Consistency in an uncertain world

Listen to this article

00:00
00:00

Last summer, a well-known UK investment company claimed that the four most dangerous words in the English language are “reversion to the mean”. Some business school professors have written in support of this claim.

Yet even in today’s uncertain world, investors should not lightly relinquish their faith in mean reversion.

“There is a seductive argument that says equity risk falls the longer the investment horizon,” write Messrs Dimson, Marsh and Staunton (DMS) of the London Business School in the latest issue of the Credit Suisse Global Investment Returns Yearbook.

“This belief that time conquers risk,” they continue, “is based on the view that equity returns are mean-reverting.”

Market valuation is not a good indicator of future investment returns, write Mr Dimson and his co-authors, who have collected data on the total returns of 20 stock markets from 1900 to the present day. Their valuation model – based on the price of 10-year trailing real dividends – shows only a weak correlation across markets between valuations and subsequent investment returns. This conclusion should satisfy other academics brought up to believe that markets are always efficiently priced.

But it is disturbing news to market practitioners who believe valuation is the best indicator of expected returns. They need not lose too much sleep. Andrew Smithers of Smithers & Co and author of Wall Street Revalued takes issue with DMS, claiming their dividend-based valuation model is inherently flawed. There is no economic reason, says Mr Smithers, why dividend yields should be mean-reverting. The payout ratio (ie share of profits distributed as dividends) varies over time. Companies have other ways of generating returns for investors – they can reinvest profits, acquire other companies or repurchase shares.

If stock markets followed a random walk, as the efficient-market hypothesis suggests, then the annualised volatility of market returns should be constant over all time horizons. In other words, it would be just as risky to invest in stocks with a 10-year holding period as over one year.

The historical evidence, based on stock market returns collected by Mr Dimson and his colleagues, suggests strongly that this is not the case.

As Mr Smithers points out, “returns become less volatile if investors hold shares for long periods; but the rate at which volatility declines is distinctly more rapid than it would be if the random-walk model were correct”.

There are few exceptions to this mean-reversion rule. Stock markets in countries whose capital stock was destroyed in war, such as Germany and Japan, experienced a one-off but permanent loss. Investors in Russia and China, who found themselves on the wrong side of Communist revolutions in the last century, fared even worse.

What is remarkable about the DMS data is how stable investment returns have been across a wide variety of countries over different time periods. Measured in constant US dollars, the average annual return of markets ranging from Australia to the US, between 1900 and 2010, was 4.8 per cent with a standard deviation of 1.3 per cent. Since 1950, the average return of these markets has been even higher at 7.3 per cent, with the same standard deviation.

The fact that we observe mean reversion in the data is not merely some statistical artefact. Rather, it is grounded in economic theory. After all, the long-run returns of the stock market are determined by the underlying profitability of listed companies.

In a capitalist economy, companies that earn excess profits attract competition and see their profit margins erode over time. Likewise, if companies do not earn their cost of capital, then competitive pressures diminish and future profitability improves. The mean reversion of stock market returns and returns on equity are one and the same.

The performance of equities in less capitalistic economies seems to support this point.

China is the most striking anomaly. Despite economic growth in double digits over the past two decades, Chinese equities have delivered an annual real return of -2.5 per cent in local currency since the early 1990s, according to the authors of the Credit Suisse Global Investment Returns Yearbook. The Chinese stock market is dominated by state-owned enterprises whose capital allocation is ultimately decided by government. Some studies suggest these businesses earn negative returns on capital once subsidies are taken into account.

The lesson of history is clear. Unless the factories have been bombed or the country is being run by crooks or, even worse, by Communists, mean reversion is one of the few constants in an ever-changing investment world.

Edward Chancellor is a member of the asset allocation team at investment manager GMO

Copyright The Financial Times Limited 2017. All rights reserved. You may share using our article tools. Please don't copy articles from FT.com and redistribute by email or post to the web.