Fabio Bertolotti and Massimiliano Marcellino, "Tax Shocks with High and Low Uncertainty", Journal of Applied Econometrics, Vo. 34, No. 6, 2019, pp. 972-993. All files are ASCII files in DOS format and are zipped in bm-files.zip. Unix/Linux users should use "unzip -a". 1) The data used in estimation are stored and described in the .m files: data_macro_series.m, data_proxy_series.m, data_switching_series.m. The first file contains data on the main macroeconomic series, the sources and modifications carried out for the estimation. The second file contains data on shock series (Romer and Romer tax shocks, Ramey government spending shocks, Romer and Romer monetary policy shocks) and the sources. The third file contains data on the series used as switching variables (uncertainty measures, mainly). 2) The .m file Plots_IRFs.m replicates the Figures reporting IRFs for both the main text and the online appendix. Estimation is carried out by the functions vvar_2s.m, varsn_2s.m, tvar_2s.m, and tvarsn_2shocks.m, tvarsn_2s_rsvcv.m, and tvarsn_2s_girf.m, which estimate and produce IRFs and other objects for a simple VAR model, a VAR with sign nonlinearity only, a Threshold VAR without sign nonlinearity, a Threshold VAR with sign nonlinearity, Threshold VAR with sign nonlinearity and regime-specific variance-covariance matrix estimation (Iterated FGLS) and Generalized IRFs for a Threshold VAR with sign nonlinearity, respectively. Output from these functions is then used as input of the functions var_2s_fp.m, varsn_2s_fp.m, tvar_2s_fp.m, tvarsn_2s_fp.m, tvarsn_2s_rsvcv_fp.m, and tvarsn_2s_girf_fp.m which combine IRFs for single variables, and finally functions plot_var_bm.m and plot_tvar_bm.m organize IRFs graphically. Detailed comments of the estimation steps can be found in the script of the tvarsn_2shocks.m function, which is the key estimation function of the paper. Other functions carrying out estimation work similarly. 3) The .m file Table1_Tests.m replicates Table 1 in the main text, by using the estimation function tvarsn_2s_tests.m, which performs nonlinearity tests described in the paper. 4) The .m file Table2_Multipliers.m replicates Table 2 in the main text, by using the estimation functions vvar_2s.m, varsn_2s.m, tvar_2s.m, and tvarsn_2shocks.m and the function multiplier_m2.m. The first set of functions performs estimation and generates inputs for the function multiplier_m2.m, which computes tax multipliers and confidence bands as described in the paper starting from IRFs and raw GDP and tax shocks data. 5) The .m file Table3_ModelFit.m replicates Table 3 in the main text, by using the function tvarsn_2s_fit.m, which computes information criteria for several specifications using different swithcing variables (the VAR specification is different from the main estimation for sake of comparability among different uncertainty measures, ads described in the main text). 6) The .m file Additional_Plots_and_Figures.m reproduces Figures 2 and 3 in the main text and performs additional empirical analyses that support specific parts of the paper, e.g. regarding the distribution of shocks over regimes, the identification of regime periods, the average length of an uncertainty regime, or the correlation between the RR monetary shocks and RR positive tax shocks in the identified high-uncertainty regime, which motivates the inclusion of RR monetary shocks in the specification.