Stylized Model of Smallholder Agricultural Households

Author: Tim G Williams (tgw@umich.edu)

A COMPLETE ODD+D PROTOCOL WILL BE RELEASED UPON ACCEPTANCE OF THE PUBLICATION.

Model overview

To be added upon acceptance of the publication.

1. Code structure

The code is divided into a few folders:

model/ contains all the model code. The model can be run as described below. All model inputs are in the base_inputs.py file.

experiments/ contains code for running various experiments (e.g., shock simulations, scenario analysis, sensitivity analysis).

calibration/ contains the code for pattern-oriented modeling.

plot/ contains all plotting functions.

NOTE: while most files are .py, some are .ipynb, which are Jupyter Notebook files.

2. Run the model

Requirements:

• Python >=3.6
• Make sure all required packages in the requirements.txt file are installed.
• Create a folder called code for the code and a folder called outputs/ for the outputs to be stored.

Navigate to the code directory.

From the command line, type:

python run.py

This will run the model and some plotting functions.

3. Conduct experiments

The experiments conducted for the main body of the paper are coded in several different files:

• experiments/analysis_shock.py examines the effects of climate shocks under the different scenarios
• experiments/analysis_poverty.py examines, under regular climatic variability, the propensity for households to have positive levels of wealth over time
• experiments/sensitivity.py conducts a sensitivity analysis to selected parameters, and visualizes them using a gradient-boosted random forest.
• experiments/explore mechanisms.ipynb generates the plots in the supplemental data to the paper