
The /Analysis folder contains all the data, code, and output of the main empirical analysis for the paper.

The code is stored in /Code
The data to be used is stored in /Input
The results of the empirical analysis is stored in /Output
Any preliminary file that is created along the way is stored in /Temp. Those may of course be deleted after running the code.

The data in /Analysis/input has been created from some raw data. The raw data is stored in /Build/Input. The code for cleaning and constructing the data is stored in /Build/Code. The cleaned data that eventually appears in /Analysis/Input is saved to both /Build/Output and in /Analysis/Input.


Below is a description of what is achieved by the various files in /Analysis/Code:

**********************************************************
“Analysis_A.do”

This do file generates the graphs and table to test proposition 1 across all firms (section III.C). 
The graphs for the rank-size relationship among large exporters, and for the link between firm size and distance of exports go into figure 1; and the statistical tests for proposition 1 go into table 1.
**********************************************************


**********************************************************
“Analysis_B.do”

This do file generates the graph to test proposition 1 across sectors (section III.D). 
The graph showing the relationship between the predicted versus actual distance elasticity of trade goes into figure 2.
**********************************************************

**********************************************************
“Analysis_A_redux.do”
“Analysis_B_redux.do”

Those 2 do-files allow to replicate all the analysis, even if the original firm level data /Build/Input/export92.dta and /Build/Input/siren.dta are not available. They use data aggregated into bins of firms, which are already pre-loaded in /Analysis/Temp/.
**********************************************************