How to Avert a Public-Pension Crisis
At the turn of the millennium, public pensions seemed to be riding high. By their own accounting, most such funds were more than fully funded, and public workers' retirement benefits were more generous than ever after a round of enhancements in the 1990s. But the two decades that followed have decimated the finances of many public-pension funds, resulting in steeply rising taxpayer costs and serious negative effects on public workers' salaries, jobs, and benefits.
The great irony is that the retirement systems that were meant to protect public workers, shielding them from the vagaries of the market, have often accentuated the effects of market swings, increasing the threats to public workers' financial well-being. Today's strong economy means policymakers are not under maximal pressure: They can see the problem, but they do not yet truly feel it. That time will come, however, and if public pensions are going to survive over the long term, funding and investment practices must improve, and benefits need to be modernized to flexibly meet the needs of today's public workforce.
There won't be a better time to avert the next pension crisis. And there are a few key steps that every jurisdiction should be taking.
GROWING BENEFITS, GROWING RISKS
To understand what must change, we must first appreciate some of the history of how public pensions got to where they are. Most public-pension systems were established in the first half of the 20th century, and for their first few decades, the value of promised benefits and their annual costs were small relative to the budgets of their sponsoring governments. In those early years, public-pension investment portfolios were also relatively safe because policymakers restricted plans' ability to invest in risky assets. The fact that public-pension plans were small and not very volatile made them easy to manage and meant they presented little or no risk to state and local budgets.
Over time, however, public-pension plans have naturally matured, and their financial situation inevitably grew more complex. As the years passed, more public employees eligible for the programs neared and reached retirement age; as a result, total plan membership grew, and there were fewer active workers for every person drawing a benefit check. Between 1960 and 2017, total public-plan membership more than quadrupled, and the ratio of active workers to retirees fell from just under 7 to 1.4.
But plan maturation was not the only thing driving public-pension membership growth. State and local governments were also adding large numbers of workers to their payrolls. Between 1960 and 2017, the United States population grew by 80%; over the same period, state and local government employment grew by 186%. As governments added more public workers and the membership of public plans aged, the total amount of promised benefits (that is, the plans' liabilities) grew precipitously.
Meanwhile, governments also significantly enhanced benefits beginning in the late 1980s through the 1990s, propelling liabilities even higher. In a recent working paper, the American Enterprise Institute's Andrew Biggs used data from the National Income and Product Accounts to document how public-pension generosity has changed over time. Up until the late 1980s, the taxpayer cost of the retirement benefits earned by public workers in any given year (the so-called "employer normal cost") remained relatively stable, hovering between 8% and 10% of payroll. But from that point forward, public workers began earning much more generous benefits for each year of work. Employer normal cost increased by 76% between 1988 and 2017, growing from 9.6% to 16.9% of payroll.
Together, the maturation of public plans, growing public payrolls, and benefit enhancements have led to a meteoric rise in pension liabilities. The federal government maintains data on public-pension assets and liabilities going back to 1945. Comparing historical data on pension liabilities to the national gross domestic product provides a reasonable measure of the size of promised benefits relative to taxpayers' overall capacity to pay. In 1960, state and local pension liabilities totaled approximately 12.6% of GDP. By 1990, the ratio had nearly doubled to 22.2% of GDP. Then between 1990 and 2017, the ratio nearly doubled again, growing to approximately 42% of GDP. The total value of retirement benefits already earned by public workers is higher today than it has ever been.
Unfortunately, state and local governments did not fully fund benefits as workers earned them. Instead, they used debt to finance a large portion of the last 30 years of pension-liability growth. In each year since 2001, governments have, on average, paid 12% less than the actuarially determined contribution, or the amount that would cover normal cost and pension debt service. Over time, this consistent underpayment adds up. If state and local governments had paid their full pension bill every year since 2001, public-pension funds would be better funded by more than $315 billion.
Responsibility for today's historic levels of pension debt cannot be fully laid at the feet of policymakers, however. Pension plans themselves deserve a hefty share of the blame. Although public-pension plans claimed to be more than 100% funded at the turn of the millennium, in reality they were well short of that benchmark. Arcane accounting practices allow public pensions to discount future benefit payments using their expected investment returns, which, in the irrational exuberance of the late 1990s, were 8% or more for most plans.
This practice creates the opportunity and incentive to understate the cost of benefits. Discounting using a higher rate reduces annual contributions in the near term, by betting that investment returns will cover a greater share of the cost. Plans have taken advantage of this fact by adopting optimistic return assumptions that make benefits look cheaper than they really are. But if returns fall short of expectations, costs can rise quickly because of the power of compounding. And that is exactly what has happened. Since 2001, average annual investment returns have totaled just 6.4%, far below expectations, resulting in growing pension debt and rising cost.
