Forecasting TVP-FAVAR

This code replicates the results in the working paper version of the article

Koop, G. and Korobilis, D. (2014). A New Index of Financial Conditions. European Economic Review, 71, pp. 101-116.

In particular, this code replicates the results in the December 1, 2013 version of the paper found in SSRN: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2374980. For the data and code of the published version, please check here.

Compared to previous working paper versions, this version also implements backward recursions using the Kalman smoother (in the early versions we only used the forward Kalman filter recursions for "online" forecasting). Compared to the final published version, which does forecasting using real data, the attached code estimates and forecasts using only final-vintage data.

Notice that the code has the option to open a MATLAB pool in order to perform parallel processing of all possible models, using a simple "parfor" loop. If you do not have the MATLAB Parallel Processing Toolbox installed in your PC just leave the code which calls the workers (i.e. CPU cores) commented out.

Exact instructions are given in a README file attached to the code. The authors do not provide further support, and the usual disclaimer applies.

Download code here: DMA_FCI.zip