EXECUTIVE SUMMARY. sector pensions nationwide face unfunded liabilities topping $3 trillion. 2

Transcription

1 EXECUTIVE SUMMARY Current pension accounting rules significantly understate state pension plan liabilities and overstate their funding health. Using accurate accounting, Washington s combined plans would face a $50.6 billion shortfall. By private pension standards Washington s pension system would be considered endangered. At market value, Washington s unfunded pension liability equals $20,141 per household. Economists Joshua Rauh of Northwestern and Robert Novy-Marx of Rochester estimate that Washington households would need to pay an extra $1,371 in annual taxes over the next 30 years to fully fund public employee pensions. 1 Even if accounting rules were reformed, lawmakers have habitually missed contributions and increased benefits, putting the plans long-term financing at risk. If historical patterns of benefit expansions and underfunding were to continue, unfunded liabilities could grow even larger than reported here. A shift to defined contribution pensions would help address both issues. Defined contribution plans are not subject to the accounting distortions and political manipulation that plague defined benefit pensions, giving the public a clear idea regarding the size and risk of public pension liabilities. 1 New York Times. A $176 Billion Gap for Public Pensions. June 21, Andrew G. Biggs Introduction Washington state employees, like most state and local government workers across the nation, are offered a defined benefit pension in retirement. A defined benefit plan differs from defined contribution 401(k)-style pension plans that most private sector employees participate in. In a defined benefit pension, an employee is offered a fixed monthly benefit at retirement. Any investment risk involved with funding those benefits is taken by the employer. In a defined contribution plan, by contrast, the employer offers the employee a fixed contribution to a retirement account, but the employee chooses how to invest that contribution and accepts the risk and return involved with those investments. Defined benefit pension plans require that governments accurately anticipate future benefit costs and make regular contributions that will be sufficient to meet those costs. Around the country, citizens have become concerned regarding the solvency, cost, generosity and risk of public sector pension plans. Washington s pension plans are well funded by public sector standards. Nationwide, public pensions report unfunded liabilities of around $770 billion. 1 But, as economists increasingly point out, public pension accounting standards are lax relative to those that private pensions are required to use and are inconsistent with how economists and financial markets would value pension liabilities. Using more realistic accounting rules, public sector pensions nationwide face unfunded liabilities topping $3 trillion. 2 These critiques apply to Washington s pensions as well. While the state s plans appear reasonably well funded relative to other public sector pensions, their funding status has slipped significantly in recent years due to missed contributions, benefit increases and poor market returns. Moreover, were Washington s plans measured using the accounting standards private sector plans are required by the federal government to use, the plans would be only 52 percent funded on average and face unfunded liabilities of $50 billion. As a result, Washington s reported funding worrying though it may already be is in many ways a mirage. Significant reforms and prolonged political discipline will be required to put these plans back on track. But lack of political discipline has been one of the drivers of pension shortfalls. Over the years, lawmakers have increased benefits in good times and lowered pension payments in bad times. But these benefit increases continue even after the good times have soured, and skipped payments are slow to resume even after tax revenues recover. In many ways, political factors are just as large a threat to pension financing and overall fiscal discipline as is the underfunding hidden by current accounting rules. Pension reforms should take both financial and political risk into account and consider options that are both more transparent and less 1 Source: Public Fund Survey, April See Novy-Marx, Robert and Rauh, Joshua D., The Liabilities and Risks of State-Sponsored Pension Plans, Journal of Economic Perspectives 23(4), Fall Biggs, Andrew G. An Options Pricing Method for Calculating the Market Price of Public Sector Pension Liabilities. Forthcoming, Public Budgeting and Finance. Biggs, Andrew G. Proposed GASB Rules Show Why Only Market Valuation Fully Captures Public Pension Liabilities. Financial Analysts Journal, March/April 2010.

