Equity gains downgraded

History tells us it pays to beware of revisionists. It has always seemed rather churlish that niggardly historians went back to find out exactly how many brokers and traders jumped to their deaths off New York skyscrapers in the Crash of 1929. Police records prove the figure might have been zero.

History tells us it pays to beware of revisionists. It has always seemed rather churlish that niggardly historians went back to find out exactly how many brokers and traders jumped to their deaths off New York skyscrapers in the Crash of 1929. Police records prove the figure might have been zero.

But the myth continues, just as most people still believe the Great Crash was the seminal event that triggered the Depression of the 1930s. In fact, the market recovered half the fall of October 1929 before resuming its decline. There were other factors behind the Depression.

Now two of the UK's leading finance academics have turned their attention to what may be another great myth, the extraordinary performance of equities over the 20th century. Professors Paul Marsh and Elroy Dimson of London Business School say in their Millennium Book (produced with ABN Amro) that the long-run performance of shares has been significantly overstated.

Received wisdom is that shares have delivered an average real (post-inflation) return of nearly 8 per cent per annum, but the actual figure is no more than 5.9 per cent, the professors say.

That received wisdom consists of the well-known BZW Equity-Gilt study first published in the 1950s, and after the demise of BZW, appearing in two rival versions from the houses that inherited BZW's business, CSFB and Barclays Capital. This annual publication is used by fund managers and the media to trumpet the virtues of equity investment.

The problem the academics found is that the data on equity returns has been significantly overstated. This is particularly true of the first half of the century. The Equity-Gilt study suggests shares returned up to 8.8 per cent in real terms between 1919 and 1954, but the professors say the correct figure for 1900 to 1954 was 3.83 per cent. This overstatement feeds through to all longer-run calculations of equity returns.

The original compilers of the data used in the Equity-Gilt study fell victim to common statistical errors. The most serious was working backwards from the companies which made up their index of leading shares when they compiled the data. This introduced a statistical "survivor bias", since it measures the performance of only shares that did well, ignoring those which went bust or underperformed. The compilers of the early BZW data also picked an unusually favourable starting point - 1919 - for their calculations.

Does it matter that equity returns, especially in the first half of the century, have not been as high as people now assume them to be? In one sense, it is a trivial and pedantic exercise. Nobody disputes, as the graphic shows, that equities have continued to offer much higher returns than their main comparators, bonds and cash (cash being represented by Treasury bills, a proxy for the nearest thing to a risk-free short-term investment you can find). The revisionist exercise does not invalidate the case for equities as long-term investments.

But the long-term record does put that case more clearly into context and becomes highly relevant when trying to project what may happen next, especially for retail investors, many of whom have to adjust their returns further for the impact of tax and transaction costs. The past 25 years have been an exceptionally good period for equities, and for bonds. The further back you check the track record, the less impressive the returns on all three asset classes become, which reminds us that the last century was violent, with two World Wars, the Great Depression and the hyper-inflation of the 1970s.

But is it realistic to project the exceptionally benign climate of the past 20 years, and what price should be placed on equities and bonds as a result? The implication of the academic study is that the price investors should pay for the higher risk/return offered by shares relative to other investment classes (the equity risk premium) is lower than most observers traditionally assumed.

There are good reasons for believing reduction is needed in the additional reward you need to earn to justify the added risk of buying shares. The world is at relative peace; central banks appear to know what they are doing (not the case in the 1930s); and the rapid increase in international trade and capital flows means investors have greater opportunity to diversify.

All that is good, but it is also already reflected in today's high stock market valuations. If you feel the rerating of shares has gone as far as it can be reasonably taken, then shares cannot outperform as markedly as they have done in the past 25 years. Risk in equities has not been eliminated. Business is inherently risky, and it will remain so, vide today's anxieties about sharply rising oil prices.

My second graph shows the results of a survey by Asset Risk Consultants, a specialist performance measurement company, which calculates the probability of being able to pick a managed fund that outperforms in different sectors. The scope for active management to outperform a comparable index fund remains much greater in overseas markets than it is in the UK.

Overall, your chances of picking a fund that will outperform consistently on a risk adjusted basis is around 35 per cent, meaning the odds are 2-1 against. For most UK funds, the figure is much lower than that.

The hardest place to get active outperformance is in the fixed-interest (bond) sector. Irony arises, because the bond sector is the one place where, as of today, you still cannot buy an index fund, although they are two a penny in the US. One or two bond index funds are in the offing. At present, most investors putting their money into bond funds are paying hefty fees for what is a very poor "odds against" proposition.

davisbiz@aol.com