Temporal Aggregation and Ordinary Least Squares Estimation of Cointegrating Regressions

The paper derives the asymptotic distribution of the ordinary least squares estimator of cointegrating vectors with temporally aggregated time series. It is shown, that temporal aggregation reduces the bias and variance of the estimator for average sampling (temporal aggregation of flow series) and does not affect the limiting distribution for systematic sampling (temporal aggregation of stock series). A Monte Carlo experiment shows the consistency of the finite sample results with the asymptotic theory.