Funding Public Pensions

Problems With Accounting Rules

Problems With Accounting Rules 

“The pension fund could run dry,” is a common enough talking point that it could be made about almost any pension fund in the country. A 2016 Google search for “pension fund run dry” provides ten hits, eight of which name Pennsylvania, New Jersey (twice), California, Chicago (twice), Texas, and Alabama with the remaining two hits leading to articles about the national pension funding crisis. Some of these plans named in these articles may be in real trouble, but reading them suggests that most of the writers are under the mistaken impression that a system funded at anything less than 100 percent is necessarily in danger of running out of money at some time in the near future.

“The state’s pension goliath, the California Public Employees’ Retirement System, had $281 billion to cover the benefits promised to 1.3 million workers and retirees in 2013. Yet it needed an additional $57 billion to meet future obligations.”26

“Before the crash, retirement systems were underfinanced (they did not have sufficient funds to pay promised benefits), but the day of reckoning was distant.”27

“ ‘Their benefits are in question,’ said Gary Wagner, a professor of economics at Old Dominion University.”28

Some experts will see these as alarmist statements. GASB members themselves might say these writers are misunderstanding the rules, even if they agree with the conclusions. The GASB Statements 67 and 68 specifically say they are only about reporting and do not dictate funding. And yet, stories like these appear across the country on a near-daily basis. The clear conclusion is that the effect of GASB rules is not just on the construction of a balance sheet, but on the interpretation of the numbers found there and the actions of the parties who make those interpretations: policy makers, citizens, or bond-rating agencies. 

If GASB itself is not the enforcer of misinterpretations, if the enforcer is city council members preening about their soi-disant fiscal responsibility, or analysts at Moody’s determined to justify a downgrade, or newspaper columnists looking for a good hook, they are doing so with tools supplied by GASB. It is disingenuous to erect a framework of tough rules and disavow their consequences. In this case, the consequences are a drive to full funding, whatever the cost.

It is crucial to understand, now and during the discussion to come, that the goal of the GASB rules is not to bankrupt governments, or to eliminate pension systems. At their root, the GASB rules are meant to create an accounting framework through which the cost of government and the cost of any individual employee are made clear. Unfortunately, the framework erected has created more problems than it has solved. The problem is not merely that the rules are commonly misinterpreted. GASB accountants have acted to insulate public plans from risks they do not face, while simultaneously failing to insure them against risks they face every day. Furthermore, by trying to bring clarity to the accounting, the rules undermine the benefits of aggregation—the whole rationale for a defined-benefit pension plan.

Again, this is hardly meant to say that there have not been improvident politicians and unwise bureaucrats, but the pension “crisis” currently affects responsible and irresponsible governments alike. Two decades of disastrous experience with the GASB pension accounting demand that we examine the rules themselves. We categorize the problems as legal, chronological, actuarial, mathematical, financial, economical, political, and philosophical. We examine them in this order.

Legal: Governments will not be liquidated

Beginning in 1994, with Statements 25 and 27, the GASB rule changes about public pensions were made to mimic rules in the private sector. However, when applied in the public sector the rule changes make pensions more expensive than necessary. 

Consider the issue of full funding. A fully-funded pension system can, at least in theory, pay off all its current debts with no further contributions from the sponsoring employer. This is vital in the private sector because at any time, a private corporation can go out of business, be liquidated and disappear. Full funding and custody by a third party is the only way to make sure a pension granted by such a business will be paid. A pension system in the private sector must be fully-funded in order for the promise of the pension to mean anything at all.

By contrast, a government will not disappear in the same way. To claim so is only to agree with, among others, GASB itself. In a 2006 paper called, “Why Governmental Accounting and Financial Reporting Is—And Should Be—Different,” they defend the difference between governmental accounting standards and those appropriate for the private sector. Early on, the authors point out that the lack of a threat of liquidation is among the primary differences:

“[M]ost governments do not operate in a competitive marketplace, face virtually no threat of liquidation, and do not have equity owners.”29

A hundred years from now there will still be a New York City, even if its area has been reduced by rising sea level. It may have suffered a bankruptcy—maybe two or three—but bankruptcy is not liquidation. It may have been split into its five boroughs, or it may have been overtaken and merged with a rapidly growing Yonkers, or maybe even taken over by the state. Unlike the liquidation of a private company, each of those possible transitions, however absurd or unlikely, leaves a successor to assume the responsibilities of the previous government. Insurance against the city’s disappearance is therefore a waste of money.30

