Wednesday, May 30, 2007

Surprise Anticipation

[6/29/07] Schwab research has found that companies that subsequently perform better or worse than current consensus expectations have some distinct traits. As the "Surprise Anticipation Matrix" graphic below illustrates, stocks with low, but rising expectation levels tend to report positive surprises. To identify these opportunities, we look for stocks with traits such as:
  • Low price multiples (e.g., low price-earnings ratios and low price-to-book ratios).
  • Management that is returning cash to investors (e.g. dividends, share buybacks, debt repayment).
  • Buying by corporate insiders and short sellers (i.e., "smart money" investors).
    Growing free cash flow (definition: cash from operations less capital expenditures and dividends).
  • Sales growth exceeding asset growth (i.e., improving operating efficiency).
  • Recent reported earnings above consensus forecasts.
  • Recent positive analyst earnings-per-share (EPS) forecast revisions.
  • Recent stock price outperformance vs. overall market.
Conversely, stocks with high but falling expectations tend to report negative surprises. These are often firms with great reputations stemming from strong past performance or firms perceived to have good future growth prospects, but these alluring attributes typically cause investors to bid up stock valuations to levels that leave no room for disappointment. To identify stocks likely to report negative surprises, we just flip the parameters of the criteria listed above.

[5/30/07] Professional stock analysts focus much of their time attempting to forecast companies' future earnings growth. Clearly, perfect foresight of future earnings would be valuable information, but that brings us to our third research principle: The fact that a correct forecast could have value is not sufficient reason to attempt such a forecast—one must also be able to make correct forecasts at a rate higher than chance alone.

To quantify the value of accurate EPS forecasting, a few years ago, Schwab measured the potential returns an investor could have earned with perfect foresight of EPS growth one year into the future. If it were possible in advance to buy the 20% of stocks with the highest subsequent one-year EPS growth, your portfolio would have grown an average of 38.7% annually from 1986 to 2004 vs. only 14.3% annually for all stocks included in our study. By contrast, a portfolio of the 20% of stocks with the lowest EPS growth (a negative growth rate in most cases) would have lost an average of 10.4% annually.

Clearly, accurate EPS forecasting could have a large payoff, but what's the probability of actually making accurate forecasts? Unfortunately, accurately forecasting EPS is extremely difficult. Schwab research has found historically that only 15% of quarterly EPS forecasts are within 1% of actual reported EPS. If at the beginning of each year, you bought the 20% of stocks with the highest consensus one-year EPS growth forecasts, your portfolio would have slightly underperformed the average stock included in our study. In other words, our findings have shown that analyst forecasts have historically been so inaccurate that basing stock purchase decisions on one-year EPS growth forecasts has been a completely futile approach to outperforming the stock market!

Forecast "surprises," not earnings
So if forecasting EPS isn't viable, what should you forecast? It's critical to understand that a company's stock price is based not just on its current fundamentals, but on all expected future cash flows, discounted back to the present according to anticipated risk. In other words, today's stock price reflects consensus expectations of future fundamentals. Consequently, meaningful stock price changes occur only when investor expectations change. And since expectation changes are triggered by the unexpected, a key element to a market-beating stock selection strategy is "surprise anticipation."

Successful equity research must be able to identify stocks that perform better than the expectations embedded in their stock prices at the time of purchase. If your stock picks experience positive surprises more often than negative surprises, you have a good chance to outperform the market in the long run. Unfortunately, investors often fail to appreciate the importance of expectations. For example, they cling to the notion that great companies will be top-performing stocks, forgetting that such greatness is generally already reflected in stock prices.

Surprise anticipation sounds nice, but is it actually possible to forecast the unexpected? The good news is that what is a surprise to most investors doesn't have to be a surprise to you! Behavioral finance researchers have discovered that investor expectations tend to err in two predictable ways.

First, investors seem to believe that successful companies will continue to succeed and that struggling companies will continue to struggle. This tendency to extrapolate past trends drives expectation levels to extremes for many companies. Yet research has shown that the forces of competition typically lead to long-term business success being more likely to reverse than to be sustainable. Therefore, one powerful way to anticipate surprises is to bet on what statisticians call mean-reversion, which is the tendency for high or low values to return to the mean, or average. When investor expectations are extremely low for a stock, you should anticipate future surprises to be positive as the company does better than pessimistic expectations. And when investor expectations are extremely high for a stock, anticipate future surprises to be negative as the company does worse than optimistic expectations.

The second way investor expectations err predictably reflects investors' reluctance to change held beliefs. For example, if a company reports a positive earnings surprise, investors typically notch up their future expectations, but they seek further confirmation by waiting to see if the company can do it again next quarter. Behavioral finance researchers have shown that this "anchor-and-adjust" process creates a tendency for short-term expectation changes to lag behind fundamental reality. Therefore, a second powerful way to anticipate surprises is to bet on systematic underreaction to newly reported information. If recent expectation changes have been positive for a stock, you should expect more positive changes in the near future as expectations play catch-up, but if recent expectation changes have been negative, expect more negative changes in the near future.

We have created a Surprise Anticipation Matrix that encapsulates these general tendencies. As the graphic below illustrates, stocks with low but rising expectation levels tend to report positive surprises while stocks with high but falling expectations tend to report negative surprises.

Schwab Equity Ratings: a surprise anticipation tool
If applying all of these research principles sounds like a lot of work, consider using Schwab Equity Ratings as a time-saving shortcut. Schwab Equity Ratings are, in essence, a sophisticated surprise anticipation tool. Indeed, since the ratings were launched in May 2002, through March 2007, A-rated stocks have reported quarterly earnings above consensus forecasts 74% of the time, while F-rated stocks have beat consensus forecasts only 41% of the time. We believe surprise anticipation is a key reason why A-rated stocks have outperformed the average rated stock by over 5% per year on a 52-week buy-and-hold basis since inception, while F-rated stocks have underperformed by a similar margin.

So in conclusion, successful equity research must be a process of discovering relevant information. When you buy a stock, you are implicitly forecasting that you know something material that other investors don't know. It seems to us that professional analysts who focus their research on EPS forecasting and investors who use those forecasts haven't accepted the fact that earnings forecasts have historically been too inaccurate to be the basis for a successful stock selection strategy. Fortunately, we believe there is a more reliable forecasting strategy available to investors called surprise anticipation.

by Greg Forsythe, CFA, Senior Vice President, Schwab Equity Ratings®, Schwab Center for Investment Research®

[6/14/07] See also Forsythe's article "Are Stocks With High Earnings Growth Good Investments?"

A recent academic paper showed that firms with above-average EPS growth in the previous five-year period were no more likely than chance to deliver above-average EPS growth in the five years that followed.2 Furthermore, since stocks with above-average historical growth tend to have above-average P/E ratios, investors in these stocks often get double-whammied: Not only does expected future EPS not materialize, but P/E ratios contract as investor expectations are adjusted downward to the new reality.

and his article in the Summer 2007 Charles Schwab OnInvesting

Investor expectations of a company's future earnings growth tend to be highly correlated to its past growth. Yet researchers have found that historical earnings growth rates have little ability to predict future earnings growth. Nonetheless, analyst forecasts reflect excessive extrapolation of past growth trends and influence investors to award high (low) price-to-earnings (P/E) multiples to stocks with high (low) EPS growth forecasts. Reflecting this overreaction tendency, research has shown that high P/E stocks tend to underperform as actual results fall short of optimistic forecasts, while low P/E stocks tend to outperform as actual results exceed pessimistic forecasts.

No comments: