By Robert E. Weintraub

Journal of Finance, Volume 18, Issue 1 (Mar., 1963)

Abstract

Do speculative price movements follow some systematic and hence predictable pattern or class of patterns? Can information about tomorrow's price movements be gained by studying today's price data? Professional traders or speculators would have to answer these questions in the affirmative, for they operate as if this period's price movements can be used to their advantage. But today it is respectable, if not fashionable, for the professional economist and statistician to answer these questions in the negative. In the past few years a number of academic papers have concluded that speculative price movements are random walks.

How should we interpret the claim that speculative prices move randomly? Following Hendrik S. Houthakker, randomness in speculative price data may be defined negatively as the "absence of any systematic pattern". Alternatively, following Harry V. Roberts, the random-walk hypothesis may be taken to mean that speculative price movements "behave very much as if they had been generated by an extremely simple chance model". Whether randomness is defined negatively following Houthakker or positively with Roberts, the implication of the random-walk hypothesis is the the speculators cannot know what they are doing when they study this period's price movements by way of seeking some sign of the next period's movements. I shall argue that the random-walk hypothesis flies in the face of common sense and the facts, and, moreover suggests a degree of naivite on the part of its advocates as to the rules of the game which professional speculators are playing.