Social Security is in the news again. Given its “pay-as-you-go” structure and demographic trends, over the years analysts have warned of the impending Social Security crisis. The difference is that now the financial day of reckoning looms quite close.

“Fortunately, there is a ‘win-win’ method that can ease the fiscal crunch while benefiting the government’s finances and program participants alike.”

The system is in the hole by $11 trillion. The government’s pay-as-you-go formula is fundamentally weak, in contrast to a private sector savings-and-investment approach. Fortunately, there is a “win-win” method that can ease the fiscal crunch while benefiting the government’s finances and program participants alike. Specifically, the government could allow Americans to opt out of Social Security, so long as they make an appropriate payment (if necessary) to prevent their opt out from worsening the government’s fiscal position. The meager rate of return that people earn on Social Security, coupled with the Treasury’s ability to borrow at very low rates, means that many people will benefit by exercising this option. By design, everyone who voluntarily leaves will either reduce or have no effect on Social Security’s funding gap. Finally, the possible increase in hours worked due to the elimination of the Social Security payroll tax may cause extra revenue to flow to the government through conventional income taxes.

The Severity of the Social Security Shortfall

According to the 2016 Board of Trustees report, in 2015, the Old-Age Survivors Insurance (OASI) and Federal Disability Insurance (DI) programs were providing some 60 million people with benefits.1 Social Security’s total expenditures were $897 billion, while its “income” was $920 billion. Of that $920 billion, $93 billion was income earned on assets held in Social Security’s so-called trust fund, consisting of special-issue Treasury securities, which, at the end of the 2015, were valued at $2.8 trillion.

Using intermediate assumptions, “the open group unfunded obligation for OASDI over the 75-year period is $11.4 trillion in present value” (Trustees Report, p. 5, italics added).2 This means that the Social Security Administration right now would need an extra $11.4 trillion in assets, rolling over and earning compound interest, so that over the next 75 years, this extra wealth could cover the gap between outgoing benefit payments and incoming payroll taxes.

It is already a fact that incoming Social Security payroll taxes do not cover outgoing (net of tax) benefit payments. This has been true since 2010. In that sense, the crisis is already upon us. If we include the interest income that the Treasury pays to Social Security on the trust fund, then we can postpone the income/outgo crunch until 2020. At that point, in order to meet existing obligations to beneficiaries, the Social Security administrators would need to start selling assets in the trust fund.

Finally, the Trustees estimate that if nothing is done, in 2034, the combined reserves of OASDI will be fully exhausted.

Why is Social Security in a Crunch?

Because Social Security is a pay-as-you-go system, its solvency deteriorated as the population aged and as the system “matured” from its inception in 1935. Consider that in 1960, the ratio of OASDI workers-to-beneficiaries was 5.1 but, by 2010, had fallen to 2.9.3 Since the federal government, in practice, takes the payroll taxes from existing workers and hands them over to existing beneficiaries, these demographic shifts obviously affect the program’s viability.

Related to the overall demographic shift is the increasing lifespan of Americans, which (other things equal) puts more strain on the Social Security system.

Cross-Sectional Pay-Go versus Longitudinal Self-Funding

Some analysts seem to misunderstand the fundamental difference between the pay-as-you-go and self-financing retirement approaches. For example, in a recent article discussing the problems of the 401(k) system, Megan McArdle wrote:

No matter how you manage your retirement system, you are ultimately expecting to depend on the labor of people younger than you. Whether that labor comes to you in the form of a dividend check or a government benefit or a saintly daughter-in-law building you a new annex in the backyard, you are still expecting someone else younger than you to make stuff, then give it to you without expecting more than gratitude in return.4

Although her other comments (not quoted) about worker productivity indicate that McArdle has some grasp of the situation, her statement quoted above is inaccurate. If a worker takes 10.6 percent (the combined employer and employee portions of Old-Age Survivors Insurance) out of his paycheck every month and buys private assets, there is an important sense in which this “provides for” his future retirement more than if the government takes the same amount (from him and his employer) and transfers the money immediately to a current Social Security beneficiary.