Chronological: Present value masks the level of urgency

Pension accounting relies on a presentation of assets and liabilities in their “present value,” the value today of a sum of money tomorrow. At a 5 percent annual discount rate, a present value of $100 today corresponds to a future value of $105 a year from now. The appropriate discount rate depends on your estimates of inflation and potential investment returns, so there is a degree of subjectivity in any present value calculation. Nonetheless, there is a time dependency of the value of money, so it makes no sense to compare 2016 assets with 2046 liabilities, except by computing the present value of the liabilities to compare to the current value of the assets.

Unfortunately, though the concept of present value is a way to translate future funding into the present day, the calculations made to describe the present value of a pension debt elide important issues surrounding a payment schedule, even beyond the issues of subjectivity. By definition, present value captures the monetary value of a future series of payments, but it fails to capture the urgency of those payments. By making all calculations in terms of present value, pension accounting rules imply an equal urgency to all debts, something that is obviously not the case. For a pension liability, the relevant payment schedule not only extends out decades into the future, but it also extends decades longer than the 30-year amortization periods required by GASB 27. The last payment owed by any pension system will not be made until the youngest current employee dies. If the youngest employee is in their 20s, this could be more than 60 or 70 years in the future. For systems that offer survivor benefits, it could be longer than that.

Consideration of the payment schedule is important to the stress of a debt because two debts with the same present value can require dramatically different plans for payment. Imagine a debtor with assets of $600 and a debt of $1000, due tomorrow, and compare him to another debtor with the same $600 in assets, but who owes 

This infographic displays the funding ratio of debt

The same present value and the same funding ratio can be a crisis or not, depending on variables that the present value of a debt does not capture. One of these persons is in serious trouble, and the other only mildly concerned, but both have debts with the same present value.

$19.72 per week for a year. Assuming a discount rate of 5 percent, these debts have precisely the same present value, and therefore precisely the same $400 unfunded liability. Yet one debtor is in much better shape than the other. The first debt could justifiably stimulate panic; the second is certainly worth more than a yawn, but much less than panic. Action is required, but the debtor has until week 32, over seven months away, to mull over action or to raise additional funds. The Illinois pension debt is $111 billion, but this need not be paid tomorrow, even if GASB 68 will have it appear on the same balance sheet as debts that are due tomorrow, or even past due. Accountants might claim these two debts are equivalent, but is this really the case?

As an incidental point, one can note that the traditional 30-year amortization schedule is only that: traditional. A pension system that dutifully follows such an amortization schedule will see its debts prepaid decades before they are actually due. Paying them in advance is not necessary to making payments, and yet it is the accepted wisdom. One searches in vain for any author presenting a reasoned justification for equating 30-year debt with immediate debt, or using a 30-year term to pay off a 60-year debt. These rules represent little more than blind acceptance of precedent.

Actuarial: Full funding is not required to pay all pension debts

The drive to full funding cannot be justified actuarially, either. Though the details depend on actuarial characteristics of the employee and retiree population, many, if not most, defined-benefit pension systems can operate forever at far less than full funding. A retired teacher in Chicago who passed away in 2014 after a long and happy retirement had every penny of her pension paid by a system far below full funding, and yet all her pension checks cleared. A system at 70 percent funding can likely pay all its obligations in a given year, and if at the end of that year it is at 70.1 percent, who is to say this cannot be repeated the following year if the actuarial facts on the ground do not change significantly? Social Security operated at what amounted to a few percentage points of full funding in its trust fund for two generations and only a very few pension plans are funded at levels so low.31

To put it more rigorously, the normal cost to a pension plan accumulated within a calendar year is the present value of the additional benefits accrued by all the employees in that year. If the contributions to the fund (employer and employee contributions, as well as investment income) are adequate to offset the normal cost and inflation, then the unfunded liability of a plan will not change from one year to the next.32 If the unfunded liability does not change one year to the next, the fund can operate indefinitely with that same unfunded liability. 

A pension fund must pay 100 percent of its debts. But it need not pay them a moment before they are actually due, and since a pension plan is constantly receiving new contributions, the fund itself need not be the only source of payments. As a result, even if all the debts are paid, at any one time, the fund itself may be at some level well below 100 percent funding. 

Recall the example above. Perhaps I took a $1,000 loan from you, promising in return to pay you $19.72 per week for a year. If I have only $600 in the bank, then I have an unfunded liability of $400. If I also have some source of income of just $7.96 per week, I will be able to pay 100 percent of this debt, down to the penny, out of the combination of my income and my savings. Every step of the way, my funding ratio—the ratio of my assets to the present value of my remaining debt— will be 60 percent or less. (See figure on page 13.)

This is a toy example and tracks the debt to only a single party. A pension system might have debts owed to tens of thousands of members or more, all owed on their own schedule. The debt estimates are also subject to considerable uncertainty, since the demographic mix of employees and retirees changes over time, too. The principle, however, is the same. So long as there is another source of income, the ratio between the fund and the present value of the debt has little to do with how much of that debt is ultimately repaid. An active pension system has three sources of income: (1) the contributions from the employer, (2) from the employees, and (3) from the returns on investment. Belaboring a point of arithmetic like this seems a waste of time, but the implication that the funding ratio has some relevance to the full repayment of the pension debt is not only a staple of public pension criticism, 33 but is enshrined in official policy statements from bond-rating agencies, actuaries, and the National Association of State Retirement Administrators.34

A 2008 report issued by the Congressional Government Accountability Office (GAO) agreed with the assessment offered here:

“Most public pension plans report having sufficient assets to pay for retiree benefits over the next several decades. Many experts and officials to whom we spoke consider a funded ratio of 80 percent to be sufficient for public plans for a couple of reasons. First, it is unlikely that public entities will go out of business or cease operations as can happen with private sector employers, and state and local governments can spread the costs of unfunded liabilities over a period of up to 30 years under current GASB standards.”35

Mathematical: Financial reports deserve more precision

Another objection to GASB 68 has to do with precision. A pension plan’s unfunded actuarial liability is a planning value, based on dozens of assumptions about market performance, population mortality, and the life choices of hundreds or thousands of employees. Comparisons of identically-calculated planning values like this are tremendously useful for identifying trends in funding progress or lack thereof. However, the accuracy of these numbers is contained in bounds much larger than is normal for other numbers found on financial statements, such as accounts payable or bonded indebtedness. Pension planning values are a guess about the future, useful primarily to compare to each other. Financial statements are a record of the past. Despite the statistical methods used to develop these estimates of future liabilities, actuarial and accounting precedent do not demand the explicit reporting of the error bounds, as one sees in other statistical estimates, such as polling data. Rendering these inherently inaccurate numbers as part of a government’s statement of net assets, as demanded by GASB 68, dramatically reduces the accuracy of that statement’s bottom line. 

The issue is not the mere uncertainty of the numbers. Accountants use uncertain numbers in many of their calculations. The problem is that uncertainty is infectious. Adding a high-precision number to a low-precision number results in a low-precision number, so the result is low-precision financial reports.

Given the size of pension debts, adding them into the total can make the actual value of a government’s net assets vary significantly from the stated value. According to its 2014 financial statements, the unfunded liability for Illinois pension funds is estimated at $111 billion, but the potential error in that estimate is as large as the absolute value of the state’s net position of minus $45 billion. Depending on the specific fund, the GASB rules will have them use an assumed rate of return between 3 to 6.5 percent. These are conservative assumptions compared to many peer systems, and to their current rates of between seven and eight percent. A difference of only a single percentage point between these guesses and the reality of the next few decades will change the state’s bottom line by over $20 billion.36 It is not typical to think of a statement of net assets as potentially uncertain by more than 50 percent of the bottom line, but that is the new standard of accounting according to GASB 68. 

