Published online by Cambridge University Press: 15 September 2016
The performance of restricted estimators such as Almon and Shiller in modeling advertising carryover is tested and compared to the unrestricted OLS estimator, using 1971–1988 monthly New York City fluid milk market data. Results indicate that in the absence of autocorrelation and multicollinearity among the lagged advertising variables, the unrestricted OLS estimator is still the preferred estimator, based on Mean Square Error and Root Mean Square Percent Error criteria. In this case, the Almon and Shiller estimators perform equally well, although next only to the OLS estimator. In the presence of autocorrelation or multicollinearity however, the restricted estimators may outperform the OLS estimator, in a MSE sense, with the flexible Shiller estimator (which subsumes the Almon) being more desirable.