When investors get stuck in the past

When past returns have been high, future returns are more likely to be low. But people don't realise that.

THE award of the Nobel economics prize to Richard Thaler is a reminder that economics has been struggling, in the past 30 years, to adapt its models to the real-life decision-making processes of actual humans. The problem applies as much in investment as anywhere else.

One of the biggest problems is the tendency to assume that the future will resemble the past. Despite all the warnings inserted by regulators, investors often believe that fund management performance will persist, but the evidence is against it. Another issue is the assumption that overall market returns will persist. 

This may be one of the factors explaining the big deficits at state and local pension funds in America. Those funds are allowed to make their own calculations at the expected rate of return on their assets; the higher the assumption the less taxpayers and employees are required to stump up.

From where do they get their numbers, which tend to be in range of 7-8% return? It is hard to believe they haven’t been affected by past returns, which were 7.8% over 25 years and 8.3% over 30 years. The median assumption has edged down from 7.91% in 2010 to 7.52%, according to a report from the National Association of State Retirement Administrators. Since the turn of the millennium, it has probably fallen around half a point. But there is a big difference in valuations between now and then; at the start of 2001, the 10-year Treasury bond yielded 5.2%. It now yields 2.4%

Why does this matter? The return from bonds is clearly the starting yield plus or minus any changes in valuation and minus any losses from default. When bonds move from a higher yield to a lower yield, there is a rise in price which gives investors an excess return. To believe that this return is repeatable, you must assume that yields will fall even further and that gets increasingly less likely as yields approach zero. 

Let us say for the sake of argument that a pension fund has a portfolio split of 70% equities and 30% bonds; assume also that with the help of corporate debt, the fund can earn an extra percentage point on its bond portfolio. At the start of 2001, then, a pension fund could expect 6.2% on its 30% bond allocation; or 1.86%. That required it to earn 8.8% on its equity portfolio in order to reach an 8% target. The equity risk premium required was 8.8% minus 5.2%, or 3.6 percentage points, slightly below the historic average. But run those numbers now and you get 1.02% from your bonds (30% of 3.4%) requiring a 9.25% return on equities to get to 7.5%. That equates to an equity risk premium of seven percentage points, way higher than history.

Is that in any way justified? As our briefing in last week’s issue pointed out, valuations are pretty high across the board. American equities are on a cyclically-adjusted price-earnings ratio of 31, according to the website of Robert Shiller, a previous Nobel winner. In the past, high valuations have been associated with low future returns.

So it could be argued that high market valuations are quite rational, if investors believe future returns will be low. But the numbers from the pension fund industry suggest that is not the case at all; cognitive dissonance is going on (a well-known behavioural trait).

The low rate conundrum

Another approach is to say that high valuations are justified by low interest rates. An equity is worth a stream of future cashflows, discounted to the present day; as long-term yields fall, so does the discount rate, and the present value of an equity rises.

But that only deals with one part of the equation, as a paper from strategist John Hussman argues. 

If interest rates are low because growth rates are also low, no valuation premium on stocks is “justified” by the low interest rates. Prospective returns are reduced without the need for any valuation premium at all.

In other words, if the discount rate falls because of slow growth, so should your expectation of future cashflows (profits). And future growth is likely to be slow given the demographics (a flat-or-falling population of working age in many western countries) and very poor productivity growth (Estimates for Britain have just been revised lower). 

What has kept profits high, as Mr Hussman remarks, is the shift in power between capital and labour in the wake of the financial crisis; a shift that has depressed wages and fuelled the rise of populism. This can’t continue, he adds:

There’s no way to make the arithmetic work without assuming an implausible and sustained surge to historically normal economic growth rates, a near-permanent suppression of interest rates despite a full resumption of normal economic growth, and the permanent maintenance of near-record profit margins via permanently depressed real wage growth, despite an unemployment rate that now stands at just 4.2%.

Again, investors seem to assume that past conditions will simply persist, despite the evidence. This behavioural bias will prove their undoing.