RDP 2013-07: An Empirical BVAR-DSGE Model of the Australian Economy 6. Conclusions
June 2013 – ISSN 1320-7229 (Print), ISSN 1448-5109 (Online)
- Download the Paper 658KB
In this paper we have extended a baseline small open economy DSGE model to include a second production sector which produces commodities. This allows the price of exports to differ from other domestically produced goods in the model. This addition is motivated by the sizable increase in Australia's terms of trade since the mid 2000s. The commodity sector we have introduced is very simple, and could be developed further; for example, by creating an explicit role for natural resources, or to consider alternative pricing assumptions. To better capture the changes in the terms of trade which have occurred, allowing permanent productivity shocks in the commodity sector, so as to capture the sluggish response of supply to the increased demand, could be examined.
We examine the forecasting performance of a BVECMX where this DSGE model is used as a prior, and the large economy variables remain exogenous, an approach which is potentially useful for forecasting for small open economies. We have concentrated on evaluating the point forecasting performance, and further research could consider the properties of the forecast distribution. We find that when estimated over the inflation-targeting period, the forecasts from the BVECMX generally outperformed those from the DSGE model. Simply combining the DSGE forecasts with those from a SOE Minnesota VAR yielded forecasts of a comparable accuracy, and forecasts from the BVECMX did not outperform the SOE Minnesota VAR alone or simple univariate models. When estimated over a longer sample, the performance of the BVECMX model forecasts improved relative to those of an AR model, particularly for inflation in the short to medium term.
In all, the Australian economy is in the midst of a once-in-a-lifetime terms of trade and investment boom. While the model in this paper is a first step, creating a DSGE model which can capture these developments more effectively is likely to be beneficial, but quite challenging.