Debate Heats Up over Public Pension Fund Discount Rates

Record-low interest rates stir controversy over the way funds value future liabilities and expose a major difference between the U.S. and Europe.


LAST SUMMER THE INDIANA Public Retirement System (INPRS) created a stir in the U.S. public defined benefit pension world by dropping its long-term investment return assumption to 6.75 percent a year, the lowest of all its peers. The move worsened the system’s funded ratio, meaning that future retirees may risk cutbacks in their pensions unless taxpayers inject more money into the system — no easy feat in these tough times. As unpalatable as those consequences are, Indiana pension officials decided they had to scale back their projections following years of historically low interest rates and the likelihood that those rates would continue to depress investment returns. Six months later not a single large U.S. public pension fund has followed Indiana’s example. Most still operate on the assumption that they can generate returns of    7 to 8 percent a year virtually in perpetuity.

Across the Atlantic a far more austere philosophy reigns. The Netherlands, which runs one of the world’s most rigorous and best-funded retirement systems, requires that pension plans discount their future liabilities using not their hoped-for rate of investment returns but a conservative benchmark tied to long-term interest rates. The standard is so strict that regulators, responding to pressure from pension funds, adjusted the benchmark in October, which raised the discount rate slightly, to 2.42 percent. Executives at Amsterdam-based ABP, the country’s largest pension fund, breathed a sigh of relief:  The move boosted their funded ratio modestly, to 97 percent. The executives can only imagine what it would be like to operate under U.S. rules — if they used Indiana’s discount rate, ABP could claim to be nearly 200 percent funded.

How can two otherwise similar pension funds use such radically different assumptions?    The $20 billion Indiana pension system and €261 billion ($348 billion) ABP, which oversees the retirement assets of Dutch public sector employees, both take in contributions, invest their portfolio and pay pensions to retirees. Yet when it comes to the all-important discount rate, the number used to calculate the present value of future pension payouts, the Indiana and Dutch pension systems part company.

U.S. public funds are free to select a rate that estimates the future returns of their investment portfolios, while the Dutch and most other European pension sponsors must by law use a rate pegged to standard market interest rates. The difference between those methods was relatively modest before the financial crisis, but it has widened dramatically as interest rates have plunged to historic lows in recent years. Today most European pension plans use a rate of     between 3 and 4 percent, with the Dutch being especially conservative, while U.S. public plans use a rate nearly twice as high, on average.


The difference is not just an arcane matter for discussion in actuarial circles. The discount rate determines whether a retirement system is sufficiently funded. The funding status, in turn, dictates the annual contribution rate, or how much money employers and/or employees need to add to the fund. It also exerts a strong influence over the fund’s investment strategy and risk appetite: what mix of stocks, bonds and alternative assets is needed to achieve a fully funded state. Ultimately, the choice of discount rate can spell the difference between a comfortable retirement or an old age spent in poverty for tens of millions of workers.

Not surprisingly, given the stakes involved, the discount rate has become a source of  heated debate in the pension fund community on both sides of the Atlantic. Critics of the U.S. public pension system are scathing about the use of expected investment returns as the discount rate. “What is unethical, unprofessional and unscientific is to make your employees believe they have something that is risk-free and make taxpayers believe it has no cost,” contends Zvi Bodie, a finance professor at Boston University and a longtime pension watcher. “It’s not just a technical issue,” he says, and it raises two fundamental questions: “What is the nature of the pension promise, and is it sustainable?” Bodie believes a 2.5 percent discount rate, similar to that used in the Netherlands, is about right.

Antti Ilmanen, head of the portfolio solutions group at AQR Capital Management in New  York and the author of the 2011 book Expected Returns, warns of the “potential dangers of relying too heavily on using the backward-looking historical average return as expected return.” Ilmanen, along with Ben Inker, head of asset allocation at Boston-based fund manager GMO, projects investment returns of  between 4 and 6 percent for the average institutional investment portfolio over the medium term. “For the whole system it’s hard to see, given current valuations, how to expect more than 5 percent,” says Inker. “If you want 7.5 percent over the next ten years, stocks will have to deliver a lot more than they did over the past ten.”

The Dutch pension system has long been viewed as one of the strongest and most secure in the world. From 2009 to 2011 it was ranked No. 1 in the Melbourne Mercer Global Pension Index. (It came in second in 2012, after Denmark; the U.S. placed ninth.) Its high ranking is attributed to its mandatory system with near-universal coverage, a high funding level, well-established regulations and a discount rate that measures pension liabilities very conservatively.

“The only people getting this right are the Netherlands,” says John Ralfe, a British pension consultant who famously reallocated the £2.3 billion ($3.6 billion) pension fund of Boots entirely into fixed-income securities in 2001, when he was head of corporate finance at the pharmacy group. “They have a very tough, strict funding regime which is not capable of being fudged,” he adds.

But that does not mean Dutch pensions are immune from funding and investment problems. Consider the case of ABP. Like all Dutch pension funds, it valued its liabilities using a flat 4 percent discount rate, until 2007. That year ABP reported a flush funding position: Its assets were worth 140 percent of its projected liabilities.

Then the Netherlands’ Pensions Act 2007 altered the rules to require that discount rates reflect market interest rates. When the financial crisis hit the following year, ABP was socked with a double whammy. The havoc in financial markets caused its portfolio to post a 20.2 percent loss in 2008. In addition, a plunge in market interest rates caused the discount rate to tumble. By the end of 2008, ABP’s once-robust funding ratio had plummeted to 89.6 percent.

“The IFRS [International Financial Reporting Standards] framework and systems we have to adhere to start from market valuation,” explains Onno Steenbeek, director of asset liability management at APG, the Amsterdam-based firm that manages ABP’s assets. “You have a choice after that. The Netherlands chose the most conservative way.” By the last quarter of 2011, when the discount rate stood at 2.74 percent, ABP’s funding ratio had recovered to 93.7 percent, well below the 105 percent level required by law. ABP responded by cutting pensioners’ monthly payments by half a percentage point and raising employer and employee contributions, which already stood at more than 20 percent. The subsequent decline in market rates further depressed the discount rate, which is recalculated quarterly. The rate would have been 2.16 percent at the end of last year if not for the October 2012 adoption of the adjusted benchmark, called the ultimate forward rate, which brought it up to 2.42 percent.

By contrast, as of  June 30, 2012, Indiana reported having only 81.2 percent of the assets it needs to fund its future pension obligations, and that’s using a comparatively generous discount rate of 6.75 percent. If Indiana used the Dutch discount rate, it would have a funding ratio of only about 45 percent, estimates Rebecca Sielman, a principal in the Windsor, Connecticut, office of actuarial firm Milliman. Many other U.S. public pension funds are in even worse shape:  The national average funding ratio stood at 77 percent at the end of 2011, according to investment consultants  Wilshire, and virtually all of those funds use even more generous discount rates than Indiana’s.

HOW DID THE DUTCH AND U.S. PUBLIC pension systems take such different paths on their way to achieving the same goal — retirement income security for their teachers, firefighters and police? In the 1950s most pension plans in the U.S. were insurance products and had very conservative fixed-income investments and liability calculations. The next decade brought a sea change. Bank trust departments, an emerging group of asset managers and, later, investment consulting firms like Frank Russell approached plan sponsors with a new pitch: Abandon the insurance approach and invest in equities for their higher returns.

Plan sponsors signed on but soon discovered that the higher equity returns injected volatility into the equation. When actuaries created smoothing formulas to correct — some say mask — equity volatility, the modern style of U.S. pension investing was off and running.

U.S. pensions built an equity culture, and their European brethren eventually followed. By 2005, ABP had just under 37 percent of its assets in fixed income, with an equal amount in equity. The remainder of the portfolio was allotted to alternative investments like hedge funds (3.2 percent), commodities (2.5 percent), inflation-linked bonds (4.1 percent) and others. The key difference: Unlike U.S. public funds, the Dutch never used expected return on assets in their asset-liability calculation.

Over the past decade, when two severe market downturns exposed the risks of equity investing, two camps emerged to debate the drastically different methods of calculating future pension obligations. On one side are defenders of the U.S. public pension system’s expected-return calculations; on the other are academics who support the Dutch, as well as the U.S. corporate pension system, which uses high-quality corporate bond rates to discount liabilities.

A global group of financial economists and academics, including Jeremy Gold, principal of Jeremy Gold Pensions in New York; Boston University’s Bodie; Theo Kocken, CEO of Dutch pension consulting firm Cardano Group; and Joshua Rauh, a finance professor at the Stanford Graduate School of Business, have been warning that the U.S. public pension system is in even worse shape than it appears. “What is obvious today wasn’t obvious in 1965,” says Gold of the discount rate determination. “Modern finance was just being born.”

The cardinal rule, say these academics and consultants, is that pension funds should value their liabilities by the nature of the liabilities themselves, not by the assets on the other side of the balance sheet. “You don’t discount liabilities at 8 percent just because assets are invested with the hope of earning 8 percent,” stresses Gold. The discount rate for valuing guaranteed liabilities is simply the risk-free rate, now somewhere around 2 to 3 percent, Kocken says.

Indiana’s recent change fell far short of those suggestions. The system, which was merged recently to combine separate plans for teachers and public employees, reduced its discount rate to 6.75 percent from previous levels of 7.5 percent and 7 percent, respectively. Given its low funded status, the system has raised contributions in recent years, although they are far below Dutch levels. The rate for public employees was 8.6 percent of salary last year, up from 5.5 percent in 2007; the rate for teachers rose to 7.5 percent last year from 7 percent in 2007. Contributions are likely to go up again next year. (Employers pay the full contribution in Indiana; employees contribute only to a separate annuity savings account.) “We decided not to drop the contribution rate until we’re 100 percent funded,” says the retirement system’s CEO, Steve Russo.

Higher contributions alone won’t get Indiana to fully funded status, so the pension system has revamped its asset allocation as well. In September 2007 the public employees’ fund had 72.3 percent of its assets invested in public equity, while the teachers’ plan had an equity weighting of 65 percent. At the end of September 2012, the combined pension system had 24.9 percent of assets in public equity, 12.7 percent in private equity, 22.3 percent in traditional fixed income, 10.3 percent in inflation-linked bonds, 8.4 percent in commodities, 4.5 percent in real estate, 7.2 percent in absolute-return strategies, 8 percent in a risk parity allocation and the balance in cash.

The biggest change has been the addition of a risk parity strategy, which allocates assets based on risk factors rather than by traditional categories such as equities or bonds. “Our job is hitting the 6.75 target with the least amount of risk,” says Russo. “If we’ve done that, we’ve done our job.” That’s where chief investment officer David Cooper steps in. The risk parity portfolio is made up of four types of assets: equities, bonds, commodities and credit spread products. “No one knows for certain what economic environment will occur next year or next decade,” Cooper says. “Thus the appropriate path is to be diversified across different types of assets that respond to different potential economic environments.”

At ABP the decline in interest rates caused the plan’s funded ratio to plunge by 25 percentage points in 2011 alone. ABP’s investment manager, APG, responded by eliminating a number of complex derivatives and trading strategies to simplify its portfolio and bolstering the risk management team (Institutional Investor, June 2012). “The risk-return trade-off is the core business of what we do,” says asset liability head Steenbeek. “Risk from interest rate exposure is dominant in all the portfolios. When it was hidden through a veil of rules — the 4 percent discount rate — it wasn’t visible.”

To build some wiggle room into the austere Dutch discount rate, regulators in October 2012 adopted a new benchmark, called the ultimate forward rate. It makes a small upward adjustment in the rate after 20 years to reflect the fact that the swap market, which provides the basis for calculating the discount rate, isn’t very liquid at longer maturities. Not everyone liked the change. “This measure is viewed by the public as a trick to avoid large cuts” in pension benefits, says Steenbeek. The move, however, only added 3 or 4 percentage points to the funded ratios of most pension plans. As a result, ABP has had to announce modest cuts in benefits and increases in contribution rates to comply with Dutch law, which gives pension plans just three years to correct any underfunding.

Unlike their public sector brethren, U.S. private pensions operate under a regime closer to the European model. In 2006 the Pension Protection Act (PPA) required that private companies use corporate bond rates as the discount rate to determine funded ratios and contribution rates. In 2012, after the 2008 financial crisis had dealt a harsh blow to most pension plans, Congress took action to ease the funding squeeze, tucking it into a highway bill known as MAP-21. Now corporations can use a 25-year interest rate average to determine their minimum funding requirements. “MAP-21 is a recognition that the PPA model for discount rates isn’t working very well,” says Neil Rue, a consultant in the Portland, Oregon, office of Pension Consulting Alliance.

The regulations have had a big impact on asset allocations. From 1992 to 2005, U.S. corporate pension funds invested more heavily in risky assets than did public funds, according to “Pension Fund Asset Allocation and Liability Discount Rates: Camouflage and Reckless Risk Taking by U.S. Public Pension Plans?”, a study published in May 2012 by three Dutch professors, Aleksandar Andonov and Rob Bauer of Maastricht University and Martijn Cremers of the University of Notre Dame. The relative stance reversed between 2006 and 2010, though, partly because of the PPA-mandated use of bond rates in funding calculations and the maturing of corporate plan members as plans were frozen and closed.

Funding rules for U.S. public sector pensions have traditionally been lax; the minimum required contribution has essentially been the same as the expense ratio. But under a change adopted in June 2012 by the Governmental Accounting Standards Board, after June 2013 public plans whose assets are insufficient to cover future liabilities will have to use a blended discount rate to determine their funded status. Under this method plans can use their expected long-term investment return rate to discount liabilities only to the point that assets cover liabilities. Beyond that, plans must discount liabilities based on the yield of 20-year tax-exempt municipal bonds with a rating of double-A or higher. (Double-A munis currently yield about 3.35 percent.)

Unaffected by the PPA, U.S. public funds have persisted in making larger allocations to risky assets, including alternatives such as hedge funds, private equity and commodities, the Dutch study found. U.S. public pensions allocate about 8 percentage points more to risky assets than Canadian funds do and 14 percent more than European funds do. The greater risk appetite among U.S. funds was also evident in a 2.6-percentage-point increase in risky assets for every 1-percentage-point decline in yields on ten-year Treasury bonds, the authors said. The study also found that pension funds investing in riskier assets use higher rates, on average, to discount their liabilities.

Milliman’s Sielman believes that some risk-taking is justified. “Should a pension sponsor never take risk at all? The answer is no,” she says. Sielman points out that public plans ask a different question from corporates in determining future liabilities. Public plans take into account factors like projected pay increases, while corporate plans do not; public funds are also viewed as perpetual entities, while corporate fund liabilities are priced as though the plan could terminate the next day and purchase an annuity to replace the pension. “There is a notable difference between corporate and public funds,” agrees GMO’s Inker.

Sielman’s study examined whether public plan discount rates were too high. Her finding: Funded ratios are reported by the 100 largest plans as 7.3 percentage points higher in aggregate, 75.1 percent compared with 67.8 percent, than Milliman’s calculation. “When we look at the public plans as a whole, the sky is not falling,” she says. “It will take a lot to recover, but it doesn’t mean we have an impending accumulated disaster from the weight of underfunded pension plans. There has been a lot of gleeful reporting on pension plan tsunamis and disaster.”

Public plan sponsors join their actuaries in defending their choice of discount rates. Kevin Murray, executive deputy comptroller at the $150 billion New York State and Local Retirement System, which lowered its discount rate by a half point, to 7.5 percent, in September, says: “What we believe is a reasonable assumption is recognizing there are peaks and valleys. Especially after the ’08–’09 crash, everybody has tended to come after us on a short-term basis.” When asked how he thinks the fund can achieve a long-term 7.5 percent return, Murray replies: “We think it’s a realistic assumption that can be achieved over 30 years. It’s an actuarial standard that an actuary has to show data to support.”

Like Murray, Gary Findlay, longtime executive director of the Missouri State Employees’ Retirement System, has been battling those who criticize public funds’ use of expected returns to discount liabilities. He contends that building a case based on a huge meltdown is not taking the long-term view.

Findlay points to the California Public Employees’ Retirement System, which recently reported returns of 7.67 percent a year, on average, over the ten years through September 30, 2012, and 7.77 percent over 20 years. Both rates exceeded CalPERS’s actuarial return assumption of 7.5 percent. Similarly, Wilshire Trust Universe Comparison Service found that median returns for the ten years ended September 30, 2012, were 8.23 percent for corporate pension funds and 7.61 percent for public funds. Public funds with assets greater than $1 billion had median ten-year returns of 8.21 percent.

Projecting returns ten years hence is a risky business given today’s volatile markets and low interest rates, but liabilities will keep accruing at a steady pace. U.S. public pension plans and their European counterparts remain far apart in how they value those liabilities. American retirees may discover the difference to their dismay in coming decades.