A Long run structural macroeconometric model of the UK

Abstract

A new modelling strategy that provides a practical approach to incorporating long-run structural relationships, suggested by economic theory, in an otherwise unrestricted VAR model is applied to construct a small quarterly macroeconometric model of the UK, estimated over 1965q1-1999q4 in nine variables: domestic and foreign outputs, prices and interest rates, oil prices, the nominal effective exchange rate, and real money balances. The aim is to develop a model with a transparent and theoretically coherent foundation. Tests of restrictions on the long-run relations of the model are presented. The dynamic properties of the model are discussed and monetary policy shocks identified.Over the past two decades, there has been a growing interest in developing macroeconomic models with transparent theoretical foundations and flexible dynamics that fit the historical time series data reasonably well. The modelling framework described in the present paper, along with the work of King et al. the first steps towards this aim. However, our work is distinguished from these earlier contributions in three respects. First, we develop a long-run framework suitable for modelling a small open macroeconomy like the UK. The models of King et al., Gali and Crowder et al. are closed economy models suitable for modelling an economy such as the US. The four-variable model of Mellander et al. is made open only through a terms of trade variable added to the three-variable consumption-investment-income model analysed by King et al. and does not have a rich enough structure for the analysis of many open economy macroeconomic problems of interest.Second, we describe a new strategy which provides a practical approach to incorporating the long-run structural relationships suggested by economic theory in an otherwise unrestricted vector autoregressive (VAR) model. Third, we employ new econometric techniques in the construction of the model and in the testing of the long-run properties predicted by the theory. The description of the modelling work not only provides one of the first examples of the use of these techniques in an applied context, but also includes a discussion of some bootstrap experiments designed to investigate the small-sample properties of the tests employed. Hence, the paper reinforces the arguments of the papers cited above by emphasising the (statistical and economic) importance of the long run in macroeconometric modelling. But it also provides a practical demonstration of

Modeling Monetary Transmission in Switzerland with a Structural Cointegrated VAR Model

“…4 A detailed description of the data can be found in the appendix. Brüggemann, 2003), Spain (Juselius and Toro, 2006) and the United Kingdom (Dhar, Pain and Thomas, 2000;Garratt, Lee, Pesaran and Shin, 2003). Baltensperger, Jordan and Savioz (2001) estimate a cointegrated VAR model for Switzerland comprising nominal M3, real gross domestic product (GDP), the GDP deflator and the government bond yield.…”

“…1 This paper investigates the transmission of monetary policy shocks in a structural cointegrated vector-autoregressive (VAR) model for Switzerland that allows us to impose and test a long-run structure as well as a short-run structure on the data. We include exogenous variables in the model, using the methodology of Pesaran, Shin and Smith (2000) and Garratt, Lee, Pesaran and Shin (2003). The five cointegrating relations among the variables in the model are interpreted as capturing money demand, the real interest rate, a term spread, uncovered interest parity and an output demand schedule.…”

Assessing Substitution and Complementary Effects Amongst Crime Typologies

“…Thus, the chosen lag length is four accordingly (Garratt et al, 2003). The cointegration test results are presented in Table 3.…”

“…However, inefficiency issues are not uncommon in core macroeconometric models (e.g. Garratt et al, 2003).…”

Identification of monetary policy in SVAR models: a data-oriented perspective

“…This approach is adopted for instance by Sims and Zha (1998), the extension proposed by Kim and Roubini (2000) with monthly data and international variables, and the macroecometric model of the UK proposed by Garratt et al (2003). These papers argue that only high-frequency data should be assumed to be in the information set of the central bank.…”