Financial: Rate of return is for the longest term GASB 68 specifies that many pension systems below a 100 percent funding ratio must use a “risk-free” rate of return as the discount rate for estimating its future liability.i This liability is the number that must appear on a government’s financial statements, alongside the more traditional components of its liabilities, like bonded indebtedness, and accounts payable.37 The risk-free rate of return is what one can count on earning on investments with no risk. More or less, this would be a portfolio that is entirely invested in US Treasury bonds or the equivalent: no stocks, no private equity funds, no commercial paper, no hedge funds.ii

Statement 68 does not insist that a fund actually be invested solely in Treasury bonds, only that it use that rate to predict its future liabilities. A fund can continue to use whatever funding strategy its managers see fit to use. But the liabilities will be calculated using this lower rate of return, thus will appear much larger than if calculated with a higher rate. This is important because GASB 68 requires the entire unfunded debt to appear each year in the government’s balance sheet. Before this statement, under the requirements of the earlier Statement 27, a government only had to acknowledge whether or not it had made the ARC, the appropriate annual payment, to the fund. 

There is an important point that is often brought up here, that the discount rates commonly in use for pension funds around the country are too high, and the modern world of low interest rates and low investment returns is here to stay. If it is indeed impossible for a well-managed portfolio to average 7.5 percent returns over the next few decades, then 7.5 percent should not be used. 

The debate about what will happen in investment markets over the next 50 years often seem unnecessarily heated. The debate is between people who correctly point out that these numbers have been achievable for decades and people who claim that things are different now. While it is true that many pension systems have been able to meet their marks over the long term, it is also true that we have suffered a decade of reduced investment returns, and the prospects for improvement are not obvious. 

In truth, neither side of this debate has any better claim than the other about the future. Both sides are defensible, and a determination about who is right can only await the coming decades. It is certainly true that a system that assumes a low but achievable rate is more conservative than a system that relies on a standard that might or might not be reached. But these discussions frequently elide an important point: few such debates are about how to design a brand new pension system. Rather, debates about such issues are debates about how to manage the systems we have already. Were one to consider establishing a new pension system, certainly choosing a low rate of return will be a good idea. It will make the system more expensive to run, but it will be more secure, too.

But what will the effect be on an existing system to dramatically lower the discount rate? This is not a small step to consider, as it will increase that overall liability considerably. A $10 billion future liability over 30 years at a discount rate of 7.5 percent will see that liability balloon to well over $15 billion with a 5 percent rate. Upon the adoption of GASB 68, the average funding ratio will decline around ten or fifteen percentage points as hundreds of billions of dollars of “new” liability is recognized under the new rules.39

Every cent of that new liability will appear on the governments’ statement of net assets. For many leaders and critics of their governments, showing a huge debt on the bottom line will be too much red ink to contemplate with equanimity and political pressure to do away with these obligations will build further—with potentially destructive consequences. A rate too high risks underfunding which may lead to higher taxes in the future, but a rate too low risks political pressure which may lead to reduced or eliminated benefits in the future. Readers will differ about which is the more salient risk.

In other words, GASB 68 creates large disincentives for governments to use a lower rate at the same time it requires them to do so. This is a recipe for unnecessary political crises across the country. It is vital that government financial reports be clear about what a government’s debts actually are, but it is also vital that these statements should not be misconstrued. For all the reasons outlined here, there is a substantial downside risk to acknowledging reductions in the assumed rate of return. If changes like these are to be made, they must be made slowly, as financial and political circumstances permit. These are systems meant to be run in perpetuity; most can afford to be patient.

One can see a cautionary tale in the woes of the US Postal Service and its retirement systems. The USPS is technically not bound by GASB accounting rules, but in late 2006, President Bush and the Republican Congress passed a law to force the system to estimate its retiree and health-care liabilities 75 years in advance. The requirements insist, that is, that the system account for the retirement expenses not only of postal workers not yet hired, but of those not yet born. Increasing the funding horizon so dramatically and suddenly is essentially the same kind of shock to the system that an abrupt increase in the discount rate would create. In the ensuing decade, the Postal Service has largely managed to fulfill its funding mandate. As of 2015, the USPS had over $335 billion saved, covering over 83 percent of the liabilities anticipated over 75 years, under very conservative assumptions about the discount rate and value of current assets.39 But the price has been high, and the service incurred $51.7 billion in operating losses between 2007 and 2014 as a result. These losses have shorted new capital investment and service expansions and left the service open to persistent charges that it is an obsolete money-loser at the same time it was forced to put aside a breathtaking sum of money. 

