Dynamic Panels, Cross Sectional Correlation, and Arbitrage in Equities Market

Abstract

This paper studies estimation in linear dynamic panel data models with multiple interactive effects when both N and T are large. We derive the bias term in an order p dynamic panel data model and the limiting distribution of the estimator. Simulation results show good finite sample properties of the estimator. An application to arbitrage in the U.S. equities market between 2000 and 2015 shows dynamic panel data model is an effective method to produce effective trading strategies.

Publication
Working paper
Date
Links