Laboratories of Insolvency

Understanding the Coming State Employee Pension Crisis.

By Jeremy Rozansky


           With the sudden awakening of the American people to the risks of personal debt, our recession politics have seen an outburst of moral financial fervor at our collective debt. The next few years will likely see budget committee confrontations over how to reduce the budget deficit, and perhaps even chip away at the national debt. Yet, the national debt may not be our biggest worry. The debt in state employee pension funds is both more immediate and possibly more difficult to remedy.
The overpromised and underfunded state pensions are likely to require future taxpayers to make up the difference between funds on hand and the promises of the past few decades. If nothing is done by the end of the decade, multiple states, including Illinois, will face unfunded obligations at around 30% of projected revenue, essentially raising taxes by a third with no new programs to show for it. On its face, this obligation is inextricable from the primary roles of state government. Few Americans oppose public education, police forces, or fire departments. State pensions are part of the competitive wages states pay teachers, policemen, and firemen, and are essential to the provision of those necessary services. But the faulty structure of most state pensions means that once insolvency hits, those paying the retirees' income will be the taxpayers of the present—as opposed to the taxpayers of the past who actually benefited from the teachers, policemen, and firemen in their productive years. This amounts to a massive transfer of wealth, and not one aimed at the poor or the elderly writ-large, but a bells-and-whistles pension package given to a certain subset of the elderly, those who decided to work for the state.
This future crisis will be a result of the unique situation of public employees and their politician employers, as well as deceptive accounting. The long-term solution will be to correct those distortions through mechanisms that mirror the pensions of the private sector. The problem, we will see, is that even if the distortions are corrected by responsible governance, the full gap between funds and promises cannot be overcome.

The Defined Benefit Swindle
            Before elaborating on why public sector pensions are so different from private sector, it is instructive to show how the pensions differ. To be sure, private sector pensions have also been insolvent, even to the point of default—Exhibit A is United Airlines' woes last decade. There are two basic types of pension plans: defined benefit and defined contribution. Ninety percent of state and local workers have defined benefit (DB) plans. An ever-increasing majority of the private sector, meanwhile, has defined contribution (DC) plans, as do all new workers in just two states that have already reformed: Alaska and Michigan. There are also hybrid plans in both the public and private sectors.
           The customary example of a defined contribution plan is a 401(k). With a DC plan, the employer and employee have a set amount to contribute (often a percentage of salary) to a fund that pays out upon retirement. The employee can choose what to invest in, and is able to make decisions about the level of risk he is willing to incur and in which financial instruments he wants to invest. Upon retirement, an annuity can be purchased as insurance against outliving one's assets. Some advantages of defined contribution plans include an inability to be underfunded (because there is no payout obligation—you get what is put in) and easy portability.
           Defined benefit plans, as the name would suggest, guarantee the annual benefits for the retired workers generally based on a formula including factors like final salary and years of work. A typical retired public worker will be annually promised the average of the three final years of their salary multiplied by the number of years they are employed by the state and a constant factor, typically .02. This annual benefit is then increased in each year of retirement by a cost-of-living adjustment. Thus a school teacher with 30 years of service and a three-year average final salary of $60,000 will be due $36,000 in his first year of retirement. Some states have reduced the .02 factor, extended the three-year average to a five-year average, capped cost-of-living adjustments, or taken other measures to reduce the amount owed to retired state workers.
           Unlike DC plans, which only pay out what they collect, defined benefit plans first calculate a payout and then try and figure out how to collect for it. The amount owed to the worker is projected by state actuaries and based on a number of assumptions: lifespan, estimated age of retirement, salary growth, and inflation. This calculated obligation lends itself to a further calculation: the Annual Required Contribution, an estimate of what needs to be contributed during the years of work to a fund that will be invested and equal the liability when it is time to pay out to the retiree. An essential element of the calculation of the Annual Required Contribution is the expected rate of return. The expected rate of return, a way of converting future liabilities into present dollars, is assumed by most states to be around 8% annually. Eight percent has, in fact, been the median annual return on pension assets over the past twenty years, but it is not guaranteed. Therefore, in the wake of the 2008 stock market crash (2008 saw 37.22% losses in the S&P 500), the combined state pension plan official estimates show $452 billion in underfunding, with assets of $2.3 trillion and liabilities estimated as a present day $2.8 trillion. Somehow, this gap is going to have to be made up, whether by increasing assets through increased contribution or by lowering liabilities in a variety of ways to be further explored.
