There was an amusing WSJ piece over the weekend talking about how most U.S. pension funds are sticking to 8% target returns on their portfolios and in their actuarial assumptions, despite not having delivered that sort of return in a decade. Kudos to them for their stubborn insistence on ignoring pesky things, like reality.
Pension plans say they take a decades-long view of potential returns. "We can't knee-jerk our way through this. Funding a retirement system is a long-term proposition," says David Stella, secretary of Wisconsin's department of employee trust funds. Last year Wisconsin's plan reviewed its expected return rate of 7.8% and remains comfortable with it, he says.
Companies have found out the hard way that their options are limited. From 2005 to 2009, S&P 500 companies with pension plans expected to generate about $475 billion in returns. The actual returns were only about $239 billion, a 50% undershoot, according to Jack Ciesielski of the Analyst's Accounting Observer.
Of course, the troubles are manifold. It's not just that the assumptions are wrong, but that changing the assumptions make the situation worse. After all, if you now say 8% is too high, and that maybe 7% is better, you just took your pension fund liability and made it dramatically larger. That, in turn, means you either need to increase contributions, decrease benefits, or, and this is the best one, make squirrely bets and take on leverage in search of higher returns, thus increasing the possibility that you blow up. Such fun.
Here is a fund manager friend who runs such things making a few practically-minded suggestions in email:
Wouldn't the best way to constrain overpromising of future benefits be to constrain the net present value return assumption target? That might be 15% current short duration government yields, 15% long term government duration yields and a long-run equity assumption of 9% on the balance (equating the risk portion of the investment portfolio).
In many respects, all of the mumbo jumbo on asset allocation policy is designed to "best" these long term targets. I wouldn't constrain mandates but the use of such a policy framework would provide political utility. The problem is the politicians have spent return assumptions (that have not come through) and so a more realistic and conservative approach to those assumptions should reduce future benefit spending (that has yet to materialize). Just thinking out loud...
So, today it might look like:
20% 1.0% short yield curve;
20% 2.7% long yield curve;
60% 9.0% long duration on the equity risk premium curve.
That becomes a 6.14% effective discount rate. The notion that leverage, hedge funds and other "things" can make up a huge premium to the risk-free rate is a big problem in actuarial assumptions. At this rate, if the markets do come close over time to the effective long term equity rate (which I still believe), then you constrain spending in the short term. This also becomes somewhat counter-cyclical.