Ever wonder how public pension systems forecast that they will earn 7.5 percent to 8 percent annually on investments? It's like making sausage: If consumers knew the process they wouldn't buy it -- or, in this case, buy into it.
Yet, when a pension board sets its assumed rate of return -- how much it expects to earn from holdings such as stocks, bonds and real estate -- the decision significantly impacts taxpayers.
The higher the rate, the less employers and employees must contribute now to fund future retirement benefits. But if projections don't pan out, the shortfall must be paid off solely by future taxpayers.
Labor unions advocate for a high rate because it leaves more money for salary and benefits now, and future shortfalls aren't their responsibility. Many government officials similarly figure they won't be around when the bill comes due.
Pension boards, usually dominated by those unions and elected officials, set the rates after obtaining recommendations from actuaries and investment advisers. Unfortunately, these experts, at best, make educated guesses that often prove wrong.
Investment shortfalls are part of the reason California taxpayers now owe at least $250 billion -- $20,700 per household -- for unfunded public pension liabilities across the state.
How are these investment forecasts derived? Let's look at an example from the California Public Employees' Retirement System, the nation's largest pension fund.
In March, for the second consecutive year, Chief Actuary Alan Milligan recommended that the board lower the assumed rate of 7.75 percent. The rate represents expected annual returns over the next 19 years. To develop his recommendation, Milligan divided the time period.
For the first 10 years, he turned to Joe Dear, CalPERS' chief investment officer. In 2010, Dear recalled last week, he solicited projections from three investment consulting firms, talked with other pension systems and conferred with members of his executive team. He then issued a 10-year forecast of 7.38 percent annually.
The next nine years were up to Milligan. He told me that he took Dear's number, looked at stock market returns going back nearly a century, and concluded he should increase the forecast for years 11 and beyond. For that period, he projected an annual return of 8.50 percent. In other words, based on past market performance, Milligan predicted investment returns would increase a decade from now.
Melding the two numbers together and subtracting administrative expenses, Milligan last year forecast an annual 7.80 percent return for the 19-year period. He said there was a 50 percent chance that the system would meet or exceed that target.
CalPERS had previously built a margin for error into its rate. Milligan advocated keeping that cushion, so, in 2011, he recommended a 7.5 percent rate. The board said no. It eliminated the cushion to justify keeping the rate unchanged at 7.75.
This year, using the same investment forecast, Milligan tried again. This time he recommended a reduction to 7.25 percent, half of the change to account for lower anticipated inflation and half to build back in the cushion. The board set the return assumption at 7.5 percent, agreeing only to the inflation adjustment.
If all of this sounds unscientific and speculative, that's because it is.
At the root of the numbers is Dear's 10-year projection. "In the near term one should never have a lot of confidence in these forecasts," he told me last week. But he firmly believes they will hold up over the long term, which he defined as 50, 75 or 100 years. That's hardly reassuring when one considers the numbers are being used for projections over the next two decades.
To many in the academic community, this is crazy. I agree. Pension systems are setting contributions by banking on investment returns that, even by their own admission, they have only a 50 percent chance of hitting.
It's backward. Pension systems should set rates based on more conservative current bond yields, which they know they can attain. If they want to invest their money elsewhere and try to attain greater returns, they should only factor in the extra money when they actually earn it.
Right now, the entire system is predicated on overly optimistic assumptions. We already know how well that's worked out.