Home / 91

Long term equity returns and mean reversion

Baron Rothschild is credited with saying that "the time to buy is when there's blood in the streets." Implicit in this statement is that markets are mean reverting and periods of extreme negative returns are not likely to be sustained. Put simply, when everyone is selling it often represents an opportunity to achieve superior future returns.

Of course being a contrarian is not easy and goes against the natural bias of human behaviour. When returns are negative there is a natural bias to shy away from investing. But this is precisely the time to invest to achieve above average nominal returns. Conversely when markets are running hot everyone wants to be greedy.

In early 2012 Alan Kohler on the ABC News commented that the current recovery was the worst on record, including the crash of 1929, the inflation-induced recession of 1973 and the bubble of 1987, and used Chart 1 to demonstrate (which has been updated to October 2014). But markets do not stay pessimistic indefinitely (with some exceptions such as the Japanese market). Nor do markets stay “irrationally exuberant”, to quote Greenspan.

Chart 1: Market crashes and recoveriesRS Chart1 051214

RS Chart1 051214

Source: Historical records based on monthly time series of the combined price history of the Sydney Stock Exchange and the Melbourne Stock Exchange from 1900 to 1970 and the ASX All Ordinaries Price Index from Dec 1970 to Oct 2014.

Chart 2: Australian equity market and length of bear marketsRS Chart2 051214Bear market is measured as a 20% decline from the market high and lasts until previous high is surpassed.

Since 1900 the Australian equity market has returned 5.76% pa (geometric average, excluding dividends and before inflation, to end of October 2014). In the short term markets are random, but over the long term markets mean revert to the long term average.

Table 1: Historical Returns 1900 to 2014RS Table1 051214The mean reversion of the Australian market can be illustrated by charting rolling five year returns (Chart 3). For simplicity standard deviation bands have been included around the long term mean of 5.76% pa. Some 68% of all returns will reside within a standard deviation of +1 and -1, and 95% of returns reside within a standard deviation of +2 and -2.

Chart 3: Rolling 5 year Australian market price index returnsRS Chart3 051214Importantly, returns significantly above or below the long term average are simply unsustainable.

The peak of 1987 was more than 4 standard deviations away from the long term mean. Statistically the chance of this occurring is 0.02%. The returns into 2007 were two standard scores above the mean. The crash of 1929, World War Two and the 1992 recession saw five year returns more than two standard deviations away from the mean and the inflation led recession of 1973 saw returns more than 3 standard deviations from the mean. Even May of 2012 saw five year returns almost two standard deviations away from the mean.

Table 2: Historical 5 Year Standard Deviations greater than +2RS Table2 051214Table 3: Historical 5 Year Standard Deviations less than -2RS Table3 051214Of course, not all companies recover from tough times. This analysis focusses on the broad market, but not all individual companies revert to a mean. Some disappear forever!

Tables 2 and 3 demonstrate that when historical returns reach positive/negative extremes, future returns have a high probability of being significantly less than/greater than the long term average. For example, in Table 2, where the market is more than two standard deviations above its long term mean, such that historical returns have been good, the future returns from that point have been poor. Conversely, in Table 3, where the market is below its long term mean, the future returns from that point have been good.

If history is a guide one would expect future returns from the low of May 2012 to be above the long term average. This does not mean the markets won’t be volatile as they never move in straight lines – and markets can remain subdued for periods much longer than expected (for example the Japanese stock market). But it can give one a glimpse of where the market may be headed.

 

Robert Stewart is a Director of Sandgreen Pty Ltd, former Head of Challenger Howard Mortgage Fund and Head of Index Funds at Colonial First State Global Asset Management. This article is for general information and is not personal financial advice. Readers should seek their own professional advice.

2 Comments

Rob Stewart

January 28, 2015

Brian

Thanks for your comment. I make the following response:

* I deliberately selected the longest time period available where monthly data was accurate and sourced from SIRCA – which was limited to price index data. The All Ordinaries Accumulation series did not start until 1979 and I am yet to find a reliable data series that includes a total return prior to this date

* I have relied upon the inflation data from the Bureau of Statistics, which starts in 1948. I am yet to discover a longer series of monthly CPI data, although annual data appears to be available if one searches hard enough.

* Finally dividend imputation was only introduced in 1987 and I am yet to discover the franking component on monthly data for the All Ordinaries


Should you have accurate monthly data for total return, CPI and franking credits for longer periods, I would gladly accept this and update my own data.

FYI, if you look at a chart of the five year returns (real) for the All Ordinaries Accumulation Index from 1984, it does not diminish the crux of the article: that the Australian market is mean reverting around a longer term mean. Markets do not stay irrationally exuberant (for example 1987, and 2005 to 2007), and nor do they stay irrationally pessimistic (for example 1992 and 2012).

Brian Salt

December 05, 2014

If markets are at all rational they will consider dividends, taxation (ie franking credits) and inflation. The above analysis excludes these so from an investors point of view how useful is it really?


 

Leave a Comment:

     

Most viewed in recent weeks

Most investors are wrong on dividend yield as income

The current yield on a share or trust is simply the latest dividend divided by the current share price, an abstract number at a point in time. What really matters is the income delivered in the long run.

My 10 biggest investment management lessons

A Chris Cuffe classic article that never ages. Every experienced investor develops a set of beliefs about how markets operate.

Seven major trends affecting Australians in retirement

Retirement planning will become increasingly complex in the face of trends in home ownership, wealth dispersion, life expectancy, health and aged care costs, work patterns and pension dependency.

Lessons from the Future Fund for retail investors

The Annual Report from Australia's sovereign wealth fund reveals new ways it is investing in fixed income and alternatives. The Fund considers its portfolio as one overall risk position with downside protection in one asset class allowing more risk in another.

Four companies riding the healthcare boom

There are strong demographic trends in ageing and consumer spending and investing in the right healthcare companies can ride this wave as well as produce better health outcomes for people. 

Five reasons SMSFs are making asset allocation changes

Substantial changes are underway in SMSFs which until recently held a narrow range of assets dominated by cash, term deposits and Australian equities. Trustees have never faced so many choices.

OpenInvest

Sponsors

Alliances

Special eBooks

Specially-selected collections of the best articles 

Read more

Earn CPD Hours

Accredited CPD hours reading Firstlinks

Read more

Pandora Archive

Firstlinks articles are collected in Pandora, Australia's national archive.

Read more