Replication code and simulation results for “Real option valuation of undeveloped oil fields: How modeling assumptions affect value and investment timing”
Description
This repository contains the Python code, model parameters, and numerical simulation results supporting the study “Real option valuation of undeveloped oil fields: How modeling assumptions affect value and investment timing”. The study compares three real option frameworks for undeveloped oil fields: the Paddock–Siegel–Smith (PSS), Brennan–Schwartz (BS), and Cortazar–Schwartz (CS) models. The numerical experiments examine how assumptions about the underlying stochastic process, operating flexibility, and construction delays affect project values and exploration and development thresholds. The package includes a self-contained Jupyter notebook for the valuation and simulation experiments, Python scripts for reproducing the numerical method schematics, reference numerical outputs, and the 18 manuscript figures. Additional files document software requirements, reproduction instructions, the correspondence between figures and generating code, and numerical verification against the reference results. All results are generated computationally using illustrative parameter settings documented in the manuscript and notebook. No new empirical dataset was collected, and reproduction does not require proprietary oil-field data. The stochastic simulations are conducted under the valuation measure; the resulting exercise probabilities should not be interpreted as forecasts of actual investment behavior. To reproduce the results, extract the archive, install the dependencies listed in requirements.txt, and follow the instructions in README.md. Newly generated outputs are stored separately from the supplied reference results.
Files
Steps to reproduce
1. Download and extract the reproduction archive. 2. Open a terminal in the extracted package directory. 3. Create and activate a dedicated Python environment. The reference calculations were performed with Python 3.12.14. 4. Install the dependencies by running: python -m pip install -r requirements.txt 5. Reproduce the calculations and figures by running: python scripts/run_all.py 6. Compare the generated numerical results with the supplied reference results by running: python scripts/verify_results.py 7. Consult the generated execution and verification reports. New outputs are written to the generated directory, while the supplied reference results remain unchanged. The notebook can alternatively be executed sequentially in a compatible Jupyter environment. Computational time depends on the hardware and software environment. No network access is required after installation of the dependencies.
Institutions
- Pontifícia Universidade Católica do Rio de JaneiroRio de Janeiro, Rio de Janeiro
- Petrobras (Brazil)Rio de Janeiro, Rio de Janeiro