It is more accurate to calculate pension liabilities using a high-grade bond yield, which reflects both the guaranteed nature of benefits and prevailing economic conditions. That's what the United States Bureau of Economic Analysis does when it tabulates public-pension liabilities. Using those data, average state and local public-pension funded ratios (the share of accrued liability that is funded) topped out at just over 83% in 1999 and fell precipitously thereafter. In fact, 1999 was the high-water mark for pension funding, and was the result of decades of progress toward full funding. Today pensions are just 47.5% funded, hardly better than they were in 1983. Pension plans' consistent underestimation of the cost of retirement benefits is a huge reason for this regression.
It is easy to understand why pension plans made such a big bet on their investment performance. The final two decades of the 20th century delivered the best market performance ever recorded. Researchers have been able to estimate the equivalent of the S&P 500 index, a broad measure of United States stock-market performance, going back to 1871. And in that nearly 150-year history, no 20-year period had higher average annual returns than the one that ended in 1999, with a whopping 16.88%. Pension plans had fortuitously positioned themselves to take advantage of that market run-up. As noted previously, in their early years, pensions faced limitations on risky investing. Those restrictions were largely removed in the 1970s and 1980s at the pension plans' behest, following which public pensions shifted out of safer assets like bonds and into riskier assets like stocks. In 1960, approximately 10% of pension investments were in risky assets; today, 72% are.
In the background, as plans reaped the market gains of the 1980s and '90s, the economy was shifting, changing the calculus of pension financing. Interest rates, which had reached new heights in the preceding decade, began to fall consistently through the 1990s and into the 2000s. Meanwhile, pension plans' investment-return assumptions, which had followed interest rates up, did not follow them back down. As a result, plans were expecting the return on their investments to exceed the "risk-free" rate (the return on the relatively safe bet on United States government bonds) by a growing margin over time. To achieve the same return, plans needed to take on more risk or give up some liquidity — and they did both.
Researchers estimate that, to get the same 7.5% to 8% return, pension plans need to take three to four times more risk today than they did 20 to 30 years ago. As risk increases, pension assets become more volatile. As a partial solution to this problem, pension plans began shifting into less liquid, alternative investments like private equity, hedge funds, and real estate. In theory, giving up some liquidity and diversifying will somewhat reduce risk for any given return target. Since 2001, public pensions have tripled the share of their portfolios devoted to alternative investments from 9% to 27%. But based on the available evidence, it's doubtful that this shift into alternatives has delivered on its promise: Returns have fallen well short of targets, fees are up, and volatility is still a problem.
The upshot of this history is that public pensions are bigger and riskier than ever. Pension liabilities and debt have never been so large relative to taxpayers' capacity to pay, and pension investments have never been so uncertain. Public pensions' finances have continued to deteriorate despite a decade-long bull market that has seen the S&P 500 index grow by more than 300% from its bottom in 2009. It's clear that pension plans are not simply going to grow their way back to fiscal health, and that more must be done to reverse the deleterious effects of rising pension debt.
THE COSTS OF DEBT
That debt is itself a major burden. Since 2001, inflation-adjusted annual taxpayer cost for public pensions has more than tripled, and most of that increase is due to increasing debt-service payments. Roughly 70% of current taxpayer contributions go to pay down debt rather than toward new benefits earned by public workers. While pension contributions still make up a relatively small share of total state government own-source revenue (not including federal transfers), that fact is driven largely by the size of state budgets relative to statewide pension plans and the plans' understatement of cost. According to an analysis by J. P. Morgan, half of states would need to devote more than 10% of own-source revenue to retirement costs if liabilities were calculated using a 6% assumed return.
And the problem of rising pension cost is more acute at the local level, in areas like education where teachers' pension costs have grown from $530 per pupil in 2004 to more than $1,300 today. Annual taxpayer contributions to teachers' pensions now represent more than 10% of per-pupil expenditures. Recent reports have shown that the rising cost of teachers' pensions has contributed to stagnant teacher salaries, decreasing benefits, and shortages of funds for important school supplies and programs for low-income students. The rising cost of pension debt more generally leaves less money available to pay for everything else, including the salaries and benefits of the current generation of not only teachers but also police officers, firefighters, and other crucial public workers.
