CW2016-ReplicationCodes

These codes reproduce the results in:

Cúrdia, V., and M. Woodford (2016)
The Central-Bank Balance Sheet as an Instrument of Monetary Policy
Journal of Monetary Economics, 84, pp. 30-65.

Technical Appendix

These replication codes are available online at:
https://github.com/vcurdia/CW2016-ReplicationCodes

Requirements

Matlab (R)

The codes were tested using Matlab (R) R2016b with the following toolboxes

• Symbolic Toolbox
• Optimization Toolbox

LaTeX

LaTeX is used by some tools to compile certain documents.

epstopdf, included in most LaTeX releases, is used by some tools.

Additional codes

Codes by Vasco Cúrdia:

• VC-Tools, version v3.0.3
• ACR-LQ joint with Filippo Altissimo and Diego Rodriguez Palenzuela, version v1.0.0

Codes by Chris Sims:

• gensys
• optimize

All auxiliary codes included in this repository in subfolders.

Description of Replication Codes

plotmgutil.m Generates Figure 1 of the paper showing the shape of the marginal utility for the two types of households.

RunAllModels.m Script that runs all models for a given parameterization. It generates simulations for all shocks under optimal policy and alternative policy rules as described in the paper and appendix.

Simulations are generated for three models:

• FF: model with financial frictions Simulations are generated by the function IntModelFF.m
• NoFF: model with heterogeneous households but no financial frictions Simulations are generated by the function IntModelNoFF.m
• RepHH: model with representative households. Simulations are generated by the function IntModelRepHH.m

Parameterization Flags:

• End Set to 0 to simulate linear financial intermediation technology, as in Figure 2 of the paper and Figures H1 through H9 of the appendix. Figure names are appended with '_Exo'. Set to 1 to simulate convex financial intermediation technology, as in the remainder of the Figures. Figure names are appended with '_End'.
• SmSigma Set to 0 in baseline calibration, implying that the ratio of sigma_b to sigma_s is 5. Set to 1 in order to generate the 'low-sigma_b' calibration, which sets sigma_b equal to sigma_s. Figures are appended with '_SmSigma'.

For each model the codes simulate responses to shocks under different policy rules. The figure names are appended according to which rule is being used according to the following:

• LQ: Optimal Policy
• Pi: Strict inflation stabilization (PiStab)
• T: Taylor rule as described in the paper
• FT: Flexible targeting criterion as described in the paper (FlexTarget)
• TYn: Taylor rule with output gap (instead of output deviations from trend) (TaylorYn) The results under this rule are shown in the appendix.

PlotAll.m Script that generates all the figures in pdf format. Figures for each parameter configuration are stored in a subfolder with the appropriate name. It requires that RunAllModels was run before hand with the same parameter flags.

MakePlotsPDF.m Script that uses the figures generated with PlotAll and creates a tex report with the following sections:

• One section for each policy comparing all three models.
  These are useful to compare how adding household heterogeneity and financial frictions affect responses to the shocks.
• One section comparing all policy rules for the FF model.
  These are useful to compare outcomes under alternative policies. Each section has a subsection for each shock considered.