Covering your local pension plan It took me three years and two - - PDF document

Covering your local pension plan

It took me three years and two dozen stories to really figure out how to tell if a public pension plan is healthy or not. Here are the tips and shortcuts to save you the time. By David Milstead davidmilstead@q.com 303-800-6423 Part One: Introduction

Defined benefit plans: Known as “pensions,” it’s the benefit that’s defined, usually as a specific

amount per month, in retirement. It’s the plan manager’s responsibility to invest the money that comes in to make sure the benefit can be paid.

Defined contribution plans: Like a 401(k), it’s the contribution – what goes in – that’s defined.

What comes out at retirement is whatever is in the account after years of investing. Typically, these accounts shift the investment risk to the individual – for better or worse. This presentation is about defined benefit plans run by the public sector – governments like states, school systems, and cities. Non-governmental entities, like for-profit corporations, run defined benefit contribution pension plans as

• well. There are special issues with corporate pension plans, like accounting rules and the tax benefits

from the contributions, that we will not cover here. Also, these non-governmental plans are covered by the Federal ERISA (Employee Retirement Income Security Act), and the employers can walk away from the plans in certain circumstances and turn the over the Pension Benefit Guaranty Corp., the federal pension insurer. Public pensions are generally regarded as contractual relationships between the public employer and the employee – meaning public pension promises must be kept.

Part Two: Pension cash flow

Contributions + Investment Income = Expenses + Benefits (hopefully) A defined benefit plan takes in cash each year from the employer and, typically, from the employee as

• well. These are the “contributions.”

Contributions are typically expressed as a percentage of pay. In Colorado, the state school employers of PERA members paid 13.85 percent of salaries into the plan, up from 10.15 percent several years ago. Employees had been paying 8 percent of their salaries; recent changes increased that figure to 10.5 for some. Other plans require higher contributions. When I did my study in 2005, one West Virginia pension plan had an employer contribution rate exceeding 22 percent – more than twice what Colorado taxpayers paid at the time. (Many pension members reject the idea that “taxpayers contribute” anything to their plan, even though they are employed by a governmental entity. Excerpt from an e-mail from a PERA member: “The 10.1% contribution of the employer and 8% from the employee are both part and parcel of the employee's overall wage and benefit compensation earned for work performed and services rendered. That money is twice removed from the taxpayer and no longer any of their business.” Former New York Times reporter David

Cay Johnston made a similar argument earlier this year in a blog post in which he lambasted journalists for using language he considered supportive of efforts to cut back pensions.) The contributions go to work as investments – stocks and bonds, and to a lesser extent, real estate, venture capital, private equity, hedge funds and timber. Some of these assets – dividend-paying stocks, bonds, real estate funds that distribute rental income or sales proceeds – throw off cash, or investment income, every year. That cash can be used to pay benefits or can be reinvested in new assets. Each pension fund has certain expenses that must be paid each year. All plans have certain administrative costs, and all plans have investment costs as well. Some plans, like PERA, manage a significant amount of their money in-house, and employ more people. This arrangement may be cheaper than paying asset-management fees to outside money managers, which many plans do to a certain

• extent. Plus, PERA, for example, also manages 401(k)-style accounts and a health-care plan, which adds

to its expenses. Expenses can be examined as a percentage of assets in the pension plan, (an “expense ratio”) and that number seems very small, perhaps 0.50 percent of assets. And the ways in which the plans are structured can cause the number to be higher, or lower, without it being wasteful or inefficient. Also, all plans have fixed costs, so a smaller plan will have a higher expense ratio. Each year, the plan must pay benefits to its retired members. A member that has retired with full eligibility for a pension gets what’s called an “unreduced” benefit. (My editors don’t like that word, so I call it a “full” pension, which is not an official pension term and, consequently, angers pension members.) A typical pension benefit is expressed as a percentage of salary, which is then multiplied by years of

• service. Or:

Benefit = Salary X percentage X years of service The salary is typically a “highest average” or “highest annual.” Some plans take the three, or five, highest annual figures. Others take the highest 24, 36, or 60 months. Others take final salary on last day of

• employment. The method of calculation is important because a pension payment can be manipulated

when the salary period is small. Some public safety plans with loose rules have “chief for a day” problems, where police officers or firefighters get a one-day promotion, a $20,000 raise for pension purposes, and a much more golden retirement. In Colorado, am unreduced pension is equal to 2.5 percent of HAS, times years of service, with a cap of 100 percent of HAS. A retiree with 30 years of service and a HAS of $50,000 would get $50,000 X .025, or $1,250, times 30 years, for $37,500. A 40-year employee would get $50,000 X .025, or $1,250, times 40 years, for $50,000. As you can see, any years worked over 40, with the 100% cap in place, yields no extra pension benefit. Benefits are then typically adjusted each year in retirement for increases in the cost of living. Some increases are set year-by-year, or are linked directly to a measure of the Consumer Price Index. Others plans, like PERA until recent reforms, are fixed. PERA retirees’ benefits were to increase 3.5 percent per year, a decision that was made in 2000 when PERA was overfunded. Recent changes now tied the COLA to the lesser of the Consumer Price Index or 2 per cent. Many plans with an older retiree population may actually be cash-flow-negative, which means they have to sell off assets in order to pay benefits. Not a good sign.

Questions to know the answers to:

• What are the employer and employee contribution rates for this pension plan? How much are

taxpayers putting into the plan each year in employer contributions?

• What are the plan’s expenses, as a percentage of assets? How does that compare to plans of

similar size? Is there something about the plan’s structure, like how much of the plan’s investments are managed in-house, that affects the comparison?

• What is the plan’s formula for calculating benefits? Is there any limit to what a member can earn

in retirement?

• How is the HAS calculated? Does the plan’s method of calculating HAS make it susceptible to

manipulation?

• Do the plan’s contributions exceed benefits paid? If not, does investment income make up the

difference? What are the trends – are benefits paid increasing at a faster rate than income?

Part Three: Pension funding - ratios and schedules

The funding ratio is the plan’s assets, compared to its liabilities. But truly understanding the ratio is little harder than you might think. Let’s start with liabilities first. A pension plan estimates all its future benefits payments to members for the rest of time. They have to make numerous assumptions – life expectancy, age of retirement, salary growth, and inflation. Then, they take that stream of payments and come up with a present value. (If you’re familiar with present value calculations, it involves using an interest rate and discounting the cash payments to current dollars, based on the premise that getting dollar in 2041 is worth less than getting a dollar in 2011.) The discount rate issue has become a major source of contention in the current debate over the pension

• issue. Public pensions typically use their assumed investment rate of return to discount the liabilities, a

choice that makes the present value of the liabilities smaller than if it used the rate on high-quality corporate bonds, which most private-sector employers do, or Treasury Bill rates. (There will be a full seminar on this issue later in the program.) That’s the plan’s liabilities. You might think the plan’s assets are the money it has in its portfolio. But that isn’t necessarily so. When a funding ratio is calculated, it typically uses what’s called “actuarial assets,” and those are different because of a process called “smoothing.” With “smoothing,” a pension plan takes one-year gains and losses and spreads them out over time. The goal is to make funding requirements more stable, so one bad or good year doesn't require employers to pump much larger or smaller amounts into the plan. In bad times, such as 2000 through 2002, or 2008, smoothing delays the accounting for losses, can make actuarial assets higher than their market value, and makes things look better than they are. In good times, such as 2003 through 2006, it delays gains, can make actuarial assets lower than market value, and makes things look worse. PERA had a great approach to this. It smoothed over four years, so its funding looked better than it truly was as the portfolio declined. After 2004, when the smoothing process had recognized the losses but hadn’t yet recognized the gains, it took a one-time markup and increased its actuarial assets by $1.5

• billion. Then it went back to smoothing again.

So, once a plan arrives at its assets and liabilities, it can calculate a funding ratio, which is Assets Liabilities If assets exceed liabilities, which is rare, the funding ratio is 100% or more and the plan is fully funded. If liabilities exceed assets, which is common, the ratio is below 100% and the plan is underfunded. The first significant drop came in the stock market decline of 2000-2002. A 2004 study of state retirement systems by Wilshire Research showed that the combined funding of 123 pension plans declined from 91 percent in 2002 to 82 percent in 2003. The median funding level - a plan's assets compared with its liabilities - was 79 percent. Gains from 2003 to 2007 helped reverse the situation; then came the recent market crash. The Pew Center for the States recently estimated that funding in state pension plans declined from 84 percent in 2008 to 78 percent in 2009.

Why is it OK for a pension plan to be continually underfunded? A good question. One PERA answer, from the plan’s spokeswoman, is “Remember that PERA does not need to be 100 percent funded since not all benefits are due and payable at one time.” Another explanation is that if a plan is fully funded, all of its assets are exposed to market risk. By being underfunded, it stands to lose less in a down year. Funding ratios are like the debt-to-equity ratio on a balance sheet – it’s important to know, but it doesn’t tell the whole story. Two companies with similar debt burdens could be in very different situations, because one is highly profitable and can pay off the debt, but the other is losing money and will have to borrow more. An easy way to look at the funding ratio is the trend. At PERA, even with the market recovery of 2003 and 2004, the funding ratio went from 100% in 2000 to 70% in 2004 – a drop of 30 percentage points in four

• years. (Even with the subsequent market recovery, PERA’s funding ratio peaked at 73.8 percent before

falling to 67.2 percent in 2009.) The pension plan may well be making internal projections for its board about the funding ratio using different assumed rates of return. PERA’s projection, which it did not typically release to its members, showed just how bad the situation was. Another way to look at the issue, one that’s more complex, is the fund’s “amortization schedule.” An amortization schedule is the number of years it will take the plan, with current contribution levels and the assumed investment return, to pay off its unfunded liability. Accounting rules say a pension plan is adequately funded if it has a 30-year amortization schedule. To create that schedule, a pension plan’s actuaries arrive at the “actuarially required contribution (rate),” or ARC, that is required to get the plan to its amortization schedule. In California, municipal pensions are required to be on an amortization schedule, and the plan sponsor must ensure contributions from employer and employee equal the ARC. But not all states have similar laws, and you may be able to detect just how much more taxpayers need to put in to pay off the liability. Colorado is an extreme example, because PERA had an infinite amortization schedule at year-end 2004, which technically meant the plan would ultimately collapse. (In this way, it was in worse shape, in my view, than the West Virginia plan with a 22 percent contribution rate and a funding ratio of 19 percent.) Here’s where PERA stood in 2005 when I did my project: PERA’s actuaries calculated an ARC of 16.91

• percent. PERA’s employers contributed 10.15 percent, but 1.02 percentage points of that went into the

health-care plan, so just 9.13 percent of employee pay went into the pension. The difference between the ARC and the actual contribution, then, was 16.91- 9.13, or 7.78 percent of employee pay. PERA’s payroll was about $5.3 billion in 2004, a figure I found in the annual report. A 16.91 percent contribution on that would be about $895 million. But there were only $453 million in employer contributions in 2004, another figure from the annual report. So this is the paragraph I wrote: “PERA's actuaries told the fund in June that to get PERA back into sound funding status, employer contributions needed to nearly double. That, according to a Rocky Mountain News analysis, would require an extra $400 million a year, starting now.” The market returns of 2004-2007 improved the situation. Using results at the end of 2007, PERA estimated that it could actually have a funding schedule for most of its employee groups that was not

• infinite. Some of the funding schedules had slipped below 30 years. The results of 2008-09 threw that all
• ut the window, and PERA had to propose a significant reform package, including higher contributions

and limits on benefits, to fix the problem. Remember, calculating a plan’s liabilities involves a lot of assumptions about age and retirement, as well as inflation and salary increases. Changing these assumptions can increase or decrease the liability, which then would increase or decrease the funding level, and consequently, the ARC.

The investment assumption – what the plan’s assets are expected to return in the long term - is the biggest factor by far. Many plans are still assuming annual returns of 8 percent - even though the portfolios may not be constructed to approach those returns. (See Part Four)

Questions to know the answers to:

• What interest rate is your plan using to discount liabilities – and have they done estimates using

different interest rates?

• What is your pension plan’s funding ratio? Is it increasing or decreasing?
• Does your plan use actuarial smoothing? Do they do it consistently?
• Is your plan on a 40-year or 30-year amortization schedule?
• What is the actuarially required contribution? Is the plan actually making the contribution? If not,

how much extra taxpayer money is required?

Part Four: Investments

In the early days of the American pension system, fund managers typically bought bonds that matured when benefits came due. It was stable, not particularly risky - and expensive. As stocks began to rise in the 1980s, more and more pension managers realized they could take advantage of the equity markets' higher returns. An employer could put less money into the plan, or pay more benefits, if investments grew more quickly. So pension funds shifted from stocks to bonds, although many still have bonds as a significant part of the portfolio. The asset mix of the pension plan’s portfolio is a key part of the annual report. Equities are generally divided into domestic and international, and there’s a category for bonds. Most pension plans have expanded their portfolios to include real estate, venture capital, private equity, hedge funds and timber. Because of this mix, it’s not entirely fair to compare a pension plan’s return to the Standard & Poor’s 500. Instead, the pension plan’s annual report likely mentions, in the annual report, a “benchmark return” for pensions of its size, and whether it beat its benchmark. There are many benchmarking reports published by industry consultants that slice and dice portfolio information and can give you a sense of how a pension plan’s asset mix compares to its peers. The following paragraphs, from one of my stories, give you a sense of how this information, looking backwards, can be used to explain how a pension plan got to its current state: “By the end of 1999, PERA had 71.2 percent of its portfolio in domestic and international stocks. Another 9.5 percent was in "alternative investments," which includes venture capital. Just 11.1 percent was in bonds and cash, with the remainder in real estate and timber. Everyone with stocks or a 401(k) knows what happened next. The market peaked in March 2000. Internet stocks went first, then other tech and telecom stocks began to follow. The Sept. 11, 2001, terrorist attacks, followed within weeks by the collapse of Enron, ensured another down year in 2001. WorldCom, and the doubts it raised about whether any big company could implode in scandal, took down all of blue- chip America in 2002. At the end of those three years, PERA's portfolio had dropped from $30 billion to $23 billion. It had a three-year annualized loss of 6.6 percent. The typical pension fund, which owned less in stocks than PERA, lost just 4 percent per year.” Pension funds typically perform asset review studies, in which an outside consultant looks at the portfolio’s asset mix, makes an inflation assumption, and calculates a likely near-term return. (Sometimes, both the staff and the board have consultants, so there are two estimates.) What makes things interesting is when a pension fund uses an investment return assumption to calculate its funding ratio, yet is given a different probable return by its investment consultants. It happened a lot more than you might think, like in Colorado, where the return assumption was 8.5 percent in the middle of the last decade, but the investment consultants said the portfolio was more likely to return 7.75 percent. Questions to know the answers to:

• What is your pension plan’s asset mix? Does it have more in stocks and other volatile

investments than similarly-sized plans?

• How does the plan’s investment return compare to the benchmark? Is it underperforming

because it wasn’t invested in the best-performing asset classes?

• Do investment consultants predict a different investment return than what’s in the actuarial

assumptions? How does the pension plan justify this? If the pension portfolio returns the lower number, what are the implications for funding?

Part Five: Governance

The financial stuff gets tough. It’s easier to figure out how your pension plan is structured, and how that differs from others. In many states, the boards of public pension plans are wholly or partially appointed through the political

• process. Some boards have educational requirements for some or all seats. A portion of many board

seats are held by workers who will someday benefit from the plan. In Colorado, when I first started covering the issue, the PERA board had 16 trustees. 14 were elected directly by the benefit recipients, active and retired; the other two board members, the state Treasurer and Auditor, are also PERA members. In a reform law several years ago, the board structure was changed. Member-elected seats were reduced to 11, and three board members were appointed by the governor instead. Virtually all pension trustees – all – travel to “conferences” that are ostensibly for education, but truly provide an opportunity for Wall Street sharpies to wine and dine gullible public servants. Many pension plans have travel or expense limits, and PERA goes further in requiring trustees to fill out an “evaluation form” after attending a conference. (A great subject of an FOI request, I’ll point out.) Pension plan executives attend, too. Pension plans also may, or may not, have their budgets set through the legislative process, and their employees may or may not be part of the state employment system. Big differences in pay and perks can result when there’s no outside oversight; pension plan managers can get intoxicated with the idea they’re

• perating in the big-money finance world, and standards for public-sector pay don’t apply.

A list of pay and perks that were available to PERA executives and employees that were not available to rank-and-file state employees as of 2005: A choice of luxury automobile or $1,175-per-month car allowance; annual bonuses, plus a special “long-term incentive” for executives; the ability to accrue 52 weeks of leave time and get paid for it on departure; health-care benefits for life for a handful of top

• execs. PERA later changed or abandoned many of these practices.

A harder governance issue to track is conflict-of-interest policies for trustees and staff. Most pension plans limit, or ban, gifts from Wall Street managers or vendors to the plan. But pension plan staff often serve as advisors or directors to the private-equity funds they invest in, and the luxurious meetings of those boards are typically considered expenses of the fund – not gifts to the pension executives.

Questions to know the answers to:

• How is the board of your pension plan structured? Do benefit recipients control the majority of

seats?

• Are there any policies that limit board travel to conferences? Who on the board has spent the

most at these conferences? How much does the board and staff spend each year on conference travel?

• How is the pension plan’s budget set? Are the plan employees state employees, part of the

personnel system?

• Is there a conflict-of-interest policy? Are board and staff members filling out disclosure forms

about gifts they receive?