           The timeline for reform to fill this substantial gap depends on the states. States will continue to pay out of their assets even if they do not grow at expected rates or project to meet liabilities. There is no automatic mechanism to assure solvency, and the attrition will begin in the next decade. Even if the states hire no new workers and the pension funds earn 8% annual returns, Illinois and three other states will run out of assets before 2020. In 2030, 31 of the 50 states will run out of assets and resort to tax revenues to fund the remaining liabilities.

Measuring our Debt
           The acknowledgement of $452 billion in present dollars in pension underfunding sounds like a disaster, but realistic accounting makes the problem harsher. States are still assuming that $1 today can become $1.08 next year simply because 8% has been a recent median of pension growth. When it has been hard to match the 8% expectation, the boards that operate the plans have invested in riskier assets like private equity, paying a steep price in 2008.
           Economists agree that risk is a factor that should lower the expectation and thereby increase the present value of the future liability, changing the methods of accounting. To demonstrate this point, consider the following two options in an investment game. In each, we are investing in the present day to try to reach a $100 liability owed in 20 years.
           In the first, we will put the investment in stocks, which carry risk. We target an 8% rate of return so we assume an 8% discount rate. In this situation, the present value of the future $100 is $21.45; we wash our hands and declare ourselves fully funded, retiring from the pension board before we ever need to be held responsible for meeting the payout.
           In the second, we will put the investment in a risk-free account like 20-year Treasury notes, implying a discount of 4.5%. To fund $100 in twenty years we will have to put away $41.46 now.
           Why is there such a disparity? Shouldn't the present value of $100 in 20 years be the same? The answer is risk. The stock buyer is not assured of $100 in 20 years. If he wanted to assure himself of $100, he could buy the equivalent of insurance. This is called a "put option." The put option would cost the stock-buyer $20.08, leading to a total cost of $41.53, a number slightly higher than the bond buyer's $41.46 in part because the stocks could earn more than $100.
           To have a responsible and realistic estimate of our actual level of underfunding, we need to convert estimates of future liabilities accounting into present value. To do so, we must apply the correct discount rate. This rate, in fact, depends on the possibility that the obligations could be defaulted on; if taxpayers could buy the insurance of a put option in our model, they would be willing to pay more for the insurance the more likely it is that they could not merely default on the debt. As we will show later, it is incredibly unlikely that the states will be able to default on these obligations—there is no precedent for such a thing. Thus the discounting should mimic a default-free promise like Treasury bonds. A fifteen-year Treasury bond will have a 3.5% discount rate; multiple economists recommend that this discount rate be used. If 3.5% rightfully replaces 8% in our equation, our $2.3 trillion in assets no longer needs to catch up to $2.8 trillion in liabilities. Instead, we will have to meet $5.1 trillion in liabilities. No longer are states underfunded by a half-trillion dollars, but by $2.8 trillion. A more recent tally suggests that there is only $1.9 trillion in aggregate state assets, so we can safely speak of a $3 trillion hole.

Few Options for Reform
           The question now before us is how we can reduce these future liabilities. Many options have been tried on a state-by-state basis, with the more diligent reformers considerably delaying the date of insolvency. Two economists, Joshua Rauh of Northwestern University and Robert Novy-Marx of the University of Rochester, have looked at the effects of various reforms on an aggregation of state pension plans. Their findings tossed cold water on even the most aggressive reformers.
