The simultaneous equation econometric model has recently come under increasing, attack as a policy analysis and forecasting tool However, traditional methods of choosing among competing models rely heavily on the use of summary Statistics It IS shown, by example, that choosing a model With relatively better summary statistics does not guarantee getting the best unconditional out-of-sample forecasts Other methods of forecasting time series that allow relative evaluation of the out-of-sample forecasting ability of econometric Simultaneous equation models are also examined