Zurich Early Retirement Simulator

This application is a financial simulation tool designed for individuals planning for early retirement (FIRE) in Canton Zurich, Switzerland.

Phase 1 focuses on the post-retirement decumulation phase for a single individual with no dependents living in Canton Zurich.

Features

• Three Comparative Simulation Methods: Compare outcomes side-by-side using:
  1. Historic Backtesting: Replay contiguous historical periods (~100 years of Swiss-adjusted data).
  2. Historic Bootstrapping: Random sampling with replacement across historical annual returns (as described in FIRE literature like The Poor Swiss).
  3. Parametric Monte Carlo: Stochastic lognormal return generator with customizable asset class means and volatilities.
• Swiss Tax Modeling: Accurately models Federal, Cantonal, and Municipal income and wealth taxes for Canton Zurich.
• AHV for Non-Workers: Models mandatory AHV contributions for early retirees before age 65.
• Pillar 2 & 3a Liquidations: Simulates growth and staggered lump-sum withdrawals of Pillar 3a accounts (up to 5) and Pillar 2 vesting accounts, including capital withdrawal taxes.
• Smart Cash Buffer: Optional defensive strategy to spend cash first during market downturns, protecting equities.
• Dynamic Expenses: Optional adjustment to reduce expenses when net worth drops below the starting watermark, plus Vanguard Dynamic Spending rules.
• Configurable Success Criteria: Set target ending net worth (e.g., preserve 50% of inflation-adjusted starting wealth) and calculate probability of success.

Data Sources & Provenance

This project incorporates historical datasets curated and maintained by Baptiste Wicht (The Poor Swiss), available in the open-source repository wichtounet/swr-calculator and described at The Poor Swiss:

1. US Stocks (USD): Robert Shiller monthly S&P 500 Total Returns dataset (1871–2025) from data/us_stocks.csv.
2. Non-US Equities (USD): Empirical ex-US stocks total return dataset (MSCI EAFE / World ex-US proxy, 1871–2025) from data/ex_us_stocks.csv.
3. Currency Exchange (USD/CHF): Historical monthly USD/CHF exchange rates (1913–2019 from data/usd_chf.csv, extended through 2025 via the Swiss National Bank (SNB)).
4. Swiss Inflation (CPI): Historical Swiss Consumer Price Index (1921–2023 from data/ch_inflation.csv, cleaned of 2022 entry error and extended through 2025 via the Swiss Federal Statistical Office (FSO/BFS)).

CHF-converted equity returns are computed annually as: $$\text{Return}{\text{CHF}} = (1 + \text{Return}{\text{USD}}) \times \left(\frac{\text{FX}{\text{End}}}{\text{FX}{\text{Start}}}\right) - 1$$

Project Structure

├── app.py                  # Streamlit frontend UI
├── data/                   # Historical CSV datasets from wichtounet/swr-calculator & SNB/FSO
├── src/
│   ├── simulation_engine.py # Core decumulation simulation loop
│   ├── tax_engine.py        # Swiss/Zurich tax calculations
│   └── historic_returns.py  # Historic return data and generators
├── scripts/
│   └── build_historic_returns.py # Pipeline building historic_returns.py from data/
├── docs/
│   ├── prd_early_retirement_calc.md # Product Requirement Document
│   └── design_doc_early_retirement.md # Technical Design Document
├── requirements.txt         # Python dependencies
└── venv/                    # Python virtual environment (ignored by git)

Getting Started

Prerequisites

• Python 3.11+
• Virtual environment tool (venv)

Installation

1. Clone the repository (once created on GitHub):

   git clone git@github.com:b-maldoca/fi.git
   cd fi

2. Set up the virtual environment and install dependencies:

   python3 -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt

Running the Application

Start the Streamlit app:

streamlit run app.py

The app will open in your default browser, typically at http://localhost:8501.