Report | Older workers

Steady contributions, affordability, and lifetime income are the building blocks of a retirement system that works for working families: Expanding Social Security is the most important step

Download PDF


Our retirement system is broken. The shift from traditional defined benefit (DB) pensions to 401(k)-style defined contribution (DC) plans was an experiment that failed, widening the gap between retirement haves and have-nots. Families at the 90th percentile in the savings distribution had $320,000 or more saved in retirement accounts in 2016, more than triple what they had in 1989. Meanwhile, the median family (at the 50th percentile) had only $7,800 in 2016, and has lost ground since the Great Recession (Morrissey 2019b, chart 7).

Roughly half of private-sector workers do not participate in employer retirement plans, in most cases because they have no choice (BLS 2018, 191). Though many people will participate in employer-based plans at some point in their lives, spotty contributions to DC plans—combined with high fees and other leakages from these plans—means that half of U.S. households are still likely to see a sharp decline in their standard of living after retiring, especially since Social Security benefits have also been cut back (Munnell, Hou, and Sanzenbacher 2018).

It is not too late to prevent this looming retirement crisis. To do so, we need to expand the part of the system that works—Social Security—and fix the one that does not: 401(k)-style plans. We also need to protect remaining DB pensions, the only type of employer-based plan that shields workers from investment risks. Finally, we need to ensure that the system is fair and government subsidies help those who need it rather than magnifying wealth inequality.

The goal of a retirement system is to smooth living standards over people’s lifetimes, especially at older ages when earning a living may no longer be an option. This requires steady contributions even before people are focused on retirement; minimizing leakages in the form of fees and preretirement withdrawals; and ensuring that retirees do not run out of money no matter how long they live.

The 401(k) system fails on all three counts. Too little money is going into 401(k) plans, mainly because many workers do not have the option to participate in a plan at work. And money leaks out before retirement in the form of high fees, loans, and withdrawals. Finally, these plans do not pool risk, so contributions need to be much higher to ensure that participants do not outlive their savings.

In contrast, Social Security has all the features of an effective system. Contributions are automatic, funded largely through payroll taxes that are split equally between employers and employees (self-employed workers pay both shares). Administrative costs are low, and participants cannot access retirement benefits before age 62. Benefits last as long as beneficiaries are alive.

But Social Security replaces only around 40% of preretirement income for the average worker—roughly half of what most probably need. In theory, workers could make up the difference with employer retirement benefits and private savings, but the typical household has little accumulated in retirement plans or in other financial assets (Morrissey 2016a; Morrissey 2019a).

Fixing this system will not be easy. There are important questions to be answered—about what to fix and what to scrap, the proper role of government, and how to navigate a shifting political landscape. Otto von Bismarck, the German statesman often credited with being the father of social insurance, famously said, “politics is the art of the possible” (Anderson 2016; Steinberg 2011). Politics will determine whether we can expand Social Security enough to make employer plans redundant, or if we need to combine Social Security expansion with a plan to fix our patchwork employer system. We do know that as long as we continue to rely on a tiered system that assumes Social Security will be supplemented with employer retirement benefits, we need to ensure that employers actually offer such plans and that these plans are managed prudently and economically. What will not work is simply encouraging people to save more in high-cost, high-risk plans.

Most reform efforts focus on encouraging people to save more rather than on fixing a broken system, placing the blame onto workers themselves. These misguided reform efforts do not confront the real problems in our retirement system because they are designed to avoid antagonizing those who benefit from the status quo: employers who contribute little toward their workers’ retirement, wealthy taxpayers who benefit from upside-down tax subsidies, and companies that charge high fees to administer plans and manage investments.

This report outlines the Economic Policy Institute’s multipronged approach for advancing retirement security, consistent with our history and mission of advancing policies that help working families:

  • Expand Social Security. Social Security is the heart of the American retirement system and should be expanded. It is efficient and provides guaranteed income to workers in retirement. However, compared with other rich nations’ public retirement systems, it is relatively stingy. The single most effective way to boost American retirement security would be to build on Social Security’s success and increase benefits.
  • Address failures of the 401(k) system with a GRA plan. DC plans, such as 401(k)s, are a failed policy experiment. The Guaranteed Retirement Account (GRA) plan offers an alternative model. GRAs are simple, low-cost, portable retirement accounts to which employers and employees each contribute at least 1.5% of pay. Employee contributions are offset by a $600 tax credit that make the plan affordable for lower-income families. GRAs could fill any gaps that remain after we have expanded Social Security as much as political limits allow, but should not be a substitute for Social Security expansion.
  • Defend traditional pensions. Traditional defined benefit pensions work for those workers lucky enough to have them. We should defend DB pensions against evidence-free, ideological attacks launched against them, especially attacks on pensions of teachers and other public-sector employees whose employers are well suited to take on long-term liabilities.

Introduction: How did our retirement system get so broken? American exceptionalism

The United States is an outlier among Western industrialized countries in its heavy reliance on employers for the provision of basic needs, including health insurance and old age pensions, in contrast with Western Europe and Canada where government social insurance programs play a larger role. This aspect of American exceptionalism, which political scientist Jacob Hacker has dubbed the “divided welfare state,” has many roots (Hacker 2002). It was essentially enshrined in the 1950 Treaty of Detroit, when the United Auto Workers (UAW) struck a deal with General Motors (GM) to deliver middle-class wages, pensions, and health benefits to workers—either directly through the GM–UAW and subsequent collective bargaining agreements or more broadly through the workplace standards those agreements helped establish.

This postwar bargain ushered in a quarter century of prosperity that was broadly shared across the income distribution. But the idea that corporations could effectively deliver large-scale social insurance may have sapped some of the political energy for more universal public programs, as Frances Perkins—President Franklin D. Roosevelt’s labor secretary and the architect of Social Security and other New Deal programs—had feared (Downey 2009). Nevertheless, for a time the employer-based welfare state seemed complementary with growth in public plans, as the UAW and other unions played an important role in marshalling support for government social insurance programs (UAW 2019). These programs continued to expand in the postwar decades, albeit more slowly than in Europe and Canada.

But from a 21st century vantage, the postwar bargain looks like a bad deal for American workers. Social Security, Medicare, and Medicaid still cover a smaller share of health and retirement costs in the United States than in most industrialized countries (Hacker 2002), and the United States still relies heavily on employers to provide basic retirement and health benefits. In theory, an employer-based system in combination with public social insurance could provide comprehensive coverage. In practice, it does not—not least because many employers do not provide these benefits in the first place. Employers have also proven less effective than government at restraining costs, and they have shifted an increasing share of these costs and risks onto workers (Hacker 2006).

The retirement security problem is big—and needs solutions to match

EPI’s policy mantra has always been to propose solutions equal to the scale of the problem, whether or not they are politically viable in the near term, while supporting some incremental reforms that move us in the right direction. There is, however, an inherent tension between pragmatism and “thinking big.” On the one hand, progress depends on acknowledging political constraints and recognizing that we are not designing systems from scratch. On the other hand, an incremental approach can serve to entrench a broken system and delay needed reforms.

EPI’s role in the health care reform debate waged in the run-up to the Obama administration offers an analogy. In this debate, EPI acknowledged that while a European- or Canadian-style “single payer” system was most efficient, it would require upending the current employer-based system. To balance these practical and political concerns, EPI partnered with Yale political scientist Jacob Hacker to propose a public option with an insurance mandate. The assumption was that the public option would eventually out-compete most private plans while minimizing the perceived disruption to the health insurance many working Americans already had.

Even though the public option was dropped from the Affordable Care Act (ACA) before its enactment during the Obama administration, EPI continued to support the ACA as an improvement over the status quo. Targeted subsidies made coverage affordable for many low- and moderate-income families, and the Medicaid expansions were a lifeline for some of the most vulnerable families. However, the ACA was less effective at restraining costs than it was at expanding coverage (Broaddus and Park 2016), largely because it did not take advantage of the bargaining power of large public plans.

