Model Reliability

This book addresses an often neglected aspect of econometrics, the question of how to assess the specification, strengths, weaknesses, limits, and sensitive features of a model. The contributions are the result of a five-year inter-university research project to improve understanding of concepts of model reliability.Contents: Introduction, David A. Belsley and Edwin Kuh. Wholesale and Retail Prices: Bivariate Time Series Modeling with Forecastable Error Variables, C.W.J. Granger, R.P. Robins, and Robert F. Engle. Recursive Stability Analysis: The Demand for Money during the German Hyperinflation, Jean-Marie Dufour. A Bayesian Analysis of the Determinant of Inflation, Edward E. Learner. The Use of Outside Information in Econometric Forecasting, Mark N. Greene, E. Philip Howrey, and Saul H. Hymans. Centering, the Constant, First Differencing, and Assessing Conditioning, David A. Belsley. Applications of Bounded-Influence and Diagnostic Methods in Energy Modeling, Roy Welsch and Stephen Swartz. A Comparison of the Michigan and Fair Models: Further Results, Ray C. Fair and Lewis S. Alexander. Linear Analysis of Large Nonlinear Models and Model Simplification, Edwin Kuh, John Neese, and Peter Hollinger.Edwin Kuh is Professor of Management and Economics at MIT and director of MIT's Center for Computational Research in Economics and Management Science. David Belsley is Professor of Economics at Boston College and Visiting Professor at the MIT Center.