Abstract

The Random Walk Hypothesis for Bank and Industrial Shares

The object of the article is a twofold one: First, for 50 frequently traded shares the random walk hypothesis is repudiated with the help of the run test and by estimation of autocorrelation coefficients. Secondly, by applying cluster analysis it is shown that the shares of German banks differ clearly from those of other companies with respect to the randomness of price fluctuations. By examining the characteristics in which banks differ from industrial firms, the author comes to the conclusion that banks can, and actually do, exploit their position on the German money and capital market to influence the price trend of their own shares. This established influence potential can be utilized by the banks also to influence other securities - though with slight or temporally limited motivation. Hence it may be concluded that one cause for the lack of efficiency of German stock exchanges lies in the position of the banks.