Do Asian Stock Markets Follow Martingales? Evidence from Spectral Shape Tests

The martingale hypothesis is examined for 5 Asian stock markets using the spectral shape tests of Durlauf (1991). Unlike the variance ratio test used in prior studies, the spectral shape tests are consistent against all stationary alternatives to the martingale null. With daily data, both tests reje...

Full description

Saved in:
Bibliographic Details
Main Authors: Koh, Seng Kee, Benedict, Fong, W. M.
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 1994
Subjects:
Online Access:https://ink.library.smu.edu.sg/lkcsb_research/2187
https://doi.org/10.1007/BF01739207
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
Description
Summary:The martingale hypothesis is examined for 5 Asian stock markets using the spectral shape tests of Durlauf (1991). Unlike the variance ratio test used in prior studies, the spectral shape tests are consistent against all stationary alternatives to the martingale null. With daily data, both tests reject the null for all 5 markets: Thailand, Hong Kong, Korea, Malaysia, and Taiwan. With weekly data, the null is rejected for Thailand and Malaysia. As expected, the spectral shape tests yield stronger rejections of the null than the variance ratio test. Departures from the martingale cannot be entirely attributed to problems caused by infrequent trading. The use of value-weighted indexes means that only the most actively traded stocks are included in the market portfolios. Plausible models of infrequent trading also suggest geometrically declining returns autocorrelations past lag one. The persistent and mainly positive autocorrelations found at distant lags may mean returns have long memory.