Random Variables

FRM Part I · FRM · 7 min read · Updated: 9 Nov 2026

Welcome to the World of Random Variables!

Welcome! If you are starting your FRM journey, the chapter on Random Variables is one of your most important milestones. Why? Because in finance, we never know exactly what a stock price or an interest rate will be tomorrow. We deal with uncertainty every day. Random variables give us a mathematical way to "package" that uncertainty so we can measure risk. Don't worry if math isn't your favorite subject—we are going to break this down step-by-step using simple logic and real-world examples.

1. What Exactly is a Random Variable?

Think of a Random Variable (RV) as a "placeholder" or a "container" for the possible outcomes of an experiment that hasn't happened yet. We usually denote them with capital letters like $ X $ or $ Y $.

Example: Imagine you are tossing a coin. You don't know if it will be Heads or Tails. If we say $ X = 1 $ if it's Heads and $ X = 0 $ if it's Tails, then $ X $ is our Random Variable.

Discrete vs. Continuous Random Variables

This is the first big distinction you need to make:

• Discrete Random Variables: These have "gaps" between values. You can count the outcomes (e.g., 1, 2, 3...).
Analogy: Think of a staircase. You can stand on step 1 or step 2, but not on step 1.5.
Finance Example: The number of times a company defaults on a loan in a year (0, 1, 2...).

• Continuous Random Variables: These can take any value within a range. There are no gaps.
Analogy: Think of a slide. You can be at any height at any time.
Finance Example: The exact daily return of the S&P 500 (e.g., 1.25%, 1.251%, 1.2515%...).

Quick Review: If you can count it on your fingers, it’s likely Discrete. If you have to measure it precisely with a scale or a clock, it’s likely Continuous.

2. Describing Probabilities: PMF, PDF, CDF, and the Quantile Function

How do we know how likely an outcome is? We use specific functions.

For Discrete Variables: The PMF

The Probability Mass Function (PMF), written as $ P(X = x) $, tells us the exact probability that the variable equals a specific value.
Rule 1: Every individual probability must be between 0 and 1, so $ 0 \leq P(X = x) \leq 1 $.
Rule 2: The sum of all probabilities across all possible outcomes must equal 1: $ \sum P(X = x) = 1 $.

For Continuous Variables: The PDF

The Probability Density Function (PDF), written as $ f(x) $, is different. For continuous variables, the probability of hitting an exact single real number (like exactly 1.000000%) is actually zero: $ P(X = x) = 0 $. Instead, we find probabilities over an interval by taking the area under the curve between two points: $ P(a \leq X \leq b) = \int_{a}^{b} f(x)\,dx $.

The Cumulative Distribution Function (CDF)

The Cumulative Distribution Function (CDF), written as $ F(x) $, is like a "running total." It tells us the probability that the random variable is less than or equal to a certain value: $ F(x) = P(X \leq x) $.
Properties of the CDF: It is non-decreasing, starts at 0 (as $ x \to -\infty $), and reaches 1 (as $ x \to \infty $). For continuous variables, the derivative of the CDF is the PDF: $ f(x) = F'(x) $.

The Quantile Function (Inverse CDF)

The Quantile Function, written as $ Q(p) = F^{-1}(p) $ where $ 0 < p < 1 $, works in the exact opposite direction of the CDF. While the CDF takes a value $ x $ and gives you a probability $ p $, the quantile function takes a probability $ p $ and returns the threshold value $ x_p $ such that $ P(X \leq x_p) = p $.
Finance Application: This is the foundational concept behind Value at Risk (VaR), where risk managers solve for the cutoff loss at a given confidence level (such as $ p = 0.05 $ or $ p = 0.01 $).

3. Expected Value: The "Center" of the Data

The Expected Value, denoted as $ E(X) $ or $ \mu $, is the long-run average or the first population moment of the random variable. It is the "center of gravity."

