FROM ISSUE NUMBER 12 ~ SUMMER 2012 GO TO TABLE OF CONTENTS

TEXT SIZE A A A

How Congress Can Help State Pension Reform

JOSH BARRO

Over the past three years, 43 states have enacted some sort of major pension reform. The immediate cause for these measures has been the economic slowdown, which has drained state coffers and left public-employee pensions precariously underfunded. Combined with the longer-term problems facing state pension plans — foremost among them the serious design flaws inherent in most pension systems — this shortage has left governors and state legislators scrambling to keep those systems solvent.

But while reform initiatives have been numerous, they have, for the most part, been ineffective. They have left untouched the fundamental problems plaguing state pension systems; as a result, many states have already had to undertake additional reforms since 2009, and are looking to make still further changes. Even the most aggressive reform of recent years — enacted in Rhode Island in late 2011 — will not do enough to place the state's pension system on sound financial footing.

The problem is that these recent fixes have not achieved either of the two goals essential to successful pension reform: first, reducing the overall costs of public pensions, and second, reducing the fiscal risks borne by taxpayers. (For further details on how states should approach the pension crisis, see my essay "Dodging the Pension Disaster" in the Spring 2011 issue of National Affairs.) Some states have made significant strides on the first of these fronts, and a few have even done so on the second. But no state has done enough to truly fix the problems affecting its pension system.

The reason is that, in order to meaningfully protect taxpayers from the investment risk inherent in pensions, states need to abandon the defined-benefit model — in which the state promises permanent fixed pension payments after retirement, and assumes the costs of providing them, even when the pension fund's assets and market performance cannot cover the expense of outgoing benefit payments. Instead, states need to move toward the defined-contribution model. Under this system, the state provides employees with some fixed amount toward retirement while they are working, allowing them to invest the money on their own. Such a system thus places the responsibility for managing public employees' retirement savings directly with workers themselves, not state taxpayers.

But while this transition is both urgent and necessary, no state has yet managed to move its existing work force to a defined-contribution system. Some states have been able to place their new hires into defined-contribution pensions, but even these reforms have been too limited in scope.

States can, and should, do better. There is no question that the politics of pension reform can be exceedingly difficult, and that this difficulty has given state lawmakers strong incentive to find ways to shirk their obligations rather than undertake painful reforms. And yet the failure to enact meaningful reforms is not entirely the fault of state politicians or the political pressures they face. Many of the barriers to reform can be removed only by the federal government; Congress therefore has a key role to play in helping states restore their pension systems to stability.

An examination of precisely why state pension reforms have failed — particularly the example of Rhode Island, which came closer than any other state to success, and yet still couldn't quite manage it — can illustrate how real progress will require some assistance from Washington. It can also highlight exactly what Congress should do to help. Federal lawmakers can and should enable many essential changes, but their work should serve three general aims: making it harder for states to skip payments into their pension systems; eliminating accounting rules that scare states away from defined-contribution retirement systems; and strengthening the incentives for retirement savings to help such defined-contribution systems succeed.

FUNDING STATE PENSIONS

Defined-benefit public pensions are, in principle, supposed to be pre-funded. The employer (in this case, the state or local government) is supposed to set aside money during a worker's career to pay for his benefits in retirement; in many cases, the employee is required to make some portion of the total contribution himself. The employer (the state or local government) then invests these assets in some mix of equities and bonds, and ideally those assets will grow in value over time and enable the employer to fund the employee's pension payments during retirement. If that does not happen, however, the state or locality (and therefore the taxpayer) is on the hook for the difference. The level of the employee's benefit is guaranteed, so if the pension plan misses its target investment return over time (and thus technically becomes "underfunded" and at risk of insolvency), retirees don't see their benefit checks shrink: Instead, those benefits are paid out of the state's or locality's general funds.

This means that defined-benefit pensions present states with two kinds of costs: those necessary to "pre-pay" the eventual benefits of today's workers, and those necessary to make up any difference between the value of a pension plan's assets and the cost of paying benefits to the plan's retirees. Every year, a state's pension actuaries basically send a bill to the state government laying out the financial contribution necessary both to cover the accrual of more benefits over the year (so-called "normal costs") and to shore up plans that are underfunded and at risk of insolvency ("amortization costs"). This is described as the annual required contribution, or ARC — but it is "required" in the sense that it is what the state's actuaries deem necessary to fill the gap, not in the sense that the state government is under an actual legal obligation to provide it, as we shall see.

