Identifying periods of market inefficiency for return predictability

ABSTRACT

The article examines the efficiency of 31 stock index series spanning 26 countries across the world, using generalized spectral test (GST) and detects departure from the martingale difference hypothesis (MDH). A moving window of 24 months was used and p-values of GST were estimated. In order to explore whether the departure from market efficiency can be used for generating profitable trades, an exponentially weighted-moving-average-based trading rule was applied and was found that average profits per trade were significantly higher when p-value of the GST was less than 0.1. These observations are in consistent with the adapted market hypothesis.

KEYWORDS:

• Market efficiency
• return predictability
• generalized spectral test
• stock trading
• moving averages

JEL CLASSIFICATION:

• G14
• G17
• G19

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Stock Index used in the study: Argentina (Merval), Australia (AORD, S&PASX200), Austria (ATX), Belgium (BFX20), Brazil (BOVESPA), Canada (GSTPSE), China (SSECI), Europe (Euro Stoxx), France (CAC40), Germany (GDAXI), Greece (GDAT), Helsinki (OMXH25), Hong Kong (HSI), India (BSESN), Indonesia (JKSE), Ireland (ISEQ), Istanbul (XU100.IS), Japan (Nikkei225), Malaysia (KLSE), Mexico (IPC), Singapore (STI), South Korean (KS11), Spain (IBEX), Switzerland (SSMI), Taiwan (TWII), United Kingdom (FTSE), United States (GSPC, DJIA, NASDAQ, NYSE).