Yohei Yamamoto, "Bootstrap Inference for Impulse Response Functions in Factor-Augmented Vector Autoregressions," Journal of Applied Econometrics, Vol. 34, No. 2, 2019, pp. 247-267. The zip file yy-files.zip contains replication files of the empirical example provided in section 6 of the paper. All files are ASCII files in DOS format. Unix/Linux users should use "unzip -a". DATA The data are stored in "nsbalpanel.txt" and are the same as the ones used in Bernanke, B.S., J. Boivin, and P. Eliasz, 2005, "Measuring the effects of monetary policy: A factor-augmented vector autoregressive (FAVAR) approach", Quarterly Journal of Economics 120(1), 387-422. The file contains 120 US macro variables from January 1959 to August 2001. See the original paper for details. PROGRAM To see the results, run "main.m" with the specifications: p: VAR lag order r: number of factors m: number of observed factors wild: 1 if heteroskedasticity is considered in the idiosyncratic errors and 0 otherwise hmax: maximum horizons nrep: number of bootstrap replications c: 1-c is the confidence level FUNCTIONS The zip file also includes 14 function files. Each of them has an explanation in the file, but here are the short descriptions: pcestim.m: estimates principal component factors, loadings, and idiosyncratic errors pcestim_nofac.m: estimates loadings and idiosyncratic errors when the factors are treated observed varestim.m: VAR estimation by least squares irfgen.m: construct impulse responses btvargen.m: implements bootstrap for the VAR model btfmgen.m: implements bootstrap for the factor model resample.m: implements iid resampling wildbtgen.m: implements wild bootstrap signcorrect.m: correct signs for the recursive restriction backlevel.m: recover the level from the transformed impulse response estimate Please address any questions to: Yohei Yamamoto Department of Economics Hitotsubashi University 2-1 Naka, Kunitachi Tokyo 186-8601 Japan