The day-of-the-week effect and conditional volatility: Sensitivity of error distributional assumptions
We test for reliable evidence of the day-of-the-week effect on both the mean and volatility for the S&P/TSX Canadian return index. Unlike previous studies, we permit several specifications for the error distribution -- GARCH normal, Student's t, generalized error distribution, and double exponential distribution. Unlike other studies, we find that the day-of-the-week effect in both mean and conditional volatility is sensitive to the particular specification of the underlying distributions. We also find that using a regression analysis assuming a Student's t distribution is a better way to investigate this effect. Our evidence demonstrates the apparent fragility of previous empirical studies on calendar anomalies. Thus, our results serve as a warning that with financial data, the error distributional assumptions are critical to correctly identifying empirical regularities in the data.
Year of publication: |
2008
|
---|---|
Authors: | Baker, H. Kent ; Rahman, Abdul ; Saadi, Samir |
Published in: |
Review of Financial Economics. - Elsevier, ISSN 1058-3300. - Vol. 17.2008, 4, p. 280-295
|
Publisher: |
Elsevier |
Subject: | Day-of-the-week effect Volatility GARCH | Error distributional assumptions |
Saved in:
Saved in favorites
Similar items by subject
-
On the robustness of week-day effect to error distributional assumption : international evidence
Boubaker, Sabri, (2017)
- More ...
Similar items by person
-
Baker, H. Kent, (2008)
-
Baker, H. Kent, (2008)
-
Random walk and breaking trend in financial series : an economic critique of unit root tests
Rahman, Abdul H., (2008)
- More ...