2 Reforming Washington State Pensions PAGE 2 susceptible to elected officials desires to promise benefits now but pay for them later. Defined contribution pension plans are used by most Americans saving for retirement and, properly designed, there is little reason that public sector employees cannot follow the same model. A final introductory note: Many of the figures cited here and much of the analysis draws upon actuarial valuations and the 2010 Risk Assessment conducted by the Office of the State Actuary. 3 Washington is unusual in having its pension analysis conducted by an actuary employed by the State rather than an outside firm retained under contract. While each approach has its pros and cons, in this case Washington s policymakers and citizens have been rewarded with some of the best actuarial analysis available in the country. While it is to be expected that an analyst coming at pension financing from an economic perspective would disagree with an actuary coming from a more traditional approach, the Risk Assessment produced by Washington s actuaries is a model of what the profession should be providing to its customers (lawmakers and the public). While this paper is critical of actuarial valuation of pension liabilities, to the degree it is critical of the work of Washington s actuaries, it is only because their work highlights almost all the relevant issues cited by economists but fails to take the final step to a full market valuation of pension liabilities. Background on Washington Pension Plans Washington state administers defined benefit pension plans for most government employees. Washington currently runs three defined benefit plans PERS (Public Employee Retiree System) Plan 1, TRS (Teachers Retirement System) Plan 1 and WSPRS (Washington State Patrol Retirement System) 1 that are closed to new entrants but have active employees and will pay benefits for several decades to come. The state runs a number of open plans, including PERS Plan 2/3, TRS Plan 2/3, SERS Plan 2/3, PSERS (Public Safety Retirement System) 2, LEOFF (Law Enforcement Officers and Fire Fighters) 1/2 and WSPRS 2. In addition, the state administers several defined contribution pension plans for public employees, although they are not analyzed in detail here. Each defined benefit plan is funded by a combination of government and employee contributions, though the mix varies from plan to plan. In SERS Plan 1 and TRS Plan 1 the employee pays a fixed contribution of 6 percent of wages while the employer pays any remaining part. This places all financial risk on the employer. In the remaining open plans, both employer and employee contributions can vary according to the funded status of the plan. Other members pay a rate Asset equal to half the normal cost of the plan, which represents the cost of benefits accruing in that year. The employer pays the other half of the normal cost as well as the cost of amortizing unfunded liabilities from prior years. Under this setup, employees bear some risk if demographic projections or changes in the assumed future investment return alter the normal cost, but they are not subject to risk based upon the current performance of plan investments. Benefits are paid using a formula based upon years of service and final pay. For instance, in PERS Plan 2 a retiree receives benefits equal to 2 percent of final earnings multiplied by the number of years of service, with final earnings defined as the highest consecutive five years of employment. Workers taking early retirement receive reduced benefits, although on advantageous terms for long-term employees, and after retirement benefits are increased annually by the rate of inflation, capped at 3 percent. In addition, the Washington plans offer disability benefits and benefits to the survivor of an eligible employee, albeit at the cost of lower initial retirement benefits. Investments The Washington State Investment Board holds a common investment fund for the various Washington pension plans. Table 1 shows both the target allocations that is, the allocation desired by the board, as well as the actual allocations, which Table 1. Asset allocations, Washington State Investment Board, as of June 30, Target allocation Actual allocation Fixed Income 20.00% 21.89% Tangible Assets 5.00% 1.16% Real Estate 13.00% 14.21% Global Equity 37.00% 34.71% Private Equity 25.00% 25.76% Innovation 0.00% 0.83% Cash 0.00% 1.44% can change from day to day as market values of assets rise or fall. Washington is unusual in devoting a large share of its investment portfolio to so-called alternative investments, in particular private equity. Private equity involves ownership stakes in companies that are not publicly traded; this can involve both ongoing companies or startup firms (so-called venture capital ). Private equity investments are generally made for the long-term. They can produce higher returns at the price of higher risk though a large private equity stake can prove problematic if funds 3 See Office of the State Actuary. Washington State 2009 Actuarial Valuation Report. October 2010; and Office of the State Actuary Risk Assessment: Moving Beyond Expectations. August 31, 2010.

3 Reforming Washington State Pensions PAGE 3 are needed in a hurry. Wilshire Consulting projects that private equity as a class will have an average return of 9.7 percent and a standard deviation of annual returns a typical measure of risk of 26 percent. For comparison, U.S. equities have a projected annual return of 7.25 percent and a standard deviation of returns 16 percent. 4 There is no denying that Washington s private equity investments have fared well over time that is one reason the private equity share has risen but as the advertisements say, past performance is no guarantee of future success. The Washington plans assume an 8 percent return on investments, which is typical of public sector plans around the country. 5 However, Washington s State Actuary has repeatedly advised using a lower return assumption of 7.5 percent. It is possible to argue for an even more conservative return estimate. Using Wilshire Consulting s projections for asset returns, risk and covariations between returns, it is possible estimate expected return of Washington s target investment portfolio. Based on Wilshire s projections, Washington s target portfolio has an expected annual return of slightly above 7 percent. Despite the Actuary s recommendations, the state s Pension Funding Council has rejected the actuary s advice to shift to a 7.5 percent assumed return. While a substantive case for an 8 percent return can be made after all, if you take sufficient risk you can achieve almost any desired rate of return in expectation it should be clear that this is not an argument about financial returns so much as it is about pension contributions. 