On the Correlation Structure of Microstructure Noise: A Financial Economic Approach

We introduce the financial economics of market microstructure to the financial econometrics of asset return volatility estimation. In particular, we derive the cross-correlation function between latent returns and market microstructure noise in several leading microstructure environments. We propose and illustrate several corresponding theory-inspired volatility estimators, which we apply to stock and oil prices. Our analysis and results are useful for assessing the validity of the frequently assumed independence of latent price and microstructure noise, for explaining observed cross-correlation patterns, for predicting as-yet undiscovered patterns, and most importantly, for promoting improved microstructure-based volatility empirics and improved empirical microstructure studies. Simultaneously and conversely, our analysis is far from the last word on the subject, as it is based on stylized benchmark models; it comes with a "call to action" for development and use of richer microstructure models in volatility estimation and beyond.