Modelling the AI Ownership Transition: Usownership, Fiscal Substitution and Monetary Constraints — Data and Reproducibility Package

Description

This dataset provides the simulation data, source code, parameter configurations and reproducibility materials supporting the article “Modelling the AI Ownership Transition: Usownership, Fiscal Substitution and Monetary Constraints.” The study develops a heterogeneous stock-flow-consistent simulation laboratory to examine how gradual contribution-weighted user ownership (“usownership”), income-support policies, conventional taxation and bounded monetary financing interact during alternative AI-driven structural-change scenarios. The archive contains the Python implementation of the model, fixed parameter sets and random seeds, policy and automation scenarios, synthetic simulation outputs, sensitivity and robustness analyses, accounting and validation tests, and scripts used to reproduce the reported tables and figures. The main experiment comprises 549 macroeconomic simulation paths, supplemented by additional diagnostic exercises covering calibration sensitivity, population resampling, fiscal-capacity regions, limiting cases, numerical tolerances and local parameter identification. All household and firm observations are synthetic. Official macroeconomic data are used only as explicitly documented calibration or reference anchors. The dataset is intended for reproducibility and methodological evaluation rather than as an empirical dataset describing any specific country. Key outputs include transfer expenditure, conventional taxation, user-held equity, monetary financing, inflation, wealth concentration, output, coverage of the income floor, and matched fiscal-equity counterfactuals. The materials allow readers to reproduce the central results, inspect accounting identities, test alternative assumptions and extend the model under different structural-change and policy configurations. Keywords: artificial intelligence; automation; economic modelling; usownership; stock-flow consistency; seigniorage; fiscal policy; redistribution; simulation; sensitivity analysis; reproducibility.

Steps to reproduce

1. Download and extract the archive. 2- Install Python 3.11 or later. 3. Create and activate a virtual environment: python -m venv .venv source .venv/bin/activate On Windows: .venv\Scripts\activate 4. Install the required dependencies and the local package: python -m pip install --upgrade pip python -m pip install -r requirements.txt python -m pip install -e . --no-deps 5. Run the validation and accounting tests: python -m pytest -q 6. Reproduce the complete numerical experiment: python scripts/reproduce_all.py --samples 128 This regenerates the main simulation outputs, policy comparisons, robustness exercises, sensitivity analyses, tables and figures used in the article. The full experiment comprises the reported macroeconomic scenario paths and uses fixed random seeds for reproducibility. 7. To reproduce the manuscript figures, tables and compile the LaTeX paper, if a TeX distribution with pdflatex or latexmk is installed, run: python scripts/reproduce_all.py --samples 128 --compile 8. For a faster verification run, use: python scripts/reproduce_all.py --samples 8 This reduced run is intended only as a smoke test and does not reproduce the full sensitivity experiment. The model outputs include transfer expenditure, conventional taxation, user-held equity, inflation, monetary financing, output, wealth concentration and coverage measures. All household- and firm-level observations are synthetic; official macroeconomic data are used only as documented calibration/reference anchors. The model is a conditional mechanism study rather than an estimated forecast of a specific economy.

Institutions

• Universitat de Girona

  Catalonia, Girona

Categories

Licence