Wise men talk because they have something to say; fools, because they have to say something. Plato

Mark Seymour of Alphen Asset Management writes that he can recommend his current reading, a book called "The Effective Investor" by Franco Busetti to anyone who is looking for a reference guide to the markets. He says that the author is a person with a recognised ability to grasp and articulate matters of investment interest in a manner which far exceeds that of the average man or fool.

I've been dipping into chapters that pique my interest, for example chapter 20 "Lessons from the first twenty five years" a contribution by Dave Foord, chapter 34 "Really Bad Research" where Rowan Williams-Short's razor sharp tongue delivers on song and a handful of other insightful chapters from which I've borrowed plenty for previous Alphen Angle ideas.

In this book, Mr Busetti refers to the eleven great investment pictures which are a series of graphs which explain certain powerful investment truths. I have omitted the graphs but where necessary I have provided a summary of the graph's message.

1) Price follows earnings

It is generally accepted that the contribution of stock returns come from: 1) a change in rating (PE ratio), 2) dividends and 3) earnings growth. Over the long term (1960 to 2008) however, a change in rating has made a negative contribution of -0.5% to the total returns of the All Share Index, dividends reinvested has made a contribution of 4.5% and earnings 12.5% which makes up the bulk of the total annual return of 16.5%.

"Without this long-term upward rating bias underpinning the market in future, returns will fluctuate more closely with the earnings cycle. Returns will therefore be both lower and more cyclical. (In addition, in the medium term leverage, gearing and cheap financing will be less than in the recent past, which will dampen returns further)".

There is a strong inverse relationship between inflation and PE ratio which suggests that if and when inflation rises the PE ratio can be expected to fall which will result in muted future returns.

There is a strong correlation between higher yields and subsequent total returns as a result of a "rerating" as higher yields revert lower. The strength of this relationship is time dependant with bonds showing a strong correlation using a 3 year time period and equities using a 5 year period.

"Always incorporate the risk of loss in your return calculations. Your expected return is the weighted probability of gains and losses".

The expected return formula incorporates inputs for both an anticipated gain over a particular period as well as a potential loss. It is calculated by summing the probability of a gain multiplied by the potential gain and the probability of a loss by the potential loss.

The trick here is that in most instances we realise gains and we tend to ignore or underplay the potential losses, however, when the loss occurs it is normally substantial and therefore has a significant negative influence on expected future returns.

"One cannot overemphasise the importance of beginning an investment plan as early as possible".

The power of compounding is illustrated by way of example where two investors' returns are compared: The first investor starts saving R1000/year in 1962 escalated at the rate of inflation between the ages of 20 and 31 and the second investor starts saving at the age of 31 (the same amount the first investor was saving - also escalated at inflation) up until the age of 65. If both savings were invested in the All Share Index, the first investor ends up with a sum of R13.9m vs. R12.8m for the second investor. This highlights the power of compounding, particularly at the early stages of the investment.

"Be skeptical of macro forecasts. They seldom forecast turning points, economists are particularly prone to herding and the economic consensus is frequently wrong".

 "The scary thing is, although we cannot forecast more than three years out, at best, in order to justify the high prices of some shares, their forecasts need to show above-average earnings growth for 10 years or more. First, this is not likely and second, even if it were, we would not be able to forecast it".

"Consensus forecasts tend to anchor, use blind extrapolation, ignore fundamentals and mean reversion and are frequently irrational".

There is a belief that we need to forecast in order to work out a valuation, however, forecasts for growth rates are often overstated, forecasts lag reality (i.e. trail as a result of extrapolating the recent past) and forecasts for cyclical data (e.g. earnings) are often understated whereas in reality the cyclical swings are often more severe. A way around this dilemma is to base valuations purely on historical data or to make use of large error bands around the forecast and be cognisant of mean reversion.

7) The bias and skewness of forecasts

"When actual earnings growth is low, analysts tend to overestimate it, and when actual earnings growth is high, they tend to underestimate it".

This investment picture is borne out of the forecasting idiosyncrasies of analysts. When it comes to earnings forecasts it can be shown that analysts show a consistent bias of overestimating earnings and secondly when growth numbers are low, forecasts are skewed to the upside and when growth numbers are large, forecasts are skewed to the downside. These idiosyncrasies more than likely lead to mispricing which provided investment opportunities for the astute investor.

"In any particular year, equities had a chance of 27% or over one in four of producing a negative return. Over a two-year period this probability more than halved to 13%. The probability of a negative return in any three-year period was only 6% and dwindled to zero for holding periods of four years or longer".

As a result of equity price fluctuation, potential losses are high over the short-term, however, as you increase your investment horizon the probability of losses decreases. The overriding attraction of equities at this point is then purely determined by its ability to grow vs. other non growth income assets.

9) Irrational volatility

"Movements in price are much larger than justified by actual subsequent events and too large to be in accord with efficient markets. The market overreacts spectacularly to both actual and anticipated events".

It can be shown that the actual price (e.g. S&P 500) deviates markedly around the rational price (calculated as the discounted value of the actual future dividends) over long periods of time. The implication is that dividends alone do not determine prices, reflecting the irrational nature of markets.

"Even if we know the precise path of future dividend streams, our valuations will be wrong most of the time, missing the large swings and roundabouts that generate the returns we would like to capture".

"Trying to guess how a periodically irrational mob will vote is a fool's game since sustainable success is impossible. However, as we have shown, there are weighting tools that work over longer periods and consequently make long-term investment both more successful and less risky than short-term speculation".

The effect Investor behaviour has on the market presents the greatest challenge to successful investment returns and therefore being able to incorporate or account for its presence is of vital importance in the investment process.

10) The forecast saddle

"So even in the most successful situations, the hit rate for analysts' growth forecasts is barely better than flipping a coin".

"Since we cannot forecast earnings growth with any confidence, we should be value investors rather than growth investors".

As was touched on in point 7, forecasting earnings growth is very difficult and there is an argument that when split into high and low growth categories, analysts' estimates for future earnings growth are overstated for low growth companies and their estimates are understated for high growth companies. The graphic of this data has a "saddle" shape of which the saddle section represents earnings growth estimates which are more-or-less correct and make up less than 50% of the estimates. The conclusion is that forecasting earnings with any degree of confidence cannot be done.

11) Luck and Skill

"The time required to distinguish between luck and skill in a manager is measured in years if not decades so most funds and fund managers do not have sufficiently long histories".

The degree of certainty that out-performance is attributable to skill is directly related to the period of analysis. For instance an equity manager which has displayed a standard deviation of returns of 16% (within the range 13% to 19%) with a median active return (i.e. annualised out-performance) of 1.1% per annum vs. the All Share Index would require 16 years of data to be 80% sure skill is present.

For further detail on these "Great Investment Pictures" seek out a copy of "The Effective Investor". Although this book is not recommended for recreational bedside reading it is a worthy reference guide for all those seeking clarity on any investment related interest.

The Alphen Angle is an electronic newsletter of PSG Alphen Asset Management. To read more about PSG Alphen please go to AlphenAM.co.za.