PsyFi Search

Saturday 27 February 2010

Behavioral Portfolios

Beer Balancing Behavioralism

Perusing the literature on behavioural finance you might be inclined to the thought that although this stuff is all very interesting and even occasionally amusing it’s not much use when you come to actually investing. It’s a bit like posture – it’s easy enough to point out to someone that they slouch like a sloth with a dose of haemorrhoids, it’s entirely another to explain to them how to retrain their muscles to start balancing pitchers of beer on their head on the way back from the bar.

Digging deeper, though, what becomes apparent is not that we're especially bad at understanding investing but that the way we go about constructing our portfolios is radically different to what theorists expect. Whether it's the theorists who are wrong or the investors is entirely dependent on your perspective. Either way recognising what's going on is an important step on the way to getting a beer balancing musculature.

Behavioral Portfolio Theory

As an example it's a strange fact that investors tend to prefer dividend paying companies to non-payers; a finding that, as we saw in The Psychology of Dividends, can't be predicted by standard portfolio theories. It can, however, be predicted by behavioural finance along with a bunch of other results that are otherwise hard to explain.

As far back as 1997 Hersh Shefrin and Meir Statman published a fascinating paper on Behavioral Portfolio Theory which starts to get into some core body building. What they pointed out is that behaviourally biased investors don't have investments which look like the mean variance portfolios that Markowitz’s Portfolio Theory suggests they should.

Markowitz's Main Variance Portfolios

Portfolio Theory has been the basis for most portfolio design for the best part of half a century. It relies on reversion to the mean and aims to create portfolios to capture maximum returns at the least risk to the investor. Importantly it treats the portfolio as a whole, not as separate parts to be individually managed. Each part is there in order to ensure that the overall portfolio meets its risk-return goal. Above all, each part has to be analysed in terms of its correlation with other parts. So, for instance, if emerging market stocks and commodities are closely correlated there would be no need to have both in an efficient portfolio, since there would be no diversification benefit to doing so.

In mean variance portfolios you would also expect to see short positions and even securities bought on margin. Individually these may be highly risky investments, but if they allow the portfolio designer to achieve the maximum-reward for least-risk position – lying on the so-called Efficient Frontier – then the individual risk is irrelevant.

Investor Zeitgeist

Shefrin and Statman contrast this mean variance portfolio design with what investors actually create. By and large you don't find many investors calculating the historical co-variances of their portfolio constituents to create efficient portfolios. In general they rely on a large amount of intitution about what's undervalued and/or what's about to catch the zeitgeist of the moment.

Moreover, investors rarely regard their portfolios as a holistic whole. At worst they account for each security separately, applying anchoring, loss aversion and framing biases to create portfolios that look like they were designed by a spider on caffeine, if they can be said to be designed at all. Shefrin and Statman suggest, however, that there is an underlying structure to such portfolios which is underpinned by what they call Behavioral Portfolio Theory.

Downside Protection, Upside Potential

They argue that investors of all kinds, private and professional, for the most part don’t integrate their portfolios to achieve the minimum risk, maximum return outcome but, rather, the standard behavioural portfolio is a simple two layer affair – a downside protection layer designed to protect against worst case scenarios and an upside potential layer which provides a chance of getting rich. Simple this may be but the architecture allows the authors to propose explanations for a wide range of behavioural biases.

The downside protection layer is supposed to be a zero risk combination of stuff like of cash and bonds, investments with low levels of risk with high certainty of future returns. The upside potential layer is where we find risky investments such as stocks and derivatives; securities with higher levels of risk and lower levels of certainty. The behavioral investor will, once they've satisfied themselves that their downside protection is in place, often regard the upside potential layer as money they can afford to lose – which leads to investments being made with low levels of analysis or understanding and often ends up being a self-fulfilling prophesy.

