June 5, 2026 – Almost to the day, today, eight years ago, was my last day at work. One thing I always looked forward to in retirement was never having to pay those dreaded payroll taxes again. Alas, eight years into retirement, I’ve picked up a few side gigs to stay involved and now run my own small financial advisory business. Not only do I pay Social Security and Medicare taxes again, but I now pay the full 12.4% Social Security and 2.9% Medicare taxes, i.e., the employee and employer portions out of my pocket. Ouch! I had hoped I would never have to put money into that stupid Social Security Ponzi Scheme again. Oh, wait, what did I just say? I must have heard this somewhere, probably from Elon Musk on Joe Rogan’s show. He probably said this mostly for the shock value without thinking too much about the financial nuances.
Nevertheless, the Ponzi Scheme comparison got me thinking: While Social Security is certainly not a literal Ponzi Scheme, where some scam artist runs off with the money, and the investors lose all their funds, is Social Security a Ponzi Scheme, at least to a degree? Are the ways in which Social Security differs from a Ponzi Scheme really only distinctions without a difference? How much better could I have done if I had invested my personal payroll contributions into the stock market or some other financial asset portfolio? All interesting questions! Let’s take a look…
What is a Ponzi Scheme?
Charles Ponzi promised a very attractive investment scheme: give him money today, and he can double that within 90 days, a rate of return of about 1,500% annualized. He claimed that he found an arbitrage scheme involving so-called international reply coupons (prepaid stamps for international mail) that would fetch this astronomical return target. Of course, none of the money he received from investors went to his alleged investment. Instead, he funded his lavish lifestyle and paid early investors with money coming in from new investors. The scheme’s popularity outpaced his cash-flow needs, and he was able to continue paying early investors from the new cash flows. Helping him keep the scheme alive was that many investors simply reinvested their “earnings.”
Eventually, of course, the system collapsed, and investors lost about $20m, which is over $300m in 2025 dollars. Charles Ponzi lived a very lavish lifestyle, and some early investors who were smart (lucky?) enough to withdraw their money from the system also benefited from the scheme. But all subsequent investors were left holding the bag. Luckily, the scheme was shut down early enough. There were some remaining assets for the investors to recover at least 30 cents on the dollar. Intriguingly, the largest Ponzi Scheme ever was Bernie Madoff’s hedge fund, where investors recovered over 90% of their initial investments. There, the court-appointed trustees clawed back large sums of ill-gotten gains that had been paid out to early investors. Of course, investors never recovered any of the imaginary and fraudulent paper gains.
Social Security has features similar to a Ponzi Scheme
The similarities between a Ponzi Scheme and Social Security are numerous. In Social Security, an initial generation of beneficiaries received above-average returns, in fact, as high as +infinity for the generation that got benefits without ever paying into the system. The politicians who passed the law got the benefits of pleasing their constituents, a.k.a., buying votes. Just like Charles Ponzi never invested anything in the international stamp scheme, Social Security is a pay-as-you-go system. Moreover, the “free money” and high internal rate of return (IRR) of the initial generation come at the cost of much lower IRR for future generations, more on that below.
Some folks criticizing Musk pointed out that Social Security has government backing, unlike the typical Ponzi Scheme. But government involvement is a two-edged sword. In fact, a government mandate makes any potential Ponzi Scheme worse, not better. With a privately run Ponzi Scheme, at least I have the option not to participate.
Moreover, the government’s backing behind Social Security is not 100% watertight. Unbeknownst to many, we have no legal ownership, no contractual right, and no property rights in the FICA taxes we pay over our lifetime. Theoretically, if the government wanted to take away our Social Security benefits, we’d have no legal recourse. The U.S. Supreme Court decided so in Flemming v. Nestor (1960). Of course, a widespread rug pull is highly unlikely because it’s politically unattractive – old folks are an important voting bloc – but there is nothing keeping the government from a gradual expropriation of your Social Security benefits. It has happened before, i.e., through increasing the age at which you can claim your full benefits. It can happen again, especially considering the current underfunding, i.e., once the Social Security Trust Fund is depleted, likely in 2033, the payroll tax revenues only cover about 77% to 81%, according to this 2025 Social Security Administration report.
The one big difference between a Ponzi Scheme and Social Security
The one major feature inherent in a Ponzi Scheme is that it will eventually collapse, causing a large group of investors to incur significant losses, up to the total loss of their investments. After that, a Ponzi Scheme has no more victims. That is very different from Social Security. That got me thinking: How could we have saved Charles Ponzi’s scheme? For example, the government could have stepped in and “recruited” another generation of investors. Take their “investments” and distribute the money among the Ponzi victims. The government no longer promises Ponzi-Style rates of return, but something much more manageable, say inflation plus the rate of population growth, and then successively forces ever-new rounds of investors to chip in and make whole the previous round of investors. None of the investments is actually put to productive use, but we have instead created a pay-as-you-go system in which one initial cohort (Charles Ponzi plus a few of his early investors) had a windfall profit, while all subsequent “investors” have a slightly positive, albeit below-market, return on their investments.