Given these realities, the pension crisis is, at its core, a solvency crisis. States have made promises that far exceed the assets they have set aside to cover them, and at some point, taxpayers will have to pay an exceedingly large bill to settle those promises. But on its surface, the pension crisis has been understood by most lawmakers as a cash-flow issue. State pensions have become an urgent problem not because long-term obligations exceed long-term assets, but because the annual contributions required to keep pensions funds in good order have, in recent years, presented immediate difficulties for tight state budgets. Those annual contributions have gone up because the value of the assets held by pension funds declined sharply during the financial crash and recession, leaving a much bigger gap to close than had existed previously.

This is why the drive for pension reform around the country has tracked the rise in the current-year cost of state pension contributions, not the extent of pension funds' overall solvency problems. Ten states enacted major reforms in 2009, but 21 did so in 2010, and 32 did in 2011. And as the number of reforms has increased, so has their aggressiveness: While early reforms tended to apply only to future hires — meaning they produced negligible short-term savings — more recent reforms have, on the whole, applied more to current workers and even current retirees, producing real near-term savings. States have, in other words, been working to fix their immediate budget gaps, rather than the fundamental problem with their pension systems.

And that real problem — the gap between promised retirement benefits and pension plans' funding — has not actually been getting worse in the past three years. Although a huge solvency gap in pension systems opened up during the stock-market crash and economic downturn of 2008 and 2009, since that time, most pension funds' investment portfolios have generally performed quite well. Indeed, the funding gap is actually smaller today than it was three years ago (though it is still larger than it was in 2007, and simply too large in general).

But even as the overall outlook for pension funds' solvency has remained constant (or improved modestly), states' required annual contributions to those funds have kept growing. And in most states, they will continue to do so through 2014. This is because most pension funds delay the recognition of unusually large gains or losses, spreading it out (usually over a period of five years). While the stock market bottomed out in mid-2009, the declines in asset values are still making their way onto pension funds' books. The pressure on state budgets is thus increasing, and it is this cash crunch that is driving states to reform pensions.

If state pension funds' cash-flow problems were a perfect proxy for their solvency problems, this might work out well: States would be motivated, one way or another, to undertake necessary reform. But states are not facing as much cash-flow pressure from pensions as they should be. That's because states — unlike private pension-plan providers — can game the means by which the annual payments into their pension systems are determined.

The key lies in an aspect of pension accounting: the calculation of the state's annual required contribution to its pension fund. In theory, the ARC is supposed to grow as the funding gap in a pension plan grows; as a plan becomes more underfunded, lawmakers are supposed to come up with more money, giving them a reason to keep pension costs low.

In practice, however, there are several serious problems with the role ARCs play in state pensions, allowing lawmakers to avoid meaningful reforms. Among these concerns, the foremost is that states can skirt their annual required contribution payments. As noted above, the ARC is a calculation of the amount necessary to keep a pension fund in the black, but there is no federal law binding states to actually make their ARC payments, and many states do not. New Jersey has failed to pay its full ARC for nearly all of the past 20 years, which has led the Garden State to have one of the country's worst pension-funding ratios. In 2010, the state faced an ARC of approximately $3 billion for its various employee pension plans, but its actual payment was much lower — $0, to be exact. And New Jersey is hardly the only shirker: In fiscal year 2009, only 61% of state and local pension plans received at least 90% of their ARC payments.

States also have a great deal of flexibility in how they determine the ARC. They can choose different methods of calculating their future pension liabilities, some conservative and some far more aggressive. At the same time, states can also choose the length of the period over which unfunded pension liabilities will be paid off. Some states choose conservative, short horizons (such as New York, which amortizes unfunded liabilities over a period of approximately 12 years, and has some of the best pension-funding ratios in the country). Other states show far less discipline. The Governmental Accounting Standards Board — a private non-governmental organization that makes recommendations regarding accounting principles for state and local governments — suggests an amortization period of no more than 30 years. But Illinois, for instance, has gone with periods as long as 50 years.

