and-sell signals are purely subjective and cannot be determined by pre-
cise numerical rules.THE RANDOMNESS OF STOCK PRICES
Although the Dow theory might not be as popular as it once was, tech-
nical analysis is still alive and well. The idea that you can identify the
major trends in the market, riding bull markets while avoiding bear
markets, is still a fundamental pursuit of technical analysts.
Yet most economists still attack the fundamental tenet of the
chartists—that stock prices follow predictable patterns. To these aca-
demic researchers, the movements of prices in the market more closely
conform to a pattern called a random walkthan to trends that forecast fu-
ture returns.
The first to make this connection was Frederick MacCauley, an
economist in the early part of this century. His comments at a 1925 din-
ner meeting of the American Statistical Association on the topic of “fore-
casting security prices” were reported in the association’s official
journal:MacCauley observed that there was a striking similarity between the fluc-
tuations of the stock market and those of a chance curve which may be ob-
tained by throwing dice. Everyone will admit that the course of such a
purely chance curve cannot be predicted. If the stock market can be fore-
cast from a graph of its movements, it must be because of its difference
from the chance curve.^3More than 30 years later, Harry Roberts, a professor at the Univer-
sity of Chicago, simulated movements in the market by plotting price
changes that resulted from completely random events, such as flips of a
coin. These simulations looked like the charts of actual stock prices,
forming shapes and following trends that are considered by chartists to
be significant predictors of future returns. But since the next period’s
price change was, by construction, a completely random event, such pat-
terns could not logically have any predictive content. This early research
supported the belief that the apparent patterns in past stock prices were
the result of completely random movements.
But does the randomness of stock prices make economic sense?
Factors influencing supply and demand do not occur randomly and areCHAPTER 17 Technical Analysis and Investing with the Trend 291
(^3) Journal of the American Statistical Association, vol. 20 (June 1925), p. 248. Comments made at the Al-
dine Club in New York on April 17, 1925.