TabU 11-1

Junk Bond Funds

1987-1997

Source: Prepared from Morningstar data

Earnings Surprise: Another Profitable Overreaction

You can frequently hamess the powerful effects of a disappointing eamings 8 8 by buying GARP (Growth at a Reasonable Price) companies after they have been blasted. A good company trading at an above average P/E will often break 8 1 on bad news. The trick is to evaluate the eamings su rise. Is it something that will alter the company permanently? Or is it simply analysts overoptimism or some other unpredictable but short-term negative event? Su rise need not be eamings related. Temporary loss of market share, industry price-cutting, a slowing of sales, and a host of other factors can also result in highly valued companies plummeting.

Still, eamings su rises are the most common investor overreaction. For example, Intel dropped 20% in three days when reported eamings in hs June 1995 quarter were up strongly, but sdll 4% below the analysts consensus forecast. Eamings continued to grow briskly and the stock almost tripled by the summer of 1997.

Is this an isolated example? Not at all.

Hewlett-Packard announced in late September 1992 that its quarterly eamings would come in below analysts estimates. The next day the stock dived 13 points, or 18%, on huge volume. The camage was out of all proportion to the cause: on a shortfall of a few tens of millions of dollars, the stock lost $3.5 billion in market value. In todays market, disproportionate drops caused by eamings shortfalls are, if anything, becoming more and more frequent.

How should the smart investor react? Buy right after the plunge? Or

The Investor Overreaction Hypothesis

Although the term "investor overreaction" is as old as markets, I put forth a statistically testable explanation of how it worked in Contrarian Investment Strategy in 1980. Based primarily on the psychological forces we examined earlier, the hypothesis states that investors overreact to events in a predictable fashion: they consistently overvalue the prospects of "best" investments and undervalue those of the "worst." They extrapolate positive or negative outlooks well into the future, pushing prices of favored stocks to excessive premiums and out-of-favor stocks to deep discounts. (Performance of "best" and "worst" stocks can be directiy compared, of course, but "best" and "worst" investments can be things other than stocks, and "best" might be in a different market than "worst.")*

wait till the dust clears? Its a tough call, but my advice would be to let the dust clear. Dont be a hero and charge into the initial panic. If you like a stock blown out by disappointing news, it pays to sit on the sidelines for a while. In all probability, you will get plenty of chances to buy it cheaper in the next 90 days.

The nature of the Streets research explains why the phenomenon occurs so consistently. Analysts react to surprises by slashing their quarterly estimates on the announcement of the disappointing news. Even so, the changes in their projections are usually not large enough for the next few quarters or the following year.

When there is a negative surprise, the poorer results, even for first-rate companies, are likely to continue for a while. It takes time to ride through an unanticipated rough stretch. As a result, the initial shock is often followed by later, if lesser shocks, which continue to pressure price.

So be patient. If a company you like comes out with an announcement that shakes the markets confidence, play the waiting game. It should pay off well over time. In the case of Hewlett-Packard, this is exactly what happened. The surprise was caused by a slowdown of sales in its European instrumentation businesses, not by its fast-growing computer peripheral lines. The stock bounced back from $14% at the time of the su rise to over $60 in December of 1997. Investors got a classic GARP double play with the P/E rising from about 16, under market at the time, to 21.6, on earnings that tripled.

If you know a company well, reacting to negative earnings su rises can make you a bundle.

In 1981, for example, "best" investments included gold-at over $800 an ounce-and "worst" tax-exempt municipal bonds, yielding as high as 15%. Bonds soon rose sharply while gold plummeted. Premiums or discounts on favored or unfavored investments can be substantial and last a long time.

Specifically, the Investor Overreaction Hypothesis predicts that after earnings or otiier surprises, investments previously considered to be "best" underperform, while those considered to be "worst" significantly outperform, as bodi regress towards a more average valuation. The hypothesis also states that the maximum price swing is produced by negative surprises on "best" stoclcs and positive surprises on "worst." On the other hand, positive surprise on favored stocks and negative surprises on out-of-favor stocks-reinforcing events-corroborate the markets opinion of these stocks and have a lesser impact on price movements than event-triggers (see chapter 6).

Finally, the overreaction hypothesis holds that even without the occurrence of an event trigger, the "best" and "worst" investments regress towards the market average. Because the investor overreaction hypothesis is based on psychological principles, it is likely to apply in other markets and in fields outside of investments and economics where risk and uncertainty exist.

The Investor Overreaction Hypotiiesis makes these predictions:

1. "Best" stocks underperform the market, while "worst" stocks outperform, for long periods.

2. Positive surprises boost "worst" stocks significantly more than they do "best" stocks.

3. Negative su rises knock "best" stocks down much more than "worst" stocks.

4. There are two distinct categories of surprise: event triggers (positive surprises on "worst" stocks, and negative 5 8 8 on "best"), and reinforcing events (negative 8 8 8 on "worst" stocks and positive su rises on "best"). Event triggers result in much larger price movements than do reinforcing events.

5. The differences will be significant only in the extreme quintiles, with a minimal impact on the 60% of stocks in the middle.

The hypothesis states that overreaction occurs before the announcement of an earnings or other su rise. A correction of the previous over-reaction occurs after the 8 8 . "Best" stocks move lower relative to