           Cost-of-living adjustments (COLAs) are a popular source of reform because they affect current retirees, where the bulk of the present problem lies. The benefits given to a retiree under the common defined benefit system are predetermined with a cost-of-living adjustment raising the annual payout to keep the benefit's purchasing power equal from year to year. In other words, there is an annual increase in payouts to a retiree of the inflation rate or another factor, say, 2%. Rauh and Novy-Marx estimate that a decrease in the COLA by one percentage point would knock liabilities down by 11%. Some states, however, have already chipped away at their COLAs. Georgia has even eliminated COLAs for new workers, while new workers in Illinois have COLAs capped at half of inflation. These are steps in the right direction that show COLA reductions are politically possible, but because COLA reductions have already been enacted, additional COLA reductions would have a more muted impact. COLA reductions are both politically tenable and they have the ability to lower benefits for current retirees (the bulk of the problem).
           The early retirements disproportionately common among public employees have given fodder to public outrage. On this issue we are quick to cite the astonishing number of ex-police officers in New York City who are collecting benefits and not yet 50. While such examples are real and demonstrate gross inefficiencies—how much are we paying them over a lifetime for 25 or so years of service?—the larger problem is white-collar public employees retiring between age 55 and 65, while still able-bodied and vigorous. Such cases of early retirement are generally incentivized by the defined benefit plan's structure. In the median case, for each additional year of early retirement, 3% is deducted from the employee's benefits. This amounts to a 3% reduction in annual benefits in exchange for an extra year of benefits. Combining the reduction in annual benefits with the increase in years retired, the average 60 year old will gain 4-6 months in total benefits for each additional year he chooses to retire. The pension fund is paying more in benefits so that the employee works less; something has been severely mismanaged. Making early retirement actuarially fair (i.e., early retirement is neither incentivized nor disincentivized) makes sense and is probably politically feasible, but only affects current workers and would reduce liabilities by just 1-2%.
           For those familiar with the Social Security debate, a basic fact of reform and a refrain of any reformer is that we are living longer now than we were when contours of the system were first worked out. Many contend that the retirement age must be lifted to keep up with the ever-growing life expectancy. However, years-of-vigor expectancy has not increased proportionally to the increase in life expectancy. Our ratio of workers to retirees will only continue to get lower and lower. Raising the retirement age should still be periodically done, especially in white-collar professions, even though it will never fix the problem completely. Many states have full retirement ages below even 65, absurdly behind not only the private sector but also the federal government. Rauh and Novy-Marx estimate that moving state pensions immediately to Social Security parameters—early retirement at 62, full at 67, and (for the most part) actuarially-fair retirement—would reduce liabilities by about a quarter. However, the pension funds would still be $1.5 trillion underfunded.

Making the Case: The Touchy Politics of Pension Reform
           All this doesn't even begin to address the political obstacles. The Social Security parameters would have to be phased in to account for public employees who are well into planning for a retirement under current parameters. This would thus lessen the impact of such radical reforms. The most aggressive pension reformer, Gov. Chris Christie of New Jersey, has proposed eliminating COLAs, increasing the retirement age to 65, reducing the discount rate to 7.5% (it should still be lower), and increasing employee pension contributions to 8.5%. That Governor Christie stops two years short of Social Security parameters is a telling sign: such parameters are not politically tenable, at least not now.
           These reforms are met with massive opposition by widely sympathetic groups: public employee unions. These unions include the nation's largest, the National Education Association (NEA), and the 2010 election cycle's biggest spender, the American Federation of State, County, and Municipal Employees (AFSCME). These groups have the ability to fund ad campaigns against the politicians who seek to shear state pensions. Democrats are particularly vulnerable considering how much they rely on public sector union's organizations, especially after the decades' long exodus from the once strong base of private sector unions. Since politicians so rarely bite the hand that feeds them, the pensions have escaped a regular trimming.
           Popular perceptions of fairness are also very important to any reform's success. Politicians who plan on reforming the pension system need to be able to change popular perceptions about public employee compensation, especially in the case of teachers. Chris Christie's example is particularly masterful. Any pension reform will, by definition, reduce the wages of public employees (wages are equal to the sum of salary and benefits). That is to say, any pension reform will decrease what schoolteachers are paid. This sounds like a grave injustice, victimizing the most altruistic among us. But for most individuals choosing between working in the public or private sector, the public sector will give higher total wages during a lifetime of work. Salaries are lower but benefits are much higher for public employees. Therefore, reducing pensions will actually bring current workers' wages on par with the private sector, even though public employees' wages still will be slightly more backloaded.
