Today I’ll continue the series on stock market inefficiency in pricing shares by looking at some academic evidence showing an investor propensity to be slow in reacting to information releases from a company. This introduces the possibility of abnormal returns following the announcement of certain types of news.
Post-earnings-announcement drift
Post-earnings-announcement drift is the observation that there is a sluggish response to the announcement of unexpectedly good or unexpectedly bad profit figures.
Bernard and Thomas (1989) found that returns on shares, after allowing for general market movements and risk, continue to drift up during the few months after the company reports unexpectedly good earnings.
There is a downward drift down for firms that report unexpectedly bad figures for up to 60 days after the announcement.
This might offer an opportunity to purchase and sell shares after the information has been made public and thereby outperform the market returns.
Why might this happen, Some ideas:
- A delay in response to new information, due to:
- Investors failing to assimilate available information – they fail to recognize fully the implications of current earnings for future earnings, there is a slow dawning (I think I might have detected this with BHP shares following their recent improved profits announcement. At first there was little response; a week or so later the share took off)
- Costs exceed gains of exploiting this – the costs could be transaction costs or opportunity costs of implementing and monitoring a trading strategy
- Researchers failing to adjust for risk properly. Thus the returns to this strategy are merely compensation for additional risk.
The method used by B&T
Shares were allocated to 10 categories of “standardised unexpected earnings (SUE)”. To do this they looked at the earnings in the last quarter or half year.
This figure was compared with the earnings for the same period the year before.
But, before comparing, allowance was made for the “normal” drift of earnings over time for that company – we are only interested in deviations from the normal trend.
This was achieved by going back 20 or 24 quarter (or 9 semi-annual periods) and conducting a “regression analysis” to obtain the drift term (Simple explanation of regression analysis: imagine a graph of the earnings with time on the x-axis. Now imagine a hand-drawn line that describes the long term trend in earnings. “Regression” is simply the mathematical line of best fit for the trend – it’s a bit more sophisticated than a crayon)
Then expected earnings for the most recent quarter is the earnings from a year ago plus the drift term.
So, the unexpected earnings is the difference between the actual earnings and the earnings as estimated by the crayon line – sorry, regression line. (They do adjust this calculation for past volatility in earnings, but we don’t need to know this).
The 10 per cent of shares with the highest positive unexpected earnings were placed in category 10. The worst unexpected return shares were placed in category 1.
84,742 firm-quarters were observed for US shares 1974-86.
Results
In a completely efficient market, once the news is released we expect the share to move to its new efficient level now incorporating that news.
There will be no pote…………………………………To read the rest of this article, and more like it, subscribe to my premium newsletter Deep Value Shares – click here http://newsletters.advfn.com/deepvalueshares/subscribe-1