For a discrete variable, you calculate it by multiplying each outcome by its probability and adding them up:
$ E(X) = \sum [x_i \cdot P(X = x_i)] $

For a continuous variable, it is computed via integration:
$ E(X) = \int_{-\infty}^{\infty} x \cdot f(x)\,dx $

Example: If there is a 50% chance of making \$10 and a 50% chance of making \$0, your expected value is:
$ (10 \cdot 0.5) + (0 \cdot 0.5) = \$5 $.

Did you know? The expected value doesn't have to be a value that is actually possible in a single trial! In the coin flip example above, you never receive \$5 in a single try; you either get \$10 or \$0. The expected value represents the average over many independent repetitions.

4. Variance and Standard Deviation: Measuring the "Risk"

In the FRM, risk is often defined as uncertainty or dispersion. Variance and Standard Deviation tell us how much the outcomes "spread out" from the mean.

Variance ($ \sigma^2 $): The second central moment, defined as the average of the squared deviations from the mean:
$ Var(X) = E[(X - \mu)^2] = E(X^2) - [E(X)]^2 $
A higher variance indicates greater dispersion and uncertainty.

Standard Deviation ($ \sigma $): The positive square root of the variance: $ \sigma = \sqrt{Var(X)} $. We use standard deviation frequently because it is measured in the exact same units as the underlying variable (such as dollars or percentage points).

5. Skewness and Kurtosis: The Shape of Risk

Beyond the mean and variance, we look at the standardized higher moments that define the distribution's shape.

Skewness (3rd Standardized Moment): Measures asymmetry about the mean:
$ Skew(X) = \frac{E[(X - \mu)^3]}{\sigma^3} $
• Symmetric Distribution: $ Skew = 0 $ (e.g., standard normal distribution).
• Positive Skew: Long tail to the right (frequent modest losses, occasional huge gains).
• Negative Skew: Long tail to the left (frequent modest gains, occasional catastrophic losses—the classic profile of credit risk and short option strategies!).

Kurtosis (4th Standardized Moment): Measures tail heaviness and peakedness:
$ Kurt(X) = \frac{E[(X - \mu)^4]}{\sigma^4} $
• A normal distribution has a kurtosis of exactly 3.
• Excess Kurtosis: Defined as $ Kurtosis - 3 $.
• Fat Tails (Leptokurtic): Distributions with excess kurtosis greater than 0. Extreme outlier events occur much more frequently than predicted by a normal distribution, making tail-risk modeling critical for risk managers.

6. Linear Transformations: Effect on Location, Scale, and Shape

Suppose we transform a random variable $ X $ into $ Y = aX + b $, where $ a $ and $ b $ are constants. Understanding how this linear shift affects each summary metric is essential for the exam:

• Mean: Shifting and scaling affect the mean directly:
$ E(aX + b) = aE(X) + b $

• Variance: Adding a constant $ b $ does not change spread, while scaling by $ a $ squares the factor:
$ Var(aX + b) = a^2 Var(X) $

• Standard Deviation: Taking the square root of the variance yields the absolute value of $ a $:
$ \sigma(aX + b) = |a|\sigma(X) $

• Median: Like the mean, the median tracks both the scale and location shift:
$ Median(aX + b) = a \cdot Median(X) + b $

• Interquartile Range (IQR): The IQR is the distance between the 75th and 25th percentiles ($ Q_3 - Q_1 $). Adding a constant shifts both percentiles equally, leaving the gap unchanged, while scaling scales the distance:
$ IQR(aX + b) = |a| \cdot IQR(X) $

• Skewness: Location shift $ b $ has no effect. Scaling by $ a > 0 $ leaves skewness unchanged, but scaling by $ a < 0 $ reverses the direction of the tail:
$ Skew(aX + b) = \text{sign}(a) \cdot Skew(X) $

• Kurtosis: Standardized by $ \sigma^4 $, kurtosis is completely invariant to both location and scale transformations (as long as $ a \neq 0 $):
$ Kurt(aX + b) = Kurt(X) $

Don't Panic! Quick Summary for the Exam

• Discrete = Countable outcomes (PMF); Continuous = Infinite continuum, probability over an interval (PDF).
• CDF gives $ P(X \leq x) $; the Quantile Function is its inverse, returning the value $ x $ for a specified cumulative probability $ p $.
• Expected Value: $ E(aX + b) = aE(X) + b $.
• Dispersion: $ Var(aX + b) = a^2Var(X) $ and $ \sigma(aX + b) = |a|\sigma(X) $; $ IQR(aX + b) = |a|IQR(X) $.
• Shape: Kurtosis is unchanged by linear transformation; Skewness flips sign if $ a < 0 $.
• Excess Kurtosis $ > 0 $ means fat tails (leptokurtic), indicating higher likelihood of extreme market moves.