           Having more backloaded pensions than an average market employer befits the political process. Politicians are judged on their budgets in the short term, not the state's long term fiscal security. To get more teachers now, a state's leaders will move a greater chunk of their employees' wages until after retirement. In many cases, state governments have even shifted the responsibility for fund contributions onto future generations by deferring contributions to funds. The ability to make promises for future governors to carry out is part of the attraction of defined benefit plans for the politicians.
           The politicians independent of the public employee unions (generally Republicans, although Republicans often cater to police and firefighter unions) will be in the best position to make such reforms. However, not all Republicans will be immune to union attacks, and few can use the bully pulpit well against the seemingly altruistic. Perhaps state governors—both Democrat and Republican—will find the unions more amenable if they wield the bully pen, namely, the power to decertify the public employee unions via executive order, thereby withholding their privilege to collectively bargain. This is not simply quid-pro-quo politics. There is a reasonable case against the existence of public employee unions. One need not look further than Franklin Roosevelt, who argued that collective bargaining could not "be transplanted into the public service" because no power in any state should dictate terms to the government, the embodied will of the people. Moreover, it might do some good for the employees themselves who are currently forced to fork over thousands to the union in annual dues that fund the union's apparatus and aid political campaigns the employee may not even support.
           After making this case with pulpit and pen, governors should pursue incremental reforms: raising the retirement age, increasing the penalty for early retirement, cutting COLAs, changing the calculation of final salary from an average over three years to five, eliminating the counting of untaken sick days, and increasing the employee contribution. This will reduce the problem; in some states, it will do so entirely.

Filling the Hole: Ambitious Reforms
           The political challenge might not be so great; some public sector unions have supported minor cuts so as to not "kill the goose that lays the golden egg," as a Minnesotan AFCSME official said. "Killing the goose" for them means ending the defined benefit plan that guarantees such a rich payout. While these incremental steps should certainly be taken for current workers, for workers not yet qualified for pension guarantees, a defined contribution plan that mimics the private sector's makes good sense. Public employee unions would not like it, and there is no reason why they would: moving state employees to defined contribution plans would end the pattern of extraordinary, bank-busting benefits for state employees. Some public employees would be happy to invest their pension, but one imagines most recognize that they have been promised something pretty lavish.
           Defined contribution plans, however, correct the major distortions that have brought the current crisis about. For one, politicians would no longer be able to put off wages in the form of generous pensions. DC might lower wages for public workers by lowering benefits (this is not necessarily the case, however; a 2006 study at MIT said that because of lower administrative costs, among other reasons, private sector DC tends to earn better than DB). Yet it is more likely to focus the conversation on public employee salaries and stop the sleight-of-hand backloading of wages that appeases the unions. Pensions would not be invested in such risky assets trying to meet guaranteed levels of asset growth; rather, individuals could choose the level of risk they felt comfortable with. No longer would states be able to hide behind gimmicky discounting rates—you have what you have. In addition, states would not be pressured to increase the benefits during bubbles, only to have them burst and be unable to retroactively decrease the guarantee—this would have protected California and other states that thought the tech bubble permanent. DC would also increase portability, allowing for greater transition of talent between the public and private sectors. The goose-killing caucus has it right.
           While moving new and unvested workers onto defined contribution plans would make sure our pension problem never reappeared, we still need to deal with the $3 trillion gap. Looking at the above reforms with a dose of prudent pessimism, it seems a nation of Chris Christies with sympathetic legislatures might approach an underfunding of $1.5-2 trillion. States would still have to divert a fifth or so of revenue to pay pension obligations, and no state is ready to cut spending by a fifth; the states that can constitutionally run a deficit would, and the states that cannot would sell poorly rated bonds as a quick fix. Thinking long term, the only thing states could do is pray for 9% investment returns every year.