Of course, nobody would want to continue investing in that Scheme at such a low rate of return, but the government’s taxing authority could clearly make that happen. That sounds like Social Security to me! Just like in the payroll tax-funded Ponzi Scheme, Social Security avoids a total loss to any one generation by spreading a large loss over all future generations. That loss may be small enough for everybody to stomach, especially when the real return is still (slightly) positive, and we factor in other attractive features of Social Security, e.g., longevity insurance and care for widows and orphans. So, how high or low is the implicit return on your Social Security cash flows? To better understand the return expectations, let’s build a simple model to study this question, which brings me to the next section…
What is the expected return of Social Security? A simple multi-generation model of a Pay-As-You-Go system
Imagine we have an infinitely lived economy, and at any point in time, there are two generations present: young and old. Let’s assume initially that both generations are of equal size in every period. Assume one time period is as long as an entire generation (e.g., about 30 years). So, the young generation in period t will be the old generation in period t+1. Also, the older generation in period t will die and no longer be around in t+1, while a new young generation enters the landscape.
Let’s assume that in period T, the current president proposes a pay-as-you-go system that taxes each worker in the young generation and gives the proceeds to the current old generation. To sell this to the young generation, he tells them, “Don’t worry, when you’re old in period T+1, you will get a transfer from the then-young generation.” Of course, this new transfer program would be wildly popular with the initial old generation, for they get money for nothing. Subsequent generations receive a return equal to the growth in the total wage base, assuming the marginal tax on labor income remains constant. But the real growth rate of the wage base is likely much lower than what you’d get in the stock market!
For example, in the chart below, I plot the cumulative gain in the wage base compared to CPI inflation and the S&P 500 Total Return index (dividends reinvested). Note that the wage data series is not per-person wages but total wage income across all workers, to gauge the potential return of an expected pay-as-you-go system according to the overlapping generations model. Quite intriguingly, the total wage base increased at a pretty attractive rate: 6.07% nominal and 2.30% if adjusted for inflation. 4.47% and 0.74%, for nominal and real growth, respectively, over the last 30 years.
Summary so far: The political appeal of a Social Security-style transfer program is that the initial generation gets a magnificent windfall, i.e., they get benefits without ever paying anything into the system. If the system had just been a one-time affair, this would have been a zero-sum game, i.e., the younger generation footing the bill would bear 100% of the burden and essentially suffer a 100% loss of their contributions. Just like in a Ponzi Scheme. However, by keeping the system alive for many more, even infinitely many generations, we spread the cost of the free gift to the initial generation over sufficiently many other people. Spreading the cost wide enough can make the loss small enough that each future generation may not even notice. But how small or big is the cost to the subsequent generations? Is the internal rate of return (IRR) of someone’s Social Security contributions and benefits much lower than expected financial market returns? This brings me to the next section…
Social Security Return Sample Calculations
Let’s consider a 67-year-old retiree who begins receiving benefits at their regular retirement age. What’s the value of the Social Security benefits? Let’s perform an actuarial exercise.
Let’s assume that several fresh retirees claim Social Security today in 2026. They look back at their earnings and payroll contribution history, calculate their benefits, and plug the Average Indexed Monthly Earnings (AIME) into the Social Security benefits formula; please see the chart below. Note the two bend points where the slope of the benefits function shifts from 0.9 to 0.32 and then to 0.15. This concave shape of the benefits function creates redistribution: retirees with higher incomes, while receiving larger total benefits, have a lower replacement ratio (benefit divided by AIME).
Side comment: Who pays the employer portion of payroll taxes? I assume that the full cost of Social Security payroll taxes (currently 12.4%, though this rate has varied between 10.4 and 12.4% over the last 45 years) is borne by the employee. Why not include only the employee portion? Simple, your employer is not a charity. They will lower your salary offer to account for that cost, of course. Similarly, without Social Security, you would likely receive a pay raise.
I consider six different work histories:
- Retiree 1: 45 years of earnings from 1981 to 2025, earning the maximum taxable earnings in each year. This is the worst-case scenario from an internal rate of return perspective because the Average Indexed Monthly Earnings (AIME) accounts for only the top 35 years of indexed earnings. In addition, the AIME is the maximum level, so you’re far into the 0.15% bracket, where the benefit as a percentage of your contributions is minimal.
- Retiree 2: 35 years of earnings from 1991 to 2025, earning the maximum taxable earnings in each year. This is slightly better than the scenario above because you avoid the ten wasted years that add essentially zero to your AIME as in the worst-case scenario. But you’re still far into the 15% bracket of the AIME vs. benefits chart. The benefit as a percentage of your average earnings will be very low.
- Retiree 3: Same as #2, but the earnings are exactly half of the maximum taxable amount every year. Thanks to the concavity of the benefits function, the monthly benefit is only about 28% lower ($3,037 vs. $4,211) despite paying in exactly 50% than retiree 2.
- Retiree 4: Same as #2, but now earnings are only 20% of retiree 2. But the benefits are significantly better than 0.2x of retiree 2. Despite paying into the system 80% less than retiree 2, the benefits are only about 60% lower.
- Retiree 5: This should be the best possible Social Security IRR, i.e., the shortest possible contribution history (you need 10 years to qualify), the latest possible history, i.e., the 10 years right before retirement, and the contributions are low enough to still land you in the 90% slope portion of the benefits.
- Retiree 6: I also want to include the typical FIRE career. Someone who claims benefits in 2026, but was already FIRE’d for more than 20 years after a 15-year high-paying career, making the maximum taxable amount every year between ages 31 and 45.
Since the value of Social Security benefits depends on your life expectancy, I also calculate the IRR for four different types of individuals. I use my actuarial sheet; see Part 56 of my SWR series for more details, and I use a real discount rate of 2.5% per annum.