Of course, pension costs can rise only so far. As a pension plan's funded ratio falls, it becomes more dependent on contributions and less reliant on investment returns. The logical extreme is a system that has no assets and whose annual contributions must equal benefit payments plus administrative expenses. Such a system is said to be "pay-as-you-go" or "pay-go." If public pensions were to reach pay-go, taxpayer contributions would need to increase by approximately 80%, or roughly another 15% of payroll. And in addition to requiring much higher contributions, reaching pay-go would also make retirees' benefit checks dependent on annual appropriations by policymakers.
As pension debt-service payments have driven costs higher, some have questioned the wisdom of trying to achieve full funding at all. But these arguments represent short-termism at its worst — seeking to lower costs today at the expense of the future. When benefits do not have to be fully funded as they are earned, there is the potential for large intergenerational inequities because policymakers can commit future taxpayers to paying for current services. Robert Costrell has shown that the degree of intergenerational inequity increases as the discount rate deviates from true returns and the funding target falls below 100%.
In fact, the pension debt that state and local governments are struggling with today is the result of a large cost shift in the late 1980s and 1990s, when benefits were enhanced and discount rates were increased. The current generation of workers and taxpayers is paying more and getting less than the previous generation because past policymakers failed to fully pay for the benefits they promised to public workers. The last 40 years of public-pension history suggests that, instead of effectively sharing risk across generations of workers and taxpayers, pensions have accentuated the effects of economic swings, with those who happen to be around when the economy is good reaping outsized gains and those on the downside paying more than their fair share.
The only way to ensure that pension costs and benefits are distributed equitably across generations of workers and taxpayers is to fully fund benefits as they are earned. And having a strong pension-funding policy is more important now than ever before because pensions are bigger and riskier now than ever before. As noted above, state and local public pensions are 47.5% funded at this point, but even that may be misleading. The last time pension funding was at roughly this level was 1983, when liabilities were approximately 18% of GDP. Today, liabilities are more than twice as large relative to GDP (42%). Because of pensions' increased scale, relatively small misses on the anticipated investment return or other assumptions (like mortality) would have outsized future budgetary implications. And because pension investment portfolios have gotten riskier, uncertainty about future returns has increased.
If state and local governments do not take steps to comprehensively address pension funding shortfalls, managing public pensions will become an increasingly odious challenge, leading to heightened budgetary pressure and more strife between public workers and taxpayers. For a generation now, the people managing pensions have tried to have their cake and eat it too, promising guaranteed benefits and constant taxpayer contributions with professional investing covering the bulk of benefit costs. But it is simply not sustainable to offer fixed benefits and contributions while also relying on risky investments.
There is a better way. Policymakers managing the most resilient and responsibly funded public pensions have not pretended they could hold back the tide of changing economic and demographic conditions, but instead have adapted by realigning benefit costs and contributions before debt piles up.
THE REFORMS THAT MATTER MOST
Much can be learned from the best funded plans, like those in South Dakota, New York, Wisconsin, and Tennessee. Plans in each of these states are better than 90% funded at this point because policymakers have been proactive in adjusting important assumptions, closing funding gaps quickly, and developing risk-sharing mechanisms that can fairly adjust contributions or benefits as needed. These states demonstrate that it is possible to sustainably manage a defined-benefit pension plan.
Unfortunately, the majority of public pensions have been far too slow to adapt, and as a result have seen their funded ratios consistently deteriorate. Many governments face daunting amounts of pension debt that can make full funding appear out of reach. Fixing public-pension funding is not technically difficult, but in most jurisdictions, fully funding pensions at this point would require significantly higher contributions or reduced benefits (or both), making a solution politically challenging to achieve. However, failing to take meaningful action to close the funding gap will only make the problem more challenging and painful to fix in the future. No matter the scale of the problem, governments that work with their plans to craft workable solutions and begin down the path to full funding will be better positioned to weather the next downturn.
The recommendations of the Society of Actuaries Blue Ribbon Panel on Public Pension Plan Funding (SOA BRP) make a great starting point for policymakers who wish to tackle the challenge of pension reform. The SOA BRP's most important recommendations involve investment-return assumptions and pension-debt amortization. The assumption plays a critical role in calculating the current value of promised benefit payments, and thus the adequacy of annual contributions to cover the cost of those benefits. The amortization schedule determines how quickly pension debt is paid off. Together with mortality estimates, the investment-return assumption and amortization policy are the most important elements of pension funding policy. Tightening rules around these three elements would dramatically improve the accuracy of public-pension cost estimates and help ensure the adequacy of annual contributions.