Specifically, when workers genuinely save some of their income (meaning it’s not immediately spent by the government), their action frees up real resources that can be channeled into productive investments. When today’s workers use (some of) their labor hours to create hammers, nails, 18 wheelers, and factories to rent and/or sell to young workers in 30 years, they amplify the productivity of the next generation. Armed with this higher stockpile of tools and equipment, the next generation of workers can produce more goods and services, out of which they can afford to pay the older, retired workers (who are now acting as capitalists). Contrary to McArdle’s claim, the whole process is a normal win-win market exchange, which is not driven by mere altruism or expectations of gratitude.

A Procedure for Reform

My suggestion is that the government, before making any major changes to the Social Security formulas, first allow Americans to opt out of the system, thereby avoiding any future payroll taxes but also forfeiting any accrued benefits. However, if the person represents a net asset to the Social Security program from the government’s perspective, then he or she must contribute this amount before being allowed to opt out.5

An opt-out option would improve upon the status quo from the government’s perspective, because Americans can opt out only if they represent a neutral or net liability to Social Security. On the other hand, since it’s voluntary, it would seem that my proposal cannot hurt the Americans who opt out; anyone worried about being hurt by this procedure can choose to remain in Social Security.

There are two important complications to my argument. First, when a given worker opts out of Social Security, his employer would no longer pay the employer portion of the payroll tax. This would eventually lead to a higher wage for the worker. By itself, substituting a higher wage for an employer-provided benefit (namely, taxes paid to a retirement fund) would increase the worker’s taxable income and, thus, the amount he owes the IRS on his personal income tax return. In order to contain the scope of the analysis and to minimize the fiscal impact of my proposal, for the rest of this article I assume that the government makes a formulaic adjustment to the income tax applicable to those workers who opt out of Social Security, so that at least, on average, workers are not paying more income tax simply because of the swap of higher wages for a now-forfeited employer-financed benefit.

The second complication to my claim that an opt out is a win-win is the possibility that even workers who remain in Social Security will see their wages drop slightly if millions of other workers decide to opt out. Specifically, what could happen is that the removal of the “tax wedge” would cause the opting out workers to provide more work effort, reducing the average wage. (An increase in the supply of labor, all else equal, would cause a drop in wages.).

Having acknowledged this possibility, I note that the empirical literature does not show that it would be a significant effect. That is, the empirical literature finds that reductions in payroll taxes do not lead employers to end up paying less per hour of labor.6 In the context of my proposal, this outcome means that workers who remained in Social Security would not see a noticeable drop in their wages.

However, even if it were true that widespread opting out of Social Security lowered the wages of even those workers who remained in the program, I would argue that this is an entirely defensible outcome. Recall that the reason this would occur (if it did occur) is that the lower tax bite leads many employees to work longer hours or to extend their working careers. Why would this be a lamentable outcome? Surely one group of workers is not entitled to higher wages via government rules that discourage other workers from competing as heavily.

Although I do not consider them to be important objections, anyone considering my proposal must consider the complications I’ve described.

The Current System Relies on Forced Loans to the Government

How can allowing an opt out represent a win-win scenario? Why would some Americans—I suspect millions—remove themselves from Social Security if they currently represent a net financial liability to the federal government? The answer depends crucially on the fact that private households have a higher discount rate than the U.S. Treasury.

Currently the U.S. Treasury can borrow money (by issuing Treasury Inflation Protected Securities or TIPS) from bond buyers for very low interest rates, such as 0.9 percent on 30-year loans and 0.4 percent on 10-year loans.7 These are “real” rates of return, meaning that the Treasury adjusts for price inflation when paying back the lenders.

On the other hand, many households currently hold a substantial portion of their wealth in assets that they expect to earn a higher real return than inflation-protected bonds issued by the U.S. Treasury. For example, using standard estimates of the equity risk premium, a recent Wall Street Journal article suggested that U.S. large-cap stocks could be expected to yield a real return of 3.5 percent (before taxes and fees) in the medium term.8 We would expect riskier assets, such as real estate or small-cap stocks, to earn an even higher (average) return.

The large scope for win-win reform occurs because many participants in Social Security are implicitly being promised a real rate of return on their payroll taxes in between these extremes. For example, Leimer (1994) estimated that people born in 1975 will “earn” an average of 1.9 percent from Social Security, while those born in 2000 will earn 1.7 percent.9

In 2015, the government effectively forced workers to lend it some $800 billion in payroll “contributions” in order to help finance the almost $900 billion it spent on Social Security benefit checks. In exchange, the workers were promised future payments from the U.S. federal government, such that many of the workers are effectively earning real returns of only 1 to 2 percent on these mandatory “loans.”

