Author: Clopper Almon

Publisher:

Published: 2017-07

Total Pages: 458

ISBN-13: 9781548489786

This book is a practical guide to building economic models both macroeconomic and multisectoral. It uses free software available from the Internet together with regularly updated databanks including the quarterly national accounts of the United States and other quarterly data. It assumes no prior acquaintance with econometrics or computer programming, but does assume a willingness to follow a mathematical argument. Much of the text has often been used in college teaching. The book begins with a very simple model that can be computed with a hand calculator or cell phone. The model has, however, a nonlinearity in the investment function and shows how a nonlinearity can lead to a model with a cycle which neither damps out nor explodes. It then moves to models built with real data in the framework of the U.S. National Income and Product Accounts and develops the important concept of identity-centered modeling. Once the identities are are working properly, the modeler can proceed to the estimation of equations by least-squares regressions. Gradually the reader is led to the construction of a model roughly comparable to those used by commercial economic forecasting firms. At every step of the way, the reader sees results of calculations with real data and is urged to estimate his or her own equations with real, up-to-date data. The second section explores making alternative forecasts with this model and devising optimal economic policies with its aid. It also explores the range of uncertainty of the forecasts due to known variability of the errors in its equations. A rather long chapter expounds some conventional econometric methods and applies them to real data to both illustrate them and evaluate their usefulness. Although most of the book uses equations which are linear in the parameters which must be estimated, some functions very useful for certain purposes in economics are non-linear in the parameters. A chapter is devoted to their estimation. Most of the book uses the relatively simple U.S. system of national accounts, but one chapter deals with modeling with the more complicated System of National Accounts used by most other countries. Like the economy itself, dynamic models such as those built here have a tendency to develop cycles. One chapter looks into the mathematical theory of why that is so. The third section enlarges the scope of the book to include multisectoral models. Although here the model actually built is a tiny one of 8-sectors with made-up data, the methods employed are exactly those of the Inforum models used in a number of countries around the world. Multi-sectoral models can use many of the same techniques used for macromodels, but there are some additional problems. One of these is the estimation of personal consumption functions where the demand for any one product depends not only on its price but those of all other products. One method which has worked better than others is explained and the results of its estimation in four countries are presented and compared. Another problem is the computation of product-to-product tables from those constructed by statistical agencies. The book does not explain Real Business Cycle models, nor Computable General Equilibrium models, nor Dynamic Stochastic General Equilibrium models, which are all, in the author's opinion, unrealistic fads, not serious ways of modeling an economy..