One thing we can do is examine the past to broaden our sense of the range of possibilities. An obvious feature of this inflation-adjusted chart of the S&P Composite is the pattern of long-term alternations between up- and down-trends. Market historians call these "secular" bull and bear markets from the Latin word saeculum "long period of time" (in contrast to aeternus "eternal" — the type of bull market we fantasize about).
If we study the data underlying the chart, we can extract a number of interesting facts about these secular patterns:
The annualized rate of growth since 1871 is 1.91%. If that seems incredibly low, remember that the chart shows "real" price growth, excluding inflation and dividends. If we factor in the dividend yield, we get an annualized return of 6.58%. Yes, dividends make a difference. Unfortunately that has been less true during the past three decades than in earlier times. When we let Excel draw a regression through the data, the slope is an even lower annualized rate of 1.70% (see the regression section below for further explanation).
If we added in the value lost from inflation, the "nominal" annualized return comes to 8.79% — the number commonly reported in the popular press. But for an accurate view of the purchasing power of our returns, we'll stick to "real" numbers.
Since that first trough in 1877 to the March 2009 low:
Based on the real S&P Composite monthly averages of daily closes, the S&P is 40% above the 2009 low, which is still 42% below the 2000 high. The 2009 low measures about 6% above the average decline for secular bear markets. Of course, this number is a bit skewed by the bottom in 1932, which saw a greater decline over a much shorter period (three years versus nine).
Add a Regression Trend Line
Let's review the same chart, this time with a regression trend line through the data.
This line essentially divides the monthly values so that the total distance of the data points above the line equals the total distance below. Remember that 1.91% annualized rate of growth since 1871? The slope of this line, an annualized rate of 1.70%, approximates that number. The 0.21% difference is largely a result of the rally over the past 16 months.
Regression to trend usually means overshooting to the other side. The latest monthly average of daily closes is 28% above trend after having fallen only 6% below trend in March of last year. Previous bottoms were considerably further below trend.
Will the March 2009 bottom be different? Only time will tell.