Saturday, September 18, 2010

The Problem With Pensions

If only it were as simple as a single "problem."  The issue of who makes up the difference when pensions are underfunded is a tricky one, and will certainly depend on the readers philosophical beliefs.  I prefer to start with what I consider the bigger problem - the reason pensions end up being underfunded is because their return assumptions are, well, to put it bluntly:  wrong.

Thus, even though pensions are being reformed so that workers contribute more to their pensions, perhaps work for more years, and may have capped benefits, the damage done by overstating long term compounded annual returns by even a little bit is catastrophic, and still results in massive underfunding. Of course, if we're lucky, we won't actually have to recognize this reality for another 10, 20, or 35 years.  Extend and pretend, bayyyy-beee!

"The median expected investment return for more than 100 U.S. public pension plans surveyed by the National Association of State Retirement Administrators remains 8%, the same level as in 2001, the association says.

The country's 15 biggest public pension systems have an average expected return of 7.8%, and only a handful recently have changed or are reconsidering those return assumptions, according to a survey of those funds by The Wall Street Journal.

Corporate pension plans in many cases have been cutting expectations more quickly than public plans, but often they were starting from more-optimistic assumptions. Pension plans at companies in the Standard & Poor's 500 stock index have trimmed expected returns by one-half of a percentage point over the past five years, but their average return assumption is also 8%, according to the Analyst's Accounting Observer, a research firm.

The rosy expectations persist despite the fact that the Dow Jones Industrial Average is back near the 10000 level it first breached in 1999. The 10-year Treasury note is yielding less than 3%, and inflation is running at only about 1%, making it tougher for plans to hit their return targets.

Return assumptions can affect the size of so-called funding gaps—the amounts by which future liabilities to retirees exceed current pension assets. That's because government plans use the return rates to calculate how much money they need to meet their future obligations to retirees. When there are funding gaps, plans have to get more contributions from either employers or employees.

The concern is that the reluctance to plan for smaller gains will understate the scale of the potential time bomb facing America's government and corporate pension plans."

Of course, they might do better than their benchmark assumptions:
"The Oregon Public Employees Retirement System has had an 8% assumption since 1989. Its actual return averaged 10.7% annually from 1970 through 2009. The Teachers Retirement System of Texas has had a similar expectation since 1986, with an annual return of 9% return since then. 

A spokeswoman for the Texas system said it doesn't change assumptions "in response to short-term situations," and currently "sees no reason to change our investment-return assumption." A spokesman for the Oregon system said there are no special plans to review its return expectation."

I guess, for me, the point is that managers are failing to recognize the economic paradigm shift we've undergone.  The last thirty years of leverage and debt expansion are finally coming home to roost, and since we don't want to recognize the reality of that, preferring instead to try to delay taking the losses that are inevitable, growth will be impaired for as far as my eyes can see.

But obviously, there's a reason why they don't want to lower return assumptions - it creates more pain!

"The Colorado Public Employees Retirement Association showed in its 2009 financial report the impact of reducing the rate. Using a 8% expected return rate, the plan faced a $23.4 billion deficit, based on market values, at the end of 2009. If the rate was cut to 6.5%, the shortfall would jump to $34 billion.

Meredith Williams, the Colorado plan's chief executive, says cutting the rate "creates pain." Nevertheless, Colorado at year-end of 2009 cut its return assumption to 8%, from 8.5%. Mr. Williams says the rate may be lowered again later this year.

Others have been more hesitant. In 2009, Matt Smith, state actuary for Washington state, recommended that its retirement system cut its return expectation to 7.5%, from 8%. That advice was rejected by the state's pension-funding council.

Mr. Smith says he thinks Washington and other states eventually will lower expected returns, but that it will be a slow process because reduced assumptions "will increase the cost of pension benefits, and right now the budgetary environment is a big obstacle to that.""

One thing that still makes no sense to me, and perhaps a reader can help me out with this, is how pension funds can be buying long term corporate bonds (I mean REALLY long term, I'll get to that in a second) with returns below their annualized bogey.  Last month Norfolk Southern issued "century bonds," due in 100 years.  Their yield was around 6%.  Last week, Rabobank did the same thing, issuing 100 year bonds with a yield of around 5.8%.  The WSJ says "Life insurers and pension funds tend to be the main buyers for such lengthy bonds because they need to match their long-term liabilities with assets of a similar duration."

Now, correct me if I'm wrong, but if you're making a super long term investment with a 6% return, your long term return assumptions probably shouldn't be in the vicinity of 8%, right?  What do you do to recoup the difference  - make up for it in volume?   wink wink...



Anonymous said...

I had this same discussion with my wife, who was just named to the investment committee at her company. One of our questions is where are the incentives for someone on the investment committee, who would approve these (too) high return assumptions, to be conservative in return assumptions? There aren't any. In fact it's easier for them to be overly optimistic. It's easier to make optimistic assumptions rather than tell your constituents (union members, employees, etc) that either they have to contribute more or that their payouts/'entitlements' will be lower. Also, the current investment committee members will all most likely be gone/dead by the time that the pension's underfunded status becomes a major problem (let's say 25 yrs from now). Of course this is *extremely* over-simplified, but my main point is that it should be as zero surprise that return assumptions are aggressive.

getyourselfconnected said...

thanks for this post! A little late to get into it but this post is well worth a read.

Holy crap, NE vs NY tomorrow!

Duff Samoa said...

There are many ways the pension funds could justify buying century bonds. It all depends on where they get their money to finance the buying. It may not all come directly from the pension fund, some of it may be borrowed.

Yangabanga said...

seriously I thought the govt made this illegal when they just fixed everything wrong with finance? really is it any different that stealing? PROTECT US!

Kid Dynamite said...

Duff - yes, leverage - that's a whole other issue.

Anonymous said...

Another issue with pensions - at least in Chicago...

The likely driver behind the super long term bonds is that it is the closest thing pension fund managers can get to matching liabilities to investments or fully hedging their portfolio. The biggest impact on the value of their long term liability is interest rates, and these long term bonds will have values that fluctuate greatly with small interest rate changes.

Kid Dynamite said...

right anon, but there's two problems with that durability mismatch: 1) as you noted, they have hundred year bonds but they may sell them well before maturity - which means they may face a sizable gain/loss on the principal. 2) as I noted: how can they justify 6% century bonds when their long term return assumption is 8%?

if you're suggesting (And I don't think you are) that they're using these century bonds as a way to express a view that rates are going lower, i think that's just as crazy.

Anonymous said...

So yeah the rate of return should match the expected return/allocations selected by the fund. It doesn't make any sense to own a 100% portfolio at 6% and then show an expected return of 8%.

I'm no pension expert, but I think the fundamental difference between a pension fund manager and an investment fund manager is their real goal - at least in today's environment. Pension managers are not focused on maximizing returns, but ensuring that liabilities can be met - that the fund can make it's annual distributions. Any fluctuation in the current value of these bonds caused by interest rate changes would be the same as the change in present value of their pension liabilities. The manager could basically just construct portfolio of long term bonds and let it ride - sort of like an interest rate hedged annuity.

Kid Dynamite said...

anon -
the WSJ article addressed this point you made:

"Any fluctuation in the current value of these bonds caused by interest rate changes would be the same as the change in present value of their pension liabilities"

the article noted that a problem is that public pensions generally use their return assumption as the discount rate, so that the value of their liabilities wouldn't change as rates change. Private pensions, on the other hand, use corporate bond yields, so would be matched up, like you theorized.

wynngolfhatguy said...

Duration. Duration is a better tool for managing fixed income portfolio and the calculating effects of interest rates on a bond portfolio. (although changes in the shape of the yield curve can affect those calculations.) While it may seem absurd to purchase a 100 year note, it may make sense for a diversified bond portfolio to keep current income high and only increase duration of the overall portfolio slightly. (Duration of a 100 year note @ 6% is probably circa 18 years.)
“Liability matching” could better be described as “duration matching”. In the article describing JCPenney, it is easier for them to duration match because the pension is now closed to new participation. There are still some problems with duration matching because they are moving targets….meaning that over time the duration of the bond portfolio and liabilities will likely change depending on shape of the yield curve, participation, called securities, reinvestments, etc…

Why would a portfolio invest in anything that is going to return anything less than 8%? Of course investments with reduced risk, carry reduced return. A Pension/Endowment manager or Private investor may invest in safer more liquid, securities to offset more volatile investments within the overall portfolio. Singling out one security because it does not return more than 8% is the wrong way to view the overall management of the portfolio.

Just my $0.02. Keep up the good work!!

Anonymous said...

Maybe they're hoping for credit quality improvements to make up that extra 2% per year? Good luck with that...

Portfolio diversification is reasonable, but not with super long dated bonds that you can't easily sell to rebalance.

Leverage is the most likely candidate... When you're in trouble, double down baby!

Kid Dynamite said...


this reminds me of a quandary I bumped into on my first job on Wall Street on a derivatives marketing desk. They had a spreadsheet model for calculating the value of a collared forward product, but it used one rate for the growth assumption, and a different rate for the discount rate! It was basically an "alpha" creator! I pointed out that this was preposterous, and needless to say, my career at this firm didn't blossom.

Sure, duration is part of the equation, but it's only one part! There's another essential part, and that's RETURN. The math doesn't work for pensions to lock themselves into 100 year paper below their target rate - unless they lever it up.

wcw said...

