Serious Money: The ratios used by analysts to value firms go in and out of fashion, writes Chris Johns

In "how to" manuals of investing, we are often exhorted to focus on the underlying fundamentals of a company's business. This sounds sensible enough, although it is rarely straightforward in practice.

One problem is that any investor needs to become acquainted with some rudimentary elements of accountancy in order to get to grips with the various ways in which financial information is presented.

A basic understanding of profit and loss statements and balance sheets is of obvious importance; from these we can glean information about many of the key numbers that analysts believe are essential inputs into any investment decision.

For a long time, the humble price/earnings (P/E) ratio held sway as the most important number in the investment toolkit. This ratio has the virtue of simplicity and it is easy to calculate. Indeed, the P/E ratio is still widely used for these very reasons. Unfortunately, it has a number of drawbacks, not least of which is the issue of precise and proper measurement of earnings and, relatedly, the ability of the company, or any analyst of the company, to manipulate (properly or improperly) the way in which profits are counted to suit a particular story.

To get a more accurate grip on underlying profitability, analysts came up with a whole raft of measures, many of which try to focus on the pure amount of cash generated by the firm in its day-to-day operations. Different bits of a company's income statement were focused on, and concepts such as free cashflow and EBITDA (earnings before interest, taxes, depreciation and amortisation) entered the lexicon; new ratios such as EV/EBITDA (enterprise value/ EBITDA) became part of the toolkit, alongside many others.

The use of new ratios mostly represented an honourable attempt to avoid the pitfalls associated with old-fashioned measures but often quickly succumbed to problems of measurement and manipulation themselves.

As the ways of looking at financial information proliferated, so did confusion grow over just how to value the worth of a company. A lot of these issues remain unresolved, no matter what some stockbrokers like to tell us. This is not to say that all, or even many, company results are distorted but more to point out some of the potential pitfalls.

A leading UK analyst called Terry Smith once famously lost his job (with investment bank giant UBS) for writing a book on how accounts can be manipulated. Some of Smith's criticisms have been dealt with by regulatory changes but much of what he wrote remains apposite.

He went on to head up a firm of stockbrokers that has championed the use of something called CFROIC (cash flow return on invested capital) as the most important measure of underlying profitability.

Other firms, notably Holt, (now owned by another large Swiss investment bank, CSFB) use essentially the same approach.

CFROIC is undoubtedly one of the cleanest measures available but, nevertheless, it needs to be used with care.

That it has yet to be adopted as the universal standard is testament to the enduring appeal - appropriate or not - of simpler measures; that it is relatively complex to calculate and often delivers subtle messages means that it does not appeal to those who would argue that investing is a simple business.

There are too many measurement issues surrounding earnings to be listed here, but a few examples will suffice. Most obviously, there are international problems - different countries employ different accounting standards.

Firms with listings on several stock markets will sometimes present different measures of earnings to comply with local rules.

A common example is a European company that presents its profits according to home country conventions as well as earnings according to US GAAP (generally agreed accounting principles) for its US investor base.

Some of these international issues will be resolved if proposals to unify standards under the auspices of the International Accounting Standards Board come to fruition.

In the US, pro-forma earnings came to notoriety during the bubble years of the late 1990s when, in one or two extreme cases, some companies simply excluded as much of the negative profits items as they could get away with. This measure, although still in use, has become rather discredited and, along with some other ways of looking at profits, has been christened in some quarters as EBBS (earnings before bad stuff).

There remain two key (simple) concepts of US earnings in common usage: "operating" and "as reported" earnings.

In order to get around measurement problems surrounding, in particular, operating earnings, S&P, the large rating and research agency, has introduced a third concept called "core earnings" which, amongst many other things, tries to account properly for pension liabilities and share options.

If you think this is not terribly clear, remember that the US is arguably the most regulated and transparent of stock markets.

Problems common to many countries include acquisition accounting, the appropriate treatment of capital spending and the appearance of so-called extraordinary items; the treatment of operating leases and R&D expenses also occasionally attracts controversy. Even accountants can get confused.

One upshot of all of this is that the ratios used by analysts to value companies go in and out of fashion. One of the latest games in the field of quantitative asset management is to develop fiendishly complex models to predict precisely which measure, or combination of measures, will next be in vogue. Some days it will be important to make informed guesses about EV/EBITDA, others less so.

What is the ordinary investor to do?

First, read a good book on the subject (as I said, good investment practice doesn't come easy); I recommend anything by Aswath Damodaran (generally from a US perspective but always an excellent overview).

Second, never concentrate on just one measure of performance.

Third, be aware of the flaws inherent in the various cash-flow based measures but realise that they represent the nearest to best practice that we have.

Fourth, whatever measures you use, come to an informed opinion about where analyst views are likely to change.

Are analysts likely to become more or less optimistic about whatever profits measures they are using?

More on the forecasting of profits, however measured, next week.