We propose a general estimation principle based on the assumption that instrumental variables (IV) do not explain the error term in a structural equation. The estimators based on the principle is inde- pendent of the normalization constraint, unlike the IV estimators.

In this paper, we study maximum likelihood estimation and Lagrange multiplier testing of a one-way error components regression model suitable for incomplete panel and including parametrically specified variance functions for both individual-specific and general error disturbances.

This paper proposes a range of parametric and nonparametric corrections to the GPH estimator of the differencing parameter in a fractionally integrated process. It is argued that the GPH estimator can suffer from an identification problem at the very low frequencies in the spectrum in small samples.

Indirect estimation methods are proposed for estimating univariate ARFIMA , as well as more complex multivariate VARFIMA models. Special attention is given to comparing the finite sampling properties of the indirect estimator with Sowell's (1992a) exact time domain maximum likelihood estimator...

The estimation of the noncentrality parameter of a chi-squared distribution, although simple to state, leads to difficulties, both for frequentist and Bayesian inferences. We propose in this paper a family of admissible improper Bayes estimators and study the minimax behavior of these estimators...

In this paper we compare via Monte Carlo experiments some methods to estimate the parameter of long-range dependence. We then apply these procedures to a financial time series to investigate its long-memory properties. The evidence of smooth long-range dependence makes the usual Efficiency...

The smoothed maximum score estimator of the coefficient vector of a binary response model is consistent and asymptotically normal under weak distributional assumptions. However, the differences between the true and nominal levels of tests based on smoothed maximum score estimates can be very...

We consider truncated processes, both in discrete and continuous time, and study their dynamic properties. When the underlying process is a diffusion process, we derive the infinitesimal generator of its truncated counterpart. This result is the basis for the estimation of the drift and...

This paper takes a computationnaly simple LS approach to develop a more efficient estimation procedure, which we call Residual Augmented Least Square (RALS), than OLS when the errors are not normally distributed. The efficiency gain is from manipulating the higher moment conditions implied by...

In this paper we examine educational data whch has a cross-classified structure. A cross-classified value-added multilevel model is proposed for these data and the problem of estimation are discussed in relation to the probllem of an endogenous regressor.