Tests of capital market efficiencythis repetition of peaks and troughs; in fact there would be a very large number of us
who would notice it. As we try to sell at the peak, so would the others. Since few
potential buyers would be interested at the peak price, the price would drop. Realising
this, we should all try to sell earlier to try to beat the drop in price, which would sim-
ply cause it to occur still earlier. The logical conclusion of this is that the price would
not in fact ever rise to the peak. Expecting the trough to be reached and eager not to
miss it, we should be buying earlier and earlier, thus keeping the price up and ensur-
ing that the trough is never reached either.
The net result of all this is that if there are sufficient investors following past price
patterns and seeking to exploit repetitions of them, those repetitions simply will not
occur. In practice, the more likely price profile of that security would approximate to
the horizontal broken line shown in Figure 9.1.Figure 9.1
Graph of the daily
share price against
time for a
hypothetical
security
If this pattern were expected to occur, investors, by their buying and selling actions, would
cause the pattern not to occur.Weak-form efficiency test results
The first recorded discovery of randomness in a competitive market was by Bachelier
when he observed it as a characteristic of commodity prices on the Paris Bourse as
long ago as 1900. His discovery went somewhat unnoticed until interest in the topic
was rekindled some years later.
Kendall (1953), accepting the popular view of the day that LSE security prices move
in regular cycles, tried to identify the pattern, only to discover that there was none;
prices seemed to move randomly.
Efficiency and randomness imply that there should be no systematic correlation
between the price movement on one day and that on another. For example, it seems to
be believed by some observers that if the price of a security rises today then it is more