Sims (1991) conjectured that the conditional maximum likelihood estimator of trend stationary models of macroeconomic time series will tend to place initial observations relatively far from the estimated trend line. This can be misleading if the entire sample has been generated by the same data generating process and there is nothing unusual about the initial observations. We use the extended Nelson Plosser data set to evaluate Sims's conjecture. We consider the weighted symmetric estimator developed by Park and Fuller (1993) as an alternative approach to this estimation problem.
Falk, Barry, "Estimating Trend Stationary Models of Homogeneously Generated Samples" (1995). Economic Staff Paper Series. 257.