Comparison of Box-Jenkins and Bonn Monetary Model Predition Performance

ISBN-13: 9783540100119 / Angielski / Miękka / 1980 / 146 str.

M.N. Bhattacharyya

Opis

The purpose of the present study is to evaluate the predictive performance of an earlier version of the Bonn .. Monetary Hodel ("Ein Okonometrisches Vierteljahresmodell des Geld- und Kreditsektors fur die Bundesrepublik Deutsch- land", by Jorn Martiensen (1975), Heisenheim, Verlag Anton Hain) against the benchmark provided by the Box-Jenkins univariate autoregressive-integrated moving average, ARI~ffi, model. Similar studies aimed at evaluating the predictive performance of the econometric models of the Horti, American economies :,ad been reported earlier. But to the best of my knowledge, no suc;, works, at least up to the time, early 1976 when I took up the present investigation, had been done with any of the European econometric mbdels, except Prothero and Wallis (1976), discussed in the introduction to the present book. The previous studies of this type generated plenty of unconstructive and sometimes unpleasant dialogues. Neither do I appreciate the over-reaction of the eco- metricians when an econometric model is evaluated nor do I appreciate when it is said that econometrics is withering. According to my opinion, econometric model building is a great experimentation and univariate time series models are not expected to be substitutes for econometric models.

Spis treści

1. Introduction.- 2. Bonn econometric model of German economy.- 3. ARIMA models for fifteen endogenous variables of the BNM model.- 4. Analysis of sample period lead 1 forecast errors.- 5. Bates-Granger composite forecast and its application in evaluating econometric model.- 6. Analysis of post-sample lead 1 forecast errors.- 7. Causal relationships between selected economic variables.- 7.1 Granger’s definition of causality and its characterization.- 7.2 Detection of causality: Pierce’s broad tests.- 7.2.1 Testing the independence of two series x and y: ?(uv;k) = 0 for all k.- 7.2.2 Assessment of different types of causality.- 7.2.3 Confirmation of unidirectional causality.- 7.3 Causal relationships between the selected monetary variables of the BNM model.- 7.4 Progressive ?2 tests for detecting causality.- 7.5 Causal relationships between short-term interest rate, 90-day money rate Frankfurt, and other selected variables.- Glossary of abbreviations used in BNM model.