The Life-Cycle Personal Accounts Proposal for Social Security, Shiller (2005)

Back in the early part of the 21st century there was a lot of talk about privatising Social Security. In that context, Robert Shiller (of Irrational Exuberance) fame performed an investigation into the historical performance of life-cycle portfolios using the historical data he had gathered.

(He also published “Life-cycle Portfolios as Government Policy” the same year but that is more of a synopsis of this paper than anything new.)

A lot of the paper is about specifics of the Social Security privatisation proposal but he does make some points that are more relevant to other applications of glidepaths.

His first point is that there seems to be no academic agreement on what glidepaths actually make sense: he references some research that young investors should only have 20% of their portfolio in equities, while other research says they should have 300% of their portfolio in equities (using massive leverage). All with reasonable sounding assumptions. Given the complete lack of academic agreement — which translates to differences in what current life-cycle funds offer.

We learn from these studies that there are a lot of difficult issues to confront in designing the optimal life-cycle portfolio, and that there is an extraordinarily wide range of possibilities, not just the possibilities that might sound intuitively plausible at first glance.

Given the differences how much faith can we put it any particular glidepath, even if we accept the idea that glidepaths are the “right” way to do things? Academics don’t even agree whether the life-cycle model implies the young should have more equities than the old.

His second point is just an observation that bonds don’t pay very much and with a life-cycle most of your money over your life will go into bonds. That’s good for low volatility but bad for returns.

He shows that with two glidepaths (one conservative ending in 10% bonds; the other more aggressive ending at 40% bonds) you have a decent chance of sub-4% lifetime internal rates of return. A lifetime of saving & investing just to earn 3.8% isn’t exactly amazing but that’s the risk you run when you put large amounts of your savings in bonds. (Of course, stocks come with their own set of issues.)

Ultimately, I find the lack of academic consensus on glidepaths a large weakness. Should they glide up or down? How quickly? From what starting point?

In the absence of a strong theoretical underpinning, I’m inclined to give a lot more weight to the performance of historical backtests instead. So our next entry in the glidepath survey will be Estrada’s two papers…