文档介绍：: .
Applied Financial Economics, 1996, 6
and that existing evidence (based on common techniques) to be the same as today's price. Symbolically, a stochastic
does not establish that financial markets are efficient. process x, follows a martingale if
This study investigates market efficiency in Canada's lar-
gest and best-known stock exchange, the Toronto Stock F (v \Ci \ V (W
Exchange (TSE), The market efficiency of the TSE is exam-
where Q, is the time t information set - assumed to include
ined from a numher of different perspectives, including unit
.X,. Equation 1 says that if .x, follows a martingale, the best
root tests, variance ratio tests and univariate long-horizon
forecast of x, ^ i that could be constructed based on current
regressions, as well as non-linearity and chaos tests.
information Q, would just equal x,. Alternatively, the mar-
The paper is organized as follows. Section II reviews the
tingale model implies that x,+, — x, is a fair game.^
voluminous literature with respect to the efficient markets
hypothesis. Section III describes the data. Section IV tests (2)
for persistence in the TSE 300 Total Return Index using unit
root tests (based upon Phillips and Perron, 1988 and Zivot Clearly, x, is a martingale only if x,.n — x, is a fair game.
and Andrews. 1992). Section V applies some of the more For this reason, fair games are sometimes called martingale
recently developed tests to see whether the TSE 300 Total differences.^
' A slochastic process z, is a fair game if z, has the prope