Although public pensions have begun to lower their investment-return assumptions, they continue to assume rates of return that significantly exceed objective projections for likely future investment performance. The median public-plan investment-return assumption has fallen from more than 8% in 2001 to 7.25% today. But using standard capital-market assumptions, Pew has estimated that the median 20-year return for a typical portfolio will be 6.4%, and that there is a one in four chance that public pensions' 20-year returns may not rise above 5%. Using a higher assumed return to calculate liabilities reduces pension cost, but only on paper. If plans' assumed returns exceed realized returns, which is likely, then they are simply delaying necessary payments, making the eventual bill much larger.
Based on Pew's capital-market projections, the median plan is understating its liabilities by between 10% and 15%, depending on duration. Plans that use higher assumed returns are understating liabilities by much more. It will be hard for governments to improve pension funding if the target keeps moving because of predictably inaccurate return assumptions. Pension plans' stubborn maintenance of overly optimistic return assumptions also feeds taxpayer cynicism regarding pensions, potentially undermining the long-term ability of governments to maintain current systems.
Given the importance of these assumptions in calculating plan cost, policymakers should remove as much subjectivity as possible from the choice of the investment-return assumption by explicitly linking it, in statute or ordinance, to the yield on United States Treasury bonds (i.e., the risk-free rate) plus some pre-specified risk premium, as recommended by the SOA BRP. Tightly constraining plans' investment-return assumptions to more closely track economic conditions would eliminate the most significant source of cost underestimation, mirroring improvements that were put in place for private-sector pension plans with the passage of the Pension Protection Act in 2006.
Mortality is the second major factor in determining pension cost, and here too public-pension plans are underestimating. The SOA recently conducted an extensive study of public-employee mortality and found that public plans' assumptions are not keeping up with their actual members' experience. The SOA reported that life-expectancy estimates, based on actual member data from 170 different public-pension systems, exceeded average public-pension estimates for every job category. The biggest deviation was for teachers, which has big cost implications because the teachers' pension plan is generally the largest public-pension system in any given state. To provide a sense of scale, when the California State Teachers' Retirement System updated its mortality assumptions in 2016 by adding just a year or two to average life expectancy, it increased the plan's pension debt by more than $6 billion, or a little less than 10%.
Unlike the investment-return assumption, addressing this problem is not straightforward. While the SOA does routinely release standard mortality tables, the variation in life expectancy across different job categories makes it difficult for public plans to use an off-the-shelf solution. At a minimum, policymakers should require their plans to base mortality assumptions on the most up-to-date SOA tables that incorporate future life-expectancy improvements. They should also require plans to conduct a significant study of member life expectancy every three to five years.
Lowering public pensions' investment-return assumptions to more realistic levels and updating mortality will increase governments' pension debt. The next step in fixing public-pension funding is to determine how the debt will be paid down over time, or amortized. Most public pensions currently spread pension-debt payments over 30 years or more and backload those payments so that they get larger as payroll grows. The result is that government payments often do not reduce the initial pension debt much over the first 10 to 20 years of the payment schedule. And it is actually common for plans to build in periods of negative amortization, where the pension debt grows because expected payments are not large enough to cover the interest on the debt.
In a recent study, Pew found that, between 2014 and 2016, less than half of states made payments large enough to pay down a portion of the pension-debt principal, and only 14 states achieved positive amortization in all three years. Statewide plans would have needed to receive $13.2 billion more in contributions in 2016 to reach the positive amortization benchmark. Governments have been doing the equivalent of paying the minimum on a credit card. So while they eventually expect to pay off the debt, the total cost will be enormous and their finances will be much riskier in the meantime. Negative amortization significantly undermines pension-plan solvency, and policymakers should eliminate the practice.
Policymakers should also consider shortening amortization periods considerably. Here it is useful to separate the past from the future. As discussed above, today's pension debt represents underpayment of benefit costs over the past 30 or more years. Given the size of most governments' pension debt, it is reasonable to spread the cost of paying it down over the next 20 to 30 years. However, for future years, pension plans should adopt more aggressive amortization schedules that ensure that any pension debt is paid off in fewer than 15 years, consistent with the SOA BRP recommendations. Since 1970, the average gap between recessions has been around five years, and to avoid a potential ratcheting up of debt over time, governments should plan to make significant headway in paying down debt between market dips.