EPI’s position in the retirement security debate is similar to our support for a public option during the health care debate. We recommend expanding an efficient government program—Social Security—to the greatest extent that is politically possible, while filling gaps in employer-based retirement coverage by combining an employer and employee mandate with progressive subsidies.

EPI’s multipronged approach meets workers where they are

EPI has a long history of engaging on retirement issues and pushing for big solutions. EPI fought back against President George W. Bush’s short-lived plan for Social Security privatization as well as a more persistent Beltway misconception that looming deficits (which are primarily driven by escalating health costs, not aging) make Social Security cuts unavoidable.

Meanwhile, EPI was an early and influential critic of 401(k) plans, using EPI board member Teresa Ghilarducci’s Guaranteed Retirement Account plan to illustrate how all workers could attain retirement security at no greater cost to the government than current subsidies to retirement plans in the form of preferential tax treatment.

The GRA plan opened the public discourse to ambitious reform efforts at a time of bipartisan infatuation with automatic enrollment as a way to “nudge” people into saving in 401(k)s and IRAs. Later, working with allies, EPI widened the range of talked-about policy options to include Social Security expansion. All the while, EPI has defended public DB pensions against critics intent on stoking “pension envy” and redefining retirement for all workers to mean account-style plans.

EPI’s multipronged approach—expanding Social Security, requiring modest employer and employee contributions into a GRA or other employer-based plan, and protecting DB pensions—recognizes that half of workers already participate in employer plans. Virtually all workers, even those with decent employer plans, would benefit from expanding Social Security benefits, which is why this is EPI’s primary focus. Beyond a certain point, however, closing the gap between workers who do and do not have employer plans may be easier with a GRA-type plan, or similar state plans, unless the goal is to replace employer plans altogether.

Social Security is the heart of the American retirement system and should be expanded

Social Security is efficient and provides guaranteed income to workers in retirement. The single most effective way to boost American retirement security would be to build on Social Security’s success and expand it. Most Americans across the political spectrum strongly support Social Security and are willing to contribute more to strengthen it, despite attempts by small-government ideologues to portray the program as unaffordable and to weaken it over the years—attempts that EPI has vigorously fought.

Social insurance has many advantages over private insurance

Market competition can spur innovation and lower costs for many goods and services. But not all markets are equally competitive, and there are advantages to government-administered or highly regulated systems when it comes to insuring people against major risks, including health and longevity risks (the risk of living longer than average and outliving one’s income).

It is difficult to predict whether we or our loved ones will develop expensive medical conditions or live to 100. Without social insurance, we would struggle to save enough to cover such expenses. Though private insurance can effectively protect against risks such as car theft that have few societal repercussions, relying on people to voluntarily purchase private insurance is problematic when risks are systemic or impose costs on others (what economists refer to as “negative externalities”).

To illustrate, medical advances resulting in rapid increases in life expectancy could cause private companies selling lifetime annuities to fail—a systemic risk if many companies collapse at the same time. If many people were counting on these annuities for retirement, this in turn would likely cause safety net spending on seniors to increase—imposing costs on taxpayers—among other economic consequences.

Social insurance can address both systemic risks and negative externalities. Though unexpected increases in life expectancy at older ages also have a negative impact on Social Security’s finances, the system’s collapse can be avoided because the government has the ability to increase taxes or trim benefits. Meanwhile, Social Security’s progressive benefit structure reduces poverty and spending on safety net programs.

Problems also arise because higher-risk individuals are more likely to participate in voluntary insurance schemes, driving up costs and shrinking coverage. Though group plans—such as employer-based health insurance and DB pensions—partly address this “adverse selection” problem, group plans need to be regulated to deter discrimination against higher-risk individuals, such as older workers.

In the retirement context, adverse selection may partly explain why few people convert 401(k) balances to lifetime income by purchasing annuities that insure them against outliving their savings. The small market for life annuities is also due to loss aversion and other psychological quirks, and the fact that purchasing an annuity is a complex transaction in which consumers find themselves at an informational disadvantage. High-pressure sales tactics masquerading as financial advice also result in bad products crowding out good ones and shrinking the overall market. Problems in the private market for life annuities are so glaring that even some Republicans who normally oppose expanding Social Security, such as Jason Fichtner, a former George W. Bush appointee to the Social Security Administration, nevertheless support converting retirement account balances into expanded Social Security benefits in order to make these savings last through retirement (Koenig, Fichtner, and Gale 2018).

Regulation of for-profit insurers is also needed because these insurers have an incentive to deny claims or take a “heads we win, tails everyone loses” approach to risks.1 These perverse incentives lead to duplicative bureaucracies at best (because the government must regulate private insurers) and predatory practices at worst (when private insurers are poorly regulated).

Another potential drawback of private insurance is its effect on the labor market. While employer-sponsored plans help employers attract and retain workers, tying benefits to particular employers can lead to “job lock” when workers who would start their own businesses or move to jobs that are a better fit for their talents and interests do not because they do not want to give up their benefits. Social insurance solves this problem by covering everyone in the population or in a target demographic. This also creates economies of scale by spreading administrative and other fixed costs over large populations.

With social insurance, costs can also be spread across individual lifetimes and generational cohorts. Thus, for example, while Social Security is mostly a pay-as-you-go system, with worker contributions going directly to retirees and other beneficiaries, a trust fund shields beneficiaries from transitory fluctuations in payroll tax receipts caused by business cycles and generational booms and busts.

That is not to say that there are no potential advantages to private insurance. In the retirement arena, for example, there are legitimate concerns about large government entities playing an active role in financial markets—or, conversely, playing a passive role and weakening the oversight function of institutional investors. Thus, large government-administered retirement plans—such as EPI’s proposed Guaranteed Retirement Account plan or similar state plans—would likely have to privatize or decentralize investment management, at least once the plans matured.

Social Security is efficient and remains popular despite constant attacks

There are clear efficiency advantages to Social Security. By covering nearly all private-sector workers, Social Security minimizes costs and risks through economies of scale and the broadest possible risk pooling. Costs are also spread over participants’ working years. While Social Security payroll taxes are somewhat regressive (earnings above $132,900 are not taxed), benefits are progressive, with lower-income earners receiving a higher share of their preretirement earnings as benefits. Social Security is also an extraordinarily popular program in a country that tends to be skeptical of government solutions.

This popularity has endured despite a decades-long misinformation campaign. A key part of the campaign to erode public support for Social Security is an attempt to convince young people that it will not be there for them when they retire, an effective line of attack first laid out in a 1983 article in the libertarian Cato Journal (Butler and Germanis 1983). Crucially, these ideological attacks have been couched as concerns about the sustainability of the program, though Social Security’s financial challenges are hardly insurmountable. An increase in the employer and employee contribution rates from 6.2% to 7.6% of taxable payroll would put the system in long-term balance (Social Security Trustees 2019, 8). As much as three-fourths of the long-term shortfall could be eliminated simply by “scrapping the cap”—subjecting all earnings to Social Security taxes while reducing the benefit multiplier for earnings above the cap on taxable earnings (Social Security Actuary 2019, 22).

EPI resisted the Beltway consensus that Social Security’s long-term shortfall was a looming crisis that should be closed wholly or largely by reducing benefits. Benefits have already been declining with the gradual increase in the retirement age and the taxation of benefits for higher-income beneficiaries—cuts enacted in 1983 when the system was actually facing an imminent shortfall. EPI has estimated that these incremental cuts will amount to a 22% reduction in benefits for the average Gen Xer and millennial.2

EPI also resisted President George W. Bush’s plan to carve private accounts out of Social Security—a bad idea that EPI had been battling for years before it became the centerpiece of the president’s post-reelection policy agenda in 2005. Among other contributions to the debate, EPI highlighted transition costs, critiqued inflated rate-of-return assumptions for private accounts, and disproved claims that black men are shortchanged by Social Security because of their shorter life expectancies (EPI 2019).

