Posterior and Predictive Densities for Simultaneous Equation Models

I. The Simultaneous Equation Model.- I.1 The statistical model ; Notations.- I.2 The identification problem.- I.3 Some previous contributions.- I.4 Review of the contents.- I.5 Technical abstract.- II. Full Information Analysis of the Two-equation Model.- II.1 Notations.- II.2 The Drèze and Morales’ approach.- II.3 An alternative approach : structural analysis.- II.4 Reduced form analysis and prediction.- II.5 Some additional remarks.- III. Limited Information Analysis of the Simultaneous Equation Model.- III.1 Some posterior joint conditional densities on ? : non-informative prior density.- III.2 Posterior moments of the reduced form parameters : limited information prior density.- III.2.1 Natural conjugate prior density (m1=1).- III.2.2 Exact prior restrictions (m1=1).- III.2.3 Some extensions for m1 > 1.- III.3. A “Limited information” prediction.- Appendix I : The matrix-t density.- Appendix II : The row-diagonal conditional matricvariate-t density.- Appendix III : The technicalities of Chapter II.- Appendix IV : Hypergeometric series ; Their computations.- Appendix V : The programs of Chapter II.- Appendix VI : An integral identity (Dickey).- IV. Empirical Illustration : The Belgian Beef Market.- IV.1 Morales’ model.- IV.2 The information content of the identifying restrictions.- IV.3 Informative approach : full information analysis.- IV.4 Informative approach : limited information analysis.- V. Conclusions.- Appendix VII : The data.- Appendix VIII : The results.- References.