References

2026 FRM Part I Learning Objectives — Global Association of Risk Professionals
2026 FRM Candidate Guide — Global Association of Risk Professionals

Quick check

Can you answer these now?

Open each question to check the key ideas from this chapter.

What is a random variable, and how are random variables fundamentally categorized?

A random variable is a variable whose possible values are numerical outcomes of a random phenomenon. It is categorized as either discrete (countable outcomes with gaps) or continuous (uncountable outcomes over an interval).

What is the difference between a Probability Mass Function (PMF) and a Probability Density Function (PDF)?

A PMF gives the exact probability $ P(X = x) $ for discrete outcomes, whereas a PDF $ f(x) $ applies to continuous variables where the probability of any single point is zero, requiring integration over an interval to obtain probability.

What is the relationship between the Cumulative Distribution Function (CDF) and the Quantile Function?

The CDF calculates the probability $ p = P(X \le x) $ for a given value $ x $, while the Quantile Function is its inverse $ Q(p) = F^{-1}(p) $, returning the cutoff value $ x_p $ for a given cumulative probability $ p $.

How is the expected value of a random variable defined and calculated for both discrete and continuous cases?

The expected value $ E(X) $ is the probability-weighted average outcome (first moment), calculated as $ \sum [x_i \cdot P(X = x_i)] $ for discrete variables and $ \int_{-\infty}^{\infty} x \cdot f(x)\,dx $ for continuous variables.

How are variance and standard deviation defined for a random variable?

Variance is the second central moment measuring average squared dispersion, $ Var(X) = E[(X - \mu)^2] = E(X^2) - [E(X)]^2 $, and standard deviation is its positive square root, $ \sigma = \sqrt{Var(X)} $.

What does skewness measure, and how do positive and negative skew differ in finance?

Skewness measures distribution asymmetry; positive skew indicates a long right tail with frequent small losses and rare large gains, while negative skew indicates a long left tail with frequent modest gains and rare catastrophic losses.

What does kurtosis measure, and what does excess kurtosis indicate about financial risk?

Kurtosis measures tail heaviness and peakedness (normal distribution = 3); an excess kurtosis greater than 0 defines a leptokurtic distribution with fat tails, indicating a higher probability of extreme outcomes.

How does a linear transformation $ Y = aX + b $ affect the mean, median, variance, and standard deviation?

The mean and median shift to $ aE(X) + b $ and $ a \cdot \text{Median}(X) + b $, while variance scales to $ a^2Var(X) $ and standard deviation scales to $ |a|\sigma(X) $ (both unaffected by $ b $).

How are skewness and kurtosis affected by the linear transformation $ Y = aX + b $?

Kurtosis is completely invariant to linear transformations, while skewness is unchanged by $ b $ and positive scaling $ a > 0 $, but reverses sign if $ a < 0 $ ($ \text{Skew}(aX+b) = \text{sign}(a) \cdot \text{Skew}(X) $).

Ready to test yourself?

Turn these notes into exam-style practice. Get unlimited AI questions on this topic with instant marking and explanations.

Practise This Topic

More FRM Part I chapters

* The content provided by thinka is generated by AI and may not always be accurate or up-to-date. Please use it as a supplementary resource and verify with official materials.

Put These Notes into Practice

Reading the notes is a great start. Now practise with unlimited AI-generated questions and get instant feedback. 100,000+ students are already improving their grades.

Start Practising Now View Pricing

Done reading? Test yourself with AI practice questions

Practise This Topic Now