Social Security’s internal rate of return is not inherently lower than that of private plans

EPI also combated a common misconception that 401(k)s provide better returns than Social Security. With both 401(k)s and Social Security, the return on contributions reflects economic growth. Social Security benefits are mostly paid directly from worker and employer contributions. Only a small portion of Social Security benefits are paid for by tapping the trust fund that was built up to pay for the large Baby Boomer cohort.

Since the average worker is decades younger than the average beneficiary, a fixed contribution rate in a pay-as-you-go system results in an implicit, or “internal,” rate of return equal to productivity and workforce growth, minus any legacy costs (Morrissey 2018a; Munnell, Hou, and Sanzenbacher 2017).

In the case of Social Security, these legacy costs are the cost of providing the first generation of recipients with benefits even though they had not had the opportunity to contribute much to the system. Given that this first generation lived through the Great Depression, the decision to extend them economic aid that would be paid for by future, richer generations seems a just one. If we are worried that these legacy costs are a drag on current participants’ internal rate of return on contributions, we could pay for these costs separately. For example, we could use dedicated estate tax revenues to pay for these costs, as proposed by Senator Chris Van Hollen (D-Md.) (Scott 2019).

Social Security’s cost-of-living adjustments reflect rising costs faced by beneficiaries

EPI has also opposed the idea of using a chained consumer price index to calculate Social Security’s cost-of-living adjustment (COLA), a proposal often advanced in the context of deficit reduction. While a chained consumer price index (CPI) more fully accounts for consumers’ ability to substitute cheaper goods and services, it increases more slowly than an unchained index and thus would reduce Social Security benefits over time.

As EPI and others have pointed out, both the chained CPI-U and the unchained CPI-W (the current index used for determining the COLA) understate inflation faced by seniors and disabled beneficiaries because beneficiaries have higher-than-average medical expenses, and health costs tend to increase faster than the price of other goods and services. In addition to spending more on health care, lower-income beneficiaries may face faster-rising costs on other goods and services than the broader population (Church 2015; Sherman and Van De Water 2019; and author’s analysis of the Federal Reserve Bank of Chicago’s Income Based Economic Index [Federal Reserve Bank of Chicago 2015]).

Social Security’s normal retirement age must take into account gaps in life expectancy

EPI has pushed back against increases in Social Security’s normal retirement age, which is already increasing in steps from 65 to 67 for workers born between 1937 and 1960. An increase in the normal retirement age of one year is equivalent to a benefit cut of 6.7%. To counter the idea that an increase in the normal retirement age is inevitable in the face of rising average life expectancy, EPI highlighted widening gaps in life expectancy by socioeconomic status; pointed out that the Social Security actuaries foresaw and planned around the increase in life expectancy over the 20th century, much of which was harmless or beneficial to the program;3 and emphasized how slow and unequal wage growth, and later the Great Recession, contributed to Social Security’s long-term shortfall. EPI pointed out that the appropriate response to slow and unequal wage growth is scrapping the cap on taxable earnings rather than raising the retirement age.

Social Security’s universality is central to its success

EPI has resisted attempts to gradually transform Social Security from a social insurance program into a safety net program targeted at low-income beneficiaries. After the failure of Bush’s privatization plan, Republican policymakers and conservative think tanks dropped their get-rich-quick talking points. Instead, they attempted to lure centrist Democrats with a combination of Social Security scaremongering and “compassionate conservative” plans to make the program more “progressive.” Specifically, they proposed cutting benefits for middle- and upper-income participants and slightly increasing benefits for very-low-income participants—invariably in ways that would shrink the program over time. EPI and other progressives pointed out that if Republicans were truly interested in alleviating old-age poverty, they could focus on expanding Supplemental Security Income (SSI) without slashing Social Security benefits for everyone else.

While it might seem strange that progressives oppose making a government program more “progressive,” this is because they resist turning a universal program into a need-based one. In a similar way, progressives support increasing education spending on low-income students without reducing spending on middle-class students and implicitly redefining the mission of public schools as serving low-income students who cannot afford private educations.

Social Security appears to be safe from outright cuts—for now. The danger of benefit cuts has abated with strong opposition from the Strengthen Social Security Coalition and its lead organization, Social Security Works; the ascendance of progressives within the Democratic party; and the loss of Republican credibility on deficits after President Trump’s deficit-expanding Tax Cut and Jobs Act (Collender 2018; Rampell 2018). Though some in the GOP are intent on shrinking “entitlements,” the threat is less serious since Trump and others have signaled their unwillingness to make cuts to this popular program.

Going on the offensive for Social Security expansion

Social Security expansion is the most important way to forestall the retirement crisis. Social Security provides low-cost, portable, risk-free, annuitized benefits and modestly progressive income redistribution. The idea of earned benefits is very popular, and Social Security has strong support among voters of all stripes (Walker, Reno, and Bethell 2014; Public Policy Polling 2018).

The main challenge to expanding Social Security is fierce political resistance from Republican policymakers and big money donors. Politics aside, one substantive disadvantage is that Social Security taxes are not, in their current form, progressive. However, considering the program as a whole—including retirement, survivor, and disability benefits—Social Security is highly efficient and modestly progressive. Meanwhile, taxpayer subsidies to employer plans, especially 401(k)s, are highly regressive.

EPI was an early and strong supporter of Social Security expansion. Surveys have found that Social Security expansion is popular among voters, even if it means they have to pay more into the system (Walker, Reno, and Bethell 2014). Support for expansion was endorsed by virtually all Democratic senators before the 2016 election. Congressional Democrats have formed an “Expand Social Security Caucus,” and in 2019 Connecticut Representative John Larson, chairman of the Social Security subcommittee, recruited over 200 Democratic co-sponsors for his Social Security 2100 Act (Altman 2016; Garcia 2016; Marans 2019). Senators Bernie Sanders (I-Vt.), Elizabeth Warren (D-Mass.), and Sherrod Brown (D-Ohio) were early and influential leaders of the expansion movement, which should receive more attention as Sanders and Warren vie for the Democratic nomination (Brown took himself out of the running) (Birnbaum 2018). Plans put forth by Sanders and Warren significantly boost benefits and extend the system’s solvency, but unlike Larson’s plan do not attempt to fully eliminate the long-term deficit (Goss 2019a, 2019b; Zandi 2019).

There are many different ways to expand benefits. Larson’s bill would increase the first multiplier in the benefit formula from 90% to 93% of the first $926 in average indexed monthly earnings. This is an example of a progressive across-the-board increase, because it would increase everyone’s benefits while resulting in a bigger percentage increase for low earners. Similarly, Warren’s more ambitious proposal to expand benefits by $200 per month would result in a larger percentage increase for low earners.

Other expansion proposals target specific groups. These include restoring benefits for college and vocational students who have a parent with a disability or who had a parent who died; providing a caregiver credit to boost future benefits for people whose covered earnings are negatively affected by caregiving responsibilities; and raising the minimum benefit above the poverty line.

Many expansion plans have both universal and targeted provisions, though the emphasis varies. A few, such as the plan put forward by Democratic presidential candidate Pete Buttigieg, only have targeted increases (Buttigieg 2019).

There are also different proposals for raising revenue. Most plans, like Larson’s, would scrap or amend the taxable earnings cap, among other changes. While some advocates view the program’s contributory structure as key to its enduring appeal, others oppose payroll tax rate increases and favor looking for other revenue sources, such as taxing investment income. For this and other reasons, most bills do not try to eliminate the long-term shortfall entirely, but rather aim to push back the date when the Social Security trust fund is projected to be exhausted. The Larson bill, which eliminates the long-term shortfall, is a notable exception. The bill does this in part by gradually raising the payroll tax rate while extending the tax to some high earners.

