Learning from Experience no Shorting

Paul Ehling, Christian Heyerdahl-Larsen, and Zeshu XU

Install dependencies

pip install -r requirements.txt

Execute the cohort simulation

python3 main.py

Linter

To have a consistent coding style, use black to lint the code before commit

black main.py

Function descriptions

1. param.py defines the parameters used in the simulations, and prepares matrix to store the results;
2. stats.py defines some small functions used throughout the simulations, including:
   1. post_var(): calculate the posterior variance, correspond to eq(2);
   2. shocks(): calculate Zt and Yt from shocks dZt;
   3. tau_calculator(): calculate tau for the cohorts in the economy;
   4. good_times(): information indicator;
3. solver.py defines some small solver functions, including:
   1. bisection(): Bisection method to solve for any the solution in any equation;
   2. solve_theta(): combines with bisection(), contains RHS - LHS of the eq(24), used to iteratively solve market-clearing price of risk, ie. theta;
   3. find_the_rich(): finds the agents that make the top 5% richest population in the economy, and they can short;
   4. bisection_partial_constraint() and solve_theta_partial_constraint() are similar to 3.i and 3.ii, but takes more arguments;
4. cohort_builder.py defines a function that builds up an OLG economy;
5. cohort_simulator.py defines a function that simulates the OLG economy forward;
6. simulation.py defines a function that wraps 4 and 5 in one function;
7. V_hat_experiment.py changes V_hat, or initial variance, and generates graphs to illustrate the simulation results;
8. main.py is the main loop that simulates the economy, and generates graphs to illustrate the simulation results.