Join our community of smart investors
OPINION

Experience tells

Experience tells
November 25, 2009
Experience tells

So, whereas classical economics once told us that companies are run by all-knowing individuals acting rationally, now we know that it's a bit more complex. Even economists have learnt by their mistakes and now tell us that within companies decisions are made by people acting on imperfect information collaborating, and sometimes competing, with each other. Where once the key assumption was that companies were in equilibrium, or tending towards it, now the assumption is that they are in flux, where tomorrow is never the same as yesterday. And the key aim for companies is to make a virtue of this by surviving - and even thriving - by gathering feedback from the jumble of signals that the marketplace throws off and building organisations that can find profitable opportunities in this state of continuous change.

The companies that thrive do so because their employees get better with practice at operating in this tough environment. As I said, experience counts. But the vital questions for us investors are: how do we measure this improvement and, once measured, can we relate it to the stock market value of companies (ie, can we use it as a tool to tell us whether a company's shares are cheap or dear)?

Quite possibly. The crucial measure, said a trio of businesspeople and academics - Jacques Solvay, Michèle Sanglier and Paul Brenton - is what they labelled "value-added per employee". This is defined as cash profits per employee or, to use accounting jargon, is earnings before interest, tax, depreciation and amortisation for a given year divided by the number of employees.

Certainly, cash profits seems the sensible figure to use. Focus on sales per employee and there is a danger that factors beyond the influence of employees will distort the ratio, such as the effect of higher input costs that are simply passed on to customers. Move down the income statement and other factors may do the distorting - why, for example, should a measure of value-added per employee be compromised by the way in which a company is financed or by accounting oddities, such as the rate at which the finance director writes off the plant and equipment?

In their book that deals with using this ratio - Modelling the growth of corporations (Palgrave) - the authors stress that sound conclusions can only be drawn from a long run of data, which irons out the effect of the business cycle. They recommend at least 10 years-worth of data.

There is a contrary argument which says that, because the nature of a company can change so dramatically over that sort of period, making conclusions from such a set of data would be worthless. That point has merits, but it really serves to remind us to be careful about the sort of companies that we subject to this test. We need to focus on well-established businesses operating in industries where the scope for radical change is limited and where demand won't evaporate. That points to capital-intensive firms supplying goods rather than services; industries such as food processing, oil production, chemicals, pharmaceuticals and engineering.

And it just so happens that I have to hand 20 years-worth of the relevant data for a company that fits these requirements - West Midlands-based engineer Castings. True, demand for the goods that Castings produces - the flexible cast iron widely used in the automotive industry - won't vanish. But it may decline gently and Castings will be subject to all sorts of competitive pressures. This much is shown by the company's efforts over this period. Although value-added per employee has increased substantially - from £5,700 in 1989-90 to a peak of £20,600 in 2007-08 - that has been insufficient to get Castings into that virtuous cycle where growing productivity drives higher employment which, in turn, feeds further productivity gains. Quite the reverse. Castings has maintained productivity by hacking away at its staff costs. Employee numbers peaked in 1998-99 and, by the end of the current year, will be down to a level not seen since 1990-91.

That's because this recession has hit Castings' value-added per employee, but not so savagely - and this is the crucial point - as the ferocity with which its share price was attacked earlier this year. This point is clearly demonstrated by running .

As an aside, purists may note that it may be odd to juxtapose the share price with employee value-added. The correct measure should surely be the market value of equity plus debt. True, but because Castings has never had debt during the whole 20-year period, market value and share price are effectively the same.

That's very convenient. It means we can skip straight through to the share price and, on a chart, the data shows a tidy fit between value-added and share price - where the former moves the latter tends to follow. However, far - very far - from the line of best fit - the line that plots the best course through the chart's dots - is the 100p price to which Castings shares sank in March for the year when value-added per employee fell to £14,000. The regression indicates the share price should have been close to 200p. And - indeed - the price has since bounced back to 175p.

Is that a coincidence or is it telling us something useful? Possibly the latter but, before I decide, I will have to run this analysis through scores of companies. Finding the data may be tricky. But, if I do, there may be a worthwhile experience waiting. I'll report back.