This research addressed the two issues central theme is the persistence and stability of two calendar anomalies. The anomalies studied are the holiday and January effects. The study use daily returns on forty stocks over ten years from January 2, 1990 to December 22, 2000. The empirical research was conducted using t-test and f-test from performing of the linear regression and the GARCH (1,1) model. The result of the GARCH (1,1) model also finds that calendar anomalies are appear that estimates of the holiday and January effects in the BEJ. There was evidence in support of the persistence of the anomalies that investors have learned of their existence and have reduced their sectors through arbitrage.