6 Lowering the assumed investment return by even a small amount has a dramatic effect on how much the government must contribute to the plans each year. For the PERS 2/3 plans, for instance, lowering the investment return by 0.5 percentage points would increase required employer contributions by almost 25 percent. Shifting to a 7 percent return would increase annual contributions by almost 50 percent. The difference in plan funding generated by even a small change in assumed returns shows how dependent upon the risk premium public pensions have become. This is particularly important when we consider that there is a very strong likelihood that Washington s plans will receive returns below 8 percent and that s true even if they ve accurately predicted future returns. For instance, consider an investment portfolio with an assumed mean return of 8 percent and standard deviation of annual returns of 11.3 percent, which appears to approximate the risk assumed by the state s actuaries. One would think that this would imply a long-term probability of actually receiving returns above or below 8 percent. In fact, because investment returns are skewed meaning that the average figure is distorted by a small number of very high returns and a larger number of lower returns there s actually only around a 40 percent probability of achieving an 8 percent average return over 30 years. But there s a 1- in-4 chance of receiving a compound return below 6 percent and a 1-in-10 chance of receiving a sub-5 percent return. If Washington s Pension Funding Council rejects a 50 basis point reduction in the assumed rate of investment return because they believe the state cannot afford it, they and the state may well be in for a rude shock. It is also worth noting that, based on the Wilshire projections for the risk of different asset classes, Washington s portfolio allocation generates one of the riskier pension investment funds in the country. Among 29 large pension funds sampled and analyzed, only the Pennsylvania State Retirement System, the Illinois Teachers Retirement System and the Minnesota Retirement System hold riskier portfolios. 7 These estimates are based upon broad asset classes equities; bonds; alternative investments; real estate; and cash rather than the large number of specific investments held on behalf of Washington s funds. If the Investment Board believes it is taking less risk than other pensions it and other public plans should disclose the assumptions by which they judge the risk of their portfolios. These are assumptions that should regularly be released as a matter of course. Plan Funding The funding health of a pension is generally measured in two ways: the ratio of assets to liabilities (the funding ratio ) and the net value of assets and liabilities (the unfunded liability ). In these measures, assets are the investments held in the plan. The value of assets is calculated on an actuarial basis in which investment returns are smoothed, meaning that the effects of changing year-to-year returns are incorporated over a number of years (eight years in the case of Washington). When asset returns have been low, as they have in recent years, the smoothed value will be higher than the market value. As of mid-2009, the combined Washington plans had smoothed assets of $ billion, versus a market value of assets of $ billion. Liabilities are calculated as present values, meaning that future benefits are discounted at present to make them comparable to the value of assets. In discounting benefits the plans use the projected rate of return on assets in Washington s case, 8 percent. This means, for instance, that a $100 benefit liability taking place in 20 years time would have a present value of $ Using this method, Washington s plans have combined liabilities of $ billion. 4 Wilshire Consulting Wilshire Report on State Retirement Systems: Funding Levels and Asset Allocation. March 3, The average public sector assumed return is slightly under 8 percent, with a range from around 7 percent to about 8.5 percent. 6 For instance, see The Wall Street Journal. Public Pension-Fund Squeeze. March 23, See Biggs, Andrew. How Have Public Sector Pensions Responded to the Financial Crisis? Forthcoming article prepared for How the Global Financial Crisis is Reshaping Retirement Security. A Wharton School/Pension Research Council/Boettner Center Conference. May 5 and 6, $100/ = $21.45.

4 Reforming Washington State Pensions PAGE 4 Table 1. Funding ratios for Washington plans, 2001 through Ratio equals actuarial value of assets over present value of liabilities. LEOFF Plan 1 LEOFF Plan 2 PERS 1 PERS 2/3 School Employees Plan 2/3 Teachers Plan 1 Teachers Plan 2/ % 100% 91% 100% 100% 94% 100% % 100% 86% 100% 100% 92% 100% % 100% 81% 100% 100% 88% 100% % 100% 77% 100% 100% 84% 100% % 100% 71% 100% 100% 78% 100% % 100% 73% 100% 100% 80% 100% % 100% 71% 100% 100% 77% 100% % 100% 71% 100% 100% 77% 100% % 100% 70% 100% 100% 75% 100% Source: 2010 Risk Assessment Combined, these asset and liability figures generate an unfunded liability of $507 million and a funding ratio of 99 percent. Table 1 breaks funding ratios down by plan and shows results from 2001 through While plans remain reasonably well funded by public sector standards, it is clear that funding for several plans has declined significantly over time. Moreover, because these valuations rely on smoothed asset values that don t fully incorporate recent market declines, it is likely that plan funding health will decline and required contributions will rise even if the economy and financial markets recover. Total employer contribution rates will rise significantly in coming years. As the Office of the State Actuary has stated, Even under an optimistic investment return scenario, biennial employer contributions still triple over the next two biennia. Under a pessimistic scenario, biennial contributions could increase over five and a half times in the next 12 years. 9 contribution increases and reductions in salaries or other benefits to make room for rising pension costs. The actuaries find that there is an almost 1-in-5 chance that pension costs will rise from the current level of 2.7 percent of General Fund expenditures to exceed 8 percent of expenditures by the mid-2020s. There is a 1-in-20 chance that pension costs will reach 9.9 percent of outlays and a similar probably that employer costs will exceed 20 percent of worker payroll, versus the current level of 7.4 percent. 10 The Actuaries model also projects a 41 percent probability of the PERS 1 or TRS 1 plans running out of money and reverting to pay-as-you-go status and a 13 percent probability of pay-as-you-go funding for an open plan. Should an open plan enter payas-you-go status, the annual costs could exceed $4 billion. An economist would ask, how much would you be willing to pay to avoid a 13 percent probability of a $4 billion cost? The answer, as we shall see below, is a lot. Declines in funding health have three main causes: poor market returns, missed contributions, and benefit increases. Financial markets have experienced significant declines in recent years. From 2001 through 2010, plan investments generated a compound average return of only 3.9 percent, versus the 8 percent expected return. This means that benefit liabilities grow faster than the assets available to fund them. One way to gauge the impact of these low returns is to ask how high returns must be over the next decade in order to bring the plan back up to where it would have been had investments yielded the projected 8 percent. To get back to projected asset levels by the year 2020 would require that annual returns over the next decade average upwards of 12.3 percent. This seems unlikely, particularly given Wilshire s downbeat projections regarding asset returns. The Washington State Legislature also has not been diligent in making its full annual contributions in recent years. As Figure 1 shows, for most of the 1990s Washington met its full Annual Required Contribution (ARC) These higher costs will impose burdens on taxpayers, beneficiaries of other government programs that may be squeezed and public employees themselves, who may see 9 Office of the State Actuary. Transmittal letter, 2009 Report to Pension Funding Council Members. September 11, Risk Assessment.

5 Reforming Washington State Pensions PAGE 5 nearly every year, while from 2001 through 2010 the government paid only 62 percent of the ARC on average. On average, from 2001 through 2009, state pensions nationwide paid 95 percent of ARCs as calculated by plan actuaries, according to data from the Center for Retirement Research at Boston College, making Washington one of the poorer funding performers during that period. Finally, elected officials instituted a number of benefit increases, which compound over time to increase plan costs. In addition to a series of small benefit increase, large benefit improvements in 1989, 2000 and again in 2007, increased total plan liabilities by almost $6.8 billion. (Figure 2.) The last of these benefit increases was a result of the socalled gain-sharing provisions, which increased benefits if the fund experienced high returns. Specifically, if the compound return exceeded 10 percent of a four year period, half of the return in excess of 10 percent would be diverted to benefit increases rather than to ordinary plan funding or to make up for unfunded liabilities from past years. It is difficult to overstate the folly of the gain-sharing policy, which creates a truncated distribution of investment returns where retirees take part of the upside while the government carries the downside. In effect, the taxpayer has been placed on the wrong side of a one-side bet. While the 10 percent returns specified as part of the gainsharing rules were deemed extraordinary, in volatile markets they are anything but unusual. In about one-third of four-year periods the average return could be expected to exceed 10 percent, thereby triggering the provision. So long as it remains in effect, the gain-sharing provision is equivalent to a permanent wage increase of around 2.74 percent for each employee over every year of their career. 11 Why policymakers did not consider the repercussions of their actions before passing the plan is not clear. Making matters more difficult is the fact that, despite being designated as non-contractual compensation at the time the rule was passed, pension participants sued upon repeal of the provision. In a September 2010 court ruling the repeal of gain-sharing was deemed an illegal reduction in benefits. If that rule stands, gainsharing could cost the budget $150 million or more over the next two years. Gain-sharing, which passed in 1998 under Republican control of the state House and Senate, is as good 11 Based on actuaries projection of an increase in the long-term TRS contribution rate of 2.74 percent.

6 Reforming Washington State Pensions PAGE 6 an example as can be found of pension mismanagement. Funding levels would have declined even more had the plans adopted changes to investment return and mortality assumptions recommended by the State Actuary. As discussed above, the actuary has several times recommended lowering assumed investment returns from 8.0 percent to 7.5 percent. Similarly, the actuary has recommended using updated mortality tables that would project longer lives for beneficiaries and consequently higher costs for the plans. 12 The Pension Funding Council adopted these recommendations, but the Legislature continued to fund based upon the older, less expensive mortality tables. 13 Using these more realistic assumptions plan funding ratios would be lower and unfunded liabilities and required contributions higher. Why Current Washington Pension Liabilities are Understated Running alongside public concern over pension financing has been a technical argument over how plan liabilities should be valued. While seemingly arcane, the results of this debate between economists and actuaries far outweigh recent declines in funding health, as they presage a fundamental reappraisal of how we analyze and fund public sector pensions. Recall that, under current practices, future benefit liabilities are discounted to the present using the interest rate that is projected to be earned by plan investments. In Washington s case, that is 8 percent. Based on the present value of plan liabilities, plan actuaries determine the level of annual contributions. Initially this approach seems correct. If plan investments generate the expected return, then today s contributions will be sufficient to fund tomorrow s benefits. But economists almost universally agree that this approach is wrong. According to economic theory as well as the practice of financial markets, the discount rate used to value a liability should reflect the relative risk of the liability itself, not of any assets set aside to fund the liability. As an article in the respected American Economic Review states, Finance theory is unambiguous that the discount rate used to value future pension obligations should reflect the riskiness of the liabilities. 14 What would this mean for pensions? Since accrued public pension benefits are guaranteed by the government in the case of Washington, through the 1956 case of Bakenhaus v. City of Seattle economists believe they should be discounted using an interest rate derived from similar governmentguaranteed investments namely, government bonds. Intuitively, since public sector pensions are government-guaranteed payments, they should be discounted using the same interest rate on municipal bonds. However, as authors Brown and Wilcox note, in practice accrued pension benefits have turned out to be even safer than municipal bonds. 15 In New York City in the 1970s and Orange County, California, in the early 1990s, full pension benefits continued to be paid even as holders of explicit debt took losses. Similarly, Vallejo, California continues to make full pension payments today despite being in bankruptcy proceedings triggered by the refusal of public employees to renegotiate labor contracts. For that reason, most economists analyzing the issue have used yields on U.S. Treasury bonds to reflect the fact that distressed governments have in the past reneged on explicit debt before they have cut accrued pension benefits. 16 Ideally, annual pension obligations would be discounted using the Treasury yield curve, such that benefits due one year in the future would be discounted at the yield on one-year Treasury securities while those 20 years in the future would reflect the yield on securities with a duration of 20 years. In practice, most public sector pensions do not release projections of future benefit obligations on a year-to-year basis, so an approximation is used. 17 Currently, Treasury yields over durations matching public pension liabilities average around 4 percent, although these values change daily. Applying a lower discount rate derived from the Treasury yield means that the present value of plan liabilities would be larger and the annual payments required to fund benefits and amortize unfunded liabilities also would grow. Economists call this preferred approach the market valuation of liabilities, or market valuation and MVL for short. Market valuation resembles the rules that private sector plans are required to follow. Private sector pension benefits are very safe, but they are not truly guaranteed. If the sponsoring corporation goes bankrupt, at least part of retirees benefits may be lost. 18 For that reason, private pension plans must discount their liabilities using the yield on a portfolio of high quality corporate bonds, currently around 5.5 percent. Since public pension benefits are safer than private pension benefits, economic 13 The mortality issue is discussed in Law Enforcement Officers And Fire Fighters Plan 2 Retirement Board. Projected Improvements in Mortality Assumptions. Preliminary Report. May 28, Office of the State Actuary. Transmittal letter, 2009 Report to Pension Funding Council Members. September 11, Brown, Jeffrey R. and Wilcox, David W., Discounting State and Local Pension Liabilities, American Economic Review, vol. 99, May See also Novy-Marx, Robert and Joshua D. Rauh, The Liabilities and Risks of State-Sponsored Pension Plans, Journal of Economic Perspectives Vol. 23, No. 4 (Fall 2009), pp Examples here include New York City in the 1970s, Orange County, CA in the 1990s and Vallejo, CA today. 17 The mid-point of public pension liabilities generally lies around 15 years in the future, so the liability can be treated as a single lump sum payment occurring 15 years hence and then discounted to the present using the yield on Treasury securities with a duration of 15 years. 18 The Pension Benefit Guaranty Corporation covers benefits up to a limit, with reductions for early retirees.