- A: A male with below-average life expectancy. A life expectancy of about 14.8 years. The actuarial calculations imply that a $1 monthly Social Security benefit, adjusted for CPI, has a discounted expected value of $143.08.
- B: A male with average life expectancy. A life expectancy of about 16.7 years. A $1 monthly Social Security benefit, adjusted for CPI, has a discounted expected value of $157.74.
- C: A female with average life expectancy. A life expectancy of about 19.1 years. A $1 monthly Social Security benefit, adjusted for CPI, has a discounted expected value of $176.48.
- D: A female with above-average life expectancy. A life expectancy of about 21.8 years. A $1 monthly Social Security benefit, adjusted for CPI, has a discounted expected value of $195.60.
- What about men with above-average health or women with below-average health? Simply use individuals C and B, respectively.
Let’s look at the results; please see the summary table below:
- In the absolute best-case scenario for Social Security, Retiree 5, not even the roaring bull market over the last 10 years would have created an equity portfolio large enough to compete with the Social Security expected benefits: $66,709 for the equity portfolio, compared to $104,332 to $142,628 in Social Security benefits. All other retirees would have done significantly better investing their contributions in the stock market.
- If equity investments were not your thing, your Social Security cash flow series could have outperformed investments in 3-month T-bills in all but the most disadvantaged retirees: Retiree 1A and 1B, i.e., with the longest contribution history and the shortest life expectancy: males with average or below-average life expectancy.
- Investing in 10-year US Treasury bonds (closely approximated by, say, the iShares ETF, ticker IEF) would have done slightly better than T-bills. But still, most retirees would have done better with their Social Security “investment.” Everyone except Retiree 1, Retiree 2A and 2B, and 6A.
- The Social Security IRRs are also interesting:
- Of course, retiree 1 did quite poorly, with an IRR that didn’t even keep pace with the average inflation rate of 3% over those 45 years. Retiree 2 also just gets roughly a zero real return, though women do a little bit better. But the low returns for retirees 1 and 2 are by design: Social Security has an element of redistribution that takes money from extremely high earners and subsidizes benefits for low earners and/or people with a shorter earnings history.
- Retirees 3 and 6 got pretty decent returns between about 4% and 6%, depending on exact parameters. That’s pretty close to the “theoretical” Social Security return, calculated as the average annual growth in the wage base.
- Retiree 4 has significantly below-average earnings and thus receives very attractive nominal returns of 5.83% to 7.4%.
- Retiree 5 has the most impressive returns, all in double digits, ranging from 19.94% to 25.59%, depending on gender and health status. Much higher than the equity market.
- This all makes sense: a more “average” earnings history should give you the average Social Security expected IRR. Higher earners do worse, and lower earners do better due to the redistribution.
Limitations
I’m aware that I took some shortcuts in my calculations. On the one hand, by basing my calculation on retirees who are 67 years old today, I ignored the possibility of someone dying before reaching that age. So, that would lower the attractiveness and the IRR of Social Security because investors who die before age 67 could have bequeathed assets to their heirs, while their Social Security benefits are forever lost. On the other hand, Social Security has some additional built-in benefits: disability benefits while working, survivor benefits, and the possibility for spouses with low or no earnings to qualify for 50% of the higher-paying spouse’s benefits. All of this would have been quite complicated to model in a short blog post, so I cross my fingers that the two effects roughly cancel out.
I also ignore taxes. Including the tax effect will be detrimental to both Social Security (because up to 85% of your benefits are taxable in retirement) and the investor. Taxes could create a drag both during the accumulation phase (dividends and interest income are taxed along the way) and in retirement. Though I suspect the tax impact on the investor is often lower because the investor pays tax only on gains, not on the cost basis. Only when investing heavily in equities over many decades, with capital gains making up 85%+ of the final portfolio, would the tax be as impactful as in the Social Security case. However, even in that case, the tax on capital gains is still lower than the ordinary income tax on Social Security.
Social Security also has a feature that’s not easily replicable with financial instruments. So, even if you had, say, $479,073 (retiree 3.B), you couldn’t so easily transform it into a $3,037 monthly benefit, adjusted for CPI inflation, even though the table above spits out the numbers for retiree 3 and an average-healthy male. That’s because there appears to be no functioning market for CPI-adjusted annuities. You will find the most competitive prices for an SPIA (Single-Premium Immediate Annuity), which is unfortunately, not CPI-adjusted. But the SPIA could still get you close: I got an SPIA quote for a 67-year-old with $479,073, and he could get a monthly benefit of $3,441. The retiree could invest the extra $3,441-$3,037=$404 each month to fund future cost-of-living adjustments, though this would not be a truly safe hedge against inflation and longevity. Also, forget about applying the 4% Rule! The $479,073 portfolio in this example would only generate just under $1,600 in monthly benefits. Social Security generates larger benefits than the 4% Rule because of the mortality risk: if you die one month after Social Security starts, your money is gone, while 4% Rule investors can give their leftover portfolio to their heirs.
