Social Security is running a permanent deficit and headed for a painful breakdown in less than ten years. The prospect of passing legislation that will fix the deficit for the long term (and avoid creating new fiscal problems) is dim in the current political environment. Utilizing a novel "proportional outcome" ballot question, Congress could push the difficult taxes vs. benefits trade-off directly to the electorate.
In the year 2033, less than a decade hence, the trustees of the social security fund are forecasting a grim milestone will be reached. The program will have paid out all of the premiums collected since its inception, and the annual revenue in that year (and likely every year thereafter) will be far from sufficient to pay out promised benefits. In short, the accounting "trust fund" will have run dry, and everyone collecting social security can expect an abrupt cut, at least 20%, to the size of their next monthly check. Benefits are only likely to erode further from there.
Few in Congress want to arrive at this jarring outcome. Every politician believes a fix, a reform, a transition, a replacement…something is needed. And yet negotiating viable legislation resulting in any combination of more revenue, lower benefits or eligibility changes has not even begun.
The longer you wait for reform, the more drastic the changes required. Of course, this math is well known, and has motivated decades of fruitless reform discussions. Unfortunately, the environment in Congress today is particularly unpromising for the passage of such a politically toxic bill. And, as the problem has become more urgent, the executive branch has only become more inclined to lead from behind. Neither Biden nor Trump (nor any anyone else likely to occupy the White House in the next five years) has demonstrated that they will champion fixing Social Security's long-term structural deficit.
So, it is very difficult to conceive how Social Security reform could be enacted any time soon. It is even more difficult to imagine that, in the unlikely event a bill claiming such a purpose passed, it would actually achieve long term fiscal balance for the program. Legislating an accounting gimmick or a short term improvement to push out doomsday by 5 or 10 years is, frankly, the optimistic scenario here. I am forced to conclude that what is needed is less a radical proposal for reform but rather a radical proposal for how to successfully achieve meaningfully reform. I offer one such proposal.
A Brief History
To understand why an unorthodox ballot question approach would even be considered, it is helpful to review the dismal history of passing legislation on this subject.
The last round of reform to shore up the system was the Social Security Amendments Act of 1983, and, sadly, this Reagan-era legislation is the only example of Congress assembling a majority in support of fiscal prudence in the entire history of the program. All of the prior amendments (save for a 1977 change to fix a costly math blunder) were focused on handing out bigger checks to more people. The 1983 changes averted an immediate crisis (it appeared the program would run out of money within the year) and ensured the trust fund would build for decades.
Interestingly, this bipartisan success did not inspire lawmakers to cooperate on further amendments over the next forty years, even though it was well recognized by the nineties that the retirement of the baby boom generation would eventually send the program into deficit again. A deficit from which it would not recover, even under its reformed structure . Instead, Social Security started to be labeled the "third rail" of American politics. Ignoring mere campaign proposals (easily jettisoned upon obtaining office), reform-related activity under recent presidential administrations can be summarized as follows:
Clinton: Together with Newt Gingrich, a secret agreement to reform Social Security was supposedly reached but never announced1. The Lewinsky scandal and its fallout derailed bi-partisan cooperation on any subject.
Bush: The President invested several months of political capital into a high-level proposal centered around creating private savings accounts. This never got past the podium speech stage. Republicans did not fully back the ideas, and Democrats strongly opposed private accounts. No legislation was introduced.
Obama: His National Commission on Fiscal Responsibility and Reform devised recommendations to increase the payroll tax and reduce benefits in a variety of ways. Within the commission itself; however, the vote fell short to send the recommendations to Congress. The tax increases cost support among Republicans, and the benefit reductions were opposed by Democrats (and progressive critics more generally).
Trump: He studiously avoided proposing any changes to Social Security.
Biden: The President has gone quiet on Social Security since taking office.
Note that the momentum toward reform has actually declined over the past decade. Not for lack of proposals. A slew of reports and books have been published on Social Security from every part of the political spectrum. Legislators have gone so far as to independently draft bills and gather sponsors. But no majority consensus has coalesced around anything, and political leadership in every branch of the system is disincentivized to negotiate.
