Authors: Ricardo C. A. Lima, Pedro Jorge Alves, Bruno Barsanetti, and Francisco Cavalcanti
You can download the replication package here (about 47 MB).
The analysis folder is subdivided into:
- code: Contains all R scripts used to generate the figures and tables.
- output: Stores the results generated by the scripts (PNG and TeX files).
In the build/output folder of the replication package, the following datasets are available for reproducing the results:
database_panel.dtaanddatabase_panel.rds: Annual panel dataset of Brazilian municipalities used in the main analysis (available in both Stata and R formats).database_three_periods.rds: Aggregated panel dataset structured into three time periods for complementary specifications.database_two_periods.rds: Aggregated panel dataset structured into two time periods for alternative empirical exercises.restricted_PSM_database.dtaandrestricted_PSM_database.rds: Sample of municipalities selected through the propensity score matching (PSM) procedure (available in both Stata and R formats).m.data_agglomeration.rds: Dataset aggregated at the urban agglomeration level, used for robustness analyses at a higher spatial aggregation.
To replicate the results, please follow these steps:
-
Download and extract the replication package: Replication Package
-
Open the main R script that sources all analysis files.
-
Set the following paths inside the script:
DROPBOX_PATH <- "INSERT YOUR DIRECTORY PATH HERE/Landslides-Urban-Patterns"
GITHUB_PATH <- "INSERT YOUR DIRECTORY PATH HERE/Landslides-Urban-Patterns"-
Set the working directory to
GITHUB_PATH. -
Run the script. All figures and tables will be automatically saved in:
./analysis/output
Below is the updated mapping between scripts and the figures/tables they generate.
analysis/code/Figure 1 - Panel A The Distribution of Landslides Across Space and Time.Ranalysis/code/Figure 1 - Panel B The Distribution of Landslides Across Space and Time.Ranalysis/code/Figure 1 - Panel C The Distribution of Landslides Across Space and Time.R
analysis/code/Figure 2 - Effects of Landslides on Urbanization Patterns.R
analysis/code/Figure 3 - Effect of Landslides on Local Labor Markets.R
analysis/code/Figure 4 - Effect of Landslides on Land-Use Regulations.R
analysis/code/Figure 5 - Effect of Landslides on the Urban Areas of High-Risk and Low-Risk Regions.R
analysis/code/Figure A.2 - Landslides Effects on Urbanization Patterns group-specific ATTs.R
analysis/code/Table B.1 - Balance Check for Treated and Control Units Before and After the Matching.R
analysis/code/Figure C.1 - Robustness Check Full Sample.Ranalysis/code/Figure C.2 - Robustness Check Considering Only Not-Yet Treated Units.Ranalysis/code/Figure C.3 - Robustness Check Southeast-Only and Southeast-Excluded.Ranalysis/code/Figure C.4 - Robustness Check Disaster Monitored Municipalities.Ranalysis/code/Figure C.5 - Robustness Check Other Natural Disasters.Ranalysis/code/Figure C.6 - Robustness Check Only municipalities treated once.Ranalysis/code/Figure C.7 - Robustness Check Inclusion of Control Variables.Ranalysis/code/Figure C.8 - Robustness Check Accounting for State-Specific Shocks.Ranalysis/code/Figure C.9 - Robustness Check Alternative Estimators.Ranalysis/code/Figure C.10 - Robustness Check Standard Errors Clustered by Micro-region.Ranalysis/code/Figure C.11 - Robustness Check The Spatial Spillovers of Landslides.Ranalysis/code/Figure C.12 - Robustness Check Alternative Time Horizons.Ranalysis/code/Figure C.14 - Robustness Check Aggregation at the Metropolitan Region Level.Ranalysis/code/Figure C.15 - Robustness Check Excluding Smallest and Largest Municipalities.Ranalysis/code/Figure C.16 - Robustness Check Effects of Extreme Rainfall in Mountain Ranges.R
(Figure C.13 – Synthetic Difference-in-Differences.do – should be run in Stata)
analysis/code/Figure D.1 Heterogeneous Effects of Landslides by Municipal Characteristics.Ranalysis/code/Figure D.2 Effects of Landslides on Population and Number of Households.Ranalysis/code/Table D.1 - Long-Term Differences.R