This paper proposes a parsimonious and model-consistent method for combining forecasts generated by structural microfounded models and judgmental forecasts. The method delivers along several dimensions. First, it improves the forecasting performance of the model. Second, it allows interpreting the judgmental forecasts through the lens of the model. Finally, it provides a framework to assess the informational content of the judgmental forecasters. I illustrate the proposed methodology with a real-time forecasting exercise using a simple neo-Keynesian dynamic stochastic general equilibrium model and the Survey of Professional Forecasters.