۱- Introduction

۲- Relation to the literature

۳- Methodology

۴- Data

۵- Results

۶- Conclusions and directions for further research

References

Abstract

Using a data set of German stocks that includes the financial crisis, this paper identifies market liquidity as the main driver of return seasonality. In comparison, the economic significance of order flow imbalance is markedly weaker. Applying panel regressions and controlling for unobserved effects, we investigate the effects of both variables simultaneously, together with dummies for calendar effects. US macroeconomic news announcements, which have been identified as one driver of return seasonality in previous studies using non-US data, are of little importance for our data set of German stocks.

Introduction

More than four decades ago, Fama (1970) published his seminal paper on efficient capital markets. Many empirical papers that appeared in the 40? years since then described systematic deviations from the Efficient Markets Hypothesis, often referred to as anomalies. Whereas many such anomalies vanished shortly after their publication, others still persist (for an overview, see Zacks 2011). In particular, seasonal patterns with stock returns on specific days being systematically higher/ lower than those on other days (e.g., the so-called turn-of-the-month effect) show remarkable persistence over time. This type of seasonality in returns is welldocumented for several countries, and it has been existing for more than 20 years (Liu 2013). The literature considered three main potential reasons for return seasonality: order flow (or order imbalance), market liquidity, and announcements of macroeconomic news (see, e.g., Zwergel 2010). The present paper documents seasonalities in returns on German stocks and tests these three potential drivers. Using a fixed-effects panel regression methodology, we investigate order imbalance and market liquidity simultaneously, controlling for unobserved effects. Previous studies have focused on only one of these explanations at a time. In addition, our focus on German stocks yields insights into a market which is relevant at the international level, but which has not yet been investigated to the same extent as, e.g., US markets. The paper contributes to the empirical literature in several ways: First, we document a relation between daily liquidity and return patterns of individual stocks, which provides evidence for market liquidity to play an important role for return seasonality. Second, in contrast to other studies, which analyze liquidity considerations of select groups of market participants, we find a link between aggregate order (flow) imbalance and return seasonality. However, despite its statistical significance, the effect of order imbalance on return seasonality is found to be negligible in economic terms. Third, whereas US macroeconomic news announcements have been documented to be an important driver of return seasonality in stock indices also outside the US (Nikkinen et al. 2007a, 2009), we find no significant effects of these announcements on return seasonality in individual German stocks.