《股指期货研究中外权威》由会员分享，可在线阅读，更多相关《股指期货研究中外权威（12页珍藏版）》请在金锄头文库上搜索。
1、Applied Financial Economics, 1996, 6, 259-269Permanent and temporary components of Canadian stock pricesAPOSTOLOS SERLETIS and MURRAY A. SONDERGARDDepartment of Economics, The University of Calgary, Calgary, Alberta T2N1N4Market efficiency in the Toronto Stock Exchange (TSE) is investigated using bo
2、th traditional tests of market efficiency, as well as tests devised in the late 1980s. In particular, unit root tests, mean reversion tests, as well as non-linearity and chaos tests are applied. The results are consistent with market efficiency.I. INTRODUCTIONThe hypothesis that stock markets are ef
3、ficient, and the notions connected with it, have recently provided the basis for a great deal of research in financial economics. Briefly stated, the hypothesis claims that asset prices are rationally related to economic realities and always incorporate all the information available to the market. T
4、his implies the ab- sence of exploitable opportunities for excess profits. Despite the widespread belief that markets are efficient, a number of studies have suggested that the prices of some assets are not rationally related to economic realities. For example. Summers (1986) argues that market valu
5、ations differ substantially and persistently from rational valuations and that existing evidence (based on common techniques) does not establish that financial markets are efficient. This study investigates market efficiency in Canadas lar- gest and best-known stock exchange, the Toronto Stock Excha
6、nge (TSE), The market efficiency of the TSE is exam- ined from a numher of different perspectives, including unit root tests, variance ratio tests and univariate long-horizon regressions, as well as non-linearity and chaos tests. The paper is organized as follows. Section II reviews the voluminous l
7、iterature with respect to the efficient markets hypothesis. Section III describes the data. Section IV tests for persistence in the TSE 300 Total Return Index using unit root tests (based upon Phillips and Perron, 1988 and Zivot and Andrews. 1992). Section V applies some of the more recently develop
8、ed tests to see whether the TSE 300 Total A slochastic process z, is a fair game if z, has the propertyReturn Index exhibits mean reverting behaviour (based on the work of Cochrane, 1988; Poterba and Summers, 1988: and Fama and French, 1988). Section VI tests (in the spirit ofSerletisand Dormaar. 19
9、96) for deterministic white chaos using the Brock, Dechert, and Scheinkman (1987) BDS nonparametric test for whiteness (independent and ident- ically distributed observations). Section VII discusses the implications of the results.II. THEORETICAL FOUNDATIONSTypically standard asset-pricing models im
10、ply the martin- gale model, according to which tomorrows price is expected to be the same as todays price. Symbolically, a stochastic process x, follows a martingale ifF (v Ci V (Wwhere Q, is the time t information set - assumed to include .X,. Equation 1 says that if .x, follows a martingale, the b
11、est forecast of x, i that could be constructed based on current information Q, would just equal x,. Alternatively, the mar- tingale model implies that x,+, x, is a fair game.(2)Clearly, x, is a martingale only if x,.n x, is a fair game. For this reason, fair games are sometimes called martingale dif
12、ferences.martingale process is a special case of the more general submartingale process. In particular, x, is a submartingale if it has the property ,(.v,41 Cl,) x,. Note that the submartingale is also a fair game where x,41 is expected to be greater than x,. In terms of the x,+ 1 X, process the sub
13、martingale model implies that ,(x,+1 x,)|i2, 0.096O-3t07 1996 Chapman all public information; and any infonnation, public as well as private. Clearly, strong-form efficiency implies semistrong-form efficiency, which, in turn, implies weak-form efficiency. However, the reverse implications do not fol
14、low, since a market easily could be weak-form efficient but not semistrong-form efficient or strong-form efficient.The martingale difference model is also less restrictive than the random-walk model - the forerunner of the theory of efficient capital markets. In particular, the martingale difference
15、 requires only independence of the conditional expectation of price changes from the available information, as risk neutrality implies. The more restrictive random-walk model, on the other hand, requires this, as well as indepen- dence involving the higher conditional moments (i.e., vari- ance, skew
16、ness, and kurtosis) of the probability distribution of price changes. By not requiring probabilistic indepen- denee between successive price changes, the martingale difference model is entirely consistent with the fact that changes in prices, although uncorrelated, tend not to be independent over time. Rather, they have clusters of volatil- ity and tranquillity (i.e., serial dependence in the higher conditional moments). This is a phenomenon originally noted for st
