27/11/2007

Behavioural Finance

While financial markets may be ‘efficient’ from time to time, there are many occasions when behavioural issues such as herding or conservatism create mis-pricing opportunities, according to research published today by leading fund manager, Standard Life Investments.

In the latest edition of Global Horizons, its regular look at long-term investment themes, Standard Life Investments examines a number of practical ways in which analysis of fund flows, investor surveys and data from derivative markets can be combined, together with more fundamental analysis, into a Focus on Change process in order to help understand better the drivers of current market movements and the triggers for future changes.

Frances Hudson, Global Thematic Strategist at Standard Life Investments, said:

“The field of behavioural science has made an important contribution to understanding individual investor behaviour. We have looked at how behavioural measures work at an aggregate level and the extent to which they can inform investors not only about current and historic market developments but also turning points and likely future outcomes. Such analysis should be an essential part of the tool kit which investors need, fitting into a Focus on Change approach, to give a better outcome for decision making.

“Essentially, behavioural finance starts where rational investing stops. For example, persistence or momentum offer the opposite side of the coin to mean reversion and fully efficient financial markets. Behavioural factors such as herding or overconfidence can manifest themselves in periodic market features, such as crowded trades or speculative bubbles.

“Trading rules based on acting when extreme values are reached, on a mean reversion principle, can be undone by persistence in financial markets. Investors, faced with evidence of a bubble, whether in UK house prices or (Chinese) Hong Kong equities, can choose to focus on fundamentals, such as increasingly expensive valuations. This brings scant comfort to those who are out of a market that continues to rise on a wave of momentum.

“There are a number of ways in which behavioural finance issues can be observed. Approaches can be as simple as drawing conclusions from the readily observed behaviour of participants, or as complex as the artificial intelligence systems developed at SLI which are based on biological neural networks that are capable of ‘learning’ relationships between data. For instance, we can look at marginal buyers and sellers as an indication of the type of market, and identify possible turning points from their changing actions. Established databases of fund flows, intelligence on director dealing, and investor surveys, are all fertile ground to aggregate and interpret information and gain valuable insights.

“Stock broker measures of risk appetite cover a range of asset classes, and use a variety of measures based on positions and prices. On examining these, it is fairly apparent that some of the measures are actually more concerned with measuring risk itself, and hence their usefulness as a proxy for risk appetite can be questioned.

“Our analysis shows that the top performing indicators across all asset classes are investor surveys. Changes in investor sentiment can trigger changes in asset prices and provide a better explanation of short term movements in asset prices than changes in fundamentals. Surveys do have drawbacks though, including investor delays in acting on impulses, or volatility reflecting different types of retail, trader or speculative investors. Hence, we take more notice of survey data when it shows extreme readings.

“Mathematical models, such as neural networks, can process financial market data very efficiently, and provide an independent view. For example, they can avoid another of the problems of human decision making – conservatism – whereby new, but valid, information is discarded in order to retain a bias towards a view already held. Such a behavioural approach should therefore act as a warning system to complement other more fundamental analysis in our tool kit”.