States can also decide how long to delay the recognition of stock-market losses. The longer the delay, the more slowly the ARC will rise after a stock-market crash. (Following the crash of 2008-09, some jurisdictions took the opportunity to lengthen the period over which they "smooth" losses to as much as ten years.) And in determining how such market losses affect their pension funds' liabilities, states may make actuarial assumptions that are not realistic — meaning that their liabilities are understated, and their ARCs inadequate.

Thus, even if a state is technically making its ARC payment, it may not be paying as much as it should to keep its pension fund healthy — and it may not be facing sufficient pressure to control pension costs. If the pressure policymakers feel is the result of the immediate cash-flow problem, and if they can address that problem by changing the way they calculate annual payments rather than by actually improving the solvency of their pension system, they are likely to take that easy way out. And doing so offers another attractive benefit: It allows states to use their pension plans as slush funds, diverting the revenues collected for the purpose of paying future retirement benefits to present-day budget expenses while putting off pension problems for later.

One state that has done so in the past is Rhode Island, which has failed to fund pension liabilities as they have accrued. The Ocean State stands out because this underfunding — along with several other problems with its pension system — recently forced state lawmakers to act. In this sense, it is an example of how the structural flaws in state pension systems can align to bring about meaningful reforms. But Rhode Island also provides an example of how the effectiveness of even ambitious reforms is limited, and how it could be improved by some policy changes at the federal level. It is therefore worth looking at the state's pension-reform efforts as a case study — and considering what instruction Rhode Island's experience might provide to the federal government.

LEARNING FROM RHODE ISLAND

In late 2011, Rhode Island enacted the country's most aggressive statewide pension reform to date. It was able to accomplish as much as it did because of a confluence of factors that overcame the usual reluctance of state legislatures to address pension funds' solvency problems. The architect of the reform — state treasurer Gina Raimondo, a Democrat with a background in venture capital who was elected in 2010 — framed pension reform as essential to ensuring that Rhode Island has the resources to provide basic government services in the future.

Raimondo started her push for pension reform by ordering a review of the state pension system's actuarial assumptions. Based on the review, actuarial estimates were revised to include more realistic (and thus higher) assumptions about life expectancy and more realistic (and thus lower) assumptions about the state's expected rate of return on assets. The result was that the state revised its pension-liability estimate significantly upward — which also sharply increased the state's ARC payment. Because of the revisions, Rhode Island faced the prospect of its ARC rising by 60% in just one year. Without reform, the state's annual pension payment would have risen from $254 million in 2012 to $407 million in 2013 and an estimated $522 million by 2016.

Meanwhile, the small Rhode Island city of Central Falls exhausted its own modest and independent pension fund. The city, which is under receivership, unilaterally cut payments to pensioners by as much as half. This drove home for Rhode Islanders that pension solvency isn't merely a theoretical problem, and discouraged lawmakers from further underfunding the state's pension system. Instead, lawmakers sought a reform that would control the rise in required annual contributions and put pensions on a path to solvency so that the problem would not recur.

All of this meant that pension discussions dominated the state's politics throughout 2011. Lawmakers and the public became immersed in the issue, with the Providence Journal assigning nine reporters and staff to cover state and local pension issues. Raimondo was an effective and sensitive communicator on pension reform, stressing that reform is a matter of "math, not politics" and that fixing the pension problem would free the state's government to focus on providing other important services to Rhode Islanders.

Many states have experienced some of these developments — major short-term cost increases, a keen awareness of the long-term risk of pension insolvency, or a leader who could successfully communicate about state pension policy. But all of them came together in Rhode Island to create a unique opportunity for reform.

And that reform was ambitious. Unlike most reforming states, Rhode Island reduced the benefits to be earned not just by new employees, but also by existing state workers for their future years of employment. This is similar to how large private firms have typically modified their pension systems in recent decades — allowing workers to keep what they have earned, while applying new rules to their pension earnings going forward. But though this approach has been common in the private sector, it has been almost unheard of in the realm of public pensions.

