THE RETURNS ON THE SINGAPORE STOCK MARKET : SEASONAL PATTERNS, OUTLIERS OR INVALID TESTS?

Abstract

Seasonal variations in mean returns have been the centre of focus in many studies of late. These studies have reported evidence of seasonal patterns in returns of the stock markets explored. Due to such market anomalies, the paradigm of market efficiency is frowned upon, as academics believe that investors could increase their expected returns by altering the timing of trades according to the seasonal patterns. In this paper, we base our analysis on the Singapore stock market to explore the existence of market seasonality. We use some tests that were employed in previous studies made on the day of the week effect and the month of the year effect. We also investigate the existence of the well-known Monday effect and January effect that have been discovered in many markets. Where such effects are discovered in our data, the study goes one step further to query if these effects are actually persistent, or if they are the results of outliers that bias the mean returns. Another possibility that may support the evidence of the market anomaly could be that results are obtained from employing invalid tests, where the required assumption of normality for those tests are violated due to non-normal rate of returns. In order to determine if the seasonal effects detected in our study are caused by the biasing powers of outliers in the data set, our analysis proceeds to censor the highest and lowest 5% of the returns in the data, defining those observations as the "outliers" in the data. It is then possible to examine if the outliers are responsible for the non-normal rate of stock returns, as well as the apparent seasonal patterns that are observed. This study provides conclusions with regards to any seasonality in the Singapore stock market. It attempts to distinguish between "genuine" market patterns from seasonality detected due to the influence of outliers, or due to employing invalid tests.