Determination of Market Efficiency, Volatility and Asymmetric Effect Using Time Series Models at Nairobi Stock Market

Stock market efficiency is an important concept, especially in a growing economy like the Kenyan one. This study empirically determines the form of market efficiency, time-varying volatility effect, and asymmetric or leverage effect of the Nairobi Stock Exchange (NSE) market. Secondary data at the daily NSE 20-share index for the period spanning from January 2001 to December 2010 was used. The market efficiency was determined basing on unit root tests, ADF test, PP test and the non-parametric Runs test. The results indicated that stock returns follow an ARMA (2,1) stochastic process with significant positive serial correlation. The ADF test and PP test clearly gave evidence that the NSE index were non-stationary (random) at level and stationary (non-random) for the first and second differences. This implies that the NSE market is informationally efficient at the weak-form level. Using the non-parametric run test, results clearly displayed NSE market is weak-form efficient. EGARCH and TGARCH model with student’s t-distribution were the best models to capture asymmetric effect of stock returns. Since (γ₁) was positive, it indicated significance evidence for asymmetry in stock returns thus presence of leverage effect (good news had a higher impact on volatility than bad news).