1.2 Why Pensions are Underfunded

Pensions can be underfunded for the most obvious reason: your employer doesn’t fill your notional-car with real gas and you end up stuck in Cleveland. But this is not what happened with most pensions and it is not what happened with the AFM-EPF.

The roots of our current crisis go back 20-years or more. Back then, most employers were paying off their yearly normal costs with real money. In many cases, the real-money accounts contained more money than the accrued liability account required: plans looked overfunded. But this was appearance only. Actuaries were underestimating the accrued liability and the plans were, in reality, underfunded.

In our travel analogy, this historical underestimation is as if previous generations of travelers, on their way to San Francisco and needing gas hopped in a time machine, traveled forward in time, siphoned off the gas from a future generation’s notional-cars and then went back and put it in their own. While they may have reached San Francisco safe and sound, they did it with someone else’s gas!

Even as actuaries have been making more accurate estimates of the accrued liability, the economic environment in the U.S. has been pushing these liabilities higher each year. This has created larger funding deficits even as employer contributions have risen and benefits have been reduced.

This is like our employers filling our notional-car with more gas than ever before. Meanwhile, the Trustees, in their role as notional-car mechanics, have arranged for the newer notional-cars to consume less gas by reducing the benefit multiplier. You’d think this should get the car there faster, but each year fewer and fewer miles are traveled. At the current rate of travel, they will never get to San Francisco. Something must be wrong with the engine.

What’s causing the engine problem is not hard to find: it’s the interest rate from the previous section.

1.2.1 The Role of the Interest Rate

In our example we used an interest rate of 4%, conveniently plucked from the air. As you all know from paying your bills or watching your savings account, interest rates have a huge impact on how much money you will spend or accumulate over a 30-year period. Table 1 shows how the interest rate would affect the amount in the accrued liability savings account for our hypothetical retiree who is expecting $500,000, 30-years from now.

Table 1: How the accrued liability on $500,000 varies with the interest rate.

Interest RateAccrued Liability

7.5% $57,111
5.0% $115,689
3.5% $178,139
2.5% $238,371

Reducing the interest rate 66%, from 7.5% to 2.5% increases the amount the employer needs to accumulate from contributions and interest today by 417%! This interest rate thing is serious!

Those lower payment amounts looked pretty good to employers. It’s human nature to prefer to pay less money today and hope for more tomorrow. This is why employers have traditionally preferred higher interest rates. Union negotiators also like higher rates because it makes benefits higher and that is popular with their constituents.

Both parties forgot the loan like obligation which is the flip side of the accrued liability savings account. This means that not only does the normal cost need to be paid each year, but the much higher interest rate obligation must be met as well to reach the total liability (San Francisco) on schedule.

If the interest rate is unrealistically high and can’t be met, disaster looms. The amount in the accrued liability account will be too low even when reliably funded by employer contributions. This places retirement benefits in jeopardy. And this typically won’t be realized for a long time, sometimes years. Once discovered, catch-up payments will be large, perhaps impossibly large.

Meanwhile, during the period when the interest rate is set unrealistically high, Trustees will view the plan as well or even overfunded. This can (and has) led to unfortunate decisions to spend more money on benefits, making the deficit that much worse when it is finally discovered.

This is, in fact, the first part of what happened to our Pension Fund (and all other pension funds) and is the foundation of our present day problems.

A Note on Terminology

We’ve been using the term “interest rate” because that’s familiar to most of you. But in the literature about pensions (and in general finance), interest rates are also called “discount rates”. There’s a good reason for that: Pension discussions usually start with a future value (the amount of money that needs to be available 30 years from now), and tries to find a present value and rate that will get us there. So if we need a future value of $500,000, a “discount rate” of 7.5% will require us to have $57,111 dollars present today. These will grow at an interest rate of 7.5%, to be $500,000 in the future.

If one is to be very precise, “interest rate” should be used when calculating future values and “discount rate” for present values. The literature is not that precise and neither are we. If you remember that “interest” and “discount” rates mean essentially the same thing you shouldn’t get confused.

1.2.2 What’s the Right Discount Rate?

When you have a defined benefit pension plan such as the AFM-EPF, where benefits are set by formula (rather than the luck of the markets) and guaranteed by law, the only discount rate that makes any sense is one which is free of risk. While nothing is truly “risk-free” in finance or life, the yield of some form of U.S. Government bonds is typically viewed as the safest. The logic behind using U.S. bonds to estimate the risk-free rate is simple: if the U.S. Government ever gets to the point where it is defaulting on bond payments, we will all have much worse things to worry about than our Pension Fund.

Unfortunately, left to themselves, the actuaries of pension plans decided to use the historical return of the stock market rather than a Treasury rate as the discount rate for calculating the accrued liability. Someone, somewhere, once calculated the historic return from the stock market to be 7.5%-8.0%, and that is what they used.

This is a massive and fundamental flaw. It confuses the rate used to estimate the return on risky-assets (such as stocks) with an actual economic liability: the money you are counting on having at retirement.1 That actual economic liability is going to demand actual money in real dollars to be plunked down on a real table at some actual real point in time. It is also enforced by a legal obligation. In contrast, the rate-of-return on stocks is what you will get “IF” everything works out as you hope for your investments. And your stock broker has no legal obligation to make up any difference between your actual return and that 7.5% estimate. That’s a big “IF” and an enormous difference. To fully appreciate this difference, print-out the table from the last section and go talk to your bank’s loan officer. Ask for a $500,000, 30-year loan, offer to pay 7.5% interest and a stock portfolio currently valued at $57,111 as collateral. If the loan officer has any sense, you will not be getting that loan. You might get it if you offered them a 30-year government bond valued at $750,000.

If this example doesn’t convince you, ask yourself what you would do if your next door neighbor showed up with a fistful of penny stocks, potentially worth millions (“higher-risk, higher-return”) on your doorstep, and asked to borrow a few hundred dollars in return for simply giving you those stocks. We hope you would have enough sense to decline that arrangement.

It is true that banks will allow you to use stock-portfolios as collateral for things like mortgages. However, when this is done, those loan agreements also contain a clause that allows the bank to demand more collateral if your portfolio drops below a certain dollar amount. If you can’t pony up the extra collateral the bank will repossess your home.

In a sense, by using the rate-of-return on risky assets to value the accrued liability savings account rather than a true risk-free rate, the actuaries and Trustees have essentially offered us a stock-portfolio as collateral on our loan to them, but left out the clause that allows us to repossess their home “IF” things don’t go as planned.

Before we continue, we wish to make it clear that we are only talking about how to value the liability. How to fund that liability is a completely separate topic. Trustees are free to take on more risk in their pension fund investments, as long as they are willing (and able) to honor the bond-like commitment owed to their pensioners by obtaining more in contributions from employers if things don’t work out as they hope.

Hopefully, this discussion has convinced you that the proper discount rate to use is a risk-free one, derived from some form of U.S. Treasury. To estimate the discount rate, the IRS requires the use of the thirty-year Treasury curve (shown in Figure 1), with narrow adjustments made for cash flows.2

Figure 1: 30-Year Treasury Rate Curve, courtesy of the Federal Reserve Bank of St. Louis


Thirty-year Treasurys were not issued between August 2001 and August 2005, hence the gap in the chart.

1.2.3 The Insanity of an Inflated Discount Rate

Using a discount rate that is too high is a serious mistake because it makes a pension plan look better off than it really is. Historically, the difference between the rates that should have been used and the ones that were actually used were so extreme that they made underfunded plans appear to be overfunded! This is arguably the single biggest factor contributing to the current plight of all defined-benefit pension plans. Table 2 reprints the table from the previous section.

Table 2: The accrued liability on $500,000 varies with the interest rate.

Interest RateAccrued Liability

7.5% $57,111
5.0% $115,689
3.5% $178,139
2.5% $238,371

Remember that this is the liability on a $500,000 pension benefit for a single person, payable in 30-years. If you compare this with the Treasury rate curve from Figure 1, it’s easy to see that while a 7.5% discount rate was reasonable at the beginning of the 1990’s, by the mid-point of that decade it should have been lowered to around 5.0%-6.0%. But it wasn’t, so in the mid-1990’s, if a pension fund appeared 100% funded by the actuarial measure, in reality it was only 49% funded. In fact, if a pension fund looked 130% overfunded by the actuarial measure, its true funded ratio would have been only 65%.

But in the halcyon days of yore, both employers and unions in all industries were happy to accept the consequences of a high discount rate. The labor representatives and union officials looked like heroes to their constituents because they had negotiated large benefit packages. And the employers were happy with the lower contributions. We don’t believe there was any deliberate intent to mislead on either side. These actions were the result of ignorance and we doubt either side realized the consequences of their actions: potential loss of benefits and risk to the pension plan’s health on the union side, and accepting the obligation of much higher payouts with the associated risk of higher contributions on the employer side. We also wish to stress that this was true for every pension plan during this era, not just the AFM-EPF.

1“Every economist who has looked at this has said, ‘It’s crazy to use what you expect to earn on assets to discount a guaranteed promise you have made. That’s nuts![9].’ ” William F. Sharpe. Nobel Prize in Economics, 1990.

2You actually need multiple discount rates, using the full curve of spot discount rates, with each future cash flow discounted by the spot discount for its time horizon (nominal yield, real, inflation, etc.). We mention this only to illustrate the issues involved.