At its core, Social Security is actually not a very partisan program, which may be the heart of the problem. Fixing its deficit does not animate either party's base. As a result, even during periods of one-party control across both houses of congress and the presidency, the issue is not a legislative priority. If you aim for the safety of a bipartisan solution, it quickly becomes obvious that this requires an unpalatable mix of tax increases and benefit cuts — just as was done in 1983. So, politicians have preferred to bandy about proposals for transformative change that play to their supporters and donors but have no chance of passage. For Republicans, a common rabbit hole has been privatization of some form. Democrats keep plugging plans to turn Social Security into a pure welfare program by scrapping the payroll tax cap, among other progressive measures.
Besides sinking any hope of bipartisan support, both of these approaches are shortsighted. I elaborate on their flaws further down. For now, let's put aside dramatic solutions to replace, reconceive or expand Social Security, and, instead, focus on how to achieve a reform that will merely balance long term revenue and expenses under its current model.
Options
The math behind funding Social Security is pleasantly straightforward. (This is in contrast to Medicare, where the spending side of the equation is much more complicated and cannot be legislated as directly). There are a finite number of well understood levers that you can pull to change the long term trajectory of the trust fund.
Helpfully, many think tanks who research policy in this area have written detailed summaries of every possible change and its estimated impact, often complete with an online calculator! The Committee for a Responsible Federal Budget (CRFB) has one such calculator2, last updated in 2021, which I have leveraged here for illustrative purposes. But, any one of them should give similar results. There are five major categories of change that could have a large impact on the program's finances.
Increase the payroll tax. It would require an across-the-board increase on the order of 3.3%, split between employees and employers, to close the funding gap completely. (Note that this figure and all figures below are as of 2021; required changes would be a bit higher now.)
Raise the payroll tax cap (with or without any additional benefits paid out for the additional payments). See below for why this is a strategic blunder.
Grow starting benefits more slowly. Growing benefits with prices rather than wages is proposed to be one method to accomplish this; although, I do wonder if this would hold true in a period of sustained high inflation. A more direct change is to alter the wage replacement multipliers used in the benefits formula. To close the gap completely requires an across the board reduction of 26% in initial benefit amounts. The CFRB estimated that if you reduced multipliers for only the top half of earners, along the lines of what was recommended by the Simpson-Bowles Commission, you could close just over a third of the gap.
Modify cost of living adjustments. Adopting a chained CPI should reduce the adjustments, closing perhaps a fifth of the gap. Means-testing COLAs to eliminate them for higher income earners would also make a dent.
Raising the starting age and/or indexing it to life expectancy. Raising the age by just one year, from 67 to 68, is estimated by the CRFB to close 13% of the gap. Or the age could be indexed to life expectancy in some way. For example, keeping a constant ratio over time of expected retirement years to potential work years could close the gap by 20%. A reasonable concern is that average longevity measures may favor wealthier retirees while low-income seniors experience lower increases in lifespan (or even go in reverse). Indexing to the expected retirement years of the bottom half of earners may allay some concerns (but may also blunt the impact of the change).
There are other, minor, revenue generators that could be adopted, such as forcing all newly hired state and local workers to participate in the system (some are currently exempt). Similarly, there are tweaks in benefits to consider, such as calculating worker benefits on the average earnings of their top 38 earning years rather than the top 35. Some of these changes seem sensible and might even garner broad, bipartisan support. Others are likely to be no more popular than the major change options (e.g., no longer having health insurance and various other savings plans deductible against the payroll tax). However, even if you adopted every one of the minor measures identified by the CRFB, their online calculator estimates that you cannot close more than half of the long term funding gap. One or more of the five major changes listed above must be employed.
A compromise touching on both raising more revenue and trimming the growth of benefits is likely most fair and acceptable to the public. But, what is the right mix? Therein lies the unpleasant trade-off that bedevils any attempt to negotiate and pass legislation. Perhaps, the citizenry can help.
The Vote
Imagine a nationwide ballot question added to the next federal election cycle which asked voters a question similar to this one:
Today, Social Security pays out more in benefits than the program collects in revenue. This shortfall is being covered by reserves (also known as trust funds), and these reserves will be exhausted within 10 years, at which time scheduled benefits will no longer be paid out in full.
