The purpose of this paper is to examine the day‐of‐the‐week effect for three primary money market instruments in Canada. The sample period is 1980‐2009.
The authors use three approaches. First, a parametric t‐test is employed to determine if a particular day‐of‐the‐week mean return is significantly different from zero, using both a full sample and a trimmed sample. Next, the Wilcoxon signed ranked test is utilized to assess whether the median weekday return is different from zero for each day. Lastly, a binary regression model is used to test if Monday's mean return is different from other days.
The traditional Monday effect is prevalent in the 1980s for corporate paper and treasury bills (TB), but not for bankers acceptances (BA). In the 1990s, the Monday effect disappears completely. However, in the 2000s the Monday effect reappears, but is positive (it reverses) for both corporate paper and BA. The authors also find strong support for Wednesday being a high return day, which concurs with related money market studies.
While the results are statistically significant, the economic significance is dubious. This study helps market participants in that it shows that they need to allow for distinct day‐of‐the‐week patterns when using yield spreads.
One practical implication for practitioners is to time purchases of Canadian money market securities for Monday when returns are low (relying on the results of the full sample period). Issuers should time sales for non‐Mondays when returns are higher and yields are lower.
This study is original in that it is the first one to analyze day‐of‐the‐week effects in the Canadian money market. The authors compare the results to studies that focus on the US market.
Emerald Group Publishing Limited
Copyright © 2011, Emerald Group Publishing Limited