Rhode Island also did not limit its reform to tweaking the defined-benefit formula by adjusting retirement ages and similar factors, as most state governments have. Instead, it shrank the defined benefit significantly and augmented it with a 401(k)-style plan, leading to a hybrid system with defined-benefit and defined-contribution characteristics. And the defined-benefit part of the system will feature a variable cost-of-living adjustment that moves with returns on pension-fund investments. As a result, part of the investment risk — even in the defined-benefit pension fund — will be borne by retirees, not by taxpayers.

But the really big savings came from a retroactive action: Like five other states (Arizona, Colorado, Minnesota, New Jersey, and South Dakota), Rhode Island reduced the cost-of-living adjustment provided to current retirees. This has been a controversial and problematic reform, as it amounts to a repudiation of a portion of the pension benefits that workers earned in the past; it may therefore constitute a violation of past employment agreements. In Rhode Island, as in other states, this move is being litigated.

Overall, the savings from Rhode Island's reform package are significant. The state will save $4 billion over 20 years, a large sum in a state with a population of just over a million and annual state spending of approximately $8 billion. And despite the fact that Rhode Island has the country's second-highest rate of unionization in the public sector (and that both houses of its legislature are controlled by Democrats), the plan passed by wide margins in both chambers.

But though these policy changes were bold, and even with all of the factors favoring real reform, Rhode Island still hasn't been able to overcome the fundamental problems plaguing its pension system. The new pension plan isn't much cheaper than the old plan, though its costs will be significantly more predictable. Severely troubled local pension systems in the state still await reform, and are swallowing ever-increasing shares of municipal budgets. And worst of all, a defined-benefit system has been retained alongside a new defined-contribution system, meaning that additional funding gaps may appear in the future.

The last problem is the most serious, because it means that the underlying cause of the state's pension problems has not been dealt with, and the potential for future mismanagement and funding problems remains. Part of the reason is Rhode Island's unusual geographical position: It is wedged between, and thus competes for residents with, two states with much higher per-capita incomes; Connecticut and Massachusetts can lure Rhode Islanders with the promise of both higher spending (and thus more generous state services) and lower taxes.

The result is a strong temptation for Rhode Island to keep its own spending high and taxes low, bridging the budget gap with borrowing. And, as many states have found, one easy place to get the money is from the pension system, which is often used as a backdoor borrowing vehicle: By failing to pay for future pension commitments as they accrue, states essentially spend money they should be putting away, and promise to pay it back later. State and local politicians in Rhode Island have a long track record of using their defined-benefit systems this way.

Rhode Island lawmakers have also made unwise decisions to sweeten pension benefits in the past. Two decades ago, the mayor of Providence entered into a consent agreement offering 6% compounded annual cost-of-living adjustments to some city retirees. At the time, the city's actuary said the increase would be affordable; over the years, however, it has in fact crippled the city's finances. Consider the example of former Providence fire chief Gilbert McLaughlin, who retired in 1991 with a salary of $63,510 and now draws a pension of more than $178,000 a year under the city's exceedingly generous system.

As in many other states, Rhode Island lawmakers sweetened pensions at the height of the tech bubble, a move that has since proved disastrous. Today, by contrast, Rhode Islanders are in a mood for pension austerity, and further giveaways would not be popular (or affordable). But the fact that the state's pension system has retained some part of its defined-benefit structure means that some day in the future — when the political and economic climates have become more favorable — legislators may well make unaffordable promises again. The very design of the system encourages such irresponsible behavior in good times, setting states up for yet more fiscal crises when the economy takes a turn for the worse.

Rhode Island's reformers understood as much, but they could not move away entirely from a defined-benefit structure. They were willing to take the political heat that comes with reducing benefits, and yet they still did not transition fully to a defined-contribution retirement system. And an important part of the reason is that the rules by which public accounting is permitted — and, in some cases, required — to work understate the benefits of moving to a defined-contribution approach while overstating the costs and risks. The rules and laws surrounding private retirement savings, meanwhile, do not do as much as they should to incentivize such savings as a viable alternative to old-fashioned pensions. The rules are thus stacked against reform — and the states cannot change those rules on their own.

ENABLING RESPONSIBILITY

If serious pension reform at the state level requires some help from the federal government, it is important to first understand what the point of that assistance should be. And the goal of any changes in Washington should be fairly straightforward: If we want state lawmakers to make good choices about pensions, we should compel them to fund their pension systems in an adequate and timely manner, and not tempt them to act irresponsibly. Enabling these two practices should be the ultimate aim of any alterations to federal policy.

