PsyFi Search

Monday 26 March 2012

Do Stocks Always Outperform (in the Long Run)?

The equity risk premium isn't a reliable forecasting tool
Funny Futures

The equity premium puzzle is one of the longest standing anomalies in finance: the finding that stockmarket investing outperforms other types of investment by a significant amount. Generally you’d expect the price of shares to rise until this advantage was cancelled out, but this hasn’t happened and has made quite a lot of futurologists look rather silly.

It turns out, though, that the equity premium puzzle may be even more puzzling than it first appears. Basically researchers aren’t exactly sure how large it is, or even if it exists at all. So is the idea that stocks always outperform other investments in the long run just another market myth used to cajole unwary investors into parting with their hard-earned cash?

Witless Aversion

The puzzle was first noted by Rajnish Mehra and Edward Prescott in 1985. They pointed out that:
“Historically the average return on equity has far exceeded the average return on short-term virtually default-free debt. Over the ninety-year period 1889-1978 the average real annual yield on the Standard and Poor 500 Index was seven percent, while the average yield on short-term debt was less than one percent”.
The value they arrived at is disputed, as we'll see below, but the general problem was that if standard conditions of competition between asset classes apply then this kind of premium doesn’t make any sense. If the risk-free asset – a government bond – is returning 1% then you’d expect equity returns to fall to the same level plus a bit of extra return for the additional risk. In order for stocks to return 6% more than the risk free asset then investors need to be incredibly risk adverse – essentially scared witless – about the possibility of losing their money. This, reckon most observers, is implausible.

The Importance of the Equity Premium

Now this 6% premium and 7% inflation adjusted return is why most investing advice pushes the ideas of stocks over less risky assets such as bonds and cash. Equities are more volatile – that is, their prices vary much more – but over the long-run a 6% premium builds up to an enormous difference in the terminal value of your investments. So we’ve gravitated to a default position that younger people with time to run to retirement should invest mainly in stocks and that they should do so with an eye for the long-term, because this smoothes out the volatility.

Dig hard enough into most financial schemes and you’ll find something that looks suspiciously like the standard equity risk premium. This is inevitable, because investment managers need to make assumptions about future returns in order to manage their asset allocation. Unfortunately if these turn out to be over-optimistic then you end up with huge pension deficits or massively underperforming endowments or social security schemes you can’t afford.

Behavioral Myopia

The behavioral explanations for the equity risk premium rely on myopic loss aversion, the concept that people have incredibly short time horizons for tolerating losses. Benartzi and Thaler calculated that if investors checked their portfolio values more frequently than once a year and sold as a result of temporary losses then this alone could explain the high level of the premium: in this view the equity risk premium is caused by investors narrowly framing their investment time horizons.

Follow up research on this by David Fielding and Livio Stracca suggested that myopic loss aversion requires even shorter timescales than might be expected, to the point where it’s not a plausible explanation for the equity risk premium on its own. Instead they propose that the main driver is disappointment aversion – the feeling aroused when an outcome is less good than you might have wished, regardless of your buying price. You’re disappointed when you make a loss on an investment, but you can also be disappointed when you expect to make 7% a year over a decade and only manage 2%. Disappointment aversion doesn’t require a fixed anchor of a buying price, but can move around dependent on your expectations.

Disappointment aversion also seems to work over quite long timeframes – up to ten years – so that if expected returns on equities are good but fall below what people expect then they will experience aversion – and desert the stockmarket. Combined with myopic loss aversion this offers an explanation for the historically high equity risk premium, and also offers the skeleton of an explanation for why cyclical bear markets can last unfeasibly long times.

A Long Flavour

This is far from the only explanation, however, and Aswath Damodaran’s recently updated paper on Equity Risk Premiums (ERP): Determinants, Estimation and Implications gives a flavour of the problems – a 95 page flavour. As he points out, the equity premium is a critical factor for individuals in determining how much they set aside for their retirement savings:
“Judgments about how much you should save for your retirement or health care and where you should invest your savings are clearly affected by how much return you think you can make on your investments. Being over optimistic about equity risk premiums will lead you to save too little to meet future needs and to over investment in risky asset classes.”
As Damodaran shows, there are wide range of ways of calculating the premium and an equally wide range of results. To add to the complexities the premium is time varying, as different market conditions result in different values. Indeed the equity risk premium is very sensitive to the period over which you calculate it which leads many to believe that it should only be used over relatively short-run periods, because investor risk tolerance is more likely to be stable over such periods. Unfortunately it can flip virtually overnight – the long-run value of the equity risk premium dropped by nearly a percent in 2008 as the banking crisis took hold.

