A Short-Run Forecasting Model of the United States Economy,
D. C. Heath and Co., 1971.
pfd files: whole book (10,866KB), toc,preface,bib,index (484KB), chapter 1 (806KB), chapter 2 (583KB), chapter 3 (717KB), chapter 4 (397KB), chapter 5 (240KB), chapter 6 (503KB), chapter 7 (92KB), chapter 8 (790KB), chapter 9 (884KB), chapter 10 (660KB), chapter 11 (1,274KB), chapter 11 (1,463KB), chapter 11 (1,126KB), chapter 11 (487KB), chapter 11 (353KB), appendix A (208KB). appendix B (168KB).
The monthly housing starts sector (Chapter 8) is estimated by one of the "disequilibrium" econometric techniques in 1972#1. The theory for this sector is reviewed in 1972#2 within the context of the literature.
The price equation (Chapter 10) is also discussed in 1970#2. The demand for employment equation (Chapter 9) is based on the work in 1969#1. The estimation technique that is used for the money GNP sector, namely 2SLS with account taken of serially correlated errors, (Chapter 2) is discussed in 1970#1.
At the time the forecasting model was developed there were many more computational constraints than there are now. The forecasting model had the advantage that it was small enough to allow many experiments to be performed using it that were not feasible to do using the larger models. The money GNP sector was particularly useful because it was linear and yet incorporated two important econometric problems: simultaneity and serially correlated errors. FIML estimates of the money GNP sector were obtained in 1973#1. The money GNP sector was then estimated by ten different estimators in 1973#2. The entire forecasting model except for the monthly housing starts sector was estimated by a variety of robust estimation methods in 1974#4. Also, the entire forecasting model except for the monthly housing starts sector was used in 1974#1 in the solution of some optimal control problems.
The accuracy of ex ante and ex post forecasts from the model is examined in 1974#3. The final use I made of the model is in Chapter 8 and Appendix B in 1976#1, where it is used for comparison purposes with the newer model.