           But what about default? United Airlines gave its unfunded pension a one-third default in bankruptcy court. Is that what the states will do, dwarfing United's puny $3.2 billion settlement? The answer is no. States are not incorporated entities like municipalities and corporations; they are sovereign entities. By their state constitutions, paying back bonds is, in most cases (with the exception of educational spending), the primary priority of state governments. Even in the municipal bankruptcies of Orange County, California and New York City, municipal pensions were untouched. There is literally no precedent of rewriting public employee pensions in fiscal peril—other things are cut to make room, including bond repayment. Most likely, states will not be able to pay their bonds and the federal government will deem states too-big-to-fail and bail them out—this is, at least, what California bond holders are banking on.
           If the federal government is destined to bail out roughly two score states, then perhaps this is an area where federal power can be used wisely to preempt the worst of the calamity. Bailing out those states that switched to defined contribution plans too late, offered to increase benefits in times of plenty, or otherwise mismanaged the requisite contributions will switch the burden from the responsible states alone to everybody, including the better managed states—a dangerous moral hazard but a likely a necessary one. If everyone is going to be penalized regardless of merit, the penalty should at the very least be lessened. Thus the federal government should provide a carrot (e.g., grants or tax-subsidized bonds for the pension accounts) in exchange for meeting several benchmarks in their pension accounts: e.g., COLAs less than or equal to inflation; full retirement age of no less than 65; and all new workers on defined contribution plans. The benchmarks should be general enough to avoid one-size-fits-all solutions. While the federal government would be increasing its debt, the federal government is a sovereign entity that also can print its own money, giving it an outlet, albeit a noxious one, from the paralysis of pension obligations.
           The Governmental Accounting Standards Board should also mandate that state employee pensions discount liabilities based on the risk-factoring market value—akin to the 3.5% discount rate mentioned above. This would, at the very least, end the systematic devaluation of liabilities for the sake of present politicians' political cover.
           There are two additional options that can certainly be combined to meet the gap in public pensions: revenue increases and retroactive reductions in benefits. Let's say every state immediately makes the reforms it needs to make and the gap between aggregate assets and liabilities is down to roughly $1.5 trillion. Tax-increases would be needed to raise revenue. Without reform, a state like Illinois would owe $2,000 per household each year after the fund goes insolvent in 2018. While reform would lower that number, the tax would still increase the pressure on many families and cause many to leave Illinois for states not set to go bankrupt for decades, like nearby Iowa and Wisconsin. There may be a place for tax increases, but states should be fearful of many negative consequences.
           Retroactive reduction is at once the most controversial and the most just solution. Based on accounting that used an 8% discount rate, multiple states fed the present crisis when the boom of the late nineties caused (according to accounting that used an 8% discount rate) pension assets to exceed liabilities. So that the state did not make money from the employees' pensions, the states retroactively increased liabilities, rebalancing the funds. Now in the lean years, the same accounting shows liabilities exceeding assets. Could we not retroactively reduce liabilities? One possible method is a constitutional amendment in the states that reduces the benefits of workers and current retirees. Another possibility is a court challenge by the taxpayers in which the contracts are deemed unfairly obtained and rewritten. The challenge could (correctly) allege that the contracts are products of self-dealing—the politician-public union relationship considered akin to a CEO who hires his wife at a bloated salary. While public interference in private contracts is often accepted by courts, public contracts are thought to be protected from public interference, as courts are wary of making all public contracts vulnerable. Precedent therefore tells us that the odds are against courts throwing out the current contracts, although it is not a foregone conclusion. What courts would decide in the eventual emergency remains the one trillion dollar question.
           While the agenda to pursue is not entirely clear, the extraordinary heft of this problem may require extraordinary means of solving it. It will require both a political program in addition to the economic one: using more stick than carrot with the labor unions, co-opting the language of fairness to properly portray the problem at hand. It will test our legal procedures. How sacrosanct are these promises? At what social cost do state obligations become untenable? We may need to live with the moral hazards we create, with justice misapplied. This will not be a fiery trial, but a mundane slog through actuarial tables. Still, we need steadfast leaders. The state pension crisis isn't about simple tweaks in accounting practices, but about past irresponsibility and the historic shame that will be ours if our offering to the next American generation is intractable debt.