Model Selection for Monetary Policy Analysis: How Important is Empirical Validity?*

Abstract

We investigate the economic significance of trading off empirical validity of models against other desirable model properties, and the potential loss from assuming model uncertainty to be higher than justified and basing monetary policy on a relatively robust model, or on a suite of models. We find that differences in model specification and in estimates of key parameters across comparable models may entail widely different monetary policy and macroeconomic performance. Our results therefore caution against compromising the empirical validity of models when selecting a model for policy analysis. We also find that potential costs from basing monetary policies on the relatively robust model or on a suite of models, even when it contains the valid model by assumption, can be substantial. This suggests an important role for econometric modelling and evaluation in exploitation of available information to reduce model uncertainty as much as possible. Our investigation is based on three alternative econometric systems of wage and price inflation in Norway.Keywords: Model uncertainty; Economic significance; Econometric modelling. AbstractWe investigate the economic significance of trading off empirical validity of models against other desirable model properties, and the potential loss from assuming model uncertainty to be higher than justified and basing monetary policy on a relatively robust model, or on a suite of models. We find that differences in model specification and in estimates of key parameters across comparable models may entail widely different monetary policy and macroeconomic performance. Our results therefore caution against compromising the empirical validity of models when selecting a model for policy analysis. We also find that potential costs from basing monetary policies on the relatively robust model or on a suite of models, even when it contains the valid model by assumption, can be substantial. This suggests an important role for econometric modelling and evaluation in exploitation of available information to reduce model uncertainty as much as possible. Our investigation is based on three alternative econometric systems of wage and price inflation in Norway.

The Role of Wage Formation in Empirical Macroeconometric Models

“…Extension to the open economy does not interfere with the model typology while it is important for the relevance of the framework to operational models. The modelling framework presented has developed over a number of years, and it has been applied to different data sets, see e.g., Nymoen (1998,2014), Nymoen (2003,2009ab), Akram and Nymoen (2009), Bårdsen et al (2012).…”

“…However, because of extended equilibrium correction and, realistically, a not completely accelerationist curve, the consequences of Phillips curve specification are more about different sizes and persistence of effects than about complete independence. Akram and Nymoen (2009) showed how optimal monetary policy could be implemented in an econometric model of the Norwegian economy called the Norwegian Aggregate Model (NAM), see Bårdsen and Nymoen (2009a). They also analyzed how much the predicted economic outcome depended on the specification of the supply side of the model, in particular as WP-ECM or PCM.…”

“…In addition to the preferences about policy horizon and gradualism captured by ( 31), the optimal interest rate response is influenced by the user's choice of macroeconomic model, which can be thought of as determining β ε . Akram and Nymoen (2009) By the use of model simulation, the authors identified a trade-off between price and output stabilization for different ranges of policy horizons. Specifically, in the case of WP-PCM and PCM the trade-off was in the range of 0 to 8 quarters.…”