International Agent-Based Trade Model (Llerena & Lorentz Replication and Extensions)

This repository contains a Python-based implementation and extension of the international agent-based model (ABM) originally proposed by Llerena & Lorentz (2004). The model simulates endogenous growth and trade across multiple economies, focusing on innovation, firm-level competition, exchange rate dynamics, and now — trade tariffs and policy regimes.

---

Purpose

• Replicate the baseline evolutionary ABM.
• Introduce FX dynamics and tariff regimes to explore trade competitiveness.
• Quantitatively assess the persistence of economic leadership, path-dependence, and regime-lock in.
• Evaluate empirical outcomes across 1,000 simulations with and without tariff intervention.

---

Contents

• sim_models.py: All simulation model logic and analysis functions.
• notebook.ipynb: Full simulation setup, results, visualizations, and interpretations.
• README.md: This file.
• requirements.txt: List of Python dependencies.

---

How to Run

1. Clone the repository

   git clone https://github.com/Stef-creator/agent-based-trade-model.git
   cd agent-based-trade-model
   

2. Create and activate virtual environment

   python3 -m venv .venv
   source .venv/bin/activate
   

3. Install dependencies

   pip install -r requirements.txt
   

4. Launch Jupyter Notebook jupyter notebook notebook.ipynb

---

Key Features

• Dynamic Exchange Rates
  Modeled based on export growth differential and random shocks.

• Tariff Module
  Tariff-imposing economy penalizes import competitiveness via a user-defined tariff rate.

• Leadership Tracking & Statistics
  Tracks:

  • Who leads initially
  • Whether they remain leader
  • Number of leadership transitions
  • Final GDP and export gaps

• Visualization Dashboard
  Distribution plots for:

  • Final leader identity
  • Regime persistence
  • GDP & export divergence
  • Leadership churn

---

Visualization Preview

Plots include:

• Final leader frequency (by economy)
• Persistent leadership share
• GDP and export dominance (boxplots)
• Number of leadership transitions (histogram)

---

Future Extensions

• Add entry and exit dynamics for firms.
• Implement endogenous tariff retaliation strategies.
• Introduce heterogeneous firm sizes, credit constraints, or sectoral shocks.
• Calibrate model to real-world trade data (e.g., WTO, IMF).
• Visualize agent-level heterogeneity (e.g., productivity dispersion).

---

Citation

If you use or adapt this work, please cite the original paper:

Llerena, P. & Lorentz, A. (2004). Cumulative Causation and Evolutionary Micro-Founded Technical Change.

---

Contact

Created by Stefan Pilegaard Pedersen

---

License

MIT License. See LICENSE file for details.