[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries

Investment of the third kind

In his recent letter (April 2004), Dr Scott suggests that the theory behind the highly successful results set out in my March 2004 article consists of little more than very unoriginal advice to buy high dividend yield and low P/E shares. This is far from being the case. If asked to paraphrase my philosophy in as few words as possible, I would describe it as ‘predictable earnings growth at an attractive price’. I first of all obtain consensus estimates of G, the growth rate of earnings per share to 12 months ahead, and R, the prospective price-earnings ratio, and then rank shares in order of attractiveness by means of the utility function: U = G - cR, where c is a positive constant. The value if c is chosen as that which maximises the predictive power of the model, and this value turns out to be significantly higher than the value obtained by a ‘least squares’ best fit on the seemingly innocuous – but quite erroneous – financial economics assumption that the equity market is in equilibrium. This means that there is indeed a bias in favour of low P/E shares, but this bias is introduced only after earnings growth rates have been taken into account as the commonsense fundamental driver of equity share returns over the long term. Also, dividend yield plays no part whatsoever in my utility ranking approach: if this were not the case, it would in particular be irrational for me to continue to hold my shares in Warren Buffett’s company, Berkshire Hathaway, which does not pay dividends.

This mention of Warren Buffett leads me to reflect that another way to describe my approach is as a shorter-term, and hence less powerful, version of his brilliantly successful investment philosophy. He uses his business acumen to assess cashflow and earnings growth to an investment horizon of five years or longer but only invests at ‘bargain’ prices. I use publicly available consensus estimates of earnings growth to an investment horizon of one year and then introduce a bias against highly rated shares.