This purpose behind this study is to develop a model for forecasting the performance of the Taiwanese economy based on monthly time series data. We first extract the useful factors through factor analysis. Next, we rank the factor scores according to the rules of the trend and interpret the scores as signals to buy or sell appropriately. Our main result is that the Sharpe ratio of out-of-sample back-testing from January 2007 to December 2010 is 0.48, indicating an ability to forecast financial crises. In addition, a Sharpe ratio of 0.95 during the 2008 financial crisis suggests that our model may have been effective in predicting this crisis. Moreover, the macroeconomic factor model can provide better forecasting skills during financial crises. To conclude, this research may be of importance in explaining the relationship between macroeconomic variables and the business cycle, as well as in providing investors with better forecasting signals of the stock market in Taiwan.