Forecasting and Policy Analysis with a Dynamic CGE Model of Australia DixonPeter B. RimmerMaureen T. 2017 The main ideas in this paper are: (i) that CGE models can be used in forecasting; and (ii) that forecasts matter for policy analysis. We demonstrate these ideas by describing an application of MONASH, a dynamic CGE model of Australia, to the Australian motor vehicle industry over the period 1987 to 2016. The key to generating behevable forecasts is to use detailed information available from expert groups specializing in the analysis of different aspects of the economy. In MONASH we incorporate forecasts by specialists: on the domestic macro economy; on Australian economic policy; on world commodity markets; on international tourism; on production technologies; and on consumer preferences. We have found that CGE forecasts incorporating such specialist information are readily saleable to public and private organizations concerned with investment, employment, training and education issues. This is partly because the economywide consistency guaranteed by the CGE approach enables users of economic intelligence to see the disparate forecasts dealing with different parts and aspects of the economy within an integrated perspective. Over the last thirty five years, CGE models have been used almost exclusively as aids to "what if" (usually policy) analysis. In almost all cases it has been assumed that the effects of the shock under consideration are independent of the future path of the economy. Thus, for "what if" analysis, a common implicit view is that reahstic basecase forecasts are unnecessary. Contrary to this view, we find that "what if" answers depend significantly on the basecase forecasts. This is not surprising when we are concerned with unemployment and other adjustment costs. However, we find that basecase forecasts are critical even when our concern is the longrun welfare implications of a policy change. For example, we find that the simulated long-run effects of a tariff cut on imported cars are strongly influenced by the basecase forecast of the rate of technical progress in the car industry relative to that in other industries.