We already have an example of what effect a federal policy that establishes serious pension-funding requirements can have — in the private sector. With the enactment of the Employee Retirement Income Security Act (often referred to as ERISA) in 1974, the federal government started forcing private firms that offer defined-benefit pensions to maintain adequate funding ratios. These requirements were strengthened by the Pension Protection Act of 2006.

These federal laws are much stricter than the standards the GASB now recommends for state and local governments. The discount rates that must be used to adjust future liabilities are lower, meaning that the estimates of present value of liabilities are higher. And unfunded liabilities must be amortized over seven years, not 30. Unlike the GASB's recommendations, moreover, the federal requirements for private companies are binding. As a result, while private and public pensions currently report about the same funding ratios, the private plans are actually significantly better funded (because the estimates of liabilities they use to arrive at the ratios are higher), and they will close their funding gaps more quickly.

The effect of ERISA was to make private defined-benefit pensions much more secure, but also more rare — a crucial point for the cause of state pension reform. While 38% of private-sector workers participated in a defined-benefit pension plan in 1980, the figure in 2008 was just 20%. Forced to incur the true cost of securely funding pension plans, many companies have chosen to abandon them in favor of 401(k) plans or similar defined-contribution models. The appeal of requiring states to abide by similar requirements is not just that they would no longer be able to shirk their obligations, but also that such requirements would make the advantages of defined-contribution pensions far clearer.

To date, the most significant support for a mandatory funding requirement for state and local pensions has come from the left. Connecticut's Democratic governor, Dan Malloy — who has struggled with a serious pension gap created by predecessors who did not provide adequate funding — has called for such a requirement. And Raegen Miller, a scholar at the Center for American Progress, has urged making federal education funding contingent on adequate contributions to teachers' pension funds.

While conservatives tend to bristle at federal interference in state fiscal matters, the option to underfund pensions simply creates opportunities for irresponsibility at the state level. It is a way for politicians to take a "buy now, pay later" approach to government spending. And the federal government does have an important interest in state-pension solvency: Allowing states to underfund creates the risk that a state will exhaust a major pension fund and look to federal taxpayers for a bailout.

If state lawmakers knew that pension promises had to be honored and, more important, had to be funded, they would be less likely to make irresponsible promises. And one obvious way to convey such knowledge is through binding federal law. Conservatives should accordingly embrace an ERISA-style funding requirement for states, both to encourage fiscally responsible pension policy and to help build state-level political support for pension reform.

TRANSITION COSTS

Denying states the option to shirk their pension payments would increase the pressure to address pension funding gaps — and to choose pension-policy options that don't generate such gaps. But there would still be major barriers to the adoption of defined-contribution pension plans, which eliminate funding gaps altogether. And on this point, too, changes in federal policy could play a useful role.

In a defined-benefit plan, the government promises a fixed benefit in retirement. While the benefit is supported by a pool of investments, the return on such investment is risky, and that risk is not passed on to the pensioners. Instead, any underperformance of pension-plan assets leads to a funding gap that must be filled by taxpayer money. In a defined-contribution plan, such as a 401(k), the employer's entire contribution is made up front, and the employee bears all risk regarding how the funds are invested. Private firms have found this shift of risk attractive and have generally adopted the 401(k) approach. But few government entities have done so, scared off in part by the "transition costs" they see arising if they close a public-employee pension plan to new entrants.

Why did such transition costs not scare off private employers? As Robert Costrell of the George W. Bush Institute has noted in a recent report, the reason is not a genuine fiscal problem, but rather a matter of government accounting conventions. In fact, the transition costs in question are largely illusory.

With an open pension plan (that is, one accepting new entrants), GASB guidance allows for unfunded liabilities to be covered over a period of up to 30 years, and the state's ARC is determined based on that liability. When a plan is closed, such liabilities are supposed to be amortized over the remaining years of service of the average active employee participating in the plan (in a typical case, that would mean about 12 years). This means the liabilities come to seem greater, and so the required annual payment does, too.

The GASB also recommends allowing open plans to set their contributions as a level percentage of payroll, meaning that if a government expects to have a bigger, more highly paid work force in the future, it can backload its payments. But a closed plan is supposed to use "level dollar" amortization, meaning that equal payments in nominal terms must be made each year.

