Testing Martingale Hypothesis Using Variance Ratio Tests: Evidence from High-frequency Data of NCDEX Soya Bean Futures

Abstract

The aim of the article is to examine the martingale hypothesis of market efficiency on high-frequency data of the soya bean futures traded in National Commodity and Derivatives Exchange (NCDEX) of India using multiple variance ratio (VR) tests from February 2015 to August 2015. The study employs high-frequency future prices of 5, 10, 15, 30 and 60 min time intervals mainly to decipher the efficiency of processing information by soya bean traders during intraday sessions of futures trading. The results of VR tests confirm that except prices of 5 and 10 min intervals which displays weak form of market efficiency, all other samples follow martingale hypothesis. The findings suggest that as information gets absorbed promptly in the intraday NCDEX soya bean futures prices, there exists fairly less opportunities to explore any trading strategy for profitable outcomes in the soya bean futures market in India.