This paper examines the return spillovers between Exchange-Traded Funds (ETFs) and stocks. While traditional approaches focus on proportional relationships between ETFs and their underlying assets, we develop a high-dimensional network framework that captures spillover effects between any ETF-stock pair, regardless of their compositional relationship. By separating idiosyncratic and systematic risks, we investigate potential drivers of contagion. We document substantial heterogeneity in spillover patterns across sectors, which is previously unaddressed in the literature. Sectors such as Utilities and Real Estate exhibit robust spillovers to both their component stocks and assets in other sectors. Conversely, in sectors such as Consumer Discretionary and Finance , cross-sector influences dominate intra-sector ETF-constituent linkages. Our results also highlight that during periods of high market volatility, sources of idiosyncratic contagion become more diverse, suggesting the need for broader market surveillance beyond the few most influential ETFs.

This article proposes a new method for examining the impact on a firm’s investment of uncertainty reflected in its stock-return volatility. We simultaneously address the endogeneity of uncertainty and mismeasurement in Tobin’s Q , but earlier empirical work often neglects one of the two issues. Our nonparametric estimates further suggest that the relation between investment and uncertainty is significantly decreasing and strongly concave. This result contrasts with the existing literature that widely adopts linear regressions. Ignoring nonlinearity or measurement error in Q can lead to a substantial estimation bias. However, the bias due to the endogeneity of uncertainty is small.