Lowering discount rates, extending life expectancy, and tightening amortization schedules will all increase necessary annual pension contributions. For many jurisdictions the increase would be substantial. In order to make the budget math work, governments may need to negotiate with workers and taxpayers to devise a shared-sacrifice solution that balances the contribution of both groups. Over the last two decades, many jurisdictions have made changes to their pension systems, but few have comprehensively fixed pension funding. Many jurisdictions have lowered benefits for new workers, and several have even reduced benefits for current workers and retirees. But benefit reductions alone will not be sufficient to repair public-pension finances. Governments do not generally face a pension-generosity problem but rather an underpayment problem. Benefit reductions without comprehensive reforms to address funding only serve to raise the ire of public workers and doom governments to increasing budgetary pressure and seemingly endless pension-reform debates.
The city of Houston's 2017 pension reforms are the best recent example of a comprehensive approach to the problem. Prior to reform, Houston had racked up more than $8 billion in pension debt owed to its three pension plans for police, firefighters, and municipal employees. As part of the solution, the city lowered the assumed return for all three plans, tightened amortization policy, and committed to putting significantly more money into the plans. In return, the city asked public workers to give a little too to make the increased budgetary cost manageable. And public workers responded by agreeing to higher annual contributions and $2.5 billion in benefit reductions. Since both workers and taxpayers contributed to the negotiated solution, it is likely to be much more durable than reforms where one group or the other bears the bulk of the cost.
Houston's reform legislation also capped future city contributions, and specified clear rules for how benefits and contributions would be brought back into alignment if costs ever increase above the cap. While limiting taxpayers' future pension-cost exposure may not be achievable in every jurisdiction, there is an important lesson in Houston's approach: It is extremely valuable, to the extent possible, to pre-define when and how plan assumptions, contributions, and benefits will be adjusted in the future if benefit costs and contributions should ever get out of alignment. Establishing clear ex ante risk-sharing plans can help ensure that cost is shared in a fair and equitable manner. Other jurisdictions, including Wisconsin and Tennessee, have adopted similar features, which have allowed them to largely avoid significant underfunding.
AVERTING THE CRISIS
No matter how much pension-funding policy is improved, if governments don't pay the bill every year then the problem is sure to get worse. Promising public workers guaranteed retirement benefits and then failing to fully pay for them is utterly irresponsible. It destines future taxpayers to pay more and get less, crowds out governments' ability to provide essential services, and leaves public workers less financially secure. While policymakers cannot fully commit their successors to making full payments, they certainly can increase accountability for doing so by making full funding the default.
Pension-fund investment performance is also very important to the long-term sustainability of public pensions. And yet, there is currently very little discussion between governments and the pension plans they sponsor about investment risk and the resulting cost uncertainty. Pension plans largely make investment allocation and risk decisions with no substantial input from policymakers whose constituents backstop any shortfalls. Given the large potential implications for taxpayers, governments should increase their oversight of pension investing by requiring routine, in-depth stress testing that details projected costs under multiple return scenarios, as recommended by Pew. Policymakers should also find ways to discuss appropriate levels of risk and investment fees, and should consider putting explicit restrictions in place that would limit pension plans' ability to exceed taxpayers' tolerance in either domain.
Finally, policymakers should also realize that funding policy is not the only challenge presented by today's public-pension system. Most public workers participate in retirement plans that base benefits on years of service and salary at the end of their careers, commonly referred to as final-average-salary defined-benefit plans (FAS DB). Such plans are heavily backloaded, meaning that workers do not earn very valuable retirement benefits until nearing retirement eligibility. While FAS DB plans are great for those who work a full career in a single jurisdiction, they can leave public workers insecure through much of their careers. As state and local governments work to address their pension-funding challenges, they should also consider modernizing their retirement benefits to better support all workers. As Marcus Winters and I proposed in this journal in 2015, policy alternatives like guaranteed-return defined-benefit plans (in other words, cash-balance plans) and well-designed defined-contribution plans can be crafted to provide important investment and mortality protections while also offering a benefit that places all workers on the path to retirement security. And such plans are more flexible and more explicitly link benefit costs and contributions, making them easier to manage responsibly.
Public retirement plans face significant challenges in the coming years. They are bigger and riskier than ever before, and their growing annual cost is crowding out spending in important areas like public safety and education. Public workers bear a significant share of increased pension cost through stagnant wages, benefit reductions, and worsening job conditions. And many jurisdictions are not prepared for the next major economic downturn. While some governments have made changes, most have been insufficient to really address the underlying funding issues.
Pensions are not yet in crisis, and there is still time to fix them without too much disruption. To do so, governments must take a more proactive role in managing the pension plans they sponsor. By taking steps to improve the accuracy of pension-cost estimates and to increase accountability for making full payments every year, and by modernizing benefits to help all workers save for a secure retirement, policymakers can leave more sustainable plans for the next generation of workers and taxpayers. But they are running out of time to set things right.
