RDP 2008-01: A Sectoral Model of the Australian Economy 5. Which Shocks Are Most Important?

Impulse responses illustrate the absolute response of domestic economic variables to various shocks, but they do not tell us anything about the relative importance of different shocks. To do this we examine the forecast error variance decomposition of our model, which identifies the proportion of the variation in each of the endogenous variables that can be attributed to shocks to the other variables.

According to Figure 8, most of the short-term variation in each of the GDP components can be explained by shocks to the components themselves, but less of their long-term variation. Take dwelling investment. At the 5-quarter horizon around 75 per cent of the forecast error variance can be explained by the shocks to the GDP components, whereas at the 20-quarter horizon this figure drops to less than 50 per cent, with most of the remaining variation explained by foreign shocks and monetary policy. Overall, there is little evidence that shocks that spill over from one sector to another are important. Real exchange rate and inflation shocks have relatively modest importance over the longer run. By contrast, foreign shocks are an important source of variation for a number of variables, particularly at longer horizons, explaining more than 30 per cent of the variation in the forecast errors for dwelling investment, machinery & equipment investment, imports, exports, inflation, the cash rate and the real exchange rate after 20 quarters.