Risk Behavior of Different Weekdays in NEPSE Index

Abstract

Background: In some societies some day of week is favored for financial transactions over other days. In this regard, in this research work, it is intended to study effect of different days of week in Nepal Stock Exchange (NEPSE) index.

Objective: This study aims to assess effect of different days of week on NEPSE index.

Materials and Methods: Data on closing indices of NEPSE, available online on official website of NEPSE are used for analysis purpose. Trading activities in NEPSE market runs on Sunday, Monday, Tuesday, Wednesday and Thursday and remains closed on Friday and Saturday. To observe the difference in NEPSE indices on different trading days at first ARIMAX model is applied with different weekdays as exogenous variable. Since the volatility in values are observed to be clustered, so GARCH model is implemented to describe variances. The means and variances predicted by the models are used to identify value-at-risk and expected-shortfall on different days of week and by observing risk in terms of these measures effect of different weekdays on NEPSE market is assessed.

Results: The optimum model resulted for Sunday as exogenous variable is ARIMAX (1, 1, 1) + GARCH (0, 2). For Monday, it is ARIMAX (3, 1, 3) + GARCH (0, 4). Similarly, for Tuesday, Wednesday and Thursday, ARIMAX (1, 1, 1) + GARCH (0, 5), ARIMAX (3, 1, 3) + GARCH (0, 2) and ARIMAX (1, 1, 1) + GARCH (2, 3), respectively are found as the optimum models. Next, values of value-at-risk on these days, calculated as 95% quantile of residuals of corresponding models are found to be 18.07, 17.86, 18.09, 17.86 and 21.86 on Sunday, Monday, Tuesday, Wednesday and Thursday, respectively. Similarly, expected shortfall on these respective days, calculated as mean of values below value-at-risk, are found to be 38.23, 38,29, 38.23, 38.28 and 38.74, respectively.

Conclusion: There is no noticeable effect of different days of week on NEPSE index when viewed with the aspect of value-at-risk and expected shortfall.