Which change to the Social Security program do you MOST support to achieve long term fiscal balance for the program? Long term fiscal balance is defined as total revenue collected will be sufficient to pay scheduled benefits for the next 75 years as projected by the Social Security and Medicare Board of Trustees.
A. I most support increasing the Social Security Tax on earnings (also known as FICA payroll tax) to achieve balance
OR
B. I most support slowing the growth of benefits paid by equally applying the list of measures below to achieve balance:
Index cost-of-living adjustments (COLAs) to a different calculation of inflation that is projected to grow benefits more slowly.
Increase the retirement age over time.
Slow the growth of initial benefit amounts by phasing in lower wage multipliers.
What would make this ballot question novel is that the outcome would not be determined by which option, A or B, received the most votes. Rather, I propose that the funding gap would be closed in exact proportion to the votes cast. If 40% of voters choose Option A then 40% of the gap will be closed through a payroll tax increase and 60% of the gap will be closed through trimming benefits. Here is a sketch of the approximate outcomes for various hypothetical vote results, based on the CRFB's 2021-era Reformer tool for fixing social security.
This "proportional outcome" approach has a number of advantages.
First, it solves the problem no matter what the outcome of the vote. Every result leads to the funding gap being closed, one way or another3.
There is no winning or losing side. Instead, every vote matters. There will be at least some votes cast for each option, and voters' preferences will be honored in proportion to their number. This makes it pretty motivating to vote, even if you expect your choice is in the minority.
The question forces the public to grapple directly with a tradeoff and accept that there is no "free lunch". (Yes, the electorate includes seniors and others with no expectation of future earned income who can, and mostly will, vote to stick it to the next generation with Option A. This is unavoidable. But, the majority of voters are personally impacted by both payroll tax increases and reduced benefits. This latter characteristic is critical to the legitimacy of a direct democracy approach.)
No one in congress needs to agree on what the right mix of tax increases and benefits cuts should be. Politicians would not even need to take a position on which way to vote. They just need to support that a vote is held, and that their constituents should go to the polls to make their voice heard.
What is the likely outcome? I expect seniors to overwhelmingly vote for increased revenue to avoid lower COLAs. Some higher income seniors may choose benefit cuts, especially if they have other FICA taxable earnings from employment. Those near retirement age would likely also vote strongly for increasing the tax. As voters get younger, though, there will be more support for trimming benefits. My gut instinct is that at least half will vote to increase the tax rate; perhaps as many as two thirds of voters. That level of increase, applied to the broad taxable wage base, remains manageable.
I am aware that a large hurdle is finding a legal basis for a federal ballot measure originating from a legislative referral. As I understand it, there are insurmountable constitutional problems with making federal law via direct democracy. However, in this case, the ballot question does not need to put legislation itself to a vote. Rather, Congress would pass a bill in the ordinary manner that amends the Social Security Act with updated formulas for the tax rate, COLA, retirement age, initial benefits, etc. The formulas in the enacted statute would refer to the results of a consultative ballot question in the calculations (funded in the same amendment) in the same way that legislation might specify that a formula refer to a specific Consumer Price Index rather than a fixed number.
Ideally, there should be no need to pass further legislation post-ballot measure. Not would it be critical to require that every single state has put the question on a ballot to consider the outcome binding on the calculations. The results are not likely to differ wildly between states (this is not a "red state vs. blue state" question). So, if a few somehow failed to execute, then there would be little impact. If some minimum threshold of participation in the ballot measure was not met by a prescribed date, then the legislation could simply default to a 50:50 tax increase to benefit reduction outcome.
But We Should Scrap the Cap!
At this point any reader of a more progressive bent will be shouting, red-faced, at their screen, "You have dismissed the best solution of all! All we need to do is remove the income cap (currently at $147K) above which the Social Security Tax is no longer applied." This could close much (though not all) of the funding gap, especially if the additional income is not counted toward any benefits (Social Security truly becomes a welfare program). Yet, I must reiterate, this is a shortsighted idea.