In a 2014 interview, the deputy director of the California State Teachers Retirement System (CalSTRS) pointed out his fund had averaged better than 7.5 percent returns for decades. Under the new rules, he said they must use a rate of about 4.5 percent, increasing the present value of their liabilities by more than 50 percent, but that the portfolio would go along, earning its 7.5 percent as before. In what way, he asked, do the new rules provide an accurate picture of that fund’s condition?40

Economical: Comparisons to pension debt should be chosen properly

Consider again the Illinois pension plans’ unfunded liability of $111 billion in 2015. This is a vast sum of money, especially for a state with an annual budget of only a bit more than half that amount. However, this debt took decades to accumulate and it will be decades before it must be paid off completely. Most, but not all, of the debt will be paid out over the next 50 years. Over that time period, using a very conservative 2 percent inflation estimate, the state’s income tax collections alone will be in the neighborhood of $1.1 trillion and the total state budget will involve spending well over $4 trillion. Therefore, this supposedly colossal debt in reality constitutes about 2.5 percent of the state budget. Personnel costs are around a quarter of the Illinois budget, so this is roughly 10 percent of the payroll costs over that time period.41

Stepping a bit further back, the Illinois economy is much larger than the state budget, and over those same 50 years, can be expected to produce around $64 trillion.42 That is, another way to look at this debt is that it is 0.17 percent of the state’s gross product over the term during which it will be paid, a much less frightening number. These are not spurious comparisons; the state’s economy and the revenue it receives are precisely the resources the state will use to pay this debt.

A roughly equivalent way to state this objection is that the GASB rules do not acknowledge as an asset the strength of the local economy and the ability of its taxpayers to pay in the future. Ensuring that the government can pay its obligations in the future is virtually the entire goal of the GASB accounting rules, but the rules are narrowly drawn so that only qualifying funds kept in trust are counted as a strength. Is it necessary to take such a narrow view of the ability to pay? Another vital difference between a government and a corporation is that a government has a claim on the potential future income of its citizens that no private corporation can make. Indeed, this works in both directions, since money withheld from the economy in the near term, say by raising taxes or cutting schools to make inflated payments to a pension fund, can reduce economic growth and thereby make pension payments more onerous in the future by reducing the size of that future economy.43

Political: Overfunding is a risk, too

A serious risk that does not get much attention at present is the overfunding of a pension plan. Such a risk may seem almost laughable given current circumstances, but it is a serious risk, worth serious concern, not least because it is the goal. The primary objective of pension funding policy choices from GASB to city hall is full funding and full funding is one good investment year away from overfunding. Indeed, full funding is arguably a synonym for overfunding. 

The primary concern is that overfunding is a tangible waste of resources, money unnecessarily diverted from other priorities, but there are other potentially disturbing consequences. A pension system—especially a well-funded one—is not insured against the depredations of politicians, for whom the near-term cost of a pension gift or skipped payment is quite low, if not zero. The GAO report cited above puts it this way: 

[S]everal [experts] commented that it can be politically unwise for a plan to be overfunded; that is, to have a funded ratio over 100 percent. The contributions made to funds with “excess” assets can become a target for lawmakers with other priorities or for those wishing to increase retiree benefits.44

In other words, it is virtually a law of nature that an overfunded pension plan—or any plan over, say, 90 percent funded—will see retiree benefits increase or budgeted contributions decrease.45 In the context of full funding, these changes will have little or no current cost. And as predictably as the sun rises, after the next investment downturn, it will be an underfunded pension plan again, but now with a government budgeting for lower payments and retirees accustomed to higher benefits. This will unavoidably cause a delay in raising payments to catch up.46 This is precisely how events worked out for CalSTRS, the giant pension fund for California teachers. Fully-funded in 1998, its investment returns plummeted when the tech bubble popped in 2000–2001, but not before the state cut its payments into the system and increased some classes of benefits.47 The system currently has a $73 billion unfunded liability and a 68 percent funding ratio, quite a fall from 1998. 

