Quantitative analysis forms 20% of the FRM Part 1 exam. Strong quantitative foundations are also essential for understanding risk models in Part 2.
Key Statistical Concepts
Probability Distributions
- Normal Distribution: Bell-shaped, symmetric, fully defined by mean and variance
- Lognormal Distribution: Used for asset prices (can't go negative)
- Student's t-Distribution: Heavier tails than normal, used with small samples
- Chi-Squared Distribution: Used in hypothesis testing of variance
- Poisson Distribution: Models the frequency of events (e.g., operational risk)
Moments of a Distribution
- Mean: Central tendency
- Variance/Std Dev: Dispersion — how spread out the data is
- Skewness: Asymmetry — negative skew means left tail is longer
- Kurtosis: Tail heaviness — excess kurtosis > 0 means fatter tails than normal
Hypothesis Testing
Key Steps
- State null and alternative hypotheses
- Choose significance level (α)
- Calculate test statistic
- Compare to critical value or p-value
- Make decision (reject or fail to reject H₀)
Common Tests
- t-test: Testing means with unknown population variance
- F-test: Testing equality of variances or overall regression significance
- Chi-squared test: Testing variance or independence
Type I and Type II Errors
- Type I (α): Rejecting H₀ when it's true (false positive)
- Type II (β): Failing to reject H₀ when it's false (false negative)
- Power = 1 - β: Probability of correctly rejecting a false H₀
Regression Analysis
Linear Regression
- Y = α + βX + ε
- β = Cov(X,Y) / Var(X)
- R² measures proportion of variance explained
- Assumptions: linearity, independence, normality, homoscedasticity
Multiple Regression
- Y = α + β₁X₁ + β₂X₂ + ... + ε
- Adjusted R² penalizes for additional variables
- F-test for overall significance, t-tests for individual coefficients
- Watch for multicollinearity
Time Series Analysis
No autocorrelation
A stationary process has constant mean, variance, and autocovariance over time. Most financial time series are non-stationary.
AR, MA, and ARMA Models
- AR(1): Y_t = φY_{t-1} + ε_t (today depends on yesterday)
- MA(1): Y_t = ε_t + θε_{t-1} (today depends on yesterday's shock)
- ARMA: Combines both
GARCH Models
For modeling time-varying volatility (volatility clustering):
- σ²_t = ω + α×r²_{t-1} + β×σ²_{t-1}
- Essential for VaR estimation and risk modeling
Practice quantitative analysis with our extensive question bank!