4. Conclusion
Ex-post analyses of forecast errors are important for both model builders and policymakers. Neil Ericsson’s paper explores the biases in the debt forecasts produced by the OMB and the CBO. We (and Ericsson and others) interpret the IIS technique to include a more general method for detecting model misspecification. In our analysis of the debt data, we find a violation of the assumption of homoskedastic errors. In other applications, the IIS technique has the potential to detect a wide variety of model misspecifications.