What about the prospect of benefit cuts? After all, current Social Security payroll tax revenue will cover only about 80% of the promised benefits. A full 20% benefit reduction across the board would certainly reduce your Social Security IRR. We can gauge this by examining the IRR deterioration when moving from the expected discounted benefits of Individual D (Female with above-average health) to those of Individual B (Male with average health), with the latter being roughly 20% lower. That reduced the IRR by well over a percentage point. So, even if retirees were to bear the entire cost, Social Security wouldn’t become a Ponzi Scheme. The IRR will merely be slightly lower. And keep in mind that not all the costs of the future Social Security reform will be borne by the current retirees. Any benefit reduction will likely be phased in slowly, probably through increases in the Full Retirement Age and the payroll tax. So, my Social Security IRR will likely not decrease by that full percentage point. I’m crossing my fingers!
What if AI takes away all the jobs? Well, that would be a concern for a pay-as-you-go retirement system because the “returns” are tied to the growth of the wage base. Being the eternal optimist, I predict that AI will cause productivity gains and some job losses, but will have an overall positive effect on all economic variables, i.e., productivity, GDP growth, earnings, etc., just like all other productivity breakthroughs before AI.
Conclusion
True, Social Security shares some features with a Ponzi Scheme. It’s also disconcerting that future governments could mess with our benefits without beneficiaries having any legal recourse. But after beating up on Social Security, I want to end this post in a conciliatory tone. Sure, I could have done much better investing my Social Security contributions in the stock market. But first, that wasn’t in the cards; we, as a society, can’t repudiate the obligations to our current retirees. And second, in absolute terms, the implicit returns on my Social Security contributions (even after accounting for both employer and employee costs) are not that bad. Especially retirees with lower incomes and/or without a full 35-year work history (the FIRE crowd!), who still fall into the 32% bracket of the benefit formula, got pretty decent returns. Less than equity returns, but comparable and in many cases even superior to nominal bond or money market returns. Heck, over the last 20 years, your Social Security contributions might have outperformed Small-Cap Value stocks, but let’s not go there. So, let’s stop beating up Social Security. By design, as a Pay-as-you-go system, it can never perform as well as the stock market. But its IRR is still high enough to make Social Security a useful tool.
Thanks for stopping by today. I look forward to your comments and suggestions!
Title Picture: WordPress AI picture generator
Great write up and modeling.
I agree with your conclusion- historically social security has paid out reasonably well vs what current and past retirees put in. The keyword there is historically.
With demographic trends (active workers per retiree in particular), most people, especially of the younger generation, are rightly concerned that conclusion will not hold. Will the person retiring in 2070 still have a decent, if lower than had they invested it, return?
I think most people believe that the real ROI will not just be lower but will be negative.
It is this precise feature that most resembles a Ponzi scheme. You describe the Ponzi scheme as having two cohorts – the winners and the bag holders. It is more of a gradient though – there are also many people who end up with a moderate or zero return in that intermediate phase before things collapse entirely.
It is my contention that we are in just such a phase now. Any Ponzi scheme doesn’t look truly horrible until it fully collapses. What most distinguishes SS from a traditional Ponzi is how long lived it can be – rather than being measured in years it is measured in generations.
Yeah, I can see that people have this concern. But with 2%+ GDP growth on average and population growth, we should still grow the wage base.
So, while a slight deterioration of the SocSec IRRs is possible I doubt it can collapse entirely.
I can see SSI basically experiencing a two-generation shrinkflation into obsolescence as budget constraints (read: debt interest payments) start to arrive
We will see how much the system delivers over the next few generations. I’m still mildly optimistic and don’t believe that SSI will become obsolete.
Great insights. I did some back of the napkin math a while back and found similar conclusions.
Especially the discounting of years over 35 and that it doesn’t matter when it was paid in was surprising.
I think that the majority may not invest that smartly. There are still a lot of people that would leave their money mostly in cash, so that this system outperforms for these people.
It also reminds of the idea to pair a FIREed person with a working person to alleviate SoRR.
Good point. This is another “redistribution,” i.e., lower-income folks who are more likely to pick low risk and low return investments (money market, CDs, savings accounts, etc.) have a much lower opportunity cost when paying into Social Security.
Also, good point about the “Retiree-Saver Investment Pact” – SWR Series Part 53. Of course, that pact involved actual investments in stocks, not a pay-as-you-go system.
I think Sebastian is actually onto something here, when the gross negative return relative to VTSAX is north of 4% for an ‘average’ investor, there isn’t really that much downside for this scheme.
https://pmc.ncbi.nlm.nih.gov/articles/PMC8695249/
I’m back to being properly appalled at how well the median/average retail investors perform (not even taking into account fees). I guess that alpha isn’t coming from nowhere, and we have an entire finance sector for a reason… but wow.
Probably something ripe for another 5000 word banger of an article… when the ‘average’ investor gets absolutely trounced by Fidelity’s self-identified top tier investors [i.e. those who forgot their passwords and/or forgot they had accounts at all, or deceased], everybody in the FIRE blogosphere is falling victim to the same precise thing Randall Munroe was on about with https://xkcd.com/2501/ that our average familiarity estimate is miles off actual conceptual familiarity of an average person.
I guess chalk up some more points for a garbage tier paternalistic system managed by politicians… because it definitely be worse.
Yeah, not everyone is as enlightened as the FIRE crowd with the index fund investing.
But I also noticed that the charlatans are at our doors: value, dividend focus, small-cap, small-cap value etc. are essentially stock picking styles that will leave many investors poorer.