The most common objection voiced to taxing additional income, with no, or limited, benefits paid out in return, is that it breaks the ethos on which the program has been sold to the public from day one. The benefits formula is already highly progressive4. By destroying the relationship at the top between what is paid in and what is paid out, you end any lingering notion that this is a forced savings or insurance program. It becomes more broadly understood as charity, and support from the upper professional class, a powerful minority constituency (and the program's new, mostly unwilling, benefactors in this instance), is likely to erode. There is also the practical consideration that such a tax increase may bring in less revenue than anticipated, as sophisticated high income earners respond by shifting their compensation into untaxed categories.
Yet, there is actually another, broader objection. Even if you fervently believe that high income earners can, and should, pay more taxes in this country than they do at present, Social Security is certainly the wrong part of the budget to apply any additional revenue you can squeeze from them.
Understand that an additional 12.4% on all income over $147K (or over $250K or $400K or any other threshold you choose) is a very large tax increase for these high earners, to put it mildly. Even under the most rosy narrative, where you assume the employer half is not coming out of the employee's pocket (not many folks in these brackets are going to fall for that fiction), the top federal tax bracket in the US would jump from a marginal rate of 39.35% (37% income tax + 2.35% Medicare) to 45.55%. Should this high income earner reside in California, they are looking at a top rate of 59.95%. The picture for the self-employed is even bleaker; they have to cough up the employer match too.
This is approaching Swedish levels of top bracket taxation and for far, far, fewer benefits5. Welcome to the top of the Laffer curve! Perhaps, an increase of this magnitude can be pushed through. But, everyone should be clear: after this point, we will be wringing no additional revenue from this bunch. And therein lies the fatal flaw with this approach.
If making Social Security solvent was the only fiscal hole that the federal government needed to plug, and no ideas for new programs and spending were afloat, then you might get away with scrapping the Social Security cap. Sadly, no description of the fiscal situation of the government of the United States could be further from reality.
Where will the additional revenue come from to pay for the deficits of decades past, now stalking us as growing interest payments on 33 trillion dollars (and counting) of built-up debt? Raising taxes on those same high income earners (and big business), of course. Where will the additional revenue come from to pay for the deficit of the present, which, even after you fix Social Security, will still be large and structural? The plan seems to be the rich and big business again. Where will the additional revenue come from to pay for the deficits of the future? Those expansions of entitlement benefits, wars, foreign aid, infrastructure investments and what not that so many legislators clamor for daily? Is there some answer other than the rich and big business? Inflation, maybe.
The notion that higher taxes on the top 1% or 5% alone can resolve the massive shortfalls in US budgets past, present and future is delusional. Any hope of moderating the growth of US debt will require pulling additional revenue out of the full swath of the tax base — top to bottom. It will also require very substantial spending cuts. You may choose to disproportionately put more burden on higher income earners, but the middle class cannot escape tax increases entirely. Helpfully, the US middle class is currently a bit lower taxed in comparison to other high income countries6. But, few believe it, and fewer care. The American middle class is as resistant to tax increases as you will find in any jurisdiction. Maybe more so! It is how they have kept a lid on their relative tax burden in the first place.
Once you accept the mathematical reality that a tax increase for the American middle cometh, it becomes a strategic question of what plausible rationale you can offer to this politically crucial demographic for why their taxes are going up. What will most contain the political fallout?
Raising the payroll tax for everyone will be very unpopular, to be sure, even if the citizenry did vote for it! But payroll taxes are at least linked to tangible and transparent individual benefits. Consider the alternative proposals you might be forced to pitch to extract the additional revenue you need out of John and Jane Q. Public. Will raising their income taxes to go into the black hole of the general budget be more popular? Or, maybe you would like to try selling a national sales tax? To foresee the electoral future of any American politician who wanders down the latter route, venture north to visit former Canadian prime minister Brian Mulroney and utter the letters G-S-T. Heck, I will save you the trip. The political consequences he will describe to you can be summarized in a single word: Martyrdom.
A payroll tax increase directed explicitly towards Social Security is actually one of the least unattractive places to increase taxes on the middle class. I would seize it. You will likely need those tax increases on the rich for funding the far less popular components of federal spending growth: most urgently, interest on the ballooning debt.