CalSTRS was hardly alone in its experience. The Chicago Teachers Pension Fund was fully-funded in 1995, which was used to justify a ten-year “holiday,” dramatically reducing payments into the fund. Even by 1999, the system was still full-funded. Along with this holiday came increased benefits and management expenses and those, combined with the two colossal incidents of financial market turmoil since then. As of 2014, the plan is 51.5 percent funded.48 The pattern is reflected in national averages. Census Bureau data shows that 1997 was a high-water mark for employer contributions to public plans, many of whom cut contributions in the warm glow of full funding, but had to restore them by 2003.49

Under the GASB rules, a mayor who chooses to skip a pension payment or settle a labor dispute with an ill-considered increase in pension benefits might incur the displeasure of a rating agency in future years. That, in turn, might raise the cost of borrowing down the road: a problem for a future mayor. Furthermore, many cities currently have what are essentially junk bond ratings. For these cities, the ratings risk only barely rises to the level of a material concern, and these consequences are only potential, not certain. In other words, the cities most likely to skimp on their pension contributions are the least likely to be harmed by the consequences in the near term. 

For CalSTRS, the Chicago schools, and so many others, reduction of the dollars flowing into the pension systems harmed the health of the funds, but in each of these cases the effect on the current budget of the sponsoring government was zero. In fact, to the extent that a skipped payment or a new police department contract relieves financial pressure on the city, the effects can be positive, in the short term. A good accounting system is supposed to provide an accurate picture of an organization’s financial health and a useful guide to action. In these cases, the accounting system guides its users to destructive and inappropriate action. 

We have already seen how the GASB rules insure against the liquidation of a city or state, a risk that does not exist. Here we see the rules simultaneously failing to insure against risks that public systems do face. In some cases, the rules even encourage those risks. The divergence between the risks insured and the risks actually faced is not merely ironic; it is a recipe for failure. The insured risks never materialize while the others always do, given enough time. In this case, failure means further increases in the cost of employee pensions, further ire directed at teachers, police officers, and other public employees, and more stress on the tottering finances of state and local governments. 

Philosophical: A pension plan is a mutual, not individual, arrangement

The last, philosophical, objection to the GASB framework requires a look into the fundamental principles behind a pension plan. 

It is typical to speak of pension plans as belonging to one of two varieties, defined benefit (DB) and defined contribution (DC). A “DB” plan is a modern name for a traditional pension plan while a DC plan is just another name for an employer-sponsored savings plan. These two categories of plan are frequently portrayed as distinguished by the level of risk incurred by the employee and the employer. The employee can be said to bear much more of the risk in a DC plan than for a traditional plan and vice versa. This is all too true, but portraying this as a question of which of two parties assumes the risk is inaccurate, because with a traditional plan there is essentially a third party to the equation: the body of plan members as a whole. 

A traditional pension plan works because not all the people paying into the system will have a long and happy retirement. Baldly put, some of the members will die before enjoying all the benefits to which they are entitled. Those who do not will see their retirement financed by those unused contributions. The consequence of this reality is that it is very difficult to separate the value of one member’s contribution from another. The value to each member is that they are all in the fund together, insuring each other.

By contrast, the principle behind the GASB accounting reforms is roughly that accountants ought to be able to match the marginal cost of each employee with the marginal contribution from that same employee. A DB plan is a collective entity, but the GASB accounting insists on looking at individuals. Quoting directly from Statement 68:

“For defined benefit pensions, this Statement identifies the methods and assumptions that should be used to project benefit payments, discount projected benefit payments to their actuarial present value, and attribute that present value to periods of employee service.”

In other words, the rules provide guidance for determining what fraction of the fund “belongs” to any individual employee, given in exchange for their work in some particular year. This is the root of the GASB insistence that current employee contributions should not be used to pay current retirees, and the insistence that all plans should be 100 percent funded. Only a fully-funded plan can have all of the employee-years allocated to employee shares of the fund assets. There is no clarity to the accounting otherwise.

A pension plan is a mutual insurance arrangement. The collective financial strength of the body of members is greater than the sum of the parts. In a pension plan, every member has an equivalent claim on every dollar coming into the system. There is no sense in which a dollar of contribution “belongs” to this or that member, and systems do not define a priority among members. The employer and the fund itself are a source of strength, but so are all the other members: three sources of security. If one is weak, the others are still available. A plan with a 30 percent funding ratio obviously is less secure than a plan at 80 percent, but under the right demographic conditions, it can still run indefinitely because of the employer and all the other plan members. By insisting the only important thing is accounting for individual contributions and expenses, the GASB rules seek to erase this source of security from consideration. Worse, the rules lead to decisions that undermine the third source of security, by making full or partial plan closures seem like a sensible idea

