On improving Almon estimation in distributed lag models with multicollinearity

The Almon procedure is a standard method for estimating distributed lag models, but it is highly vulnerable to multicollinearity among lagged regressors, reducing estimation reliability. Although several biased Almon-type estimators have been proposed, they remain sensitive to strong correlations, limiting their effectiveness. This study introduces a new Almon-type biased estimator specifically designed to improve robustness under multicollinearity. We derive the necessary and sufficient conditions under which the proposed estimator surpasses existing alternatives using the matrix mean squared error criterion. Monte Carlo simulations and two empirical applications confirm its superiority, showing consistently lower mean squared error values.