GASB recommendations are just guidelines, but in this case most states have accepted them. The upshot for reformers is that, when a state or locality closes a pension plan (as it would if it were moving new or existing employees to a new defined-contribution plan), it would need to start making sharply higher pension payments. Opponents of states' switching to a defined-contribution model point this out in order to argue that pension reform, intended to save money, is actually very costly to taxpayers.

But in fact, these "transition cost" arguments misinterpret the states' circumstances in several ways. First of all, following the GASB's guidance to accelerate the amortization of a funding gap does not constitute a real increase in cost. Higher payments in the near term are offset by savings in future years. It is essentially like the difference between a 15-year and a 30-year mortgage: If both loans have the same interest rate, the 15-year mortgage is not "more expensive," though it will require coming up with more cash in the near term.

Even more important, GASB guidance is voluntary. As noted above, states routinely ignore the GASB's recommendations about pension funding. They can also ignore the guidance to speed up paying off unfunded liabilities after closing a plan. If a state doesn't want to face transition costs when moving employees to a new pension plan, it doesn't have to.

Ignoring GASB guidance to accelerate amortization might sound irresponsible, but it isn't. The rationale for speeding up amortization is that an unfunded liability should not be allowed to outlive the payroll base that is used to service it. But as Costrell argues, the cost of amortizing an unfunded liability can be responsibly spread over the payroll of employees participating in the old defined-benefit plan and the new defined-contribution plan. After all, unfunded liabilities in public pensions are ultimately serviced by the taxpayers, who (unlike a private employer) are not going away even if a defined-benefit plan ceases to contain any active employees.

This problem of illusory transition costs — which have played an oversized role in the liberal case against using the defined-contribution approach in public retirement systems — is already on its way to being addressed. The GASB will within the next year be making two important changes to its recommendations that will make it easier for states to close defined-benefit pension plans. First, the recommended calculation of the ARC for all plans will be based on an amortization period equal to the weighted average remaining years of work for active employees. In other words, the amortization period that the GASB currently recommends for closed plans will also be recommended for open plans, meaning that closing a plan will not shorten the amortization window. Second, the GASB will make clear that this standard for calculating the ARC is divorced from actual policy decisions about funding. Plans that currently use amortization periods close to 30 years will not be expected to significantly shorten them or to make the additional payments that such shortening would entail.

These clarifications of the GASB's recommendations will be helpful. But it would also be helpful if Congress — perhaps as part of a move to apply ERISA-like requirements to states, as proposed above — were to codify this new understanding. If the federal government imposes a new pension funding requirement on state and local governments, it should not penalize those that choose to close their defined-benefit pension plans to new entrants.

SAVING FOR RETIREMENT

Advocates of preserving defined-benefit pensions have one other significant argument against the defined-contribution approach, and this one — unlike the dire warnings of transition costs — is not based on an accounting fiction. These advocates note that most 401(k) plans are woefully inadequate: Workers do not save enough for retirement; they use bad investment strategies; and they pay too much in management fees. A system of 401(k) plus Social Security, these people argue, is leaving most workers with too little in the way of retirement savings.

Alicia Munnell and her colleagues at Boston College's Center for Retirement Research have quantified these problems. In 2009, at the trough of the recession-induced stock-market decline, the researchers found that 51% of American households were not on track to be able to replace their working incomes in retirement, up from just 31% in 1983.

Several factors have driven this trend — low interest rates, poor equity performance, growing longevity, and increases in the Social Security retirement age. But the shift away from defined-benefit pensions has also been an important factor. As Munnell notes: "In theory 401(k) plans could provide adequate retirement income, but many individuals make mistakes at nearly every step along the way. As a result, according to the 2007 Survey of Consumer Finances, the median 401(k) / IRA balance for participants approaching retirement was only $78,000."

Retirement under-saving is a major problem, but it is not one that states can fix by retaining their defined-benefit pension systems. So while this concern is a legitimate criticism of the defined-contribution approach to retirement security, it is not a case for the defined-benefit approach. The under-saving problem is economy-wide, and government defined-benefit pensions cover only the less than one-fifth of workers who are employed in the public sector. The cost of financing these systems, meanwhile, falls on the very same taxpayers who are struggling to save adequately for retirement. Public defined-benefit pensions thus stand to exacerbate the difficulty that Munnell and her colleagues identify, not to solve it.

