The long and short of it

tauringhuang.

Aug 3rd 2008
From Economist.com

Getting it right, in two years and ten


TEN years ago, GMO, an American fund-management group, was losing clients. GMO was sceptical about the dotcom boom and thought that equities offered poor value. Its performance lagged that of other groups who appeared to be more in tune with the “new paradigm” of the late 1990s.

At the time, GMO forecasted likely future levels of returns from ten separate asset classes. Emerging-market equities came first, followed by American real-estate investment trusts, emerging-country debt, international smallcaps, index-linked Treasury bonds, American government bonds, EAFE (Europe, Australia and far east) equities, Treasury bills, overseas bonds and the S&P 500 index (American equities).

Remarkably, ten years later that prediction was proved almost entirely correct, from 13.4% real for emerging market equities all the way down to 0% real for American equities. The only differences were that foreign bonds and Treasury bills swapped places, as did government bonds and EAFE stocks.

The chances of getting that forecast exactly right were less than one in 500,000. Does this suggest that GMO’s fund managers are exceptionally lucky? Perhaps. But it also suggests that long-term market movements may be rather easier to predict than short-term ones. In the short-term, markets can be pushed to extremes, and thus away from fundamental values, by fear and greed. But bubbles don’t tend to last for a decade.

In other words, if you buy shares on a price-earnings ratio of 30-40, there is always the chance that they will shift to a rating of 50 or so over the following 12 months, allowing you to make money. But in ten years’ time, the shares will likely trade on around the market average of 15. You need a lot of earnings growth to make up for that potential halving in valuation. The odds are not on your side.

This is roughly how GMO goes about the process: it looks at the relationship between valuations and long-term returns. The return from equities, for example, is equal to the existing dividend yield, plus future dividend growth, plus or minus changes in valuations. Ten years ago, the dividend yield on the American market was low while valuations were high. The likely long-term return looked low, and so it has proved.

Using similar reasoning, GMO has a very gloomy outlook for the American and British housing markets at the moment. By using the ratio of the median house price to the median family income, GMO reckons that prices in America need to fall by 17% instantly or stay flat for four years to return value. In Britain, prices need to fall by 38% or stay flat for seven years. And of course, there is no guarantee they will stay at fair value; in the mid-1990s, they dropped well below it.

More generally, Jeremy Grantham, GMO’s chairman, thinks that the equity bear market will continue for another couple of years, with the S&P 500 dropping by around 10-15% from here. But he warns that the chance of a meltdown—a drop well below fair value—has increased.

But there is a big difference between now and 1998; equity valuations are a lot lower. GMO might prove to be right about the outlook for equities, but, as it has shown before, it can be wrong over the next year or two. Its forecast about the Anglo-American housing market, based on long-term valuation measures, looks rather more soundly based.