Modern Computational Economics and Policy Applications

A workshop for the IMF's Institute for Capacity Development

Abstract

Open source scientific computing environments built around the Python programming language have expanded rapidly in recent years. They now form the dominant paradigm in artificial intelligence and many fields within the natural sciences. Economists can greatly enhance their modeling and data processing capabilities by exploiting Python's scientific ecosystem. This course will cover the foundations of Python programming and Python scientific libraries, as well as showing how they can be used in economic applications for rapid development and high performance computing.

Times and Dates

• Dates: March 25-27, 2024
• Times: 9:30 -- 12:30 and 14:00 -- 17:00
• Location: room HQ2-3B-748 (in-person participants)

Instructors

Chase Coleman is a computational economist based at New York University where he is a visiting assistant professor. He was an early contributor at QuantEcon and, along with other members of QuantEcon, has given lectures and workshops on Python, Julia, and other open source computational tools at institutions and universities all around the world.

John Stachurski is a mathematical and computational economist based at the Australian National University who works on algorithms at the intersection of dynamic programming, Markov dynamics, economics, and finance. His work is published in journals such as the Journal of Finance, the Journal of Economic Theory, Automatica, Econometrica, and Operations Research. In 2016 he co-founded QuantEcon with Thomas J. Sargent.

In addition, 2011 Nobel Laureate Thomas J. Sargent will join remotely and run a one hour session on the 27th.

Syllabus

• Monday morning: Introduction
  • Scientific computing: directions and trends (intro_slides/sci_comp_intro.pdf)
  • Python and the AI revolution (ai_revolution/ai_revolution.pdf)
  • A brief tour of Python's massive scientific computing ecosystem (scientific_python/main.pdf)
  • Working with Jupyter (free coding)
  • A taste of HPC with Python (fun_with_jax.ipynb)
• Monday afternoon: Python basics
  • Core Python (quick_python_intro.ipynb)
  • NumPy / SciPy / Matplotlib / Numba (quick_scientific_python_intro.ipynb)
  • Exercises: Simulation (simulation_exercises.ipynb)
  • Exercises: Lorenz curves and Gini coefficients (lorenz_gini.ipynb)
• Tuesday morning: Markov models in Python
  • Markov chains: Basic concepts (finite_markov.ipynb)
  • Intermezzo: A quick introduction to JAX (jax_intro.ipynb)
  • Wealth distribution dynamics (wealth_dynamics.ipynb)
  • Exercises: Markov chain exercises (markov_homework.ipynb)
• Tuesday afternoon: Dynamic programming
  • Job search (job_search.ipynb)
  • A simple optimal savings problem (opt_savings_1.ipynb)
  • Alternative algorithms: VFI, HPI and OPI (opt_savings_2.ipynb)
  • The endogenous grid method (egm.ipynb)
• Wednesday morning: Heterogeneous agents
  • Heterogenous firms (hopenhayn.ipynb)
  • The Aiyagari model (aiyagari.ipynb)
• Wednesday afternoon: Further applications
  • Sovereign default (arellano.ipynb)
  • The Bianchi overborrowing model (overborrowing.ipynb, bianchi.pdf)

Software

The main interface to Python will be either jupyter-notebook or jupyter-lab.

Access to the ipython REPL will also be useful.

Some work will be done remotely using Google Colab --- a Google account is required.

Required Python libraries (much of which is found in the Anaconda Python distribution):

• numpy
• scipy
• matplotlib
• pandas
• scikit-learn
• statsmodels
• numba
• f2py
• quantecon

Useful References

• https://www.anaconda.com/download
• google/jax#13684 (for JS)