A Monte Carlo study of the effect of design characteristics on the
inequality restricted maximum entropy estimator
Randall C. Campbell
Department of Finance and Economics, Mississippi State University, Mississippi State, MS 39762, USA
R. Carter Hill
Department of Economics, Louisiana State University, Baton Rouge, LA 70803, USA
In this paper, we conduct a set of Monte Carlo sampling experiments to examine the effect of design characteristics on the inequality restricted maximum entropy (RME) estimator. We generate data under varying design characteristics, and estimate the parameters using maximum entropy and least squares estimation, both with and without parameter inequality restrictions. As part of the experimental design, we vary the distribution of the regressors, the distribution of the errors, the degree of collinearity, and the signal-to-noise ratio. We compare the alternative estimators on the basis of mean square error.
The link below contain sample GAUSS programs to carry out some of the described estimations: gme3m13.gau, rme3m121.gau, cme3m225.gau