The Privatization Path
It is not worth dwelling long on plans to replace Social Security in whole, or in part, with individual, private, retirement accounts. It has little traction in the Republican Party these days, and Donald Trump, today's leading presidential candidate, has a decades-long history of opposition to privatization of the program. In the short to medium term, private accounts do not solve Social Security's fiscal imbalance and may even exacerbate it. Then there is the sticky question of what happens to workers whose private investment choices do poorly. There will be political pressure to make them at least partially whole, creating the terrible incentives and high cost of any system with private gains and public losses. For fiscal conservatives, just maintaining the current program on solid financial footing without it turning a welfare program or expanding in other directions should be considered success.
A less radical reform would be to allow the trust fund itself to invest money outside of Treasuries. This change, alone, cannot resolve the funding shortfall at this late hour (after all, there are no new excess contributions to invest these days). But, in theory, it should reduce the magnitude of the tax increases and/or benefit cuts required to restore the program to the black for the long term. This is not a wacky idea. Many other countries manage their social security funds in just this way (e.g., Canada made this change in the late nineties with the creation of the Canada Pension Plan), and it has brought them superior returns. Why not the US?
Admittedly, the last thing the US government wants right now is fewer buyers of Treasuries. But, as the trust fund dwindles to nothing, it is an opportune time to move the Social Security program out of the general budget. Once a shored up program begins to rebuild savings, keeping the surplus out of the hands of Congress will more explicitly expose continued, large, deficits in all of the remaining spending categories (hopefully applying some pressure for better fiscal discipline). This accounting advantage is real even if all the new trust fund did was continue to invest in public issue US Treasuries. However, the stage would be set for broader investment strategies.
One model is a broad portfolio like that of the Canada Pension Plan (CPP) or CalPERS: investments around the world in stocks, bonds, real estate, infrastructure and private equity. The potential is easy to see. Since its inception over twenty years ago, the CPP has posted a net nominal annual return of 7.2%. This is a large improvement over mere government bonds. Had the US Social Security Trust Fund experienced comparable returns over the same period, we would be looking at a very different fiscal situation for the program today. But, lest we forget, higher return always comes with higher risk.
To illustrate, consider that the Canada Pension Plan is actually one of two, massive, government pension funds in Canada that are externally invested. The other being the Caisse de Dépôt et Placement du Québec (CDPQ). In the midst of the financial crisis, the CDPQ closed the year 2008 with an eye-watering loss of 25% of the fund's assets. It was a vivid reminder of the volatility of market investments and how even seemingly safe bets can go bad (they experienced huge losses in asset backed commercial paper). It also exposes how management skill, governance and luck will drive differential performance of otherwise similar funds. In 2008 the CPP lost only 13%, and has reported higher returns than the CDPQ nearly every year since.
The woes of the CDPQ should not be exaggerated. Measured in decades, the fund's performance remains very positive. Indeed, it would take extraordinary incompetence for a professionally managed, diversified, pension fund of this nature to do worse than Treasuries in the long run. However, the tiny risk of catastrophic losses is not the only concern raised by external investment.
Granting the Social Security Trust Fund investment discretion does not guarantee decision-making based purely on balancing risk with return. Partisan, ideological and plain-old, political survival instincts of the new fund managers could manifest in various ways. They may be persuaded to weigh ESG-like metrics, to divest from profitable companies or industries under activist pressure (e.g., fossil fuels), to be mindful of spreading the wealth (is every state getting equal investment?) or to employ DEI criteria in hiring investment managers. Investment in domestic infrastructure projects is a category that would seem particularly susceptible to political pressure.
Then there are the potential landmines found in foreign investment. At one point the CPP, as a small example, found itself facing uncomfortable questions in the Canadian Parliament for its investments in Chinese manufacturers of surveillance equipment supplied to the Chinese government for use in Xinjiang (home to most of the Uyghur population). As an arm of a global power, I expect a US fund to face greater controversy and pressure while investing abroad. The US Congress, Executive and usual list of three letter agencies are all possible sources of meddling to exploit foreign investment as a tool of foreign policy.
The sheer size of the trust fund could also be a challenge. At its height the Social Security trust fund held $3T dollars, exceeding the size of the largest pension or sovereign wealth funds in the world today. It is being rapidly depleted, of course, but under a reform scenario it is likely to regain a multi-trillion dollar balance. Such a behemoth of a fund can move markets and would require exceptionally careful governance.