Instead, the federal government should take steps to raise rates of retirement saving throughout the economy. This move would not only ensure greater retirement security for the public: It would also allow state and local governments to dramatically reduce the cost of providing retirement security to their employees, just as private firms have done by closing their defined-benefit pension plans.

There are several ways in which the federal government could increase retirement-saving rates. One is through the direct federal subsidization of retirement saving. Tax advantages for saving could be made more generous, or the government could provide direct cash payments into workers' IRAs, perhaps on a matching basis. But these options would be costly, and would entail adding to the tax burden.

Alternatively, the federal government could mandate additional retirement saving. This could take the form of required IRA contributions, similar to proposals for voluntary private accounts in Social Security. Another option would be to require workers to participate in a professionally managed retirement fund, similar to the system used in Singapore, where a portion of workers' paychecks is diverted into a sovereign-wealth fund managed by the government. Such a system would be cheaper to operate than individually managed accounts — meaning that less money would go to management fees and more to returns for retirees. But such approaches would take a heavy hand, and a co-mingled fund in particular also raises public-choice problems: Would the managers be subject to undue political influence in their investing choices, or try to exert government control over the companies they invested in?

Finally, the federal government could take steps to "nudge" workers toward saving more, without mandating or funding those savings. For example, it could automatically enroll every worker in an IRA with an automatic payroll deduction, but allow individuals to opt out if they chose. It could incentivize employers to use automatic enrollment in their 401(k) plans, and to increase default savings rates as employees age. The government could also incentivize employers to offer 401(k) plans with more appropriate investment options and lower management fees. These nudges wouldn't increase the savings rate as much as mandates would, but they would be less restrictive of individual freedom.

The federal government is in a much better position than state governments are to implement these reforms. States cannot afford to take on major new fiscal commitments to workers' retirements, and the national labor market makes it difficult for them to impose savings mandates or even effective nudges. After all, insufficient savings result from the free choices of individuals: Offered the choice to spend or save, many people spend today even if tomorrow they might wish they had saved. Pushing people to save more makes people better off in the long run, but in the short term, they may prefer to live in jurisdictions that let them save less — meaning that states that get aggressive on mandatory saving may lose residents to their neighbors.

Without some federal action, defenders of traditional pensions will be left with their strongest argument: We have not yet developed an effective alternative to defined-benefit pensions that can enable most Americans to save sufficiently for retirement. But were some sensible federal reforms along these lines to be enacted — ensuring that all workers, including state employees, saved enough — the burden on states to pay for workers' retirements through pension systems would be lightened significantly.

ENABLING REFORM

Until about 2009, the key problem with public pensions was that lawmakers were asleep at the wheel. Pensions are complicated, and when they are not causing huge problems, they are boring; it is thus not surprising that they usually take a back seat to more urgent, hot-button political issues. Consequently, it has been easy for lawmakers to ignore the long-term risks to which they have been exposing taxpayers; it has not been hard for public-employee unions to talk these lawmakers into providing ever more generous benefits.

The good news is that, as demonstrated by the fact that nearly every state is now engaged in some sort of pension reform, lawmakers are no longer asleep at the wheel. But the mere understanding that they need to act is not enough. State lawmakers need to put in place policies that will address the structural flaws and backward incentives that put states in this funding crunch in the first place — flaws and incentives that, if left unchanged, will simply return states to the same position in the years ahead. The problem, of course, is that, at least for now, states do not have all the tools they require to accomplish effective reforms.

Those tools need to come from the federal government. Washington has enormous power to change the pension-reform landscape — by preventing states from engaging in irresponsible pension-underfunding practices, by clarifying the lack of steep transition costs to defined-contribution systems, and by taking action to improve retirement-saving rates across the entire economy. Were federal authorities to make good pension reform easier, states would be in a better position to tackle their pressing pension problems — thereby preventing a catastrophe for their own finances, and the nation's.

Josh Barro is a fiscal-policy analyst based in New York and a blogger for Bloomberg View.