In short, Social Security has all the features of a successful retirement program: steady contributions, low costs, and benefits retirees can count on no matter how long they live. In addition to retirement benefits, the program also protects workers and their families against the loss of earnings due to death or disability. The only serious challenge to Social Security expansion is political—a powerful anti-tax and anti-government movement that has successfully dissuaded Congress from increasing the payroll tax rate or implementing other changes to bring more revenue into the system. Such increases, which had once been fairly routine and enjoyed bipartisan support, have not occurred since 1983 (SSA n.d.; Campbell and Morgan 2005).

401(k)s are a failed policy experiment

In contrast to Social Security, the product of a long deliberative process, 401(k)s are an accident of history. It should not come as a shock that 401(k)s failed to deliver more broad-based retirement security—they were never meant to. They were created in 1978 by a change in the Internal Revenue Code clarifying the tax treatment of deferred compensation. They were designed as a perk for bankers, not as a replacement for pensions (Tong 2013). Seemingly overnight, however, ordinary workers became responsible for their own retirements.

Initially, 401(k) and similar DC plans were promoted based on the idea that individuals were best equipped to invest their own money, a claim that has been soundly debunked (Ayres and Curtis 2015; Benartzi, Previtero, and Thaler 2011; Brown et al. 2008; Beshears et al. 2018; CEA 2015; Choi, Laibson, and Madrian 2011; Lusardi and Mitchell 2007, 2014; Society of Actuaries 2018; Sun and Webb 2012). Most retirement savers are not financially savvy. Many fail to diversify across asset classes, adopt an all-or-nothing approach to risk, engage in counterproductive market timing, and naively assume “you get what you pay for” when it comes to mutual fund fees. Most have trouble estimating how much they will need to save, in part because they tend to underestimate their life expectancy and are unsure how to spend down their savings in retirement. Evidence that people are steered to high-cost investments by salespeople masquerading as financial advisers led the Obama administration’s Department of Labor to issue a rule that would have effectively outlawed the practice of offering conflicted “investment advice” to retirement savers, a rule later abandoned by the Trump administration (CEA 2015; Morrissey and Shierholz 2017).

As evidence has mounted against the “smart investor” case for participant-directed accounts, DC plans have been promoted in the public sector by stoking pension envy: Why should government workers have pensions while most taxpayers have 401(k) plans? This line of argument ignores the fact that traditional DB pensions are more cost-effective than DC plans and partly compensate for lower wages in the public sector (Morrissey 2015, 2016b, 2018b; Oakley and Kenneally 2015, 2019; Rhee and Fornia 2014).

The last remaining argument in favor of DC plans—a valid one, at least in the private sector—is that many employers are not in a position to take on long-term pension liabilities. As will be discussed later, this has sparked interest in hybrid plans that give workers some of the advantages of traditional DB pensions without saddling employers with long-term obligations. The Guaranteed Retirement Account plan, discussed below, is one such plan.

Why our do-it-yourself system does not work

While much attention has been paid to the failure of individuals to save and invest optimally, many problems with the 401(k) system have less to do with individual behavior and more to do with a flawed system. What is wrong with 401(k)s? First, too little money is going in, mainly because many workers are not covered by a plan at work. Second, tax subsidies mostly flow to wealthy households who save regardless. Third, money leaks out before retirement in the form of high fees, loans, and withdrawals. Last but not least, these plans do not pool risk, so contributions need to be much higher to ensure that participants do not outlive their savings. Each of these problems is fleshed out below.

The problem of too little money going in is a multidimensional one, but it is largely due to differential access. First, many workers work for employers that do not offer plans, or workers may not be eligible for an employer plan because they do not meet age and service requirements. Second, some workers who do have access to an employer plan do not participate, though this is less common than not having access to a plan in the first place. Third, workers in a plan may not contribute enough themselves or may receive meager employer contributions. Though a typical employer match is 50 cents per dollar contributed by workers up to 6% of pay, employer contributions are optional unless specified in a collective bargaining agreement.

A major problem with our retirement system that receives too little attention is that tax subsidies for retirement savings are upside down. Eighty percent of tax subsidies for retirement savings go to households with incomes above $100,000 (Tax Policy Center 2017). The tax deferral for 401(k) and IRA contributions provides little or no benefit to low-income households, since the subsidy is a function of the income tax that would otherwise be owed on investment earnings, and many low-income families do not owe income tax.

Contrary to popular belief, the tax deferral does not subsidize contributions to retirement accounts, because taxes are owed when funds are withdrawn. Rather, the tax advantage stems from investment earnings not being taxed annually. There is evidence to suggest that the tax deferral has the largest incentive effect on lower-income workers who may receive little or no actual benefit but believe that they do (Benjamin 2003; Chernozhukov and Hansen 2004; Engen and Gale 2000; Engen, Gale, and Scholz 1996; Heim and Lurie 2014).4 And while many low-income workers do not benefit from the tax deferral, they still face a penalty if they need to access these savings before retirement.

Tax “incentives” are mostly wasted in terms of achieving the social goals they supposedly promote. Though the lion’s share of the tax benefit goes to upper-income households, these households do not appear to save more because of it—they simply steer savings to tax-favored accounts. In practice, much of the wealth in 401(k) plans and IRAs may be intended for heirs, which means that taxpayers are promoting wealth inequality, especially since heirs do not pay taxes on these inheritances if the funds remain in tax-favored plans.

Another major problem with 401(k)s is leakage. Participants pay high fees and often borrow or withdraw money before retirement. Fees are high due to a principal-agent problem (employers choose the plan but employees shoulder most of the fees), a lack of transparency, and other market failures. Individual accounts are also inherently more expensive than pooled and professionally managed pensions. The same is true for actively managed funds versus passive investments. Individual investors also earn lower risk-adjusted returns than pension funds and other institutional investors due to poor market timing and inadequate diversification.

These 401(k)-style defined contribution plans are also riskier and therefore less efficient than DB pensions because individual savers do not benefit from intergenerational risk-sharing or longevity risk pooling. Because DC plans have lower net returns than DB pensions and no risk pooling, EPI and others have estimated that contributions to DC plans need to be almost twice as large as contributions to DB pensions to provide similar retirement security (Morrissey 2009; Rhee and Fornia 2014).

Unlike participants in DB pensions, which spread investment risk across workers who retire at different times, 401(k) participants bear the full risk of retiring in the wake of a bear market, as many baby boomers learned the hard way during the financial crisis of 2008–2009. Contrary to popular belief, cumulative investment returns do not average out over longer investment horizons, and popular “target date” funds, which shift to more conservative portfolios as workers approach retirement, are not inherently safer than portfolios with fixed asset allocations.5 The only protection for individual savers against market risk is amassing precautionary savings or opting for more conservative investments. Either strategy lowers lifetime living standards.

A thought experiment: Do-it-yourself health care

Our retirement system relies heavily on individually directed 401(k) accounts that require workers to decide how much to save, how to invest, and how to spend down their savings when they retire. These decisions are complex and the system is expensive. Costs are high both because the system lacks economies of scale and risk pooling, and because many employers and participants have difficulty assessing whether the financial services they are paying for in the form of mutual fund fees and other fees are worth the expense.

To understand how retirement savers are being set up to fail, health care offers a useful analogy. Imagine if we replaced medical consultations with online advice, touting the supposed advantages of people taking charge of their own health care. Then we spent decades wringing our hands about how people get confused about medications and do not follow healthy lifestyles. People are human—that is nothing new. The question is: Who thought a do-it-yourself system would work as well as or better than a professionally managed one?

Of course, 401(k)s have worse problems than just expecting too much from people—they are also a clear rip-off in many ways. Imagine replacing most doctors with websites chosen by employers but paid for mostly by workers, and the employer cost is further minimized if the website promotes certain prescription medications. Finally, imagine that we scrapped the insurance aspect altogether—no risk pooling at all. Have a chronic condition that is expensive to treat and your plan’s balance cannot cover it? You are out of luck.

That is the 401(k) system in a nutshell—401(k)s appear to offer choices, but these choices are limited, expensive, and tricky to navigate. They do not offer what most people need—and what Social Security and DB pensions provide: a cost-effective way to build a secure retirement that does not require people to have financial expertise.

The Guaranteed Retirement Account plan changed the conversation around retirement

For many years, policymakers in Washington focused on encouraging workers to save. This focus on individual behavior persisted despite the fact that most workers who do not participate in a retirement plan do not have access to one at work, and 401(k)s are a poor fit even for workers lucky enough to have access to a plan.

The Guaranteed Retirement Account plan, proposed by Teresa Ghilarducci as part of EPI’s Agenda for Shared Prosperity in 2008, was groundbreaking because it would ensure that all workers and employers contribute something toward retirement, whether to an existing employer plan or to a new portable account called a Guaranteed Retirement Account (Ghilarducci 2008). The GRA system would be administered by a government agency, but funds would be pooled and invested by private-sector managers, similar to the Thrift Savings Plan for federal employees. GRAs would have low administrative costs, provide lifetime benefits, and offer some protection from investment risk. The GRA plan, a public alternative to employer plans, was proposed alongside Jacob Hacker’s Health Care for America plan, which called for a public health insurance option (Hacker 2007). EPI staff and allies worked with the authors on both plans.

A primary goal of EPI’s GRA plan was to draw attention to regressive tax subsidies for 401(k)s and IRAs by calling for replacing them with a revenue-neutral flat tax credit that would fully offset employee contributions for low-income workers. Though a flat tax credit was included in later versions of the GRA plan, it is not a revenue-neutral replacement for other tax subsidies if participants in 401(k) and IRA plans are given the option of remaining in these plans. However, the cost of the GRA tax credit can be offset by phasing in lower contribution limits and new account balance limits for tax-favored retirement accounts. Annuitizing benefits also ensures that the tax credit is not used to subsidize intergenerational wealth transfers.

Unlike some of the hybrid plans it was modeled on (see the section “Hybrid plans may be a good compromise for certain nonpublic employees,” below), EPI’s GRA plan did not share risk between employers and employees, but rather spread investment risk across retiree cohorts using a reserve fund. Any remaining investment risk was borne by taxpayers in the form of a 3% inflation-adjusted government-backed rate-of-return guarantee. Longevity risk, meanwhile, could be shared with annuity providers, including life insurance companies hedging mortality risk.

How the GRA model has evolved

A later version of the GRA plan reduced the 3% real rate-of-return guarantee to a guarantee against investment losses in order to make it easier for the plan to garner political support (Ghilarducci and James 2016). Such a guarantee of principal provides little protection to long-term participants, who are unlikely to experience cumulative investment losses after 30 or more years, and does little to smooth outcomes across retiree cohorts. However, it offers peace of mind to all participants and some real protection to those who begin participating late in their careers. For this reason, the cost of such a guarantee would be higher in the plan’s early years but modest once the plan matures.6 Conversely, a 3% real rate-of-return guarantee helps smooth outcomes but entails a cost or risk to taxpayers, since the yield on long-term inflation-indexed Treasury bonds is currently below 1%.

A rate-of-return target, which can be adjusted if necessary, could be a pragmatic compromise. EPI is modeling a reserve fund that could provide a targeted real return above a guarantee of principal. As a point of reference, the Canada Pension Plan has a long-run real return target of 3.9%, which, if not met, triggers benefit cuts. A reserve fund with a target interest credit would provide more smoothing of retirement outcomes than a guarantee of principal alone, but without the cost or risk to taxpayers of a 3% real guarantee.

The GRA plan as a model for the National Retirement Security Project and other efforts

The GRA model spawned coalitions supporting similar mandatory plans. EPI participated in the Retirement USA initiative convened by the Pension Rights Center and the AFL-CIO. The Retirement USA initiative developed a set of principles that included mandatory employer and employee contributions, lifetime benefits, and other features modeled on the GRA plan. Similar principles were later adopted by the Retirement Security for All coalition, spearheaded by public-sector unions to promote state-level initiatives for private-sector workers who do not have access to employer plans.

EPI incubated the National Retirement Security Project in December 2017. The project, headed by former Maryland Lieutenant Governor Kathleen Kennedy Townsend, who chaired the state’s retirement initiative, promotes a GRA-like system of universal, funded, and portable retirement accounts with lifetime benefits. The American Federation of Teachers provided seed money and helped spearhead the initiative by funding surveys to gauge the appeal of a GRA-like plan.

Going forward, the GRA model should be refined and used to fill in gaps in retirement security that remain after we have expanded Social Security as much as political limits allow. Well-designed GRAs can be a vital part of an improved retirement landscape for American workers.

Remaining defined benefit pensions should be protected

Traditional defined benefit pensions are still an important source of retirement income for some workers. Unlike DC plans, DB pensions pool risk, have professionally managed funds, and benefit from economies of scale. In contrast, DC plans have lower risk-adjusted returns, which reduce retirement wealth by as much as 20% compared with DB pensions.7

Pension sponsors can eliminate individual longevity and investment risks by risk pooling. However, they still shoulder cohort longevity and market risks, since future lifespans and long-term investment returns cannot be perfectly predicted from past experience. Though pensions work well in the public sector, many private-sector employers are not in a position to take on long-term risks, a challenge that motivates the design of hybrid DB–DC plans, which are discussed below.

Multiple-employer pensions are portable and can outlast individual employers. Plans like TIAA are common in sectors with mobile workers, such as university professors and clergy. Similarly, multiemployer “Taft-Hartley” pensions jointly sponsored by employers and unions are common in sectors where workers often have longer-lasting relationships with unions than with individual employers, such as the trucking and construction industries. The challenge with these plans is that the combination of a shrinking pool of active workers and lower-than-expected investment returns can lead to a “death spiral,” as what might otherwise be a manageable funding gap grows larger in relation to a declining payroll. This is why in some countries, unlike the United States, participation in industry pension schemes is often mandatory or quasi-mandatory, which lessens risks associated with a shrinking pool of active participants. Though multiple-employer pensions have worked well in Europe and Canada and in some sectors in the United States, severely underfunded Taft-Hartley plans have put a damper on expanding this model in the United States.

One sector that is clearly well positioned to take on long-term pension obligations is the public sector. Yet public-sector pensions, which are geared toward attracting and retaining career civil servants, are under fire from conservative think tanks and funders intent on shrinking the size of government, deprofessionalizing teaching and other public service careers, and privatizing or introducing “choice” in education and other public services in ways that undermine well-functioning institutions (Farmer 2017; McGee 2011; Morrissey 2017; Recknow, Jacobsen, and Henig 2019).

EPI has pushed back against attempts to exaggerate the cost of public-sector benefits and the extent of public plan underfunding. As EPI research has noted, public pension critics magnify the cost of pension liabilities by arguing that the present cost of future benefits should be estimated using a low “risk-free” rate of return rather than the expected return on pension fund assets. By this measure, any pension fund that would be unable to pay for projected benefits if Treasury bonds were substituted for pension assets appears significantly underfunded even if required contributions are made in full and actuarial assumptions are met. As EPI has shown, using a risk-free rate to estimate the cost of retiree benefits also greatly exaggerates the cost of public employee compensation, which critics have used to argue that teachers and other public-sector workers are overpaid. EPI has challenged the use of a risk-free rate in these contexts, which is often misrepresented as the consensus view among economists.

