Fund managers differ, just as individual securities do. Style analysis is one of several methods for classifying managers into broad groups, according to their performance.

One way of looking at style is as a way of grouping fund managers by the characteristics of the portfolios they select. Experience shows that most portfolios of securities behave in more predictable ways than the individual securities themselves. Therefore it is possible to make useful statements about the factors likely to influence a manager's future returns, without analysing each security the manager holds.

The following two-stage approach was developed by Nobel laureate William Sharpe.

The first stage is to subdivide a market into groups of securities with similar characteristics. There are many ways to do this:

q Market capitalisation - It can be argued that stocks of similar market capitalisation tend to be affected by economic conditions.

q Value and growth - This approach is normally based on the concept of subdividing a market equally. All the securities within the market will be assigned to either value or growth depending on some financial measure, eg, price-to-book ratio. Value stocks are usually defined to have low price-to-book ratios, while growth stocks have high ones. Value stocks are generally companies in mature industries, while growth stocks tend to be in emerging and dynamic industries, so can be expected to react differently to economic factors.

Although many further subdivisions could be made, it seems appropriate to restrict the number to as few pertinent style factors as possible. For example, in theory we could subdivide the market into growth large companies, growth small companies, value large companies and value small companies. However, small companies currently account for only 6% of the market, so more usually the market is subdivided into value large companies, growth large companies and smaller companies.

Once the market is subdivided, indices are created to reflect the various styles. The value index, for instance, would consist of alllarge companies' stocks classified as value stocks; the growth index would inclu de all the growth stocks from large companies; and all the rest would go in the small caps index.

A manager's style is then obtained by considering which combination of the various style indices most closely replicates the manager's performance. This combination is obtained using multiple regression analysis. Suppose, for example, a manager has a style classification of 60% value, 30% growth and 10% smaller companies. This classification emerges because the regression analysis shows that this particular combination of the various style indices most closely matches the manager's performance.

In the figure we show style classifications for three hypothetical managers. Manager A would be deemed a value manager, as its style classification is overweight in that factor relative to the market weighting of approximately 45%. Similarly, manager B would be deemed a growth manager (being over the market weighting of growth" of approximately 45%) and manager C has a smaller companies bias. It should be noted that cash has been included in the style classification, as most managers hold some cash.

Use of style analysis is likely to grow in the UK, largely because of its value in the following areas.

First, it can be a useful tool in assessing the reasons behind the performance figures published by UK managers. The performance of the various style indices can be significantly different, depending on market conditions. For example, in 1997 large companies markedly outperformed small companies. Even within the large-company definition, there were notable variations in performance.

By the same token, style is one of the risk factors that should be identified when choosing a fund manager, because it can be used to identify the market environments in which a given manager is likely to do well or badly. Often two or more fund managers are required, to diversify the investment manager risk by spreading it across several investment houses. Style analysis will help to ensure that this is done efficiently, by identifying complementary managers with different styles. Lastly, style analysis can help in assessing whether a fund manager is actually implementing the investment process it claims to use, if the fund manager has an investment process which should lead to a style bias.

Brian St John-Hall is actuarial director at Sedgwick Noble Lowndes Investment Strategy in Croydon"

Topics