7 Reforming Washington State Pensions PAGE 7 theory says they should be discounted using a lower interest rate, not a higher one. These differences in accounting approaches mean that pension funding ratios calculated for public and private sector plans are fundamentally incompatible. Now, the market valuation approach does not mean that pensions must invest only in corporate bonds; in fact, most private pensions do hold stocks. But it does mean that plans can credit themselves with the higher return on stocks only after those higher returns have been generated. If investments in stocks do produce higher returns, then those returns are reflected in the asset values used to calculate funding health. But under market valuation, the chickens cannot be counted before the eggs have hatched. Using a market valuation approach, public pensions are not nearly as well funded as they claim. Overall, public pensions nationwide would not be around 78 percent funded, as they report, but only around 44 percent funded under market valuation. If we used the market value of assets rather than the smoothed actuarial value, the average funding ratio would be even lower. Likewise, under market valuation total unfunded liabilities pension liabilities would not equal around $770 billion, as stated, but over $3 trillion. To put things in perspective, the Department of Labor judges that any pension plan that is less than 80 percent funded should be considered endangered, while plans less than 65 percent funded are considered critical. 19 If judged using private pension accounting, which does not allow for increased risk-taking to create a false perception of better funding, almost all public pension plans across the country including all but one of Washington s would be considered in critical status. Accurate accounting for public pension liabilities would have a similar impact on Washington s plans. Using the 2010 Risk Assessment s sensitivity analysis with regard to investment returns, it is possible to estimate employer normal costs and the cost of amortizing unfunded liabilities for a selection of Washington plans under the market valuation approach. Normal costs and costs of unfunded liabilities are estimated based on the assumption that costs are linear with regard to the natural log of the discount rate. Table 2 shows that the normal cost of funding accruing benefits is roughly 4 times higher under a market valuation approach than under the plans expected cost accounting. For instance, the normal cost for PERS 2/3 would rise from 4.6 percent of salaries under an 8 percent assumed return to 21.3 percent of salaries under market valuation. For comparative purposes, this higher figure is the amount a private sector employee would need to save in a defined contribution plan holding government bonds in order to match the guaranteed benefit provided by PERS 2/3. The cost of amortizing unfunded liabilities from prior years would rise even more. While actuarial analysis with access to all details of plan design is desirable, the cost increases shown in Table 2 are qualitatively similar to a recent actuarial analysis of the Florida Retirement System conducted at the request of Gov. Rick Scott. 20 As of mid-2010, Washing plans together have actuarial assets of $55.5 billion and actuarial liabilities of $60.2 billion. (Table 3.) This leaves them 92 percent funded, with an unfunded liability of around $4.7 billion. Using a market valuation approach, however, plan liabilities are around 1.75 times higher at $106.1 billion. This higher amount is how markets would value guaranteed government payments of the size and timing of Washington s pension plans. 19 Life and Health Insurance News.com (2009). The Big Pension Freeze. May See DuZebe, Robert S. Study Reflecting Impact to the FRS of Changing the Investment Return Assumption to one of the following: 7.5%, 7.0%, 6.0%, 5.0%, 4.0% and 3.0%. Milliman. March 11, The increase in employer costs as the interest rate declines is larger in the Washington plans than in the Florida Retirement System. This may be due to employee contributions being capped in several Washington plans while all FRS plans appear to distribute increases in normal costs to both employer and employee contribution rates.

