This paper is organized as follows: parameterization problems endemic to time series models are discussed in section one. New approaches to the parameterization problems are summarized and then applied to the problem of simulataneously estimating the length and the coefficients of a distributed lat regression model. The asymptotic properties of the new estimator of the distributed lag model (DLM) are examined in section two, and the small sample properties of the estimator are examined in section three using Mote Carlo experiments.