Realized correlation analysis of contagion

•Magnitude and significant correlations.•Existence of textbook asymmetries.•Volatility of the cost-of-carry model assumptions.•Lead-lag asymmetries for the cost-of-carry model.•Usefulness of heterogeneity and jumps by introducing two cost-of-carry HAR (COC) models.

This paper investigates the cross-market contagion between spot and futures US stock markets by examining the significance and properties (textbook and lead-lag asymmetries) of realized correlation, testing the assumptions of the cost-of-carry model, as well as testing the in-sample predictive significance of heterogeneity and jumps to realized correlation. Evidence from the US stock market suggests realized correlation can be very helpful analyzing contagion. There is strong evidence of statistically significant cross-market contagion in the US stock markets, when realized correlation is used as conditional correlation, across all methods employed. To the best of my knowledge, this paper is the first to nonparametrically analyze contagion based on realized correlation.