Testing calendar effects of international equity and real estate markets

Abstract

There are a lot of previous studies on calendar effects. However, most of them use traditional methods like regression. Hui et al. Habitat International 48, 38–45, (2015b) incorporated Shiryaev-Zhou index with logistic regression to study the Halloween and January effects of eight securitized real estate markets, but they fixed the moving-window size to be 130 days. How the change in moving-window size affects the calendar effects cannot be seen. In this study, we also apply the Shiryaev-Zhou index, but we allow the moving-window size to vary. Furthermore, we incorporated Shiryaev-Zhou index with analysis of mean (ANOM) and logistic regression to examine calendar effects of general equity and securitized real estate indices of Hong Kong, Japan, US, UK, France and Germany during the period 1996 – 2014. The results show that our new methods can detect additional channels of significant calendar effects of which normal methods fail to show. Furthermore, the general equity indices show significant Halloween and January effects. However, for the securitized real estate indices, the Halloween and January effects are less significant or even go into reverse in some cases. This study has two main implications. Firstly, investors can formulate a better trading strategy to earn more profits. Secondly, trends and phenomena found in equity markets may not be applicable to real estate markets, so investment rules on equity markets may not work on real estate markets.