A general framework for using Monte Carlo methods in dynamic systems is provided and its wide applications indicated. Under this framework, several currently available techniques are studied and generalized to accommodate more complex features. All of these methods are partial combinations of three ingredients: importance sampling and resampling, rejection sampling, and Markov chain iterations. We deliver a guideline on how they should be used and under what circumstance each method is most suitable. Through the analysis of differences and connections, we consolidate these methods into a generic algorithm by combining desirable features. In addition, we propose a general use of Rao-Blackwellization to improve performances. Examples from econometrics and engineering are presented to demonstrate the importance of Rao-Blackwellization and to compare different Monte Carlo procedures. Keywords: Blind deconvolution; Bootstrap filter; Gibbs sampling; Hidden Markov model; Kalman filter; Markov...

This paper looks at the problem of performing likelihood inference for limited dependent processes. Throughout we use simulation to carry out either classical inference through a simulated score method (simulated EM algorithm) or Bayesian analysis. A common theme is to develop computationally robust methods which are likely to perform well for any time series problem. The central tools we use to deal with the time series dimension of the models are the scan sampler and the simulation signal smoother. Keywords: Disequilibrium models, Markov chain Monte Carlo, Scan sampler, Tobit model. 1. INTRODUCTION 1.1. General problem Suppose s 2 R np is jointly Gaussian with a density f S (s; #) indexed by a finite dimensional parameter # . The task is to use likelihood methods to infer about # in problems where there is a form of time dependence in s. The difficulty will be that we observe only a coarse version of s, h(s), which we call y, where the support of y is a proper subset of the support of s. Recently, there has been extensive work in the econometric literature on the use of simulation to estimate these types of coarse models in contexts where the time dimension n is typically small. This very interesting literature is reviewed by Hajivassiliou and Ruud (1994). In this paper we will focus on time series problems. We take n as the length of the p- dimensional time series and we suppose that n is typically large compared with p. Our desire is to develop methods which deliver likelihood-based estimators in O(n) computations. We will be able to carry this out by exploiting the common structure of time series models. To focus ideas we will write down a general Gaussian model for s, which will be assumed to follow a Gaussian state space form (see, for example, Harvey (1...

In this paper three models for data generated by different linear regression distributions with Gaussian errors are discussed: Finite mixture models with random and fixed covariates and a fixed partition model. The models are compared with respect to the adequacy of their assumptions for various data situations. The interpretation of parameters is discussed. The emphasis is put on the identifiability of the parameters. Identifiability is a necessary condition for the existence of consistent estimators. It turns out that the models treated here cause other identifiability problems than simple Gaussian mixtures. This was ignored up to now and thus there are no satisfying consistency proofs in this area. Counterexamples and sufficient conditions for identifiability are given. The identifiability concept is used for fixed partition models for the first time in this paper. The concept is generalized to "partial identifiability", i.e. identifiability of only a part of the parameters. 1. INTR...

This paper has not undergone the review accorded official NBER publications; in particular, it has not been submitted for approval by the Board of Directors. This research was supported by a grant to NBER from NSF—RANN (Grant No. APR76—025ll). The views expressed herein do not necessarily reflect those of the National Science Foundation. The authors are indebted to George Kaufman, Sherman Maisel,

This paper proposes a general empirical framework to estimate the protection-for-sale model, where the protection regime shifts according to a sector’s market structure (perfectly or monopolistically competitive). We base the protection structure on Grossman and Helpman (1994) for the subset of perfectly competitive sectors and on Chang (2005) for the subset of monopolistically competitive sectors. The two protection regimes are simultaneously estimated with joint constraints. The results of the J-test consistently reject the homogeneous (perfect competition) protection-for-sale model often adopted in previous literature and suggest a direction of improvement toward the proposed heterogeneous protection structure model.

Credit rationing is by many Keynesian economists nowadays seen as one of the most important examples of market failure in a modern capitalistic economy. Credit rationing occurs if in equilibrium the demand for loans exceeds the supply at the ruling price (interest rate). Since (Belgian) SMEs heavily rely on bank finance, this restricted access to bank finance could have a negative effect on the investment in new profitable projects and growth of the economy. Due to the lack of empirical research in this domain, we study the empirical significance of credit rationing for SMEs in a Belgian (bank based) context for the period 1993-2001. The novelty of our research consists of the distinction we make between credit rationing for long and short term bank credit. Moreover, in contrast to the numerous previous studies, pioneered by Fazzari et al. (1988), we opt for an endogenous classification into ‘credit rationed ’ and ‘non credit rationed’ firms, allowing for switching between both groups throughout time. An extensive panel data set consisting of 2.698 SMEs reporting data over the period 1993-2001 is used to estimate the demand-supply disequilibrium model for long and short term bank debt. This is done by means of the instrumental variable technique 3 Stage Least Squares (3SLS).

This paper proposes a general empirical framework to estimate the protection-for-sale model, where the protection regime shifts according to a sector’s market structure (perfectly or monopolistically competitive). We base the protection structure on Grossman and Helpman (1994) for the subset of perfectly competitive sectors and on Chang (2005) for the subset of monopolistically competitive sectors. The two protection regimes are simultaneously estimated with joint constraints. The results of the J-test consistently reject the homogeneous (perfect competition) protection-for-sale model often adopted in previous literature and suggest a direction of improvement toward the proposed heterogeneous protection structure model.

part of this publication may be reproduced, stored in a retreival system, or transmitted in any form or by any means, electronic, mechanical, photocopy, recording or otherwise, without the prior written consat of the publisher. International Standard Book Number: O-88410-268-8