Do Stock Prices Move Too Much to be Justified by Subsequent Changes in Dividends?

A simple model that is commonly used to interpret movements in corporate common stock. price indexes asserts that real stock prices equal the present value of rationally expected or optimally forecasted future real dividends discounted by a constant real discount rate. This valuation model (or variations on it in which the real discount rate is not constant but fairly stable) is often used by economists and market analysts alike as a plausible model to describe the behavior of aggregate market indexes and is viewed as providing a reasonable story to tell when people ask what accounts for a sudden movement in stock price indexes. Such movements are then attributed to "new information" about future dividends. I will refer to this model as the "efficient markets model" although it should be recognized that this name has also been applied to other models. It has often been claimed in popular discussions that stock price indexes seem too "volatile," that is, that the movements in stock price indexes could not realistically be attributed to any objective new information, since movements in the price indexes seem to be "too big" relative to actual subsequent events. Recently, the notion that financial asset prices are too volatile to accord with efficient markets has received some econometric support in papers by Stephen LeRoy and Richard Porter on the stock market, and by myself on the bond market. To illustrate graphically why it seems that stock prices are too volatile, I have plotted in Figure 1 a stock price index p, with its ex post rational counterpart p* (data set 1).' The stock price index pt is the real Standard and Poor's Composite Stock Price Index (detrended by dividing by a factor proportional to the long-run exponential growth path) and p* is the present discounted value of the actual subsequent real dividends (also as a proportion of the same long-run growth factor).2 The analogous series for a modified Dow Jones Industrial Average appear in Figure 2 (data set 2). One is struck by the smoothness and stability of the ex post rational price series p* when compared with the actual price series. This behavior of p* is due to the fact that the present value relation relates p* to a long-weighted moving average of dividends (with weights corresponding to discount factors) and moving averages tend to smooth the series averaged. Moreover, while real dividends did vary over this sample period, they did not vary long enough or far enough to cause major movements in p*. For example, while one normally thinks of the Great Depression as a time when business was bad, real dividends were substantially below their long run exponential growth path (i.e., 10-25 percent below the