The building of macroeconomic models has been a traditional task for econometricians but it seems to have been receiving less attention in recent years than was so historically. This is surprising, given that data is better now and more plentiful and computing is much easier. To build such a model one has to have a clear-cut viewpoint about specification in general, and then to test between a few alternatives, and to consider smaller systems of equations. The old-fashioned problems of dealing with large-scale systems, modeled jointly, and having estimation difficulties is now replaced by smaller, much more dynamic and possibly non-linear systems. In this framework, institutions, facts about local knowledge, relevant economic theory, and some non-linearity, is easier to include. This book gives an excellent insight into the process of forming a high-quality, real-world macro-model then evaluating and using it. This is an important topic and makes this an important book.

Clive W.J. Granger

Research Professor

Nobel Laureate