structure that provides higher income replacement rates for low earners than high earners. However, much of the progressivity in lifetime benefits has been eroded by growing disparities in life expectancy, so that internal rates of return are converging across income groups. Because women earn less but live longer than men, their internal rate of return is higher, though they’re still at much higher risk of experiencing hardship in old age.

Differences by sex, family type, and birth cohort are discussed in an actuarial note published by the Social Security Administration, though it’s necessary to add 2.6 percent expected inflation to compare their inflation-adjusted estimates with the nominal rates of return discussed here. The Social Security actuaries also include disability costs and benefits in their calculations, whereas the focus of this blog post is on comparing different ways to pay for retirement.

Some readers may be surprised that Social Security’s internal rate of return is on the order of 5 percent, when Treasury bonds in the Social Security trust fund have been earning only half that over the past decade. One reason is that Social Security actuaries expect interest rates to rise now that the economy has recovered from the Great Recession. But the much more important reason is that Social Security’s ability to pay for benefits has almost nothing to do with the interest earned on Treasury bonds. Social Security is mostly a pay-as-you-go system, with each generation of workers supporting previous generations in retirement and the trust fund mostly serving to smooth baby booms and busts.

Social Security’s pay-as-you-go structure allowed the program to pay for benefits for the first generation of beneficiaries who began reaching retirement age toward the end of the Great Depression and had not had a chance to contribute to the program, which was signed into law in 1935. It also has the underappreciated benefit of providing participants with an implicit return on contributions that is based on economic growth rather than the vagaries of financial markets. Thus, retired beneficiaries are supported by workers whose careers occur roughly one to 65 years after their own, and whose earnings and contributions reflect the population and productivity growth that occurred in the interim.

Over the long run, investment returns and beneficiaries’ internal rate of return are both constrained by economic growth and should converge. Over shorter periods, investment returns can exceed economic growth rates if profits grow at the expense of wages, as has occurred to some extent in recent decades. Social Security’s internal rate of return can also lag behind investment returns if wage growth is concentrated at the top of the earnings distribution above the taxable earnings cap. Investments in other countries can also deliver returns that exceed domestic growth rates.

But there’s no reason to expect investment returns to exceed Social Security’s internal rate of return over the long run. And investment returns are more volatile than economic growth rates because asset markets are speculative and forward looking—you may have the misfortune of retiring in a long bear market. When stock market returns do outpace economic growth, price-earnings ratios increase and future returns are likely to lag.

Last but not least, Social Security benefits go to targeted beneficiaries—retired workers, people with disabilities, and their spouses and dependents. These benefits continue until older beneficiaries die, disabled worker beneficiaries are able to return to work (or not), and younger dependents reach working age.

In contrast, despite weak rules intended to discourage their use as tax shelters, wealthy retirees tend to hoard savings in 401(k)-style plans to take advantage of tax subsidies. Retirees are also loath to draw down savings because they don’t know how long they will live, and annuities are complicated and expensive to buy on the individual market. Thus, tax-favored retirement savings are often bequeathed to heirs and contribute to growing wealth inequality rather than economic security.

In short, for the vast majority of Americans, Social Security is a great investment and an insurance plan in the bargain. Multimillionaires who might think they don’t need it can easily afford to participate, and many—like investors who lost their life savings with Bernard Madoff—will be glad they did.