Price on Value
December, 1999

Surprise! Surprise!

John Price,  Ph.D.

Have you ever been embarrassed by a surprise birthday party? It might help you to get over it if your shares had positive earnings surprises.

What is an earnings surprise?

One of the main tasks of a share market analyst is to forecast the earnings of a company. These forecasts can just be for the next accounting period. More often they run out for at least a few years. They are generally reported as earnings per share EPS meaning the net profit of a company divided by the number of outstanding shares. When a company has more than one analyst following it, the figures can be stated as consensus or average forecasts.

An earnings surprise is when the reported earnings differ from the consensus forecasts. If the difference is positive it is called a positive surprise. Otherwise it is called a negative surprise.

When there is a positive surprise the price of the share will generally rise. The opposite will take place when there is a negative surprise. The reason for this is that earnings of a company are closely tied to the price that the market is willing to pay for it. Also, valuation models for shares usually involve estimates of the future earnings of a company.

Just as a bond can be valued by discounting its stream of dividends back to present time, so a share can be valued by discounting its stream of earnings back to present time. Of course, things are more complicated with shares. For one thing, you don’t actually receive the earnings. Nevertheless, this ‘discounting’ idea is the core of most valuation models.

Another piece of evidence regarding the importance of the earnings of a company is that, after the share price, usually the first figure to be looked at when considering a purchase is the price-earnings ratio. This is the ratio of the price of the share divided by the earnings per share. A simple way to think of this is that it is the number of years for the earnings to equal the price of the share. If the P/E ratio is 10, then it will take ten years of earnings at the current level to equal the price. If the ratio is 20, it will take twenty years, and so on.

The idea of an earnings surprise explains why a share can drop in price even though there has been an increase in earnings. The market was expecting an even bigger increase.

Research on earnings surprises

There are many studies describing the relationship between earnings surprises and share price movements. David Dreman and Michael Berry studied the effect of earnings surprises on USA shares over 20 years from 1973 to 1993 (Financial Analysts Journal, 1995). They divided shares into three groups: shares with high, medium and low P/E ratios. Within each group, shares with positive surprises outperformed the group by a healthy 12 percent in the quarter following the surprise. Enough to keep you smiling during the most embarrassing surprise party.

But there is more. Over the following year the shares continued their outperformance by 4.3 percent.

The results were the exact opposite for negative earnings surprises. For the first quarter after the negative surprise they underperformed the shares in their respective groups by 12 per cent for the first quarter and by almost 5 percent for the first year.

Predicting an earnings surprise

Jia Ye tells us that you can improve the predictability of earnings surprises. In the Journal of Portfolio Management (1999) he presents data showing that firms with a two-for-one split have 40 percent more positive surprises than negative surprises in the year following the split. For non-splitting firms, the number of positive and negative earnings surprises were about equal.

Consider the following simple trading rule: each month buy an equal amount of stocks that announce a two-for-one or three-for-two split during the previous two years, and sell stocks in the portfolio that do not meet this requirement. Backtesting over the years from 1979 through 1996 showed that this rule outperformed the S&P 500 by a healthy 5.29 percentage points per year.

Using Earnings Surprise Data

There are a number of ways that you can use earnings surprise data. The most obvious one is to buy stocks that have a positive earnings surprise. The difficulty with this is that you have to be quick to get the full advantage. There is usually a sudden upward movement the moment the earnings are announced so you are not going to be able to buy at the pre-announcement price. Still, studies such as the one above by Dreman and Berry show that the returns are above average for the next year or so.

Earnings surprise data also gives us useful information about the management of the company. Even though analysts are not employed by the company, they generally attend press conferences and other meetings arranged by the company. At these meetings the management of the company will discuss the future prospects of the company. In the jargon of the industry, companies guide analysts towards earnings figures.

So what can we deduce if there are very large positive or negative surprises? Either the analysts are not paying close attention or the company itself is unsure of the profitability of its own operations. Assuming the second alternative, investors looking for long term value might be advised to skip over companies which regularly have very high positive or very low negative earnings surprises.

On the other side, if earnings forecasts and the actual earnings tend to be close, then we have a better chance of evaluating the worth of a company. Investing in such a company is just a little closer to investing in a bond with a predicable stream of payments.

Is there a catch? You bet! According to Arthur Levitt, chairman of the SEC, many companies are now working the system so that they come up with a positive earnings surprise quarter after quarter. Have a look at the history of Microsoft, Dell and Lucent. Last quarter well over half the companies in the top 500 reported positive surprises. And if a company can’t come up with positive surprise, then likely as not it will take a ‘big bath’. The theory goes that Wall Street will not take too much notice of a one-time loss. The aim is to stash all the losses, and perhaps a bit more, into a single quarter. Then trickle out the resulting savings at a rate just above analysts forecasts.

Happy birthday!
______________

Next month’s article "Earnings forecasts made easy" will look at some of my recent research on how to improve your ability to forecast earnings. It is based on a new measure I call STAEGR^TM which is short for stability of earnings growth.