8 Reforming Washington State Pensions PAGE 8 Given these higher liabilities, Washington pensions together would be only 52 percent funded, with an unfunded liability of $50.6 billion. To put this in perspective, Washington s unfunded pension liability equals $20,141 per household. Separately, economists Joshua Rauh of Northwestern and Robert Novy-Marx of Rochester estimate that Washington households would need to pay an extra $1,371 in annual taxes over the next 30 years to fully fund public employee pensions. 21 If the calculation were based on the $44.2 billion market value of plan assets rather than the larger smoothed actuarial value, Washington s unfunded liabilities would approach $62 billion. The $46 billion difference between plan liabilities at the expected return of 8 percent and the market value of liabilities simply represents the value that markets place on the risk that Washington plans have taken. This risk is effectively off balance sheet, meaning that it is nowhere disclosed in pension reports or other government documents. But it remains real and should be understood. Understanding Market Valuation Given the dramatic changes to measured pension funding under a market valuation approach, it is worth explaining the method and its justification in greater detail. Market valuation can be confusing at first and is subject to a range of (generally non-germane) objections, so it is worthwhile to establish both the credibility and the intuition of the economic approach. To begin, it should be understood that market valuation is mainstream economics and is almost universally believed by economists to be the appropriate way to measure pension liabilities, including the liabilities for public sector plans. It is the only framework that is consistent with how economic theory and financial markets value liabilities and risk. As the Vice Chair of the Federal Reserve Board put it, While economists are famous for disagreeing with each other on virtually every other conceivable issue, when it comes to this one there is no professional disagreement: The only appropriate way to calculate the present value of a very-low-risk liability is to use a very-low-risk discount rate. 22 President Truman s complaint about two handed economists who never reach consensus does not apply here. The broad agreement in the economic community should be taken seriously. Indeed, the market valuation approach is absolutely central to modern economics and corporate finance. 23 Market valuation, true to its name, is also central to the way financial markets work in practice. For instance, a public pension could purchase annuities from private life insurance companies to cover its future benefit liabilities. 24 An annuity is a financial product that pays a stream of benefits for as long as the holder lives. Based on current actuarial accounting, if Washington s pension liabilities equal $60 billion then presumably a payment of $60 billion to private insurance companies would be sufficient to purchase annuities ensuring that those payments would be made. In reality, no insurance company in the world would take this deal. Rather, they would demand an amount roughly equal to the market value of liabilities in Washington s case, around $106 billion to provide those annuity contracts. The reason is that private markets care only about the liability, not whether you have financed it with safe or risky assets. 21 New York Times. A $176 Billion Gap for Public Pensions. June 21, Kohn, Donald L., Statement at the National Conference on Public Employee Retirement Systems Annual Conference. New Orleans, Louisiana, May 20, The basic premise derives from the Modigliani-Miller theorem in corporate finance, which holds that the value of a liability is independent of the way the liability is financed. This theorem was cited as one reason Modigliani was awarded the 1985 Nobel Prize in Economics. 24 It is not uncommon for private DB pensions in the UK to follow this practice. Under US public sector accounting rules a pension would be severely and unfairly penalized for doing so.

9 Reforming Washington State Pensions PAGE 9 Why is this the case? Since Washington must pay its pension liabilities regardless of how its plan investments fare, there is a contingent liability on Washington taxpayers. If the pension takes an aggressive funding strategy by holding risky assets, that means smaller pension contributions today but a larger contingent liability should the plans investments not pan out. A private insurance company would value both the upfront costs and the contingent liabilities should the investment returns fall below expected values. The total cost of meeting the plan s liabilities that is, the cost of current contributions plus the taxpayer backstop should market returns fall short of expectations is the same regardless of what the plan invests in. 25 That is why there is a single discount rate to be applied to the liability, despite a seemingly infinite number of funding strategies. Indeed, public pensions could not invest properly if their asset managers did not understand how to value liabilities. A pension s asset managers would not pay more for a company s stock if that company invested its corporate pension funds in stocks than in bonds. Yet pensions claim that they themselves become better funded by shifting to equities. Unfortunately, pension funds separate the asset and the liability sides of their programs; there is essentially nothing in the way of asset-liability management, meaning that none of the insights of their asset managers benefit those who must value the plan s liabilities. One could argue that investing in risky assets goes both ways: the plan could just as easily end up overfunded as underfunded. It is on points like this that market valuation shows its merits. As the 2010 Risk Assessment notes, State revenue and investments moved together. Good returns will happen when the economy, incomes, jobs and revenues are high, while bad returns correlate with bad times. The Risk Assessment continues, Weak economic environments were correlated with weak investment returns. Lower investment returns created the need for increased contributions at a time when employers and members could least afford them. This is exactly what we are seeing today. At a time when taxpayers are already hard-pressed, pension funding is asking them for more. Unlike the Risk Assessment, however, market valuation doesn t simply describe this pattern. It quantitatively captures the fact that protection against bad outcomes is far more valuable and expensive than profits from good outcomes, because the bad outcomes tend to occur when everything else is going wrong and the relative value of a dollar is highest. 26 A standard response to these kinds of arguments is that government is simply different from a private insurance company, a private corporation or private individuals. More specifically, because government has the power to tax it is also infinitely-funded. Because the government will be around forever, it can rely on the so-called long term rate of return on investments while ignoring year-to-year fluctuations. Like the basic argument for discounting liabilities at the expected return on assets, this argument is attractive at first. But it breaks down very quickly under scrutiny and few economists take it seriously. First, government is not a person or entity that bears risk by itself. As the Congressional Budget Office (CBO) has pointed out, The government does not have a capacity to bear risk on its own. 27 Rather, government transfers risk between different stakeholders, such as taxpayers, public employees, bondholders and beneficiaries of government programs, just as a corporation transfers risk to stock and bondholders, employees and, where possible, customers. To realize the truth of this, we need merely to look around us. All over the country, taxpayers are contributing more to pensions, other programs are starved of revenues, and bondholders are beginning to worry about the safety of their investments. It is these individuals, not the government, that is paying the cost for market declines that have hit pension investments. The implication of this, the CBO has argued in contexts ranging from student loan guarantees, to bank deposit insurance, to guarantees against market risk for Social Security personal accounts, is that governments should value risk the same way that their stakeholders do, using market signals and market prices. 