Economists tell us that accounting is a mere subsection of economics and forms only one of the tools they need to make decisions about the market. So is there any reason to believe that accounting information is any more relevant than all other forms of information available to the market?
Economists also argue that the stock market is a near-perfect market. All known information is built into the price of any share at any one time and any share will go on a "random walk"--its price could go up or down and the direction of this movement is unpredictable. When you buy a share, the risk is that its price will fall.
There are two categories of risk. The first relates to the market as a whole and is known as portfolio, market or systematic risk. The second relates to individual shares and is known as diversifiable, unique or unsystematic risk. Harry Markowitz (1) suggested that the latter could be diversified away, meaning that by buying more than one share you must reduce the risk of outright failure. Given that he won a Nobel prize in economies for his work on modern portfolio theory in 1990, who are we to argue? Meir Statman (2) argues that a portfolio of about 30 stocks will achieve optimal diversification, but, of course, a smaller number still adheres to the principle.
This leaves portfolio, market or systematic risk that we cannot control, so we need a risk premium over and above the risk-free rate obtainable from government securities and building society accounts to persuade us to invest in shares. Financial institutions tend to issue unit trusts that 'adhere to these principles. For example, a unit trust might invest in every FTSE-100 company, varying its investment by size of company, so that the return achieved always reflects the index.
The key question is: do you accept the received wisdom? Half of me thinks that most people do, so it must be right, while the other half argues that it must be illogical. Take portfolio theory, for example. Diversification mast reduce the risk of diversifiable, unique or unsystematic risk, but it must follow that you average down the return by reducing the risk. If you back every horse in a race, then you must back the winner, but if you are correctly confident that half the field cannot win and back the other half, you will increase your return. So portfolio theory holds good only if shares do indeed go on a random walk and are totally unpredictable. But are they? Surely an analysis of published financial accounts can find flaws that the market fails to spot?
Experience tells us that the market tends to take note of what companies' directors have to say, but misses vital bits of information hidden in the accounts. For example, shares in SFI plc traded at over 2 [pounds sterling] for weeks after its accounts showed stock days as 20 days. Anyone analysing the accounts of restaurants and bars will know that stock days are usually around five days, so 20 days was not believable. It took several more weeks before a black hole in the accounts was announced and the shares became worthless. Cedar Group pie, when its shares were trading at between 5 [pounds sterling] and 7 [pounds sterling], was showing debtor days in excess of 365 days. The suspicion must have been that sales were being taken early and, although this was never proven, the shares collapsed to 5p.
Of course, examples can always be found in hindsight, but are these representative? I have evidence to show that it is sometimes possible to predict disaster a long time before it happens. A couple of years ago I concluded that three companies (Amey, Capita and Jarvis) involved in private finance initiative contracts had weak balance sheets and that it was inconceivable that all three would survive (see "Losing the initiative", November 2002). As it turned out, Amey did not have the resources to survive by itself and was taken over. Jarvis had trouble, with its share …