Choosing the Moment

Warren Buffett, as usual, looks at this differently. Taking a hindsight view of the twentieth century in US markets:
“To break things down another way, we had three huge, secular bull markets that covered about 44 years, during which the Dow gained more than 11,000 points. And we had three periods of stagnation, covering some 56 years. During those 56 years the country made major economic progress and yet the Dow actually lost 292 points.”
This happened despite the US economy growing in every decade of the last century.  Buffett’s point is that stockmarkets are poor proxies for the underlying state of the economy, although they’re probably very highly geared to the general state of fear, uncertainty and doubt in the wider world. Generally stockmarket rises are coupled to improvements in corporate profitability and interest rate falls. Again, as Buffett points out:
“As Keynes would remind us, the superiority of stocks isn't inevitable. They own the advantage only when certain conditions prevail.”
Based on Buffett’s own rule of thumb estimate of market value of all publicly traded securities as a percentage of GNP the market is currently modestly overvalued: which doesn’t mean it can’t get cheaper, but gives an estimate of about 4.4% a year market return for the next eight years. Which implies an equity risk premium a bit lower than generally assumed.

Myth or Miracle?

As we saw in Triumph of the Pessimists, the general view that you can’t go wrong with stocks in the long-run is almost certainly wrong. Elroy Dimson’s research implies an equity risk premium nearer 4% than 6% and a wide, wide range of possible outcomes. Furthermore it points out that US and UK investors’ favourable view of stockmarket returns is likely an issue of survivorship bias because most other countries required an investment horizon of over twenty years in order to guarantee an inflation adjusted positive return.

So, do stocks always outperform (in the long run)? No, they don’t. And don’t expect long-run market returns in excess of 7% going forward. It might happen, but it probably won’t.


Related resources: If you're interested in the intricate workings of securities valuation you could do a lot worse than visiting Professor Damodran's website Damadoran Online.  The Buffett Total Market Cap to GNP ratio for the US can be found at gurufocus. The Schiller-Case cyclically adjusted PE ratio is at Robert Schiller's on-line data site.

Related articles: 

4 comments:

  1. What about transaction costs? I have an intuition that a sizeable proportion of the historical equity premium may be explained by them. Tim, you seem to have encyclopedic knowledge of the academic literature, have you never stumbled on this idea before? I can't possibly be the only one who came up with the idea (however right or wrong).

    ReplyDelete
  2. Hi cig

    There is a related argument, but it’s not quite the same. The argument is about liquidity – if stocks are illiquid (have high transaction costs) then investors will demand a high equity risk premium. As stocks become more illiquid during crises (more difficult to shift) then the premium goes up.

    It’s discussed in the Damodaran paper quoted above, but it’s only one of a lot of potential factors. Basically, yes your intuition is good and you could be right, but it’s unlikely this is the whole story.

    ReplyDelete
  3. You need to read the works of Antti Ilmanaen, who shows that the equity premium is low, as Eric Falkenstein and I point out.

    http://www.wiley.com/WileyCDA/Section/id-302475.html?query=Antti+Ilmanen

    Eric's book "Finding Alpha" builds on that.

    ReplyDelete
  4. Hi David

    Thanks for that. I haven't read Antti Ilmanen, but Rob Arnott and Peter Bernstein suggested that the long-run equity premium might be as low as 2.5%, with a wide variance in returns for specific periods. See Arnott: 40 Years of Bonds Beating Equities".

    In the end I don't actually think there's a single "correct" number. Probably we'd all be better off assuming that it's on the low side and operating accordingly. Ultimately saving a bit too much is safer than saving a bit too little.

    ReplyDelete