Yet both parties can improve on this status quo. The government can borrow at lower yields by issuing TIPS and, thus, pay today’s Social Security beneficiaries with money obtained on cheaper terms. At the same time, many (probably millions) of today’s workers would prefer to take the Social Security withholding of each paycheck and devote that gross income to something other than lending it to the government at an official 1 to 2 percent, especially when the return will likely be adjusted downward after a major “reform.”

There are even win-win outcomes for those who currently may be net assets to the system. Note that the above estimated rates of return calculated for cohorts are average figures that overstate the returns to high-income earners. For example, a 2013 analysis estimated that a woman who reached and maintained the Federal Insurance Contribution Act (FICA) maximum income threshold by age 37 would earn an implicit real rate of return of only 0.8 percent (while someone who paid half as much to Social Security from every paycheck would earn a return of 2.2 percent).10 This woman is exactly the type of high-income worker who presumably would be eager to leave Social Security and invest her savings elsewhere. But since Social Security effectively forces her to lend money long-term to the government at a very low real rate of 0.8 percent, it would hurt the financial footing of the U.S. Treasury to simply allow her to opt out. It is for cases such as this that we could include a requirement to “settle up” with the system before opting out. Even in this type of scenario, the worker benefits from the new option.

For a back-of-the-envelope illustration of my perspective—which ignores some of the nuances in projections of tax and benefit levels over time—consider a 32-year-old man who “contributed” the maximum $12,561 (which is the 10.6 percent combined employer and employee portions of the payroll tax dedicated exclusively to Old-Age Survivors Insurance) into Social Security in 2016, and will continue paying this amount (in today’s dollars) until he retires at age 67. He then draws out the maximum benefit, which, for simplicity, we will take as a fixed $38,000 (in today’s dollars) per year until he dies at age 80.11 With these assumptions, the real return for this hypothetical man is a mere 0.8 percent.

Under my proposal, if the Treasury uses an outside inflation-adjusted borrowing rate of 1 percent, then the cash flow generated by this man’s (forced) participation in Social Security has a present value of about $21,000 to the government. So, to opt out, the man could either write a lump-sum check for $21,000 or, if he lacked the financing, continue to pay the maximum amount into FICA for another year and eight months, at which point he would no longer be a net asset to the government.

Note that this arrangement benefits the man because he would surely do something better with the flow of $12,561 maximum OASI contributions than earning the 0.8 percent he’d get if he kept it in Social Security, even accounting for the $21,000 upfront charge. (Note, though, that throughout this example, we are assuming the man’s gross salary rises by the full amount of the OASI taxes previously paid by his employer.) If the man used his previous OASI-dedicated cash flows just to buy TIPS bonds earning 1 percent, he’d mimic his original outcome; he effectively would have just re-contracted with the government to lend through the TIPS market instead of through Social Security. But if he let his original OASI contributions cover his “opt-out charge,” and then invested the remainder of the same cashflow stream in the private sector earning (say) just 1.5 percent in real terms, then he could fund the same retirement income that Social Security had been promising, and still have an extra $77,000.

For more on these topics, see Thomas R. Saving, “Social Security”, in The Concise Encyclopedia of Economics. See also David R. Henderson, “David Stockman’s Time Horizon”, EconLog, June 16, 2012 and David R. Henderson, “The Privatization Shell Game”, EconLog, September 13, 2010.

As mentioned earlier, the cut in Social Security taxes could lead to an increase in the supply of labor. (Recall that that possible effect is what’s behind the potential small drop in wages, even for workers who don’t opt out.) The upside of that possibility, from the government’s viewpoint, is not only would the financial footing of the Social Security program improve, but also the government’s other tax revenue would increase somewhat. The lower effective marginal tax rate on labor income would lead some people to work more, through more hours per year and also by extending their careers. To reiterate, the consensus in the empirical literature is that a reduction in payroll taxes would accrue almost entirely to workers’ benefit (through higher wages and salaries paid for the same quantity supplied of labor). Yet now I am pointing out that to the extent that this might not hold for all workers—perhaps for retired professionals who might consult for additional income in their later years—then their extra work effort (brought forth by lower marginal tax rates) would result in additional tax receipts for the government through normal income tax channels.


