In autumn of the year 2000, Jeffery Abarbanell and Victor Bernard published a thoughtful paper investigating whether financial markets appeared to have a preference for the timing of accounting earnings. In other words, does it really matter to investors whether net income will be earned this year or next, provided that the net present value of earnings is equal in both instances? According to the efficient market hypothesis, the timing of earnings should be irrelevant and investors should be ambivalent to earnings achieved either now or in five years, provided that the present value is equal. Yet we often hear in the popular press of managers feeling the pressure to 'deliver' accounting earnings to meet an analyst forecast every quarter. Is this pressure that managers feel actually coming from investors, or are investors willing to wait for accounting earnings?
Abarbanell and Bernard test this question by modeling firms' market caps as a function of their book value of assets and analyst forecasted earnings one year ahead, two years ahead, three years ahead, and further into the future (i.e. 'terminal earnings' or 'long-term earnings'). Their regression results indicate that one-year-ahead accounting earnings are much more related to a firm's price than either the book value of assets or the firm's terminal earnings.
To validate that one-year-ahead earnings are truly more significant than long-term-earnings or the book value of assets, the authors develop a trading strategy to take advantage of this mispricing. If investors place too much focus on short-term-earnings and later correct this mistake, then a firm's stock price will decline after the realization of short-term-earnings as price reverts to a normal level. This trading strategy test fails, though, and Abarbanell and Bernard don't find the decline in stock price that they would have predicted. Because they can't demonstrate this trading strategy, the authors conclude that the analyst forecasts of earnings in their first model must just have been estimated with too much error to be used reliably.
Abarbanell and Bernard conclude, then, that they really haven't identified a trait of the financial market after all (at least not from 1978 to 1993). They don't find that stocks are mispriced because of overvaluing short-term earnings, they can't find a trading strategy that would reliably exploit an overvaluation of these earnings, and what they really claim to have discovered is that investors don't pay a premium to get accounting earnings in the next year.
So what do you think? Are accounting earnings something you take into account when assessing the price you'd pay for a stock? For instance, do you use the P/E (price to accounting earnings) ratio when figuring out whether to buy a stock? Or do you think accounting information is relatively unhelpful in your investment process?
If accounting earnings are important to you, do you look for analyst forecasts of earnings to judge a company's growth prospects/health? Would you rather have a firm who's predicted to have higher accounting earnings next year, or is it perfectly fine with you to wait on earnings later, provided that you'd get more of them? Let me know!
----------------------------------------------------
Papers Discussed:
Abarbanell, Jeffery, and Victor Bernard. 2000. “Is the U.S. Stock Market Myopic?” Journal of Accounting Research 38 (2):221–42. https://doi.org/10.2307/2672932.
