The estimator of the coefficient covariance matrix proposed in White (1982) can be used to robustify the classical Wald test. Sampling experiments recently performed on linear regressions and simultaneous equation models, however, suggest that such an estimator tends to underestimate the...

Through Monte Carlo experiments, this paper compares the performances of different gradient optimization algorithms, when performing full information maximum likelihood (FIML) estimation of econometric models. Different matrices are used (Hessian, outer products matrix, GLS-type matrix, as well...

At tbe IBM Pisa Scientific Center an interactive package has been developed under CP-67/CMS, which is particularly helpful when the data to be processed are time series. The interactive facilities of the operating system CP-67/CMS are strenghtened in such a way as to allow an easy interactive...

DMS/2 (Decisional Models Solution, version 2) is a computer package for solution of nonlinear econometric models. This technical report describes the new features that improve over the DMS-package.

The autoregressive conditional heteroskedasticity (ARCH) estimation procedure provides a specification of the error terms as well as estimates of the coefficients. A simple interest rate equation is estimated using least squares and also using ARCH. Then the stochastic simulation methodology is...

Several methods have been proposed in the last few years for evaluating uncertainty in forecasts produced by nonlinear econometric models. Some methods resort to Monte Carlo, while others resort to different simulation techniques. This work aims at comparing these methods by means of experiments...

When econometric models are used as forecasting tools, forecast errors can be decomposed into several components, one of which is due to estimation errors, while another one is due to the stochastic nature of the variables to be predicted. Conditional on model's specification and on the...

Most of the methods proposed in the literature for evaluating forecast uncertainty in econometric models need an estimate of the structural coefficiencs covariance matrix among input data. When estimation is performed with full information maximum likelihood, alternative estimators of such a...

Full information maximum likelihood estimation of econometric models, linear and nonlinear in variables, is performed by means of two gradient algorithms, using either the Hessian matrix or a computationally simpler approximation. In the first part of the paper, the behavior of the two methods...