Secrets Of Success: History doesn't always repeat itself

So it is too, it seems to me, with investment. For the long haul, those who have seen it all before have a built-in advantage over those younger guns who are willing to take more risks and often reap the greater short-term rewards.

One market observer who has been following the business for more than half a century is the New York commentator Peter Bernstein. In his latest circular to clients, he offers a long-term perspective on the current valuations of shares and bonds.

He points out that the way investors regard these two asset classes has been undergoing a profound and important change over the past few years - though not all the implications, he thinks, have yet been fully thought through.

His starting point is a simple graph that plots the relationship between two simple metrics over the entire course of modern market history (1871 to the present) - the US bond yield andthe earnings yield of the US stock market. The latter is simply the inverse of the market's price-earnings ratio and measures how much quoted companies in aggregate are earning in profits as a proportion of their market capitalisation.

Anyone who has been exposed to broking literature in recent years will know that one of the key arguments that has been used to justify buying equities recently has been the fact that the earnings yield on the market is currently higher than the bond yield. That, so the argument goes, is an important reason why you should be buying shares rather than bonds at the current time.

For the greater part of market history, however, as Bernstein's simple graph shows, the earnings yield has always been higher than the bond yield. It was not until 1981 that the earnings yield on the stock market fell below that of the bond yield for the first time. For more than 100 years before that, in other words, investors expected the earnings yield on shares to exceed bond yields as a matter or course.

This in turn reflected the idea that shares were riskier than bonds, which in a low inflation world is what theory tells us they should be. It was only the high and unexpected inflation of the 1970s onwards that persuaded investors to abandon this long-held tenet of investment life.

Until central bankers set out on their long and eventually successful assault on inflation in the early 1980s, bond returns were shredded by inflation, but those days are now long gone.

If you believe that inflation has now been largely tamed, it follows that we should not be surprised to see earnings yields above bond yields. Might that not be the logical state of affairs once more? In Bernstein's view, the answer is yes.

His argument for some years has been that the bond market's behaviour in the high inflation years of 1971 to 1991 (which first saw yields rise dramatically a peak in 1981 and then fall all the way back down again) is a huge historical anomaly that should be discounted when discussing the relationship between shares and bond valuations now.

As he puts it: "The relationships between bond and stock valuations, or returns, over the past 30 years are irrelevant, obsolete and dangerously misleading."

To see why, you also need to look at the correlation between earnings and bond yields. For most of the period from 1970 to the late 1970s, the two numbers tended to move in the same direction.

In other words, when bond yields were rising, so too were earnings yields (meaning investors required lower p/e ratios to invest in shares), and vice versa when they were falling. Once bond yields started falling in 1981, it marked the start of the great equity bull market of the next 20 years, with earnings gains compounded by expanding p/e ratios.

Yet going back to the first 100 years of market history, this cosy relationship was also the other way round. The two numbers tended to move in opposite, not similar, directions. Conventional wisdom in the 1950s, Bernstein reminds us, was that stock and bond prices should move in opposite directions - "expanding business was good for stocks and bad for bonds, while recessions were good for bonds but bad for stocks".

Until this relationship broke down in the inflationary years, it meant that bonds were widely seen as a safe haven alternative to riskier equities - as indeed they were. Not only did bonds produce positive annual returns more often than shares, but the returns were also more predictable and consistent.

Over most of market history, the current yield on bonds has been a very reliable forecaster of their future returns, while equity returns have been much more volatile and difficult to predict.

Put all this long-run perspective in the mixer and what do we have now? That is the interesting question. Now that it has returned to its old pattern, will the earnings yield stay above the bond yield? Unless you believe that inflation is coming back in a big way, which seems unlikely, there seems good reason to think that it might. But if so, what should the spread be? From 1870 to 1980, the difference averaged about 3.5 per cent, against around 1.0 per cent now. Given how much more liquid and diversified world stock markets are today, you could argue that this disparity is not as far out as the historical comparison might suggest.

If bonds and equities are going to return to their historical relationship (that is, moving in opposite directions), it also implies that bonds will again become more effective and reliable portfolio diversifiers. First, however, bond yields may have to rise to make them more attractive on absolute grounds. Today's ultra-low long term bond yields suggest that while investors have re-embraced the idea that bonds are safe investments, they may have been taking the idea a little too far.

As for shares, this week's strong advance shows that they are still much in demand, but don't forget the other historical fact, that the end of bull phases in the market are always the most exciting.

jd@intelligent-investor.co.uk