Data & Coding

The data and codes for this paper are available on Open ICPSR here along with a detailed README file. The paper is mainly theoretical. It solves an optimal control problem maximizing a welfare function under constraints, with the behavior of key variables being governed by an epidemiology (SEIR) model.

Following COVID, many economists used SEIR models nested in an economic model, but only a handful solved an optimal control with both testing and lockdowns as control. Relatedly, since SEIR models come in different flavors, as they are usually tweaked to fit the objectives of the paper, calculating R0 becomes quite complex. As a result, the majority of papers do not derive R0. I am attaching here a method for deriving R0 that we used, applying an approach used by mathematicians in the field. The references are in the attached document.

The online appendix of the paper here, among other things, makes available hand-collected data on the time it takes to complete a foreclosure across states. It also provided a detailed explanation of how we put together the merged dataset from publicly available sources. The dataset is very large, we will soon break it down into pieces and upload it here [COMING SOON].

This paper is currently under review. The STATA code that runs all tables and figures in the paper can be found here This may be geeky, but the code may help you with automatically formatting the tables from within STATA to be exported directly into Latex.

The paper merges together very large datasets on banks (see figure). Cross-walks between these datasets are not readily available. The Python code here helps in that regard. More detailed explanations will be added soon.