Ultimately, the tax and/or benefit formulas underlying Social Security need reform regardless of whether the Trust fund is allowed to invest outside of Treasuries to increase returns. If this latter change is also enacted, it seems prudent to be conservative in its implementation. Allowing a portion of the trust fund to be invested in a diversified portfolio of domestic bonds held to maturity, for example, would be low risk and offer a modest improvement in fund performance. If nothing else, I would completely bar foreign investment.
The existence and details of this negotiation appear to originate from one or more books. I have seen the "The Pact: Bill Clinton, Newt Gingrich and the Rivalry that Defined a Generation" by Steven Gillon cited in support of this assertion. Since this history is rather peripheral to this article, I have not read the book to confirm this detail.
The online calculator can be found at: https://www.crfb.org/socialsecurityreformer/.
Whether the funding gap will really be solved for 75 years is uncertain, since it is dependent on the accuracy of the projections of the Trustees. I have not waded too deeply into their very detailed report outlining the assumptions behind their projections. But, even a brief perusal of the assumptions reveals some that seem a tad optimistic. For example, "birth rates are assumed to return to a level of 2 children per woman for 2056 and thereafter". Um, why? And we can only have the lowest confidence in anyone's GDP estimates for the year 2097. It may be that a funding gap re-emerges in 40 years. This will still represent a large improvement over the status quo.
The CBO once calculated that "the ratio of lifetime benefits to lifetime taxes for the bottom quintile [of income earners] is nearly double that for the middle quintile, which is in turn about 40 percent higher than for the top quintile." See https://www.cbo.gov/publication/18266.
Sweden's top marginal rate in its most heavily taxed municipality is 55.15% (20% national income tax + 35.15% local income tax). The employee social security contribution is capped at a certain income threshold, as it is in the US, so it does not impact the top marginal rate. The current cap is equivalent to only $55K. Sweden does collect a much higher payroll tax from employers of 31.42%, which leads to an overall marginal tax wedge on high income labor that is quite high. Thus, if you consider the employer taxes as well, then the top marginal tax wedge in Sweden becomes 66% versus 63% in California (were the US federal payroll tax cap to be removed entirely). However, Swedish payroll taxes include full health insurance, and there is effectively no property tax in Sweden nor any estate tax. So, I submit that it is not clear that high income earners would face lower taxes in California than in Sweden, under this scenario. The real difference in taxation in Sweden is that the structure is very flat. Workers hit the top marginal income tax rate at a much, much lower income than in the US. The OECD details the tax system in Sweden here: https://www.oecd-ilibrary.org/sites/3aebdb4e-en/index.html?itemId=/content/component/3aebdb4e-en.
The average worker in the US faces a net tax rate that is about average for the OECD. Considered as an overall tax wedge on labor income that includes the employer taxes paid (a more accurate measure of the real burden of taxation), the average US worker is below the average for the OECD. In general, the entire Anglosphere clusters below the EU on this measure, along with various other non-EU countries, such as Israel, Costa Rica, Switzerland and Korea. See https://www.oecd.org/tax/taxing-wages-20725124.htm.
Congrats! This is the most creative suggestion I have ever seen, although it obviously won't be implemented.
However, I disagree about not lifting the cap. You don't have to destroy the relationship at the top between what is paid in and what is paid out - instead, you could couple lifting the cap with abolishing the cap on benefits and add a third bend point to the PIA formula at the current maximum.
Something along those lines:
Current formula: 90 percent of the first $1,174 of his/her average indexed monthly earnings, plus 32 percent of his/her average indexed monthly earnings over $1,174 and through $7,078, plus 15 percent of his/her average indexed monthly earnings over $7,078, maximum benefit $4,873 per month.
New formula: the same+(0.2*X) percent of average indexed monthly earnings above $14,050 (current cap), where X=years since lifting the cap, maximum 8% (40 years after lifting the cap - people who retire probably paid payroll taxes above the cap all the time), no cap at benefits.
Obviously such a raise wouldn't need to happen immediately (that would be too much of a shock), but gradually (e.g. raised 1% per year), and it could be complimented with general tax cut&introducing VAT tax.