Public pension critics also claim that pensions are unsuited for a modern, mobile workforce, even though most workers, including mobile workers, are better off with cost-efficient and low-risk pensions than with account-style plans, as will be discussed below. Rather than dismiss traditional pensions as outdated and out of reach, EPI has shown that they are a good fit for many employers and workers in the public sector and in unionized industries. They also serve as a useful model for hybrid plans that combine DB and DC elements to give workers more retirement security without requiring employers to take on long-term liabilities.

Hybrid plans may be a good compromise for some private-sector workers

Hybrid plans are plans that combine features of DC plans and traditional DB pensions. Some hybrid plans look like DC-style retirement savings accounts, but the funds in these notional “accounts” are pooled and professionally managed, as with DB pensions. Other hybrid plans, including those known as “target benefit plans,” provide annuitized benefits similar to traditional DB pensions, but participants bear some of the risk if investments underperform.

Cash balance plans are the best known type of hybrid plan in the United States. With cash balance plans, the employer is responsible for investing the funds and guaranteeing a minimum interest credit, which may be tied to yields on Treasury bonds or other variable benchmarks. Cash balance plans must offer participants the option of converting their savings to a lifetime income stream at retirement, usually on more favorable terms than annuities purchased by individual investors.

Many public pension critics favor replacing pensions with DC plans or hybrid cash balance plans, saying that benefits in traditional DB pensions disadvantage teachers who move between school districts and other mobile workers. Traditional pension benefits are usually a multiple of years of service and final salary (typically averaged over 3–5 years) and tend to be higher for teachers who stay in one school district.8 However, the fact that traditional pensions promote employee retention does not mean that mobile workers would be better off with higher-cost, higher-risk plans. EPI has pointed out that the vast majority of teachers, including mobile teachers, are better off with traditional DB pensions than with DC plans. EPI has also shown that cash balance plans, far from being fairer and more transparent than DB pensions, as public pension critics often claim, are often biased in favor of younger workers and perversely encourage turnover (Morrissey 2017).

Though EPI questions the wisdom of replacing traditional pensions with cash balance plans in the public sector, these and other hybrid plans may be a good compromise in other sectors where employers are not in a position to take on long-term liabilities. Hybrid plans can spread risks usually borne by DB sponsors while providing more retirement security than DC plans. In addition to cash balance plans, hybrid plans include “notional” (a.k.a. “nonfinancial”) defined contribution plans such as those in Sweden and Italy and “target benefit” and similar plans found in the Netherlands, the United Kingdom, and Canada.

With most types of hybrid plans, pooled and professionally managed investments provide a better risk-adjusted return net of fees than individually directed DC plans, but may not achieve the same return as a traditional DB pension if the investment horizon is shorter. In addition, most hybrid plans are designed so employers can avoid taking on long-term liabilities and therefore still expose participants to market risk in the form of a variable interest credit or adjustable pension benefit.

Mandatory versus voluntary saving

EPI’s multipronged approach—expanding Social Security, mandating minimum contributions to a GRA or other employer plan, and defending traditional employer-provided pensions—ensures that workers are preparing for retirement throughout their working lives. In contrast, proponents of the 401(k) system say it offers choice and flexibility, including the choice to put off saving to the future.

In theory, mandatory contributions to Social Security or a GRA-like plan could cause some people to save more—or earlier—than is optimal. Some young workers, for example, may want to delay retirement contributions until they have completed their education, paid off their student debt, raised their children, or seen their earnings increase with experience. A one-size-fits-all approach could also cause low earners to oversave because Social Security has a progressive benefit structure with a higher income replacement rate for low earners. High earners, meanwhile, may prefer to build wealth in other forms, such as investing in family businesses.

While concerns about mandatory oversaving sound plausible in theory, it helps to remember that savings-to-income ratios have been flat over the lifecycle despite strong evidence that the need for savings has increased in recent decades. Thus, Social Security expansion or a mandatory GRA could be thought of as a way to restore lost income and meet rising expenses in retirement rather than shifting consumption toward older ages.

Though concerns about oversaving may be overblown for now, it is a point to keep in mind when considering how much to expand mandatory contributions to retirement plans and whether to offset the cost for lower-income workers. Thus, the GRA plan calls for a refundable tax credit that fully offsets the cost of employee contributions for workers earnings up to $40,000. Similarly, higher Social Security contributions could be offset by expanding the Earned Income Tax Credit.

The original GRA plan, which predated serious Social Security expansion efforts, called for a combined 5% contribution rate. The later version of the plan called for a 3% combined rate, which would provide an inflation-indexed GRA benefit replacing an estimated 14% of wage-indexed career earnings for an average worker.9 Combined with Social Security, even this more modest GRA plan would allow retirees to replace an estimated 69% of wage-indexed earnings for a prototypical low earner, 55% for a prototypical medium earner, and 48% for a prototypical high earner. Since 70% is often considered a minimal target replacement rate for most workers (though most experts recommend a higher replacement rate for lower-income workers), a 3% GRA is unlikely to lead to much “oversaving,” even in combination with expanded Social Security benefits.


The American retirement system is often described as a “three-legged stool” made up of Social Security, employer pensions, and personal savings. This is not an accurate description, since Social Security is by far the most important leg of the stool, while personal savings (aside from home equity) have never played a major role in retirement for most families. In addition, the distinction between employer plans and personal savings has become blurred in the 401(k) era.

Realistically, people need help preparing for retirement. Our first priority should be expanding Social Security to the fullest extent possible. As social insurance advocates Benjamin Veghte and Theda Skocpol remind us, “Social Security is America’s most effective and beloved social program—because everyone contributes and all families can count on it” (Veghte and Skocpol 2012).

But as calls for Social Security expansion have grown louder, conservatives have pushed back against the idea of a looming retirement crisis, pointing to discrepancies between tax data and household survey data in measuring senior incomes. While it is true that retirement income is underreported in household survey data, government researchers with access to tax records have found that the largest source of retirement income outside of Social Security remains DB pensions (Bee and Mitchell 2017). This is a surprising finding, since participation in DC plans surpassed DB pension participation almost 30 years ago (EBSA 2018, Table E4).

Though DB pensions remain an important source of retirement income for many public-sector and unionized workers, most future retirees will not be able to count on these benefits. Meanwhile, even the most ambitious Social Security plans proposed to date by our elected leaders do not eliminate the need for employer benefits. The GRA plan is a model for filling the gap in employer coverage with a fair and affordable plan.

The GRA plan paved the way for other plans with mandatory contributions, including plans recently put forward by centrist Democrats and at least one Republican (Berkowitz and Moller 2019; Coons and Klobuchar 2019; Koenig, Fichtner, and Gale 2018). These plans recognize the importance of universal access, steady contributions, pooled funds, and annuitized benefits. They show that policymakers understand the need to do more than encourage workers to save for retirement in a system that was not designed for this purpose. However, we need to guard against the view that these plans are a substitute for Social Security expansion.

Social Security, DB pensions, and GRAs have the three building blocks of a good retirement system: steady contributions, low costs, and secure income in retirement. While these are the solutions to most people’s retirement problems, challenges remain. One is that some families need more income now, not later. Retirement must be affordable for low-income families, and families should be able to repay college loans, handle medical bills, weather unemployment spells, and deal with other surprise or necessary expenses without draining their retirement savings or finding themselves at the mercy of high-interest lenders. Another is that it has become harder to count on the steady wage growth and investment earnings that are needed to reliably fund the different tiers of our retirement system. Going forward, EPI will look for ways to address these and other challenges in order to make retirement as affordable and worry-free as possible for all Americans.

About the author

Monique Morrissey joined the Economic Policy Institute in 2006. Her areas of interest include Social Security, pensions, savings, tax expenditures, older workers, public employees, and unions.


1. Under a limited liability system, shareholder losses are limited to what corporations can pay, which is often less than the full societal costs.