When critics complain that “generational equity” demands that one age cohort must not subsidize another, they are demanding less security than pension plans were invented to provide. One age cohort does provide security for another and the same security will be provided to them in turn, as well as those after them. This is how these plans were designed. But this kind of equity is not an absolute good. Presumably if generational equity were valued above all other considerations, then teacher salaries would only be financed with debt, so the children who reap the benefit would eventually pay the expense. Perhaps it is best to describe this as a different kind of generational equity, where one generation receives the benefits it enjoyed in the past. The kindness we extend to our children is different from the kindness we expect from them, but that does not excuse them from extending the same kindness to their children in turn.

The invention of mutual insurance redefined life among the working and middle classes in the nineteenth and early twentieth centuries. Insurance was not the only such innovation, but was part of a movement toward the democratization of finance that included life insurance, disability insurance, and unemployment insurance, not to mention savings banks, savings bonds, and mutual funds. The push to old-age insurance was a part of a political movement that began years before 1911, when the pioneering director D. W. Griffith produced a film called “What Shall We Do with Our Old?” The answer the movie proposed was to establish old-age insurance to provide pensions to alleviate poverty among the elderly. It took decades of effort, but eventually politicians and industry responded. Private pension systems were established, and between 1914 and 1934, when Social Security was established, twenty-eight states had established experiments with old-age pension plans for their poorer residents.50

The plot of Griffith’s film was a tragedy, meant to illustrate a preventable irony: old-age penury in a wealthy society. The refinement of mutual insurance into “old-age insurance,”—which subsequently became known as pension plans—has all but erased this kind of poverty in developed countries. It seems an odd claim that “clarity” in accounting should be deemed more important than that. 