28 Likewise, Nobel Prize winning economist Robert Merton has specifically cited federal guarantees for private sector pensions which are equivalent to the taxpayer guarantee of public sector pensions as an appropriate area for market valuation approaches. 29 The argument that a government program should be immune from market pricing simply doesn t fly with economists. Second, even if government is truly infinitely lived, that doesn t mean it can ignore market risk. After all, a defined benefit plans has to pay the bills when they come due, not out in the infinite future. And the majority of the typical pension s liabilities occur, not in the distant future, but over the next 15 years. 30 No one can guarantee an 8 percent return over that period. Moreover, to the degree that government s permanence matters, that makes it even more likely that pension benefits will be 25 This argument is developed more fully in Biggs, Andrew An Options Pricing Method for Calculating the Market Price of Public Sector Pension Liabilities, Public Budgeting & Finance, forthcoming. 26 In economic terms, pension deficits occur when the marginal value of a dollar is highest while surpluses occur when it is lowest. 27 Congressional Budget Office, Estimating the Value of Subsidies for Federal Loans and Loan Guarantees, August See Lucas, Deborah and Phaup, Marvin. The Cost of Risk to the Government and Its Implications for Federal Budgeting, National Bureau of Economic Research, Merton, Robert. Applications of Option-Pricing Theory: Twenty-Five Years Later, American Economic Review, vol. 88, no. 1 (March 1988), pp Waring, M. Barton, Liability-relative investing, Journal of Portfolio Management 30(4).

10 Reforming Washington State Pensions PAGE 10 paid. Under standard financial economics, the more likely a benefit is to be paid the lower the discount rate to be applied. Third, as economists Paul Samuelson and Zvi Bodie have shown, stock investments actually don t become less risky over time, such that a longer-lived entity need not worry about risk. 31 It is true that the standard deviation of annual investment returns declines over longer holding periods; for instance, the standard deviation of stock returns over a single year is around 20 percent but over 15 years it falls to around 5 percent. But the longer compounding period dominates the smaller variation in returns. For instance, an investor who receives a 1-year return on stocks that s one standard deviation below the mean will end the year with a balance around 19 percent less than if he had received the average return. Over 5 years, however, that investor would receive around 35 percent less and over 15 years around 50 percent less. Markets understand this: The value of a put option which insures against falling stock values doesn t get less expensive over time, it gets more expensive. Fourth, economic theory does say there are certain instances in which government can ignore risk, namely when those risks are both small and uncorrelated to the government s other assets and liabilities. 32 Something like building a public university might qualify, but holding billions in equities and other risky assets is an entirely different story. Despite the prevailing story that government can ignore risk, the academic research supporting this claim is very limited. Finally, if you believe that government can guarantee everyone the expected return on stocks without having to account for risk, then virtually every investment in the economy should be carried out through the government. For some, this isn t too troubling a thought. But the logic lays out the fundamental problem: even if we all bear risk together, that doesn t make the risk go away. Ultimately it is people, not government, who pay the bills and bear the risk. The Relative Generosity of Washington Pension Benefits As state and local governments have considered how to address their budgets, including rising costs of public sector pensions, the adequacy of public employee compensation will play a role in shaping policy choices. One way to measure relative generosity is to compare public sector benefits paid to a full career employee to those paid by a typical private sector worker with the same earnings. To test this, we begin with stylized lifetime earnings patterns generated by the Social Security Administration, in which age-based earnings factors are multiplied by the Average Wage Index for a given year to simulate the typical shape of earnings over the life cycle. We then track how an individual with these earnings would fare under the Washington PERS 2 system and under a typical private sector 401(k) plan. We assume a 30-year working career spanning ages 27 to 56, with retirement in 2010 at age 57. Both workers are assumed to participate in Social Security on top of their individual pension coverage. An employee participating in the PERS 2 plan contributes 4.6 percent of earnings to the plan, with the employer covering remaining costs. PERS 2 pays a benefit equal to the average of the highest 5 years of earnings multiplied by years of service multiplied by a replacement factor of 2 percent. Individuals wishing to include a survivor benefit for a spouse receive 78.3 percent of the base benefit. Given the assumptions above, an individual retiring in 2010 would receive an annual benefit of around $19,305. In addition, he would be eligible to receive a Social Security benefit at age 65 of around $18,500 in today s dollars. Together, they would replace approximately 95 percent of his final earnings prior to retirement. Most financial advisors, who recommend a replacement rate of 70 to 80 percent of pre-retirement earnings, would consider this income level generous. The most common 401(k) employer match is 50 percent of employee contributions, up to a maximum match of 3 percent of earnings. 33 We assume that individuals contribute 6 percent of pay in order to receive the maximum employer match, making the total contribution 9 percent of pay. For the 401(k) plan, one must choose an assumed rate of return at which contributions are compounded through retirement. To match the riskless nature of public sector pension benefits we follow the CBO s practice of assuming investment in U.S. Treasury Securities, currently yielding around 4 percent over 20 years. While the typical individual does not invest in such a portfolio, to compare a riskier but higher yielding portfolio to the risk-free benefit given to public sector employees would cause the public and private sector pensions not to be comparable in terms of risk. 34 At retirement the private sector employee s 401(k) balance equals $124,514, sufficient to generate an annual retirement benefit of $3,900 per year, on top of which he would receive a Social Security benefit at age 65 of around $18,500 in today s dollars. Together, they would replace 31 Samuelson, Paul A., Risk and Uncertainty: A Fallacy of Large Numbers, The Collected Scientific Papers of Paul A. Samuelson, ed. Joseph E. Stiglitz, MIT Press, Bodie, Zvi, On the Risk of Stocks in the Long Run, Financial Analysts Journal, May-June Arrow, Kenneth J. and Lind, R.C., Uncertainty and the Evaluation of Public Investment Decisions, American Economic Review, Vol. 60, Source: 50th Annual Survey of Profit Sharing and 401k Plans, Published by Profit Sharing/401k Council of America. 34 We assume no administrative costs for the 401(k) plan; costs in the range of 0.5 percent to 1 percent of assets managed are typical, meaning that the figures here would slightly overstate the 401(k) balance. Assuming an administrative cost of 0.5 percent of assets managed would reduce benefits at retirement by around 7 percent.