2. An increase in the normal retirement age from 65 to 67 is equivalent to a 13.3% reduction in benefits. The Social Security Administration estimates that income taxes levied on Social Security benefits will reduce benefits by an average of around 10% in 2030 and 11% in 2040. Multiplying 86.7% by 90% equals approximately 78% (Source: Author’s analysis of SSA 2010 and Purcell 2015).

3. Reductions in infant mortality and longer work lives increase the growth rate of the working-age population. In a pay-as-you-go system, faster growth of the working-age population makes it easier to pay for the cost of current beneficiaries.

4. Lower-income workers still owe payroll taxes and often file tax returns in which 401(k) contributions are reported on an attached W-2. Like many participants, they may not be aware that contributions and accumulated investment income are taxed upon withdrawal, so that the potential tax advantage, at least for people who are in the same tax bracket after retirement, is limited to not paying income taxes on “internal buildup” in the intervening years, taxes these workers may not owe in the first place.

5. The main advantage of target date funds is that younger workers have more time to make up any losses by increasing contributions.

6. Alternatively, participants who begin participating at older ages could be given the choice of investing in risk-free assets, such as Treasury securities, or waiving the guarantee.

7. Author’s estimate based on contributions increasing 2.6% annually with inflation over 40 years and a 100 basis point difference in returns (5.5% rather than 6.5%, as assumed in CEA 2015).

8. Teachers who change school districts mid-career will have some service credits multiplied by their mid-career salaries rather than by their higher final salaries. In addition, some teachers leave before their benefits vest, though this can be true with DC plans as well.

9. Author’s estimate based on a 3.4% real (6.0% nominal) rate of return and an inflation-indexed benefit. Other assumptions are based on the long-run assumptions of the Social Security Trustees Report (SSA 2019) and Social Security’s prototypical “medium earner” (Clingman, Burkhalter, and Chaplain 2017).


In addition to the cited references below, this paper draws heavily on EPI’s work on Social Security, public pensions, and defined-benefit/defined-contribution plans, all available on EPI’s Retirement issue page (EPI 2019).

Altman, Nancy. 2016. “Expanding Social Security Is the Official Position of the Democratic Party! Will Republicans Continue to Support Cuts?HuffPost, July 10, 2016.

Anderson, Elizabeth. 2016. “Common Property: How Social Insurance Became Confused with Socialism.” Boston Review, July 25, 2016.

Ayres, Ian, and Quinn Curtis. 2014. “Beyond Diversification: The Pervasive Problem of Excessive Fees and Dominated Funds in 401(k) Plans.” Yale Law Journal 124 (2014): 1476.

Bee, Adam, and Joshua Mitchell. 2017. “Do Older Americans Have More Income Than We Think?” U.S. Census Bureau, Social, Economic, and Housing Statistics Division (SESHD) Working Paper #2017-39, July 2017.

Benartzi, Shlomo, Alessandro Previtero, and Richard H. Thaler. 2011. “Annuitization Puzzles.” Journal of Economic Perspectives 25, no. 4: 143–164.

Benjamin, Daniel. 2003. “Does 401(k) Eligibility Increase Saving? Evidence from Propensity Score Subclassification.” Journal of Public Economics 87 (May): 1259–1290.

Berkowitz, Kelsey, and Zach Moller. 2019. Universal Private Retirement Accounts: A Lifetime of Work Should Mean a Great Retirement for Everyone. Third Way, March 2019.

Beshears, John, James J. Choi, David Laibson, and Brigitte C. Madrian. 2018. “Behavioral Household Finance.” In Handbook of Behavioral Economics: Foundations and Applications 1, edited by B. Douglas Bernheim, Stefano DellaVigna, and David Laibson. Amsterdam: Elsevier.

Birnbaum, Emily. 2018. “More Than 150 Dems Launch Caucus to Expand Social Security Benefits.” The Hill, September 13, 2018.

Broaddus, Matt, and Edwin Park. 2016. Affordable Care Act Has Produced Historic Gains in Health Coverage. Center on Budget and Policy Priorities, December 15, 2016.

Brown, Jeffrey R., Jeffrey R. Kling, Sendhil Mullainathan, and Marian V. Wrobel. 2008. “Why Don’t People Insure Late-Life Consumption? A Framing Explanation of the Under-Annuitization Puzzle.” American Economic Review 98, no. 2: 304–309.

Bureau of Labor Statistics (BLS). 2018. National Compensation Survey: Employee Benefits in the United States, March 2018. September 2018.

Butler, Stuart, and Peter Germanis. 1983. “Achieving a ‘Leninist’ Strategy.” Cato Journal, Fall 1983.

Buttigieg, Pete. 2019. Dignity and Security in Retirement. Accessed December 2, 2019 at

Campbell, Andrea Louise, and Kimberly J. Morgan. 2005. “Financing the Welfare State: Elite Politics and the Decline of the Social Insurance Model in America.” Studies in American Political Development 19 (Fall 2005): 173–195.

Chernozhukov, Victor, and Christian Hansen. 2004. “The Impact of 401(k) Participation on the Wealth Distribution.” The Review of Economics and Statistics 86, no. 3 (August): 735–751.

Choi, James J., David Laibson, and Brigitte C. Madrian. 2011. “$100 Bills on the Sidewalk: Suboptimal Investment in 401(k) Plans.” Review of Economics and Statistics 93, no. 3: 748–763.

Church, Jonathan. 2015. “The Cost of ‘Basic Necessities’ Has Risen Slightly More Than Inflation over the Last 30 Years,” Beyond the Numbers 4, no. 10 (Bureau of Labor Statistics), June 2015.

Clingman, Michael, Kyle Burkhalter, and Chris Chaplain. 2017. “Replacement Rates for Hypothetical Retired Workers.” Actuarial Note, Social Security Administration, July 2017.

Collender, Stan. 2018. “On The Deficit, GOP Has Been Playing Us All for Suckers.” Forbes, April 15, 2018.

Coons, Chris, and Amy Klobuchar. 2019. “Most Americans Can’t Save for Retirement. We Want to Fix That.” CNN Business, April 25, 2019.

Council of Economic Advisers (CEA). 2015. The Effects of Conflicted Investment Advice on Retirement Savings. February 2015.

Downey, Kirstin. 2009. The Woman Behind the New Deal: The Life of Frances Perkins, FDR’s Secretary of Labor and His Moral Conscience. New York: Random House.

Economic Policy Institute. 2019. “Retirement” (web page). Last accessed November 2019.

Employee Benefits Security Administration (EBSA). 2018. Private Pension Plan Bulletin Historical Tables and Graphs, 1975–2016. December 2018.

Engen, Eric, and William G. Gale. 2000. “The Effects of 401(k) Plans on Household Wealth: Differences Across Earnings Groups.” National Bureau of Economic Research Working Paper no. 8032, December 2000.

Engen, Eric M., William G. Gale, and John K. Scholz. 1996. “The Illusory Effects of Saving Incentives on Saving.” Journal of Economic Perspectives 10 (Fall): 113–138.

Farmer, Liz. 2017. “John Arnold: The Most Hated Man in Pensionland.” Governing, April 2017.

Federal Reserve Bank of Chicago. 2015. Income Based Economic Index (IBEX). Last modified June 23, 2015; accessed June 2019 at

Garcia, Eric. 2016. “Democrats No Longer Divided over Social Security; Expansion, Not Cuts, Has Been the Theme in Washington and on Campaign Trail.” Roll Call, September 12, 2016.

Ghilarducci, Teresa. 2008. Guaranteed Retirement Accounts: Toward Retirement Income Security. Economic Policy Institute, March 2008.

Ghilarducci, Teresa, and Tony James. 2016. Rescuing Retirement: A Plan to Guarantee Retirement Security for All Americans. New York: Columbia Univ. Press.

Goss, Stephen C. 2019a. Letter from the Social Security Administration Office of the Chief Actuary to Chairman John Larson, Senator Richard Blumenthal, and Senator Chris Van Hollen. September 18, 2019. In response to a request for an update of estimates of the financial effects on Social Security of enacting the Social Security 2100 Act.

