The Information Content of Financial Aggregates in Australia
This paper examines the information provided by financial aggregates as predictors of real output and inflation. We employ vector autoregression (VAR) techniques to summarise the information in the data, providing evidence on the incremental forecasting value of financial aggregates in a range of forecasting systems for these variables. The in-sample results suggest significant predictive power in only a small number of cases. We then test the forecast performance of the VAR systems for two years out-of-sample in order to mimic more closely the real-time forecasting problem faced by policymakers. Overall, both in-sample and out-of-sample results suggest no robust finding of exploitable information for forecasting purposes in any of the financial aggregates under examination. There is some evidence that the aggregates yield improved forecasts late in the sample period, but there is insufficient subsequent data to draw robust conclusions from this.
What people are saying - Write a review
We haven't found any reviews in the usual places.
aggregate under consideration Bank of Atlanta Bank of Australia block exogeneity tests causal ordering Choleski decomposition content of financial correlations corresponding model currency data series Error Bands estimates exchange rate explanatory power F-tests Federal Reserve Bank financial aggregate relative five variable systems forecast accuracy Forecast Error Statistics forecast error variance forecast horizon forecast of inflation forecast of output full sample Granger causality growth and inflation growth rate Horizon 3 Variable improvement at steps improvement over steps Initial Ordering innovation associated interest rate lags mean square error Model Measure Step monetary aggregate monetary policy Monte Carlo integration Notable improvement orthogonal output and inflation OUTPUT GROWTH Forecast p-value Period Horizon policy makers price level Ratio Theil real GDP real output growth RMSE root mean square Sailesh Ramamurtie sample period Slight improvement specification Tallman Tao Zha Theil U statistic Uniformly worse Variable system containing variables of interest variance decomposition results variance-covariance matrix vector autoregression