  • 26. Marc Lifsher. “California pension funds are running dry.” In: Los Angeles Times (Nov. 2014). 6/27/15. URL:
  • 27. Roger Lowenstein. “The Next Crisis: Public Pension Funds.” In: New York Times Magazine (June 2010). 6/28/2015. URL:
  • 28. Eric Boehm. “Pennsylvania pension funds could run dry in as little as 10 years.” In: Daily Local News (Apr. 2015). 6/25/15. URL:
  • 29. GASB. Why Governmental Accounting And Financial Reporting Is—And Should Be—Different. 6/21/15. Governmental Accounting Standards Board. Norwalk, Connecticut, Nov. 2006, revised 2013. URL:
  • 30. For more about the difference between private and public entities, and the differences in risk, see Peng, State and Local Pension Fund Management, section 4.2.
  • 31. Author’s calculations from (Social Security Trustees. Trustee’s Report. 6/23/15. Social Security Administration. 2014. URL: The system was unable to sustain this low level of funding because of demographic changes. That is, the actuarial facts on the ground did change, as the baby boom worked its way through the system.
  • 32. This is equation 7.5 of (Howard E. Winklevoss. Pension Mathematics with Numerical Illustrations. Second edition. Pension Research Council of the Wharton School of Business / University of Pennsylvania Press, 1993), chapter 7. Winklevoss was writing about private pension plans, so he subsequently points out that plans will seek full funding. The federal law governing those plans requires them to do so. The same restriction is not true of public systems, for the reasons noted here.
  • 33. Borenstein, “Labor perpetuates pension myth that 80 percent funding goal is OK.”
  • 34. Keith Brainard and Paul Zorn. The 80-percent threshold: Its source as a healthy or minimum funding level for public pension plans. Tech. rep. 6/28/15. Lexington, KY: National Association of State Retirement Administators (NASRA), Jan. 2012. URL: percent_funding_threshold.pdf.
  • 35. Barbara D. Bovbjerg. State and Local Government Pension Plans: Current Structure and Funded Status. Testimony before the Joint Economic Committee. 6/19/15. Washington DC: United States Government Accountability Office (GAO), July 2008. URL:
  • 36. Author calculations from 2014 Illinois financial statements. See also note 16 of the 2015 CAFR. The first publication of this report stated that the absolute value of the state’s revenue shortfall was $45 million. I also clarify here that GASB rules would require the use of a 3 to 6.5 percent assumed rate of return which is lower than current rates of return that are between seven and eight percent.
  • i. Since this paper’s first publication, it has come to our attention that some pension plans that are not fully funded are not using a risk-free rate of return as the discount rate. This may be attributable to an alternative interpretation of GASB 68 requirements, specifically paragraphs 27-29. GASB 68 paragraph 26 explains what discount rate should be used. The discount rate of a “long-term expected rate of return” can be used if the plan’s net position is projected to adequately pay benefits. GASB explains the specific circumstances where this rate of return can be used. If those circumstances aren’t met, then the discount rate “should be the single rate that reflects” what is sometimes called a risk-free rate of return—“[a] yield or index rate for 20-year, tax-exempt general obligation municipal bonds with an average rating of AAA/Aa or higher (or equivalent quality on another rating scale).” Projected benefit payments are “all benefits to be provided to current and active employees” according to GASB 68 paragraph 24-25. Paragraphs 27-29 say that “the pension plan’s projected fiduciary net position," explained in these 3 paragraphs, should be compared with the “amount of projected benefit payments, calculated according to paragraphs 24-25 already discussed here. There is a possible matter of interpretation in paragraph 27 that has emerged. The first sentence of paragraph 27 states that those two calculations (the calculations of the plan’s fiduciary net position and the amount of projected benefit payments) “should be compared in each period of projected benefit payments.” The period over which these two calculations should be compared can be taken to characterize a pension plan that will be depleted. This can be accurately described, as per Sgouros interpretation, as “not fully funded.” This interpretation is consistent with that in this paper and is resonant with the paper’s interview with of Deputy Director of CalSTRS and with others not directly included in the paper. An alternative interpretation, is that the rule would only apply to pensions actually running out of money in the short-term. It’s not surprising that this interpretation may be emerging since the issuance of GASB 68, as it would exempt many pension funds from having to use the risk-free rate of return as the discount rate.
  • 37. See GASB 68 paragraphs 27 and 28.
  • ii. The first publication of this paper implied that returns on municipal bonds could not be used as a risk-free rate of return. This is not correct. The rate of return on 20-year-tax-exempt general obligation bonds with an average rating of AA/Aa or higher (or equivalent rating in another scale) may be used as a risk-free rate of return. It is common to take one of two yields as a risk-free rate of return--the index rate of treasury bonds or the rate of 20-year tax-exempt general obligation bonds with an average rating of AA/Aa or higher (or equivalent rating in another scale). The latter is what GASB 68 paragraph 26 says must be used as the discount rate if the comparison between a plan’s fiduciary net position is not balanced with the projected benefit payments.
  • 39. a. b. John E. Cihota. Considerations in Structuring Estimated Liabilities. Tech. rep. FT-WP-15-003. Washington, DC: Office of the Inspector General, United State Postal Service, Jan. 2015. URL:
  • 40. . Interview with Ed Derman, deputy CEO of CalSTRS, June 6, 2014.
  • 41. Further calculations from 2014 Illinois financial statements.
  • 42. Author estimates from Bureau of Economic Analysis data.
  • 43. This kind of argument is a staple of anti-tax argumentation, and it is remarkable how seldom it is deployed in this context.
  • 44. Bovbjerg, State and Local Government Pension Plans: Current Structure and Funded Status.
  • 45. The risk of a fully-funded, or overfunded,pension plan is not only the political risk of increased benefits and reduced contributions, but also that policy makers will perceive an opportunity to close the plan entirely. According to the GASB framework, this is a rational choice for a fully-funded plan. There may be a political cost to such a decision, but the accounting says there would be no financial cost. In reality, closing a plan substantially increases the risk to the taxpayers, and experience shows that few such decisions have turned out to be good ones. (See page 22.)
  • 46. Peng, State and Local Pension Fund Management, chapter 6.
  • 47. California Assembly Legislative Analyst’s Office, Addressing CalSTRS’ Long-Term Funding Needs.
  • 48. Ted Dabrowski, John Klingner, and Tait Jensen. CPS pensions: From retirement security to political slush fund. Tech. rep. Chicago and Springfield, IL: Illinois Policy Institute, Aug. 2015. URL:
  • 49. Peng, State and Local Pension Fund Management, p.22.
  • 50. Dora L. Costa. The Evolution of Retirement: An American Economic History, 1880- 1990. 6/26/15. University of Chicago Press, 1998. URL: