Multivariate Random Variables

Welcome to Multivariate Random Variables!

In previous chapters, we looked at single random variables (like the return on one stock). But in the real world of finance, assets don't live on islands! They interact with each other. If the tech sector crashes, it's likely that both Apple and Microsoft will drop. Multivariate Random Variables is simply the study of how two or more variables behave together. Understanding this is the secret sauce for building diversified portfolios and managing risk.

Don’t worry if this seems a bit abstract at first. We will use simple analogies and step-by-step breakdowns to make it click!

1. Joint Probability Distributions

When we look at two variables at once (let's call them \( X \) and \( Y \)), we use a Joint Probability Distribution. It tells us the probability that \( X \) takes a certain value and \( Y \) takes a certain value at the same time.

Discrete Joint Distributions

Think of this as a simple table. Imagine \( X \) is the change in Interest Rates (Up or Down) and \( Y \) is the change in Stock Prices (Up or Down). A joint probability would be the number in the cell where "Interest Rates Up" and "Stock Prices Down" meet.

The Golden Rule: The sum of all probabilities in a joint distribution table must always equal 1.0.

Continuous Joint Distributions

For continuous variables, we use a Joint Probability Density Function (PDF), denoted as \( f(x, y) \). Instead of a flat table, imagine a 3D mountain. The total "volume" under this mountain must equal 1. To find the probability of a range, we would technically use calculus (double integrals), but for the FRM, focus on the conceptual understanding: the total area/volume is always 100%.

Quick Review: Joint probability is just the "AND" probability—the chance of two things happening together.

2. Marginal and Conditional Distributions

Sometimes we have the big picture (the joint distribution) but we only want to focus on one specific part.

Marginal Probability

Marginal Probability is the probability of one variable occurring regardless of the other.
Simple Trick: Think of a probability table. If you add up all the values in a single row or a single column and write the total in the "margin" of the paper, you’ve found the marginal probability!

\( P(X = x) = \sum P(X=x, Y=y) \) (for all possible \( y \))

Conditional Probability

This is the probability of \( X \) happening given that we already know \( Y \) has happened. We write this as \( P(X|Y) \).

The Formula: \( P(X|Y) = \frac{P(X, Y)}{P(Y)} \)

Analogy: Imagine a room of people. The probability that someone is wearing a coat (X) given that it is snowing outside (Y). Knowing it is snowing changes your "base" population, making it much more likely they are wearing a coat.

Key Takeaway: Marginal = The total for one variable. Conditional = The probability updated with new information.

3. Independence of Random Variables

Two variables are Independent if the occurrence of one tells you absolutely nothing about the occurrence of the other.
In math terms, \( X \) and \( Y \) are independent if:
1. \( P(X, Y) = P(X) \times P(Y) \)
2. \( P(X|Y) = P(X) \)

Common Mistake: Students often confuse "Mutually Exclusive" with "Independent." If two events are mutually exclusive (like a coin landing on Heads vs. Tails), they are not independent—in fact, they are highly dependent because if one happens, the other cannot happen!

4. Covariance and Correlation

This is a core topic for the FRM. We need to measure how variables move together.

Covariance

Covariance measures the direction of the linear relationship between two variables.

• Positive Covariance: They move in the same direction.
• Negative Covariance: They move in opposite directions.
• Zero Covariance: No linear relationship.

The formula for Covariance is:
\( Cov(X, Y) = E[(X - \mu_X)(Y - \mu_Y)] \)
Or the easier "computational" version:
\( Cov(X, Y) = E(XY) - E(X)E(Y) \)

Correlation (\( \rho \))

Covariance is hard to interpret because it depends on the units (e.g., dollars vs. percentages). Correlation standardizes this into a number between -1 and +1.

The Formula: \( \rho_{XY} = \frac{Cov(X, Y)}{\sigma_X \sigma_Y} \)

Did you know? A correlation of +1 means the two variables move in perfect lockstep. A correlation of -1 means they move perfectly opposite. A correlation of 0 means there is no linear relationship.

Key Takeaway: Correlation is just "scaled" covariance. It makes it easier to compare the relationship between different pairs of assets.

5. Expectations and Variances of Combined Variables

In finance, we often create portfolios (sums of variables). You need to know how to calculate the expected return and risk of these combinations.

Expected Value of a Sum

This one is easy! The expected value of a sum is just the sum of the expected values:
\( E(aX + bY) = aE(X) + bE(Y) \)

Variance of a Sum (The "Risk" Formula)

This is very important for the exam. When you combine two variables, the variance isn't just the sum of their variances—you have to account for their covariance!
\( Var(aX + bY) = a^2 Var(X) + b^2 Var(Y) + 2ab Cov(X, Y) \)

Since \( Cov(X, Y) = \rho_{XY} \sigma_X \sigma_Y \), we can also write it as:
\( Var(aX + bY) = a^2 \sigma_X^2 + b^2 \sigma_Y^2 + 2ab \rho_{XY} \sigma_X \sigma_Y \)

Memory Aid: Think of the algebraic expansion of \( (a+b)^2 = a^2 + b^2 + 2ab \). The variance formula follows the same pattern, just with the "Covariance" term attached to the \( 2ab \) part!

6. The Bivariate Normal Distribution

A Bivariate Normal Distribution describes two variables that are both normally distributed. To fully describe this relationship, you only need 5 parameters:
1. Mean of X (\( \mu_X \))
2. Mean of Y (\( \mu_Y \))
3. Variance of X (\( \sigma_X^2 \))
4. Variance of Y (\( \sigma_Y^2 \))
5. Correlation between them (\( \rho \))

Key Property to Remember:

For most distributions, "Zero Correlation" does not necessarily mean the variables are independent. However, for the Normal Distribution, if the correlation is zero, the variables are independent. This is a favorite "trick" question on the exam!

Quick Summary: Multivariate variables help us understand the "co-movement" of assets. By mastering Joint, Marginal, and Conditional probabilities, and learning the Variance of a Sum formula, you are well on your way to mastering portfolio theory!