Measuring Monetary Policy in the UK : A Factor-Augmented Vector Autoregression Model Approach

This paper investigates the determinants of UK interest rates using a factor-augmented vector autoregression model (VAR), similar to the one suggested by Bernanke, Boivin and Eliasz (Quarterly Journal of Economics, Vol. 120 (2005), No. 1, pp. 387-422). The method allows impulse response functions to be generated for all the variables in the data set and so is able to provide a more complete description of UK monetary policy than is possible using standard VARs. The results show that the addition of factors to a benchmark VAR generates a more reasonable characterization of the effects of unexpected increases in the interest rate and, in particular, gets rid of a 'price puzzle' response present in the benchmark VAR. The extra information generated by this method, however, also brings to light other identification issues, notably house price and stock market 'puzzles'. Importantly the out-of-sample prediction performance of the factor-augmented VARs is also very good and strongly superior to those of the benchmark VAR and simple autoregression models