================================================================================================================

README: Description of codes and data for the paper: "Learning, Capital-Embodied Technology and Aggregate Fluctuations" by C. Gortz and J. Tsoukalas.

For any questions plese email: john.tsoukalas@glasgow.ac.uk

================================================================================================================ 

The files in this folder can be used to generate the results in Tables 4 and 5 
as well as Figure 1.

The MATLAB file policyFunctions_HPC.m uses value function iteration to compute the policy functions
of the model. It calls Tauchen.m.

The MATLAB file noLearningModel.m uses policy functions computed in policyFunctions_HPC.m to generate the statistics in Table 4, Panel B (case without learning). It calls Markov.m and hpfilter.m.

The MATLAB file learningModel.m uses policy functions computed in policyFunctions_HPC.m to generate the statistics in Table 4, Panel C (case with learning). It calls Markov.m and hpfilter.m.

The MATLAB file recession_analysis1.m uses the policy functions computed in policyFunctions_HPC.m to generate the statistics in Table 5, columns titled "Recessions due to noise". It calls Markov.m, simulate_learning_model.m and identify_recessions.m.

The MATLAB file recession_analysis2.m uses the policy functions computed in policyFunctions_HPC.m to generate the statistics in Table 5, columns titled "Recessions due to fundamentals". It calls Markov.m, simulate_learning_model.m and identify_recessions.m.

The impulse response functions shown in Figure 1 are generated using the MATLAB pre-processor Dynare. IRFs for the model with and without vintage capital can be generated using vintageCapitalModel.mod and NoVintageCapitalModel.mod (works with Dynare version 4.2.0). The steady state of these models is calculated in functions vintageCapitalModel_steadystate and NoVintageCapitalModelLoading_steadystate. The output generated by vintageCapitalModel.mod and NoVintageCapitalModel.mod is loaded in plot_IRFs.m to generate Figure 1.

The excel file Data.xlsx contains the data and calculations to generate the statistics for the data in Tables 1, 4, 5 (sheet US Data) and 2 (sheet G7 Data). The p-values shown in Table 1 and 2 can be calculated using the eViews add-on TSNorm.aipz. which is also contained in the folder.

The Matlab version used is R2009b.