11 Reforming Washington State Pensions PAGE 11 around 56 percent of his final earnings. 35. PERS 2 benefits are around five times more generous than those paid by a typical private sector pension; once the guaranteed nature of defined benefit pensions is accounted for. In effect, Washington taxpayers are financing generous guaranteed pensions to public employees while their own pensions are smaller and riskier. It is not clear why public sector employees require pensions that are larger and less risky than those offered to the private sector workers who finance them. Most public employees are not public safety officers, where there is at least a plausible argument for early retirement and relatively generous pensions. On the contrary, generous pensions may be a device to hide public employee compensation from a public that would react negatively to high public sector salaries. This is one more reason by why an accurate accounting for public sector pension costs is an essential tool for public policy. Reforming Washington Pensions The first step to any reform is knowing how big the problem is and what steps will and won t get you out of it. That s why accurate accounting for pension liabilities is so important. If Washington pension liabilities are measured in a way that is consistent with economic theory, financial markets and private pension accounting, funding shortfalls are nearly twice the level disclosed by the plans. And investing in riskier assets with higher promised rates of return won t provide an easy way out of the problem. Having calculated the shortfall, however, how should it be closed? The preceding section, which shows that Washington s plans provide significantly more generous benefits than are paid to private sector workers, implies that higher taxes on private sector workers or the businesses that employ them are not the answer. Higher employee contributions will probably have to play some role, though perhaps they should not be the largest factor. As noted above, pensions provided by Washington plans tend to be generous relative to what most financial advisors would recommend. Public employees are fine with that at the moment, since they pay only a small part of the total cost. But were costs accurately calculated and public employees required to pay their share, it is likely that many would prefer lower contributions and lower benefits as a path to reform. A number of other changes are possible. These could include: Reducing Cost of Living Adjustments (COLAs) is probably the most effective single way of lowering pension liabilities; a 1 percentage point reduction in annual COLAs would reduce liabilities by around 10 percent. 36 Legislation to eliminate COLAs for PERS 1 and TRS 1 has been passed, with protections for individuals receiving the minimum benefit and with reductions in the minimum employer contribution for these plans. Without an actuarial analysis of the legislation it is difficult to say how savings from reduced COLAs and dis-savings from lower employer contributions will net out. As in other states that have proposed reducing COLAs, press reports indicate that retired public employees may sue to have the legislation reversed. 37 Benefits could be calculated based on career-long average earnings rather than final earnings. This both would reduce benefits by including lower-earning years in the calculation and reduce incentives to raise final earnings to artificially boost retirement benefits. Increasing the retirement age would reduce liabilities by 2 to 4 percent per year of increase, while encouraging longer work lives. 38 Using an actuarially fair formula to calculate early retirement benefits would reduce liabilities by 2 to 5 percent. 39 It is possible, albeit politically very difficult, to construct a package of increased contributions and reduced benefits that would make pensions financial sustainable under honest accounting for plan liabilities. It is questionable how attractive they would appear to employees under these conditions, however. Much of the appeal of public sector defined benefit pensions, for employees and employers alike, is that they promise something for nothing. Once that illusion is gone, they appear significantly less attractive. Moreover, while incremental reform measures would strengthen pensions for the short term, for the long term a shift toward a defined contribution model likely makes more sense. Defined contribution plans are preferable in part for policy reasons, but perhaps more so because they are far less susceptible to underfunding and accounting abuse than defined benefit programs. Why fundamental change is needed Public pension financing fundamentally understates pension liabilities and erroneously argues that taking more risk would make those liabilities go away. It might seem that so long as we correct that accounting flaw, defined benefit pensions could continue in more or less their current form, albeit at a far higher cost. But 35 The employee is assumed to convert his 401(k) balance to an annuity that includes a 100 percent survivor benefit. Annuity prices are those offered by the federal Thrift Savings Plan. 36 Norcross, Eileen and Andrew Biggs. The Crisis in Public Sector Pension Plans: A Blueprint for Reform in New Jersey. The Mercatus Center. June 23, Associated Press. State government retirees may sue over pension. May 31, Rauh, Joshua and Robert Novy-Marx. Policy Options for State Pension Systems and Their Impact on Plan Liabilities. October Journal of Pension Economics and Finance, Forthcoming. 39 Rauh and Novy-Marx (2010).