KD, I'm with wghg here. You match duration while aiming for your hurdle rate on a portfolio basis. Don't throw out centuries worth of lessons in diversification because a pension portfolio accountant assigned an absolute return target.

The root of the accounting problem is banal, not malign. Defined benefit plan sponsors suffer during periods of disinflation, since they almost universally promise returns in nominal rather than real terms. Social Security, notably, does not, and as a result does not suffer this problem.

Kid Dynamite said...

WCW - i'm not throwing out diversification theory, I'm saying that this 8% absolute return target is nonsense. and their own holdings (6% century bonds!) are proof of that.

I understand the concept of duration matching. But it does no good to match duration without return to accompany it.

If their long duration assets are yielding 6%, then their long term return assumption should be... wait for it... right around 6%!

Steve said...

@KD 11:35am "If their long duration assets are yielding 6%, then their long term return assumption should be... wait for it... right around 6%!"

Their long-duration BOND asset of that particular credit is yielding 6% now. That is not a good proxy for their entire portfolio return. Although equities don't really have a duration, in a low-inflation, low-growth, lowish-dividend environment it should be long - 20+ years.

why not use that historical return as a better gauge of future return, if you are attempting precise duration-matching? Focusing on current 100-year corporate bond yields as the best estimate of 50-year future portfolio returns is a distraction.

Kid Dynamite said...

Steve -

This is what I'm saying:

If I have $1mm in the bank, and I need $40k annually to live, I need a 4% annual return. Let's pretend I'm going to live 100 more years.

If I can buy 100 year bonds that yield 2%, which is well below my necessary target return, do you know how many of those 100 year bonds I'm going to buy? None. Zero. Anything I buy at below my target portfolio return rate must be offset by something I then buy with an expected return ABOVE my target portfolio rate.

Now, what I think that you, WCW, Wynnguy, and others are saying is that you want some fixed income exposure in order to compensate in deflationary times - where I'll show profit on my bond principal, I'll have diversification, I'll have possibly non-correlated returns. Fine. I also need to adjust my return assumptions. That's all I'm saying.

If you're making long duration investments with an X% return, your discount rate (or your long duration return assumption) can't be X+2%, unless you expect to do much better than your net portfolio rate on the other portion of your portfolio.

Maybe all I'm saying is that I believe that 8% is the number that's generally used as the long term equity annualized return. Currently, I would expect the future equity return number to be even less than that, not the "more than that" it would need to be to make up for the deficit that my portfolio would already be in for locking up 6% in long duration fixed income (BELOW MY NECESSARY TARGET! - the 8% annualized return target for the portfolio is the same as that of equities - that's why you can't have long duration assets at below 8%)

Kid Dynamite said...

everyone should read Paul Kedrosky's post on the subject too:

I think he makes the point that others are trying to make in this comment thread: you mix your asset allocations and duration allocations, and weight the return assumptions. (but comes to my conclusion - that return assumptions are too high!)

we can make up some numbers:
20% long duration century bonds: 6% yld
20% short duration corporates: 3% yld
60% long duration equities/other: 9% yld

portfolio avg: 7.2%, even with the aggro equity assumption.

so basically, to Steve, WCW, Wynnguy - yes - I understand that you can have a 6% return on a long duration portion of your portfolio and still achieve an 8% overall return, but my point was that in order to do that, you need to have even crazier assumptions for the balance of your portfolio, which I believe will prove impossible. that's all.

wynngolfhatguy said...

My first point was to drive home that a 100 year note @ 6% has only the interest rate sensitivity of 17 year note. A twenty year zero has more sensitivity. Therefore, purchasing such a security could make sense for in an overall portfolio strategy. Again, it is common to assume that fixed income securities (even when rates were at higher levels) will return less than the target rate when developing an overall portfolio strategy. (There is a reason that Endowments/Pensions don’t place 100% of their assets in Private Equity and Hedge Funds….liquidity & non-correlation among them)

You could certainly argue that 8% is too high an assumption. Please recognize that the rub should NOT be buying 6% long term bonds (assuming an 8% target rate) but INSTEAD that the assumptions for riskier assets are too high. Which I think is what you stated in your last post.

Another option would be to adjust an annual pension according to the performance of the underlying assets. My home state of Wisconsin does this. It has allowed our pension to remain in strong shape relative to other States’.

Anonymous said...

There's a difference between public pensions and corporate pensions. Corporate pensions do not discount their liabilities using their return assumptions--they use corporate bond rates--there is in fact little to no impact on corporate plans regarding investment assumptions beyond year to year assumptions and contributions. Corporates may be buying these bonds to extend their overall durations but not to meet a longterm funding goal, for the same reasons as insurers. If it's public plans, who do discount based on investment assumptions, then it doesn't make sense. But to make the argument you need to make the distinction.

Kid Dynamite said...

Anon - that's explicitly mentioned in the article