1. Definition
- In finance, a return is the gain or loss of an investment over a period, expressed as a fraction of the initial value.
$R_t = \frac{P_t – P_{t-1}}{P_{t-1}}$
where $P_t$ = asset price at time $t$.
- The return distribution describes the probability distribution of returns over time.
- Instead of saying “average return = 5%,” we capture the entire range of possible returns and their probabilities.
2. Key Characteristics of Return Distributions
- Mean (Expected Return): average return investors expect.
- Variance / Volatility: spread of returns → risk measure.
- Skewness: asymmetry (positive skew = large gains possible; negative skew = large losses possible).
- Kurtosis (fat tails): likelihood of extreme events compared to a Normal distribution.
In finance, returns often have fat tails and negative skewness, meaning:
- Extreme losses are more likely than a Normal distribution predicts.
3. Common Models for Return Distributions
Normal Distribution (simplest assumption)
- Early finance theory (e.g., Black-Scholes) assumed returns ~ Normal($\mu,\sigma^2$).
- Easy to work with, but underestimates extreme events (crashes).
Student’s t-Distribution
- Heavy-tailed alternative.
- Better for capturing extreme losses.
Empirical Distributions
- Using historical returns directly to build a distribution (non-parametric).
Mixture Distributions
- Combine several normals to model volatility clustering.
GARCH Models
- Model time-varying volatility of returns → conditional return distributions change over time.
4. Examples
Stock Market
- Average daily return ≈ 0.05% (small).
- Volatility ≈ 1% per day.
- Distribution: centered near 0, but fat-tailed (crashes more likely than Normal).
Portfolio Returns
- Built from weighted sum of individual asset return distributions.
- Diversification shapes the overall distribution.
5. Why Return Distributions Matter
- Risk Management:
- Need distribution to compute VaR and CVaR (Expected Shortfall).
- Option Pricing:
- Payoffs depend on full return distribution, not just average.
- Portfolio Optimization:
- Balancing expected return vs risk (variance, skewness, kurtosis).
- Stress Testing:
- Assess likelihood of extreme losses.
6. Visualization (Intuition)
- Normal curve: symmetric, thin tails.
- Actual stock returns: fatter tails, slightly negative skew (losses more extreme than gains).
- This mismatch explains why “crash risk” is higher than classical models predict.
Summary:
A return distribution describes the probability distribution of investment returns. Key features are mean, variance, skewness, and kurtosis. While the Normal distribution is the simplest assumption, real-world returns often show fat tails and asymmetry, requiring heavy-tailed or time-varying models (t-distribution, GARCH, empirical distributions). They are essential for risk management, option pricing, and portfolio optimization.