Goss, Stephen C. 2019b. Letter from the Social Security Administration Office of the Chief Actuary to Senator Bernie Sanders and Representative Peter DeFazio. February 13, 2019. In response to a request for estimates of the financial effects on the Social Security (OASDI) program assuming enactment of the “Social Security Expansion Act.”

Hacker, Jacob. 2002. The Divided Welfare State: The Battle over Public and Private Social Benefits in the United States. Cambridge: Cambridge Univ. Press.

Hacker, Jacob. 2006. The Great Risk Shift: The Assault on American Jobs, Families, Health Care and Retirement and How You Can Fight Back. New York: Oxford Univ. Press.

Hacker, Jacob. 2007. Health Care for America: A Proposal for Guaranteed, Affordable Health Care for All Americans Building on Medicare and Employment-Based Insurance. Economic Policy Institute, January 2007.

Heim, Bradley T., and Ithai Z. Lurie. 2014. “Taxes, Income, and Retirement Savings: Differences by Permanent and Transitory Income.” Contemporary Economic Policy 32, no. 3: 592–617.

Koenig, Gary, Jason J. Fichtner, and William G. Gale. 2018. Supplemental Transition Accounts for Retirement. AARP, January 2018.

Lusardi, Annamaria, and Olivia S. Mitchell. 2007. “Financial Literacy and Retirement Preparedness: Evidence and Implications for Financial Education.” Business Economics, January 2007.

Lusardi, Annamaria, and Olivia S. Mitchell. 2014. “The Economic Importance of Financial Literacy: Theory and Evidence.” Journal of Economic Literature 52, no. 1: 5–44.

Marans, Daniel. 2019. “House Democrats Unveil Social Security Expansion Bill with Unprecedented Support; Rep. John Larson Believes His Plan to Increase Benefits Will Pass the House.” HuffPost, January 30, 2019.

McGee, Josh. 2011. Creating a New Public Pension System. Laura and John Arnold Foundation.

Morrissey, Monique. 2009. Toward a Universal, Secure, and Adequate Retirement System. Retirement USA Conference Report, October 21, 2009.

Morrissey, Monique. 2015. Will Switching Government Workers to Account-Type Plans Save Taxpayers Money? Economic Policy Institute, March 2015.

Morrissey, Monique. 2016a. The State of American Retirement: How 401(k)s Have Failed Most American Workers. Economic Policy Institute, March 2016.

Morrissey, Monique. 2016b. Unequal Public-Sector Pay in Connecticut? Yes—Taxpayers Are Getting a Bargain! Economic Policy Institute, December 2016.

Morrissey, Monique. 2017. Teachers and Schools Are Well Served by Teacher Pensions. Economic Policy Institute, April 2017.

Morrissey, Monique. 2018a. “Social Security Is Looking Like a Pretty Good Investment These Days.” Working Economics Blog (Economic Policy Institute), April 23, 2018.

Morrissey, Monique. 2018b. “Why Is Wealthy Westport Trying to Gut Police Pensions?” Working Economics Blog (Economic Policy Institute), June 11, 2018.

Morrissey, Monique. 2019a. “Policy Solutions for the Retirement Crisis.” Generations, October 2019.

Morrissey, Monique. 2019b. The State of American Retirement Savings: How the Shift to 401(k)s Has Increased Gaps in Retirement Preparedness Based on Income, Race, Ethnicity, Education, and Marital Status. Economic Policy Institute, December 2019.

Morrissey, Monique, and Heidi Shierholz. 2017. “EPI Comment to the SEC Regarding the Fiduciary Rule.” Comments submitted to Jay Clayton, chairman of the U.S. Securities and Exchange Commission, October 12, 2017.

Munnell, Alicia H., Wenliang Hou, and Geoffrey T. Sanzenbacher. 2017. “How to Pay for Social Security’s Missing Trust Fund?” Center for Retirement Research at Boston College Working Paper no. 2017-18, December 2017.

Munnell, Alicia H., Wenliang Hou, and Geoffrey T. Sanzenbacher. 2018. “National Retirement Risk Index Shows Modest Improvement in 2016.” Issue in Brief 18-1. Center for Retirement Research at Boston College, January 2018.

Oakley, Diane, and Kelly Kenneally. 2015. Retirement Security 2015: Roadmap for Policy Makers; Americans’ Views of the Retirement Crisis. National Institute on Retirement Security, March 2015.

Oakley, Diane, and Kelly Kenneally. 2019. Retirement Insecurity 2019: Americans’ View of the Retirement Crisis. National Institute on Retirement Security, March 2019.

Public Policy Polling. 2018. National Survey Results. Survey of 584 registered voters commissioned by Social Security Works, March 9–11, 2018.

Purcell, Patrick J. 2015. “Income Taxes on Social Security Benefits.” Social Security Issue Paper no. 2015-02, December 2015.

Rampell, Catherine. 2018. “A Balanced-Budget Amendment Is Always Stupid. Right Now, It’s a Joke.” Washington Post, April 12, 2018.

Reckhow, Sarah, Rebecca Jacobsen, and Jeffrey R. Henig. 2019. “Why Are Teachers Protesting in ‘Blue’ Cities?Washington Post, March 8, 2019.

Rhee, Nari, and William B. Fornia. 2014. Still a Better Bang for the Buck: An Update on the Economic Efficiencies of Defined Benefit Pensions. National Institute on Retirement Security, December 2014.

Scott, Victoria. 2019. “Van Hollen Proposes Raising Estate Tax to Boost Social Security.” The Hill, June 25, 2019.

Sherman, Arloc, and Paul N. Van De Water. 2019. Reducing Cost-of-Living Adjustment Would Make Poverty Line a Less Accurate Measure of Basic Needs. Center on Budget and Policy Priorities, June 2019.

Social Security Administration (SSA). n.d. Social Security & Medicare Tax Rates. Accessed online on November 27, 2019, at

Social Security Administration (SSA). 2010. “Effect of Early or Delayed Retirement on Retirement Benefits” (online data table). Last modified August 19, 2010.

Social Security Administration Office of the Chief Actuary (Social Security Actuary). 2019. Summary of Provisions That Would Change the Social Security Program. November 2019.

Social Security Trustees. 2019. The 2019 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds.

Society of Actuaries. 2018. Retirement Survey Report Key Findings and Issues: 2017 Risks and Process of Retirement Survey.

Steinberg, Jonathan. 2011. Bismarck: A Life. Oxford: Oxford Univ. Press.

Sun, Wei, and Anthony Webb. 2012. “Should Households Base Asset Decumulation Strategies on Required Minimum Distribution Tables?” Center for Retirement Research at Boston College WP no. 2012-10, April 2012.

Tax Policy Center. 2017. “T17-0112 – Tax Benefit of Certain Retirement Savings Incentives (Present-Value Approach), Baseline: Current Law, Distribution of Federal Tax Change by Expanded Cash Income Level, 2017” (online data table). April 18, 2017.

Tong, Scott. 2013. “Father of Modern 401(k) Says It Fails Many Americans.” Marketplace, June 13, 2013.

United Auto Workers (UAW). 2019. “A Battle Fought” (web page). Accessed November 2019.

Veghte, Benjamin W., and Theda Skocpol. 2012. “Social Security Is in Good Shape—and Does Not Drive the Deficit.” Scholars Strategy Network, April 2012.

Walker, Elisa A., Virginia P. Reno, and Thomas N. Bethell. 2014. Americans Make Hard Choices on Social Security: A Survey with Trade-Off Analysis. National Academy of Social Insurance, October 2014.

Zandi, Mark. 2019. Evaluating Senator Warren’s Social Security Reform Plan. Moody’s Analytics, September 2019.

See related work on Retirement | Social Security | Older workers

See more work by Monique Morrissey