Charting Your Way to Stock Gains

The human eye sees a straight line and projects it to eternity.

The investing world is more complicated. Companies may grow earnings "above trend," only to "revert to the mean" or even "fall below trend." Venture capitalists search for "hockey stick" markets, in which sales go from flat to infinity. In 1999, investors saw Internet trends as "trees growing to the sky."

Related Stories

Six Free Web Sites Investors Should Bookmark

But let's look beyond these buzz phrases at some straightforward techniques for projecting and interpreting trends. Although we'll be looking at charts, this is not technical analysis, a method of evaluating securities by relying on the assumption that market data, such as charts of price, volume, and open interest, can help predict future (usually short-term) market trends.

These techniques are applicable not only to stock prices, but any data series, such as sales of DVD players. This piece will show how graphical display of data can inform or mislead. With such knowledge, an investor is better equipped to make savvy investment decisions.

Let's start with a chart of closing prices in the S&P 500 since 1950:

S&P 500 Since 1950

Source: Heron Capital Management

I've chosen a time period of about 50 years so you can see expansions and recessions. As we learned in 2000, projecting S&P 500 returns by using data from just the previous five years got investors into a lot of trouble. I've added two lines: a smoothed one that shows what the S&P 500 would have looked like if it had grown at a constant 8.4% rate (the average growth rate for those 51 years), and a logarithmic regression line, which estimates a "best-fit" curve and has a growth estimate of 7.3%.

The "best-fit" curve offers the smallest average gap between actual and hypothetical data points. (Technical note: Logarithmic regression is applied to growth series such as the above index series, population or GDP growth. Linear regression is applied to data that charts more of a straight line: for example, weight vs. height in men -- the taller the man, the heavier the weight.)

A casual observer could draw invalid conclusions from looking at data displayed in this fashion. From 1950 through 1986, the the S&P 500 hardly seems to budge. Even the October 1987 stock market crash appears no more than a tiny blip. By comparison, the past 15 years appear to be an enormous boom, followed by an enormous bust. Since the human eye seeks symmetry, one might conclude that the S&P 500 is heading down toward 500.

Converting the same data to a log chart eliminates this concern:

S&P 500 Since 1950: Log Scale

Source: Heron Capital Management

The log scale presents percent changes rather than point changes. With such a presentation of data, a move in the S&P 500 from 10 to 20 appears the same as a move from 100 to 200, or 1000 to 2000. This chart reveals that the S&P 500 has a long-term uptrend with short-term rallies, plateaus and dips. The bear markets of 1973-74, 1987 and 2000-01 are noticeable, but don't appear nearly as painful as we recall.

Why does this presentation of data make sense? The S&P 500 index is dependent on growth in S&P 500 earnings, which is dependent on growth in the economy. This chart shows that even after shocks to the U.S. economy, such as September's terrorist attacks, it remains capable of continued growth. Therefore, taking all your money out of the stock market based on the negative returns of the past two years flies in the face of a substantial 50-year trend.

How aggressively should you be invested in stocks right now? This chart shows the year-over-year change in the S&P 500:

S&P 500 Year Over Year Since 1950

Source: Heron Capital Management

Even though the average year-over-year gain is 8.4%, the average year-over-year change in the S&P 500 swings from extremes, with little time spent around the average. (I call this a "sine wave" pattern.) The worst hit was during the 1973-74 bear market and the second-worst was the most recent bear. The chart also shows that annual gains were sustained for an unusually long time during the 1993-2000 bull market.

What was happening during the peaks and valleys of data in this chart? Most of the sharp increases in the stock market index correspond to the end of recessions (examples: 1975, 1983 and 1994). By comparing this chart to one that tracks changes in corporate earnings, I noticed a high correlation between growth in corporate earnings and that in the stock market.

We've just finished a period where S&P 500 earnings fell more than 50% from the third quarter of 2000 to the same period in 2001. In 2002, earnings growth estimates range from 5% to 15%. By combining the likely return of earnings growth with the end of the recession, we can project an upswing in the S&P 500 based on the sine wave pattern above. As a result, it is currently a moderately low-risk time to be investing in the stock market.

---

There are many graphics products and stock charting tools to calculate smoothed trend lines and convert the data into log scale. In the above examples, I used the "average" function in an Excel spreadsheet to derive the smoothed curve, the "logest" function to derive the "best-fit" curve, and created the charts in default and log scales.

The Visual Display of Quantitative Information by Edward R. Tufte is a good reference book on this subject. You also may want to review my column Island of Foxes and Hares: How the S-Curve Forecasts Stock Behavior.

---

David Edwards is a portfolio manager and president of Heron Capital Management, a New York management firm. At the time of publication, Edwards didn't hold a position in any of the stocks mentioned in this article, although holdings can change at any time. Under no circumstances does the information in this column represent a recommendation to buy or sell stocks. Edwards appreciates your feedback and invites you to send it to David Edwards.