Says the guy who promotes option trading strategies??? Lots of respect for your work, but calling those who who have a different opinion than you on the Small cap value premium as charlatans is a bit much don’t you think? Dr. William Bernstein and Rick Ferri have indicated that they have a small cap tilt, are also very clear in explaining the risk/benefits of said tilts, and are also quite clear that this tilt is not necessary. Same goes for Ben Felix at PWL. I wouldn’t call these rather accomplished and intelligent individuals charlatans.
Dividend investing – I’d start to agree with the Charlatan label here.. Still a bit harsh though.
My options strategy has made money consistently. With an Information Ratio that puts many hedge funds to shame. Compare that to the sad record of SCV over the last 20y (more risk, lower return than the S&P 500). So you probably don’t want to go there, buddy.
I maintain that people who are unconditionally promoting these stock picking schemes as free and reliable alpha are charlatans. I have met and talked to Bill Bernstein and Rick Ferri (have you, by the way?), and I want to stress that they are good guys. They are obviously not charlatans. The fact that you want to tie them to my quote says more about your investing knowledge than mine.
I’m glad that we can agree on the dividend/yield charlatans.
Is it really fair to directly compare SS to stock market gains? Even so, I guess you’ve shown it still looks decent for large cohorts of contributors. I wish our system was based on a sovereign wealth fund instead, with guardrails for minimum/maximum payouts when participants are “annuitized” at withdrawal time. I’m sure other countries (Australia ?) do something similar already. Plenty of fodder for political bickering still, but at least people would be more accustomed to the concept of variable benefits and the never-ending discussion would be more about what a contemporaneous fair minimum and maximum is. Like everything in life nothing is truly guaranteed.
There was a time under George W. Bush when we ran Social Security surpluses and someone suggested investing them in the stock market. It didn’t go over well. Too much push-back.
What we got is the American approach: cold individualism. If you’re dedicated and motivated, you can achieve FI through tax-advantaged accounts (IRA, 401 (k), Roth, etc.), but you have to do it yourself. The government wouldn’t invest the money for you.
Thank you for an excellent post as always. It appears that the assumption that makes this look better for most cohorts is the rate of population growth and thereby the growth of the workforce. With the current political situation on immigration and the advent of AI wouldn’t that assumption be under jeopardy?
Not all immigration is equal. We need immigrants who come here and work hard and contribute: grow GDP and pay taxes. I’m 100% fine with kicking out unproductive immigrants that come here to raid the social safety networks, commit crimes, etc. They are a net-negative and should be removed. But we still have a ton of productive immigration, even after getting rid of the bad apples.
About AI: I’ve mentioned it in the post: AI will not make everyone unemployed. AI will increase productivity and wages.
Dear Big ERN, excellent analysis as always. In this case however, I am not convinced by your conclusions. Please allow me to elaborate (sorry for the long comment):
[1] You concluded that Social Security (SS) has decent returns because its return is comparable and sometimes superior to treasury bond/bill. That would be a good return for a short term investment, but SS is a retirement product with 30+ years of investment horizon. If your investment horizon was 30+ years, would you really invest in treasury bond/bill? Of course not, you would invest mostly in stocks. So, for a retirement product, SS has horrible returns for everyone except Retiree 5.
[2] For a FIRE aspirant who is already investing say 50% of his income in S&P 500, another 12.4% of his income going to a retirement product (SS) that has a return comparable to 10 year treasury essentially means an asset allocation of 80% stock + 20% treasury during accumulation phase, which is not terrible. But for an average American (not Retiree 5) who is investing say 6% of his income in stock, 12.4% of his income going to SS essentially means an asset allocation of 33% stock + 67% treasury during accumulation phase, which is absolutely terrible. For him, that 12.4% invested in S&P 500 instead would have made enormous difference.
[3] Let’s say an investor had extremely low risk tolerance, and actually wanted to invest mostly in 10 year treasury. Even he (except Retiree 5) would be better off buying treasury bonds instead of “investing” in SS for several reasons: (a) if he bought treasury bonds the government would be obligated to pay him the principal and interest, but as you yourself noted government is not obligated to pay him SS benefits; (b) if he died, in case of treasury bonds his wife would get 100%, but in case of SS she might only get 50%; (c) if she died too, in case of treasury bonds their children would get 100%, but in case of SS their children might get 0%.
[4] Yes, SS cares for widows and orphans, disabled, but couldn’t that be more efficiently done through term life insurance and disability insurance respectively (made mandatory if necessary)? Yes, SS can be a hedge against inflation and longevity due to lack of CPI-adjusted annuities, but an immediate annuity where the monthly payout compounds 3% annually could pretty much solve that too (such annuities already exist in other countries). SS indeed benefits Retiree 5, but that is just redistribution of the return Retiree 1 & 2 would have gotten, so Retiree 1 & 2 were essentially taxed for the benefit of Retiree 5. The government could do that directly like any other welfare program, they don’t need SS for that. So SS doesn’t really seem necessary.
[5] Yes, US shouldn’t repudiate the obligations to those currently receiving SS or already near retirement. The government promised them these benefits and they have paid into SS for years, so of course the government should pay them these benefits. But the government should formally recognize that obligation by adding it to the $39 trillion government debt and should handle it the same way. It is not a justification for forcing everyone else to “invest” in SS – which seems to be a weird mix of bad retirement product, insurance, welfare, and annuity. For everyone else, SS should be replaced by a Thrift Savings Plan where the 12.4% goes to an account in that taxpayer’s name and that taxpayer decides what it is invested in.
[6] Also, you seemed to suggest that SS is not a Ponzi scheme because Ponzi schemes collapse. If Charles Ponzi or Bernie Madoff could force everyone to “invest” in their schemes for generations then their schemes would not have collapsed either (the returns would go down) – does that mean they weren’t running Ponzi schemes? Of course not, those would still be Ponzi schemes. The only difference between SS and their scheme seems to be that they used deception to get investments because they couldn’t force people to invest, whereas the government didn’t need to use deception because they can force people to “invest”. Thrift Savings Plan is not a Ponzi scheme because the money taken from an investor actually gets invested in an account in that investor’s name. In case of SS however, the money taken from an investor doesn’t get invested at all, instead money is taken from someone else to pay that investor, and so on. So SS seems to be a Ponzi scheme to me, orchestrated through force rather than deception, but Ponzi scheme nonetheless.
Thanks again for sharing this excellent analysis.
Yes, all good points!
1: Agree, over long horizons, we should be 100% equities initially an only roll down the glidepath later in retirement. But not everyone has that kind of stomach. I know too many folks who have 100% bonds + money market.
2: Yes, that’s a bond allocation way too high for most folks.
3: Retiree 5 is a very oddball retiree. Probably less than 0.1% of retirees fall into that category.
4: Exactly my point: the government sells SocSec by bundling it with other services that are really cheap to come by in the marketplace. But some voters will get distracted by these “shiny objects.”
5: Agree. The government should be more explicit and have at least estimates on the implicit liabilities. Every business would have to do so!
6: Agree. Social Security is in effect like a Ponzi Scheme that has carried on and spread the losses over future generations. So, in the broad sense, not in the narrow sense, SocSec could be considered a Ponzi Scheme.
Retiree 5 is me! A stay at home parent for 15 years, an early retiree to take care of aged parents with special needs, with a brief career as a wildly successful small business owner (that allowed me to retire early). So yeah, an oddball. I’m grateful SS will theoretically be at least mostly there for me at some point.
Nice! Congrats on getting the potentially highest possible IRR!
The redistribution in Social Security is not as great as it might seem from the table. Longevity is greatest among those with college educations, who, on average are also those with higher income.
Correct. That was one of the reasons for the redistribution. However, if we look at retirees 2C=highly paid white collar worker, 3B=moderately paid worker, and 4A=low-paid worker with low life expectancy, there is still some remaining redistribution in there: 2C=3.27%, 3B=4.68%, 4A=5.83% IRR.
Sounds like ‘not aggressively priced and non-optional longevity insurance’ is still the right mental bumper sticker to put on this for the FIRE set – but I hadn’t fully appreciated that those of us who can saturate SSI contributions for a decade are actually much of an advantageous edge case (even though we could do better on aggregate through equity investing, as an insurance product this isn’t a completely terrible deal).
I guess my only remaining frustration is that it could have been rather non-controversially anchored to life expectancy decades ago – which would have absolutely reduced just how much generational wealth transfer phenomenology was prominent, and would also have pressed out the ‘glide slope to empty lockbox’ substantially.
When a scheme has paid out well *historically*, but for current generations paying in who can read the demographic tea leaves realize that just meeting CPI adjusted TIPS returns may become untenable, it honestly feels like we’ll still wind up in a place where the needed reforms are going to arrive far to late, and need to be far too drastic to be comfortable or even necessarily politically viable to implements.
Yeah, the best time for Social Security reform has always been many years ago. But politicians have no incentives to fix anything that will pay off 10+ years down the road.
You could also argue that the stock market also displays Ponzi like dynamics. Stock prices today embed expectations of future growth that cannot be physically realized indefinitely on a finite planet. If the economy must eventually stabilize within ecological limits, then long-run growth rates may have to approach zero.
I don’t share your Mathusian/Paul Ehrlich view of the world. The best thing to happen to the planet is if we get richer and then become rich enough to care about the environment. Some of the cleanest places on earth are Norway, Switzerland, Luxembourg, etc.
Also see here: https://www.latimes.com/opinion/story/2026-03-17/paul-ehrlich-wrong-everything
Nice analysis looking at typical returns.
I would like to see some accounting for risk in the market returns category.
What would happen to the hypothetical generation of retirees who didn’t have social security and had instead invested in the stock market if the stock market drops 90% (like it did in the Great Depression)?
Social Security effectively guarantees that elderly people who don’t have a lot of savings, or who are prone to making high risk investment decisions, are not freezing to death homeless and hungry in the winter.
Your prior analysis of the 4% guideline notes the probability of invested assets being completely consumed over 30, 40, and 50 year retirements. Social security forms a backstop against absolute destitution for frail elderly people no longer able to work who run out of investment assets.
Those advocating that they could likely “get better returns investing their social security contributions” don’t seem to consider what the loss of that social safety net would look like in the rare(but inevitable) severe market crash for those who are most vulnerable.
Could you comment on the potential for assets to drop to zero for retirees who invest social security in the stock market instead of getting a guaranteed lifetime, inflation adjusted, government payout?
Obviously, the 12/31/2025 end point of the equity market investment created some insanely good investment results. But even if we experience a big drop of 79% in real terms, most retirees would still come out ahead of Social Security. And the market recovered pretty quickly afterward.
