Three Years of Pension Returns Do Not Tell Us Much About Thirty Years

Pensions saw strong double-digit returns in the last years of the 90s. They had negative returns in the first years of the last decade. If anyone extrapolated from the last three years of the 90s and predicted double digit returns for the decades ahead or the first three years of the last decade and predicted flat or negative returns, they should have been taken far away from any position with any responsibility for pensions.

Unfortunately, the world does not work this way. Such absurd extrapolations are common practice as Sacramento Bee columnist Dan Walters showed us in a discussion of the returns on California pensions. Not only is it foolish to assume that a recent past of low returns will translate into low returns in the future and vice versa, the opposite is in fact true.

The plunge in trend price to earnings ratios (PEs) associated with the collapse of the stock bubble in 2000-2002 meant that it was possible for the stock market to provide higher returns over a sustained period. In the same way, the run-up of PEs to unprecedented levels at the end of the 90s ensured that returns would have to be lower in the future.

Those familiar with arithmetic understand this simple point. Unfortunately, relatively few such people are admitted to the debate over pension fund accounting.

California did it..
written by pete, July 18, 2012 10:01

State employee benefits were increased big time during the dotcom bubble based on assumptions of average market returns of 8 or 10%, which is more or less the average from year 0 till that time. Then this genius was put in charge of the study of the 2008 crash. Many folks feel those days are toast. With GDP growth at around 2%, maybe, hard to see firms getting 8 or 10% return on capital for very long. There are no studies showing the benefits of market timing based on the PE mean reversion you are suggesting. The opposite is in fact true, many studies showing attempts at market timing are ill fated. Here is a nice example, but there is a lot of research on this subject. Buy and hold trumps market timing. Only brokers want investors to trade.

http://www.nber.org/papers/w4890

...
written by leo from chicago, July 18, 2012 10:19

MarketPlace had the same dumb discussion:

http://www.marketplace.org/topics/your-money/calpers-earnings-fall-flat

PEs are still elevated
written by john, July 18, 2012 1:04

Dean has been right on about the housing bust. But, his weakness is when it comes to discussing public pensions. They are in horrible shape precisely because they figured 8% returns would continue forever. This is how California justified their 3/30 pension formulas for safety workers which is now destroying the finances of cities throughout the state. The UC system did a similar thing when they justified not having ANY contributions to their pensions system for over a decade.

And their is no evidence whatsoever we are poised for a return to double digit pension returns. When you have 10 year bond yields at 1.5% and Shiller PEs very elevated compared to historical averages you can say quite confidently it is very likely that future returns will be much be very poor over the next decade.

Not market timing
written by Dean, July 18, 2012 1:04

Pete,

please read my note and the work cited. There is absolutely nothing in there about market timing. Long-term returns will depend on the price you pay for stock. This doesn't require extensive study, it is arithmetic.

the other extrapolation
written by tew, July 18, 2012 7:11

Did Dean protest when states like California dramatically (and retroactively) boosted pension benefits and reduced or eliminated employee contributions based on temporary gains and an unsustainable stock market bubble in the late 1990s?

If it is wrong to extrapolate the past few years' returns why was it OK to saddle the taxpayers with pension benefits similarly derived?

Three Years of Pension Returns Do Not Tell Us Much About Thirty Years

About Beat the Press