Secrets of Success: The wonderland of information ratios

An interesting piece of research this week from the Edinburgh investment consultant The WM Company takes us deeper into the issue of what information can usefully be gleaned from the analysis of fund performance statistics. This, it turns out the more you look, is an Alice in Wonderland world where nothing is quite what it appears, and just as you think you may have found something of value, its meaning starts to evaporate.

An interesting piece of research this week from the Edinburgh investment consultant The WM Company takes us deeper into the issue of what information can usefully be gleaned from the analysis of fund performance statistics. This, it turns out the more you look, is an Alice in Wonderland world where nothing is quite what it appears, and just as you think you may have found something of value, its meaning starts to evaporate.

The WM study takes as its subject "the information ratio", a now-common statistical tool for assessing past performance of fund managers. It measures the excess return a fund has achieved per unit of risk adopted. Risk in this context is defined as the extent to which a fund's performance deviates from its benchmark index. The excess return is similarly the extent to which a fund's performance has exceeded (or fallen short of) the same benchmark index.

Once you accept the logic of using a market index as a benchmark, the idea that you are taking "active risk" by doing anything different from the benchmark has some obvious intuitive value. The information ratio is just the next step in the line of thought. If you want a fund manager to beat the market for you, it makes sense to go for one who can demonstrate the ability to make returns that justify the "risk" involved.

In practice, the information ratio is widely used as a measure for comparing and ranking fund managers across their sectors. Funds that can boast an information ratio of 0.5 or more are the ones that will typically find themselves winning the business in professional pitches. Indeed, a positive information ratio (a reading of more than 0) is often interpreted as the truest single measure of fund manager skill.

In aggregate, the universe of professionally managed funds has a negative information ratio for the same reason that most actively managed funds underperform the market over any reasonable period of time. For every fund that beats the benchmark, there has to be another that fails to do so. Collectively, the industry will underperform the market by the degree of the costs it incurs. A positive information ratio is, therefore, the exception rather than the rule, and much valued in marketing terms as a result.

The trouble, as the WM study shows, is that the information ratio, while an interesting tool of analysis, is not really of much value for the purposes to which it is usually put, that is to say in choosing whom to hire to look after your money. The study looks at the behaviour of information ratios across a wide range of pension-fund portfolios managed by fund management companies, and concludes broadly that there is no real evidence that strong performance on this measure persists into the future.

In fact, comparing what happened to funds' information ratios over the two periods 1997-1999 and 2000-2002, the trend if anything appears to be the other way round. Funds with the 25 per cent of highest information ratios in one period were less likely to rank in the same quartile in the next period than you would expect even if they were randomly distributed. The funds with the worst information ratios were more likely to be in the top 25 per cent in the next period.

In general, as the table shows, very few funds consistently achieve an information ratio of more than 0.5 over time. In the three periods studied by WM, only two out of 96 pension funds analysed by WM managed an information ratio of more than 0.5 in all three periods, and only a quarter of them managed it in two out of three periods. For funds investing in other markets, the figures were even lower. (This study highlights something else that comes through a number of performance studies, namely that UK fund managers appear to be particularly poor at producing above average returns investing in the United States).

Another research paper about information ratios that I have seen recently, from Vanguard, makes a similar point. Its analysis suggests that information ratios have no more predictive value for US fund managers than they do in Britain. With the usual notable minority, the data suggests that the numbers only tell you who did well in the past. They cannot tell you who will do well in the future.

Why is that? One reason is our old friend "style" factors. The WM study suggests that fund manager style is much more consistent than performance. In other words, a fund that follows a "value" approach to investing tends to go on doing so regardless of market conditions. The same goes for a fund that has a bias towards say large-cap rather than small-cap stocks. In one sense this is a good thing: it means that those who are engaging a fund manager's services know what they are getting: in many cases the styles are what the pension fund or adviser in question has actually asked the fund manager to follow.

All these findings are context dependent. The periods that the WM study looks at include some dramatic turning points - the switch from bull to bear market between 1997-1999 and 2000-2002 and the simultaneous switch in style from growth to value and large-cap to small-cap stocks. These trends, as I was saying last week, have dominated the performance tables for all investors.

But in another sense the data is less comforting. The implication is that the skill of the fund manager is a less important contributor to the performance of the fund than its style. As style factors vary over time, calling the style trends is an important potential source of outperformance - yet most fund managers are either unwilling, unable or simply prevented from taking on that task for you. It begs the question: what actually are you paying for? As for information ratios, while they are useful, they don't really seem to tell you as much as at first appears.

jd@intelligent-investor.co.uk