Lotteries on the Upside

This two layer structure can explain a remarkably large number of behavioral finance phenomena. Consider, for instance, the oft-noted tendency of investors to buy both insurance and lottery tickets. As the writers put it:
“Lotteries contain no fundamental risk, meaning risk that is related to economic events. Instead, they have risk that is manufactured artificially. Behavioral [investors] buy lottery tickets for their upside potential layers when their aspiration levels are very high relative to the amount they allocate to upside potential layers. Investors with $1 cannot have a shot at a $5 million aspiration level other than through lottery tickets. Investors who allocate more money to the upside potential account and investors who have lower aspiration levels might satisfy their aspiration levels by buying call options rather than lottery tickets. Of course, mean-variance investors never buy lottery tickets.”
Mental Accounting of Portfolios

The behavioral portfolio layering concept is completely foreign to mean variance investing of the Markowitizian kind where all elements of the portfolio lock together. At root, of course, the behavioral portfolio is Mental Accounting writ large and holders of such portfolios will go to extreme lengths to avoid the losses in the upside potential layer impacting the value of their downside protection layer. This leads to the avoidance of investments in the upside layer that can overrun such as short positions and buying on margin - both of which are essential to achieving efficient mean variance portfolios.

Dividends in Behavioral Portfolios

What then of dividends? Well, the behavioral portfolio structure suggests a reason why dividends continue to exist despite economic logic dictating they shouldn't, provides a hint as to why changes in dividend policy can provoke significant changes in stock prices and explains why dividend cuts from major corporations can create a furore amongst private investors. Shefrin and Statman suggest that the reason is simple – although the capital gains from stocks are allocated to the upside potential layer the income from dividends is placed in the downside protection layer. It's the violation of this sacred protection that triggers the excessive response.

If the idea behind behavioral portfolio theory is correct – and it certainly feels like it touches a lot of investing nerves - then it's fairly obvious why Harry Markowitz's ideas are largely ignored by swathes of private investors. It's not that Portfolio Theory goes wrong surprisingly often – it does, due the breakdown of historical co-variances between asset classes leading to inadvertent acceptance of excessive risk at the extreme ends of market variation – it's that most people can't see beyond their beyond their own psychological desire to protect themselves.

Value Investing on the Downside

It's just unfortunate that these attempts at protection often to lead to portfolios that offer greater risk for less return than anything standard Portfolio Theory could come up with. Mean variance investing may go wrong more often than you'd expect but basic logic has, so far, always re-asserted itself. Not so in the behavioral portfolio, where crashing and booming markets often leads to investors changing their upside/downside ratios at exactly the wrong time: just ask all those people who took on larger mortgages and riskier investments right at the top of the market and then did the opposite at the bottom.

So we have a situation where the analytical approach to investment is ignored by the masses in favour of bias driven intutition, but where the former often fails due to the biases of the latter causing unpredictable and extreme market behavior. The irony is that if more people invested analytically then the analytic models would fail less often. Of course, between these two stools lies the value opportunity, if you're knowledgable enough to take advantage of it.

Forearmed is forewarned, as they say. Time to practise some beer balancing.

Related Articles: Markowitz's Portfolio Theory and the Efficient Frontier, Correlation is not Causality (and is often Spurious), Mental Accounting: Not All Money is Equal


  1. I don't think that's right, Tim.

    The "analytic" approach is filled with biases and subjectivity. It is made to look analytic. But that's primarily done for marketing purposes. I see very little genuine objectivity in what you refer to as the "analytic" approach.

    I view the "analytic" approach as a pure disaster. I don't say that investors get it all right either. They would benefit from the insights that could be generated by a genuine analytic approach. But I think you are too soft in your assessment of what you call the "analytic" approach while being in relative terms too harsh in your assessment of the investor approach.


  2. There are confounded factors regarding dividends, though, aren't there. In the real world (as opposed to MPT), even the inhumanly rational Homo economicus is working with imperfect information. The realized cashflows of dividends are harder to fake than accounting profits, and so the information about dividend-paying stocks is less imperfect. Whenever the market price of risk is positive, it is therefore rational to value a dividend-paying stock higher than an otherwise equivalent non-dividend one. In that respect the behaviors of the rational investor and the "2-layer" investor cannot be distinguished.