So, my recommendation: it can never be a corner solution (only SocSoc or only 100% equities). Optimally, we have a point in between where we have financial assets to fund a part of our retirement plus Social Security as a floor.
There’s an alternative angle to be taken – I’m sure we all appreciate the value add proposition that is having a social safety net as the overall good it is.
There is still an opportunity cost, and from the perspective of a low income family, having a still considerable chunk of what would be investable income put into this setup basically precludes them from participating in higher growth markets (even if their participation would be unsurprisingly suboptimal in terms of IRR maximization).
What this results in is concentration of risk-taking ability into those who already have disproportionately more resources, so in the truly most ironic sense of a progressive system, social security has both established a floor, but also broadened the gap between the upper class and lower class simply through differential participation in the (admittedly cyclical) wealth creation process
Wish there was some way to nail that retiree #5 situation without having to only earn $810 per month for 10 years 🙂 hmm
Well, there can’t be too many folks with that earnings history. Also, $810 is the AIME averaged over 35 years, with 25 entries counting 0. So the earnings of 205 at the full maximum taxable income would have been higher: 810 x 35/10 = $ 2,835 per month on average over the entire history. Or 0.20 x 184,500/12=3075 in the year 2025.
SS is a system for everyone, not only for a selected group. Therefore, you have to consider how the majority of population can comfortably rely on when they no longer have the ability to earn income in the retirement. You can’t run IRR for everything in your life, at least 90% population do not have ability to run IRR to understand it. SS and Medicare still a place to keep most people in the shelter when they are old. Looking around the world, no any system is significantly better. Germany requires you to have to work for long long time which FIRE is almost impossible.
That’s a lame argument. Just because most folks don’t understand gravity, doesn’t mean gravity doesn’t apply to them. We could all benefit from understanding the IRR behind Social Security. We should propose more IRR calculations not less.
But I agree: our SS system is still better than many other examples in the world. Sadly, Germany, which invested the system, now has one of the world’s worst systems.
Financial evolutions of the performance of Social Security as an investment vehicle are fundamentally misplaced, in my opinion.
Social Security is a welfare system. Its primary function is not to supplement retirement income for upper middle class Americans. It’s primary function is to minimize the social cost of tens of millions of elderly and disabled Americans living in abject poverty, where begging, stealing, and committing other acts of violence would be the norm for the sake of survival.
One doesn’t rule out the other. We can have a transfer program like that and still calculate the IRR.
Social Security is an insurance policy against letting our old people starve in the streets, not an investment scheme. You’ve demonstrated that it’s progressive, which could be argued allows employers to get away with paying lower wages than they should to the lower ranks of workers. There is not the political pressure of old working class people destitute in the streets. Perhaps higher wage more educated workers could have the knowledge and take the risk of investing the funds on their own, but that’s not what the SS program is about.
I want to commend your SWR series. Going through the whole thing is like taking a college class and could (should!) be a book! This is the most comprehensive and specific drawdown information anywhere. The spreadsheet is a great gift to the public and I thank you for your generosity.
I do have questions for a future post, unless I’ve missed it – how is the CPI adjustment to be calculated? Monthly or annually? Where do we get the CPI number? Isn’t it always a month or two behind? I’m quite vague on when and how to implement the “real” numbers.
Or perhaps I am just confusing myself. Maybe CPI inflation adjustment is for the purpose of SWR simulations only? I re-read part 5 and am not looking for a CPI adjustment in calculating the SWR. I am looking for the specifics on how you figure out your spend each month in retirement when you sit down at the computer.
Is it as simple as
1 – calculating 1/12 of initial SWR % on the total portfolio at that moment in time
2 – subtract social security or other income payments from amount
3 – sell assets/withdraw funds in accordance with rebalancing asset allocation including any glidepath
Am I missing steps? A procedural step-by-step post would be very useful.
Sorry for hijacking your comments. Wasn’t sure where to post a question on your body of work as a whole. Thank you.
The SWR is done in real numbers. The spending is real, CPI-adjusted. The returns are CPI-adjusted.
Correct: You run your simulations every month exactly with the 3 steps you describe.
Correct: SSI is to make sure nobody needs to feel guilty about starving, destitute old folks. As I’ve written elsewhere in the comments: that doesn’t mean I’m prohibited from calculating an IRR. It’s still a useful exercise.
The CPI adjustments are done monthly always for the current month. I don’t care if the CPI numbers are released with a roughly 14-day lag. What matters is that the average retiree spends his/her consumption basket every month. Your nominal cost will rise exactly with the rate of inflation. You don’t have to wait for http://www.bls.gov to release the numbers to find out how much you can spend.
SS is not an annuity or an investment. It’s a tax we pay so that our society does not feature tens of millions of old folks starving to death in “Hooverville” shacks.
That would absolutely happen without SS, because most people are for whatever reason unable to provide for their own retirement. It was the case when SS was enacted. It was the problem they were trying to solve.
So it’s not a Ponzi Scheme. It’s just a taxed redistribution from workers to those who cannot work. It’s welfare. The badly named “trust fund” is merely a float budget so that the system stays solvent during recessions. The vast majority of our taxes are not invested in anything, so it is odd to try to calculate an IRR. It’s like calculating the IRR on the gasoline tax used to fund road repairs.
We have to be very careful not to use rhetoric in a way that creates confusion on the above points. I feel like this post does create that confusion.
And of course we’re far into a decades-long bull market in which the S&P500’s PE ratio has risen above 30, in part because of unsustainable deficit taxing/spending by governments.
