Company earnings are, you’d have thought, a pretty straightforward measure of a company’s health. After all, earnings are a statement of profits, are they not? Of course this is finance and, therefore, nothing is quite as it seems.
You see earnings are not necessarily cash and contain a mysterious component called accruals, the calculation of which keeps accountants the world over gainfully employed. Companies that rely heavily on accrual based earnings tend to have more earnings surprises and this anomaly is exploitable by investors: or at least it was, until it mysteriously vanished.
Corporate earnings have two components: cash and accruals. Cash is usually easily measurable: it’s the amount the company has received in cash in return for services and products. Cashflow is king, as they say, and all investors should take a hard look at a corporation’s cashflow before they invest: if they’re spending more hard cash than they get then trouble may be looming.
Accruals, however, are also part of a company’s earnings but they’re not quite the same as cash in the bank. They include, for instance, IOUs, known as accounts receivable, and also changes to inventory, stock held for future sales. These end up on the balance sheet as net operating assets and you can calculate the accruals for a given year by figuring out the change in net operating assets between one year and the next. Add cashflow to accruals and you have earnings. Simple, eh?
OK, this is vastly simplified, but the point is clear: accruals may make a huge difference to a company’s earnings and their calcluation is based on some predictions about the future which, as we all know, is riddled with uncertainty. Given two companies with the same earnings we should prefer the one with more cashflow and less accruals. In fact, in some cases we can find companies with negative cashflow and positive earnings and that should be a big warning sign.
When Richard Sloan finally got his paper on the accruals anomaly published in 1996 it created a bit of a furore. Here’s the conclusion:
“The persistence of earnings performance is shown to depend on the relative magnitudes of the cash and accrual components of earnings. However, stock prices act as if investors fail to identify correctly the different properties of these two components of earnings”.
Put simply company earnings growth tends to be persistent over time – higher growth companies continue to grow fast and slower growth companies continue to grow slowly. However, this is dependent on how much of the earnings is actual cash and how much is estimated accruals. A growth stock whose earnings are dominated by accounting estimates is highly likely to suffer a reversal in fortunes, while those whose earnings are dominated by cash are likely to continue to be highly profitable.
Risky or Irrational?
A lot of economists spent a lot of time trying to prove that Sloan’s finding could be explained rationally under the Efficient Markets Hypothesis: after all, rationally, such an obvious anomaly should be arbitraged away before investors can take advantage of it. It’s a testimony to how powerful the EMH is that it could even be leveraged to explain such an obviously irrational mispricing anomaly, but it also indicates the theory’s fundamental weakness: a theory that can’t be disproved isn’t a theory at all, it’s a statement of faith.
In The Accrual Anomaly: Risk or Mispricing the researchers, David Hirshleifer, Kewei Hou and Siew Hong Teoh look at whether the accrual anomaly is actually a rational response to additional, but as yet unperceived, risk. It’s certainly the case that the existing rational risk models don’t capture the accrual effect, but the underlying argument is that there are some additional risk factors that, if added, will restore order to the efficient market universe.
This, of course, is akin to the medieval astronomers we described in Copernicus, Muddling Through, who frantically kept adding ever more complexity to their Earth-centric models of the Solar System to maintain the fiction that the Earth didn’t move. In the end a good scientific explanation should be as simple as possible, and adding more and more risk factors to the efficient market models looks increasingly like an act of desperation. As the researchers, wryly, remark:
“A naïve strategy of proposing new factor structures until the anomaly vanishes can ‘work’ even if the anomaly in fact represents market inefficiency rather than a rational risk premium”
Anyway, Hirshliefer and colleagues concluded, within the bounds of the testing that was possible, that what they were seeing wasn’t the result of some unknown risk factor, but actual market mispricing of stocks – people simply were failing to adjust their expectations based on the difference between real cash based earnings in the bank account now and future based estimates of accounting earnings sometime in the future.
The general theory here, that investors overestimate the value of accrual based earnings somewhat begs the question of why this is. It clearly suggests that accountancy estimates of future earnings aren’t particularly reliable – which we shouldn’t be too surprised about – but it also leaves open the issue of why this is. There are a couple of related explanations. The first is that accruals allow company managements the opportunity to manipulate earnings. The second is that companies with higher levels of accruals are more vulnerable to changes in the future.
Support for the first point comes from research by Strydom and colleagues who show that firms with high accruals and good corporate governance tend to have a lower level of accrual anomaly. In essence transparency leads investors to correctly evaluate the company’s valuation. This doesn’t rule out the possibility that even well managed companies with high accruals are more likely to suffer earnings reversals but it does suggest a more detailed level of analysis of quality of corporate accrual statements might yield investors an edge.
This edge might well be needed because, in line with the research we discussed in Pricing Anomalies: Now You See Me, Now You Don’t, it appears that the accruals anomaly is shrinking. In Going, Going, Gone: The Demise of the Accruals Anomaly the researchers find that the abnormal returns generated by the anomaly have disappeared and suggest that the causes of this are:
“A decline in the size of the mispricing signal (as measured by accruals in the extreme accrual deciles and the relative persistence of cash flows and accruals) and an increase in the capital invested by hedge funds into exploiting the signal (as measured by hedge fund assets under management and trading volume in high accrual firms)”
Basically, in the wake of the great accounting scandals of the first decade of the twenty first century both accountants and managements have become more cautious about using accruals to manage earnings, while hedge funds have hired up all the academics that researched the anomaly to exploit it, and have promptly arbitraged it away. Which is a nice example of how adaptive markets work: it took academic research to identify the anomaly and then practitioners acted on this research to change the way markets operate.
Accruals are not Cash
Of course, it’s not quite true that the effect disappeared overnight. If you ever wanted a big red flashing warning sign for the 2007 crash the accrual anomaly is it. Banks issued loans on little collateral but capitalized the future loan payments onto their balance sheets and thus boosted their earnings dramatically through the use of accruals. Yet despite the fact that accrual anomaly is well known this doesn’t seem to have given most analysts, managements or investors a moment’s concern, as they happily equated possible future earnings with actual cashflow.
Food for thought, if you’re of an inclination to trust analyst forecasts or lend any credibility to the predictions of company managements.
I'd also recomment Accounting For Growth by Terry Smith if you can find a copy.