Statistical analysis of the overnight and daytime return

TL;DR

This study analyzes the separate overnight and daytime returns and volatilities of NYSE stocks over 20 years, revealing their statistical properties, correlations, and contributions to total daily returns.

Contribution

It provides a comprehensive statistical analysis of overnight and daytime returns, highlighting their similarities, differences, and stability over time, which was not extensively studied before.

Findings

Both return components exhibit power-law tail distributions.

Long-term memory exists in volatilities but not in returns.

Daytime return has a greater impact on total return.

Abstract

We investigate the two components of the total daily return (close-to-close), the overnight return (close-to-open) and the daytime return (open-to-close), as well as the corresponding volatilities of the 2215 NYSE stocks from 1988 to 2007. The tail distribution of the volatility, the long-term memory in the sequence, and the cross-correlation between different returns are analyzed. Our results suggest that: (i) The two component returns and volatilities have similar features as that of the total return and volatility. The tail distribution follows a power law for all volatilities, and long-term correlations exist in the volatility sequences but not in the return sequences. (ii) The daytime return contributes more to the total return. Both the tail distribution and the long-term memory of the daytime volatility are more similar to that of the total volatility, compared to the overnight records. In addition, the cross-correlation between the daytime return and the total return is also stronger. (iii) The two component returns tend to be anti-correlated. Moreover, we find that the cross-correlations between the three different returns (total, overnight, and daytime) are quite stable over the entire 20-year period.