Project Overview

This project uses a GARCH(1,1) model to forecast the future daily volatility of the Nasdaq Composite Index (^IXIC). The goal is to estimate how much price movement (volatility) to expect in the coming days, which is useful for risk management, trading strategies, and options analysis.

The project includes:

  • Downloading daily Nasdaq data from Yahoo Finance
  • Calculating log returns
  • Fitting a GARCH(1,1) model
  • Forecasting 30 days of volatility
  • Comparing model forecasts with historical rolling volatility
  • Wrapping the workflow into a reusable Python class

Key Outputs

✅ Nasdaq Price vs. Rolling Volatility (Last 250 Days)

  • Rolling 30-day volatility spikes during market drops.
  • Calm price trends are usually matched by low volatility.

price_vs_volatility

✅ 30-Day Forecasted Daily Volatility (GARCH)

The GARCH model captures persistence in volatility and forecasts a slight increase in the coming days:

garch_forecast

📊 Simulating Future Price Paths

Using the GARCH forecast, we simulated 500 possible future price paths to visualize the uncertainty over a 30-day horizon.

  • Price paths are based on cumulative log returns with volatility scaling.
  • The average path stays relatively flat (no return drift).
  • The shaded region represents the 90% confidence band (5th to 95th percentile).

simulated_paths

📈 Trade Signal Generation

We compared recent actual prices to the simulated volatility cone:

  • Buy if actual price falls below the 5th percentile path.
  • Sell if it exceeds the 95th percentile.
  • Hold if it stays within.

This created a basic statistical trading rule based on extreme deviation from expected price movement.

trade_signals

Insights

  • ✅ Most prices stayed inside the 90% cone - a sign that the GARCH forecast captured realistic risk.
  • 📉 Only a few buy/sell signals were triggered - consistent with conservative statistical strategies.
  • 🧠 Confidence intervals expand with time - showing how uncertainty grows in longer forecasts.
  • 🔄 The approach is symmetric and agnostic to direction - it only reacts to large deviations, not trends.

How GARCH Helps in Trading

Volatility doesn’t predict direction, but it helps shape trading decisions:

  • 📏 Position sizing: Reduce exposure in high-volatility environments
  • 🧠 Risk management: Anticipate large moves during volatile periods
  • 💸 Options trading: Use volatility expectations to choose strike prices or entry points
  • ⚡ Breakout filters: Combine price and volatility to avoid whipsaws

Code Structure

The full workflow is encapsulated in a class:

model = GARCHForecaster("^IXIC")
model.download_data()
model.fit_model()
model.forecast_volatility(horizon=30)
model.plot_forecast()

🔗 View notebook and source code on GitHub