So it’s tempting to look back in hindsight and compare our SS taxes to what might have been earned instead.
But doing so would only exacerbate SORR. Social Security is part of the cure for SORR because its payments are not based on market returns. The addition of even a modest SS payment makes a dramatic difference to any FIRE/SWR calculator.
But again, it is not an annuity or investment! It is a tax we pay out of a sense of generosity and an awareness that most of us are vulnerable to being financially wiped out if a really bad economic event came along.
The “Money & Macro” channel on YouTube features a video comparing defined benefit transfer programs like the U.S. Social Security system with national annuity retirement systems that are becoming more popular (again, see the bull market for why their popularity is up right now).
George W Bush ran on a platform that included investing the SS trust fund but by the idea died after a couple years of poor stock returns and falling bond yields illustrated why it was not set up that way to begin with.
My question is if Social Security is a welfare provision and not a pension/ponzi scheme, then why don’t we fund it out of standard federal tax revenues and apply means testing like we do for all other welfare provisions? It would greatly reduce government expenditure, be significantly more progressive in nature, and it wouldn’t crowd out investing as much.
I would support that idea!
The politically practical answer is that universal benefit programs are more popular with voters. Social Security and Medicare are close to untouchable because just about everyone gets to take advantage of them, and therefore everyone likes and supports them.
A means-tested welfare program might make more fiscal sense. It would also be the first thing on the chopping block when rich people with well-paid lobbyists come calling.
I’m not sure it would be the rich who would have an issue with this solution. Their IRR is very low and with means testing their cost may be significantly lower even if its coming from general tax revenue. They are definitely the ones who most benefit from a reduction in the crowding out of investment. Its more likely the middle class (retiree 3 and 4) would have the biggest concern.
It seems like you could come up with a means tested solution that closely mimics the IRR curve of the current system to make everyone satisfied. Regardless, as a voter, I find it extremely helpful to know the effectiveness of the current system, and I have read no other article that does a better job of that than this.
Just because it’s welfare doesn’t mean it never be a Ponzi Scheme. The Trust Fund is a “Robbing Peter to Pay Paul Scheme.”
I can calculate an IRR on any cash flow where I pay something in and receive something back at different times. Your example with a gasoline tax is silly.
A government-run, government-endorsed, fully funded system to overcome SoRR would be the system I described in Part 53 of my series: https://earlyretirementnow.com/2022/06/06/hedging-against-sequence-risk-swr-series-part-53/
In hindsight, the proposal by G.W. Bush would have been a blockbuster deal for the participants. But politicians think too short-term. Which is another reason why they shouldn’t be in charge of anything valuable.
Politicians reflect voters, and voters definitely think short term. In a different, more rational universe we might be demanding plausible five year plans, but as it is most voters don’t have the cash to cover a simple car or home repair without going into debt. So we vote for slogans that feel good, and keep thinking the politicians are going to make us rich.
That’s also why most voters are not taking care of their own retirements. The vast majority of people are too low on the personality trait of conscientiousness to optimize the care of their own lives. My dentist told me that most grown-arse adults don’t brush their teeth very well.
Talking about who “deserves” what is a dead end, as is talking about how people “ought to” be saving in their 401k plans, which have been around for about 40 years now. Nobody deserves anything, and if we all did what we “ought to” do we’d be living in such a utopia that SS wouldn’t even be an issue.
The point is that SS prevents a major social, ethical, and political problem from developing, and that’s the non-financial benefit we get for our payroll taxes. “Nobody starves” is a part of the US social contract and if that deal abruptly ended, it’s hard to imagine democracy, capitalism, or rule of law continuing to prevail for long.
In 3rd world countries, coups are common because people are desperate for any kind of change, and because they have nothing to lose, as the existing regime does nothing for them. Those who talk about overthrowing the form of government in developed countries hit a wall when people think about all the benefits programs they might lose.
Maybe we could say people in places like Chad or Burkina Faso should just buy stocks and finance their own retirements, but they aren’t doing so or can’t do so. The majority of Americans living paycheck to paycheck with insufficient retirement savings, even if they have nice debt-financed things, are in a similar position without SS.
Even most people who are FIRE with stock-heavy portfolios would be thrown into abject poverty if a repeat of the Great Depression series of returns was to occur. SS provides a baseline of support that is irreplaceable with risk instruments for most people.
There’s always the option of not participating in Social Security by working at a job that is exempt from that 6.2% OASDI payroll tax. I’m a bit surprised that more people aren’t talking about that possibility. From my perspective, I always chuckle when people keep sweating over when to take Social Security when it makes more sense for me to take the pension right away after retirement. The pension plan offers to cover 100% of medical insurance premiums (both regular insurance for the under-65 retirees and Medicare for the 65+ crowd) after working for at least 25 years, which makes taking the pension immediately at retirement a no-brainer. More importantly, the pension has the nuclear option to pull all of your contributions over the years plus interest at the cost of being ineligible for all benefits. Social Security doesn’t offer that option to be 100% self-sufficient!
That being said, there is a major downside to my pension. Annual COLA increases are fixed at a maximum of 2% with no COLA bank, which means that the pension is all but guaranteed to lose purchasing power during retirement.
This is why I am and will continue to hold 100% global stocks (VT) in my governmental 457b account. My investing time horizon spans both my working years and my retirement.