Abstract

Hasbrouck (2018) takes advantage of the fact that U.S. equity market data are timestamped to nanosecond precision, and explores models of price dynamics at resolutions sufficient to capture the reactions of the fastest agents. The paper therefore addresses the econometric analysis of multivariate time series models at sub-millisecond frequencies and relies on long distributed lag models to alleviate the computational complexity while still taking advantage of the inherent sparsity of price transitions.

I think that the fundamental econometric problem of interest in the paper pertains to the analysis of mixed frequency data. In my comment I focus on the connections with the MIDAS literature.

View Publication on Journal Site