The Trouble With Pensions

fatal error computerOne of the most captivating presentations at Benefits Canada’s annual Pensions and Benefits Summit in April was given by Malcolm Hamilton, a member of the Canadian Investment Review editorial advisory board, and a Principal with Mercer Human Resource Consulting Limited. It was a session not to be missed. Herewith, some excerpts:

“I glommed on in the 1980s to asset and liability modelling and of all the work I ever did in my consulting career, I suppose it was the thing that I enjoyed most because it was very complicated, you had to develop programs and it was hard to communicate. You could just see the awe on the face of the audience as you would present all of these graphs which had no obvious conclusion.”

Therein lies the problem, he thinks, for DB pensions: lots of stats, and no obvious conclusions, or rather a conclusion that hasn’t held up over the passage of time.

“I noticed early on that all the studies ended with the same conclusion: that it would be roughly 60/40 for no apparent reason. And this sort of undercut the perceived value of the work. And I eventually came to realize it was because actuaries were so excited about getting into the modelling that there are two questions that you really should ask before you start managing somebody’s risk. You should ask, before you do the modelling, not after. The first question is whose risk are you managing. The second question is why are they taking risk.”

Not knowing the whose risk you were managing meant that actuaries simply borrowed from economic models, put forth by Nobel Laureautes, that examined the ideal savings and consumption pattern over an individual’s life time. “That was basically the holy grail of personal financial planning,” Hamilton explains. But “it turned out to be not a terribly practical way for people to save for retirement.”

For better or worse, the actuaries got there first, at least with DB plans. “Unfortunately they probably got it wrong,” Hamilton notes. “[I]f a defined benefit plan means guaranteed benefits, then the cost of the benefit that you’re promising to the employees is really a riskless benefit. It doesn’t depend in any way on whether you fund the benefit or how you invest the pension fund. You are just basically offering annuities and they have values and you can estimate those values without knowing how you’re funded, how well you’re funded and what your business is or how you plan to invest in pension funds. So their conclusion was that the defined benefit plan was really two different things. There’s a promise that employers make to employees to pay guaranteed things. That’s easily be priced. And then there’s this second exercise which is more bewildering, which is legal requirements, how the money gets assigned, how the money gets invested, the gains and losses seem to largely accrue to shareholders, so how should that be invested. That really doesn’t have any thing to do with retirement savings? It has to do with where shareholders are told by management that the company has to set aside a pot of money in trust.”

But, at least in the corporate sector, investments do matter. The basic issue, Hamilton argues, is whether shareholders want the company to take risk. At first, economists thought, “[i]f you take risk, shareholders will expect to get better returns because they expect to get better returns when they take risk. If you don’t take risk, they’ll be content with lower returns because whatever you as a corporation do, however much risk you decide to take or not to take, your shareholders can largely undo that by changing their own portfolio.” However, the later and enduring conclusion was that the DB fund was a tax shelter and therefore should be invested entirely in fixed income.

“What we’re seeing today is that we’re finally catching up with th[is] 1980 insight,” he adds. “A DB plan is there forever, so a DB plan is forever young and has an infinite horizon and is a sophisticated investor. In essence, it took the lifecycle model and said if you put in ageless individual who live forever, how will the model work and the answer was you basically increase your expected return by taking on risk. The actuaries looked at that and realized that by increasing the expected return, they can set less money aside and reduce the cost of the pension by taking risk. Better still, since we have an infinite horizon, if we just ignore what’s going on in the short-term and rely on the randomness to average out over our very long investment period, we can pretty much ignore the risk.

That, believe it or not, was pretty much the way DB plan investment started. It was very lucrative to take risk because you assumed you were getting a pleasant risk premium without really having to bear the risk. The actuaries could deal with the risk by ignoring it or smoothing it or amortizing it. The plans in the early days were small, the fund did pretty well – nobody asked any questions; it’s not as if it was on the balance sheet. Nobody paid any attention. So basically the whole thing started down that road. It didn’t work. It took a long time not to work.”

The three problems with plans today

Hamilton identifies three issues. One is that plans today are mature. Second, “[i]t’s well known in the investment community that you can’t distinguish the harebrained scheme from the good scheme in a bull market because all the hare-brained schemes work in a bull market. It’s only when you get to bear markets, it’s only when you have mature pension plans that you find out what works and what doesn’t.”

The third issue is risk. “The accounting system basically told corporations that if you take risk you can tell your shareholders that your pension is less expensive and you don’t have to tell them, by the way, that the risk is higher. You don’t have to tell them that they’re not being compensated for bearing that risk. You just have to tell them that you’re making that pension plan less expensive. With the passage of time in the private sector, the accountants have eventually concluded with the economists have it right.”

How sizable is that risk? Plans “are now confronted unavoidably with the very high cost of guaranteed pensions in a low interest rate environment. They’re looking at that and they’re doing exactly what they should do. They’re looking at that and saying, you know, guaranteed pensions cost a fortune. We don’t think employees would pay for it. If it would cost 30% of pay. Giving employees a guaranteed pension, do you think they would voluntarily pay the 30%? Do we roll the salary back by 30% and say don’t worry about the rollback because you have a guaranteed pension. And the conclusion, I think rightfully, is the cost of pensions today.” So, if DB pensions were properly priced would employees voluntarily pay for them?

In the private sector, the solution is a targeted benefit plan. “[T]he target benefit plan basically looks like a DB plan, but the message to each of the members is different. It isn’t: this is your guaranteed pension and you get it no matter what. The message to members is that this is the pension we would like to deliver and we will deliver as long as we can get returns in this order of magnitude. But if we can’t get the returns for a period of time, whether because interest rates are low or stock markets do badly, then we’re going to reserve the right to reduce the pensions and we won’t reduce them permanently. It’s not for all time, we just reduce them for the time the plan is underfunded and if things return to normal, if interest rates go up, if stock markets improve then pensions will go back to where they were, perhaps even more.”