Temporal aggregation and the finite sample performance of spectral regression estimators in cointegrated systems - A simulation study

The finite sample performance of spectral regression estimators in temporal ly aggregated cointegrated systems is investigated via the use of simulatio n experiments. The simulations address issues such as "optimal" choice of b andwidth parameter and effects of smoothing kernel in constructing estimate s of spectral densities that are used by the spectral regression estimators ; the effects of stock and flow variables and mixtures of the two, includin g the relative finite sample efficiency of the estimators under different c ombinations of stock and flow variables; and the effects of conducting iter ations of the spectral estimators. A striking feature of the results is the crucial role that correct choice of bandwidth and kernel function plays in producing accurate estimates of the unknown parameters. Furthermore, estim ates obtained using flow data alone are found to be more efficient, in the sense of having smaller variance, than those obtained using stock data alon e or mixtures of stocks and flows, thereby confirming in finite samples the ir relative asymptotic properties.