The Social Security program is in the hole by $11 trillion. The government could ameliorate this gap by letting Americans opt out of the system, so long as they make a payment to cover the present value (if positive) of their net obligations. Because the pay-as-you-go system has an implicit rate of return that is lower than private investments but higher than what the government would pay in the bond market, there are “win-win” moves that allow workers to build up a larger retirement fund while reducing the government’s financial hole. Economists generally think that most workers would capture most of the gains (in the form of higher wages and salaries) from elimination of Social Security taxes, but to the extent that this weren’t true, it would mean that workers were supplying more total hours because of the lower tax bite. In that case, the reduction in the implicit labor tax rate would boost the wage base, increasing federal receipts from normal income taxes, and would likely lead to greater rates of saving.


If we include the analysis of the Medicare Trustees, then the combined shortfall exceeds $40 trillion in present-value terms. See “2016 Annual Report of The Boards of Trustees of the Federal Hospital Insurance and Federal Supplemental Medical Insurance Trust Funds,”, p. 216. Note that the $40+ trillion figure includes $28.6 trillion in obligations for Medicare Parts B and D, which have different procedures for financing that OASDI and the Medicare Hospital Insurance (HI) programs. Also note that if we exclude the $3.1 trillion in reserve assets held by the Social Security and Medicare trust funds, then the 75-year shortfall (in present value terms) is $43.6 trillion.

Figures from the Social Security Administration, presented in Table 3-1 (p. 75) of Doug McGuff and Robert P. Murphy, The Primal Prescription (Malibu, CA: Primal Blueprint Publishing), 2015.

Megan McArdle, “The 401(k) Problem We Refuse to Solve,” Bloomberg, January 3, 2017.

If desired, we might also insist that those opting out buy adequate life insurance, and contribute to a safe retirement account. This would be analogous to the government mandating that drivers obtain auto insurance if they are to be allowed on the roads; the government doesn’t itself collect the premiums and pay out accident claims, which is how the government currently handles “old age insurance.”

An important paper in this literature that presents both theoretical and empirical results is J.A. Brittain, 1971, “The Incidence of Social Security Payroll Taxes,” American Economic Review 61:110-125. While this classic paper relied on cross-sectional analysis of 64 countries, a more recent analysis in J. Gruber (1997), “The Incidence of Payroll Taxation: Evidence from Chile,” Journal of Labor Economics 15:S72-101, performed a time-series study of Chile during the 1979-1986 to assess the impact of its large 1981 payroll tax cut. Gruber found “the reduced costs of payroll taxation to firms appear to have been fully passed on to workers in the form of higher wages” (p. S99). Furthermore, according to Don Fullerton and Gilbert E. Metcalf’s March 2002 “Tax Incidence,” NBER Working Paper 8829, “In a survey of all labor economists at top-40 U.S. institutions, Fuchs et al. (1998) find that the median belief about the payroll tax is that 20 percent of the burden is borne by employers” (p. 28, fn 39). In this quote the authors referred to Fuchs, V.R., A.B. Krueger, and J.M. Poterba (1998), “Economists’ Views About Parameters, Values, and Policies: Survey Results in Labor and Public Economics,” Journal of Economic Literature 36:1387-1425.

David Blanchett, “What Stock Market Return Should Your Financial Plan Assume?” WSJ blog post, March 31, 2016.

D.R. Leimer, “Cohort specific measures of lifetime net social security transfers,” 1994, Working paper 59 (Office of Research and Statistics, Social Security Administration, Washington, DC), cited in Martin Feldstein and Jeffrey Liebman, “Social Security,” Chapter 32 in A.J. Auerbach and M. Feldstein, Handbook of Public Economics, Vol. 4, 2002 (Elsevier Science).

See Jim Kessler and David Brown, “Is Social Security Regressive?” Third Way, July 23, 2013.

Projected future Social Security payments based on: Estimated Annual Benefit Amount Payable.


*Robert P. Murphy is Research Assistant Professor with the Free Market Institute at Texas Tech University. He is the author of Choice: Cooperation, Enterprise, and Human Action (Independent Institute, 2015).

For more articles by Robert P. Murphy, see the Archive.