What numbers can you trust? Public companies publish a lot of information about their financial health and that has to be the key to valuing them. But accounting scandals, starting with Enron, have shaken confidence in the numbers accountants produce.
Instead, the zeitgeist is that "cash is king". There is a prevalent idea that investors should ignore accrual-based accounts and instead look at companies' cash flow.
Bankers and private-equity buyers often rely on cash when deciding on deals and so, this logic goes, should everyone else.
Another phrase was added to conventional wisdom at about the same time: "Don't trust analysts." The scandals at the beginning of this decade made clear that many analysts were, to all intents and purposes, just extensions of investment banks' other operations, existing primarily to generate business for dealmakers.
This new received wisdom showed that the investment community was learning the right lessons from the accounting debacles.
But new research from the business schools of the University of California at Los Angeles, Columbia University and Yale University, suggests that conventional wisdom is now ahead of itself. They studied valuations across a wide range of companies in Australia, France, Germany, Hong Kong, Japan, South Africa, Taiwan, the UK and the US. What, they wondered, would be the best predictor of a company's valuation? Earnings (as determined by accrual accounting), operating cash flow or dividends (for those that had them)?
They used complicated statistics to see how reliably different measures could be used to predict the market's valuation of a company. The answer, in every country, was earnings - and by a statistically significant margin.
In the US, where they had access to more data, they also looked at multiples of sales, book value and ebitda (earnings before interest, taxation, depreciation and amortisation, one of the most popular measures for screening out accrual accountants' assumptions).
With this wider field, earnings were again by far the best predictor of a company's valuation. Sales and operating cash flow, which put the greatest weight on the "raw" amounts of cash coming in to the business, were the worst.
They also looked at forecasts, using data of analysts' predictions in various countries, provided by Thomson's I/B/E/S earnings consensus database. Earnings forecasts were better predictors than cash flow or dividend forecasts.
Further, given a choice between forecast and historic earnings, it is the forecast earnings that prove to be the better predictor, in all the countries covered. And the longer-term the forecast, the more confidently you can use it to predict a company's valuation.
Forecasts for two years hence, although less likely to be accurate, are still more central to a company's current market value than one-year forecasts.
Earnings are also a better predictor than dividends, although only by a small margin in Hong Kong, Japan and South Africa.
Again, dividend forecasts are better than current dividends, but not to the same degree by which earnings forecasts are better than actual reported earnings. This is probably because dividends tend to be "sticky", so their current level is a good predictor of dividends in the future.
But the trend is clear - the assumptions made by accountants still seem closer to the market's fundamental view of a company's worth than any of the alternative cash-based measures that are harder to manipulate.
Why is this the case? A reminder of the main assumptions required under accrual accounting demonstrates this clearly enough.
Accountants need to decide on depreciation. It would plainly give an inaccurate account of the health of a company to assume that its assets will last forever, so the value of big investments, particularly machinery, should be written down each year. This avoids an apparent sharp fall in profits when a company buys major new plant or equipment.
As for revenue recognition, it often makes sense to "book" as income money that has not yet been received as cash - if a company has good reason to believe it is coming.
Flashpoints remain. If a company claims to be earning money but bleeding cash, that is a bad sign.
Tax is another danger signal. Governments are not stupid. If a company declares a big profit, but persuades the government that it need not pay any tax, that is a signal that it is not as profitable as its numbers make it appear.
And the cash flow statement is published for a reason. When accountants make assumptions that do not show up in the company's bank accounts, those assumptions will show up in the cash flow statement (barring outright fraud). Cash flow statements are worth reading, and not only when a firm is in trouble.
But the bottom line, as accountants might put it, is that for all the problems with accrual accountants, and with investment bank analysts, there are reasons why they exist. When deciding whether to invest, you are better with them than without them.
Cash Flow is King? Comparing Valuations Based on Cash Flow Versus Earnings Multiples, by Jing Liu, Dorron Nissim and Jacob Thomas.