Stay informed with free updates
Simply sign up to the Pensions myFT Digest — delivered directly to your inbox.
Two key principles run through how we save for retirement. One is that we should mix different asset classes — exciting stocks and resilient bonds — in an effort to diversify and balance out risks. The other is that as we get older, we should cut back on stocks and instead build up on that safe, reliable debt so nothing can go wrong when we’re no longer earning.
It’s simple. It’s elegant. It’s distilled into formulas that most of our pensions follow. But an academic paper published late last year in the US sets out an awkwardly strong argument that it’s rubbish, and that in fact we should stick all our retirement wealth — 100 per cent of it — in stocks. And leave it there. The paper looks at how a typical couple, a woman and a man, starting to save at age 25 can hit three critical targets: amassing as much money as possible before retirement at 65; providing financial security through retirement; and leaving a bequest for the next generation.
It examines markets data going back to 1890 in countries where it is available, or else for as long as possible, to figure out the optimal balance of asset classes. To consternation in some quarters, it finds the best mix is 50 per cent in domestic equities and 50 per cent in international equities. That’s it. No bonds, no shifting the mix around with age, just 50:50.
The study assumes that the hypothetical couple saves 10 per cent of gross income, and then starts off draining 4 per cent of those funds a year at retirement. If the money runs out, they rely on social security, which is based on the US model, as is longevity.
On average, parking entirely in stocks produces some 30 per cent more wealth at retirement than stocks and bonds combined. To get the same cash pile at retirement, someone blending into bonds would need to save 40 per cent more than an all-stocks saver. That is striking, but not so surprising — stocks have beaten bonds, and bills, everywhere every year since 1900, as UBS’s recent long-term markets study noted.
But the really odd bit concerns the chance of so-called ruin, or running out of funds in retirement. A portfolio that bulks up in bonds as we get older comes with a 17 per cent chance of ruin. For domestic equities, the risk is about the same. But a half-and-half strategy of domestic and international stocks produces an 8 per cent chance of ruin.
Over short periods, like a month, stocks and bonds do tend to move in opposite directions and provide some stability and balance. But “long-term returns can differ quite a lot”, said Professor Scott Cederburg from the University of Arizona, one of the authors of the report. “They become highly correlated for long-term investors so bonds stop providing so much diversification benefit.” These basic findings are applicable to the US and other countries, he said.
There are two problems with this analysis. One is humans in general. Pullbacks, or drawdowns in market parlance, are more severe with just stocks. Cederburg — who is following this approach himself — said the “psychological pain” of seeing savings shrink is brutal, and people often pull out when markets are falling, to their detriment. Missing the upswing after a downturn is a classic mistake. Ideally, he said, savers would be better educated about the ups and downs of long-term investment.
The other problem is one human in particular: Cliff Asness — heavyweight quant battler-in-chief. He never shies away from an argument, and this is no exception. In a response to the paper in January, he described the 100 per cent equities approach as “finger painting”. “Both the new and old version of this argument . . . come down to the rather trivial observation that the asset with a higher expected return [stocks] has, on average, a higher realised return,” he wrote. “Staggering stuff.”
A classic mix of stocks and bonds does not deliver a higher unconditional expected return, he said, but “we believe it has a higher return for the risk taken”.
Cederburg has his own pushbacks to that, in particular that the long and international nature of his data series helps to overcome any outsized influence from several decades of upbeat US stock market performance. The data may have its own problems for direct international comparisons but, he said, “We think it’s better to have more data”.
In any case, he is now fielding lots of inquiries from investment managers and the like, keen to pick apart his conclusions or learn from his findings, or both. That is healthy — market practitioners should not be shy about questioning long-held practices, particularly now the extended era of low inflation and low interest rates has ended. In our sunset years, we all depend on them getting this right.
