How Lead-Lag Correlations Affect the Intraday Pattern of Collective Stock Dynamics

Properly estimating correlations and understanding how they change under different economic conditions plays a key role in asset pricing models, risk management, and many econometric models. In this paper we introduce a robust framework to identify a meaningful correlation relationship, address different types of correlations and their interplay, and address correlations across different time scales. First, we present a methodological framework to estimate synchronous, lagged, and autocorrelations for stock price return time series, and validate their statistical significance across different time horizons. Second, we explore the interplay between these different co-movement relationships, using a model to uncouple the factors contributing to the intraday pattern of contemporaneous correlations, including volatility, autocorrelations and lagged cross-correlations. Third, we use the methodological framework to investigate correlations between stocks traded on the New York Stock Exchange in the periods 2001-03 and 2011-13, and provide insights on how correlations and their dynamics have changed over time