Girsanov Theorem

Hey there students! 👋 Today we're diving into one of the most powerful tools in mathematical finance - the Girsanov Theorem. This theorem is like having a mathematical superpower that lets us transform how we view random processes, especially when dealing with financial markets. By the end of this lesson, you'll understand how to change probability measures, work with Radon-Nikodym derivatives, and see why this theorem is absolutely crucial for pricing financial derivatives. Get ready to unlock the secret behind risk-neutral pricing! 🚀

Understanding Probability Measures and Why We Need to Change Them

Before we jump into the Girsanov Theorem itself, let's understand what we mean by a "probability measure" and why changing it matters in finance. Think of a probability measure as a way of assigning probabilities to different outcomes in our mathematical world.

Imagine you're watching a stock price move over time. In the real world (what we call the "physical measure" or "P-measure"), the stock might have an expected return of 8% per year because investors demand compensation for risk. However, when pricing derivatives like options, we often want to work in a "risk-neutral world" (the "Q-measure") where all assets have the same expected return as the risk-free rate, say 3%.

This is where the magic happens! 🎭 The Girsanov Theorem allows us to mathematically transform our view from the real world to this risk-neutral world. It's like having special glasses that let you see the same random process from a completely different perspective.

The key insight is that while the probabilities change, the actual paths of the process remain the same. It's similar to how the same movie looks different when viewed through different colored filters - the story is the same, but your perception changes.

The Mathematics Behind Measure Changes

Now let's get into the mathematical meat of the theorem! The Girsanov Theorem deals with something called a Radon-Nikodym derivative, which sounds scary but is actually quite intuitive.

The Radon-Nikodym derivative, often denoted as $\frac{dQ}{dP}$, tells us how to convert probabilities from one measure to another. Think of it as an exchange rate between two different probability "currencies." If we have a random variable $X$ under measure $P$, its expectation under measure $Q$ is:

$$E^Q[X] = E^P\left[X \cdot \frac{dQ}{dP}\right]$$

For the Girsanov transformation, this derivative has a very specific form. If we want to change the drift of a Brownian motion $W_t$ by adding a deterministic function $\theta(t)$, the Radon-Nikodym derivative is:

$$\frac{dQ}{dP} = \exp\left(-\int_0^T \theta(s) dW_s - \frac{1}{2}\int_0^T \theta(s)^2 ds\right)$$

This exponential form might look complex, but it has beautiful properties. The first integral represents the "signal" we're adding, while the second integral is a "correction term" that ensures our new measure is still a valid probability measure.

The Girsanov Theorem in Action

Here's where everything comes together beautifully! 🌟 The Girsanov Theorem states that if $W_t$ is a Brownian motion under measure $P$, and we define our Radon-Nikodym derivative as above, then under the new measure $Q$:

$$\tilde{W}_t = W_t + \int_0^t \theta(s) ds$$

is a Brownian motion under $Q$. This means we've successfully changed the drift of our original Brownian motion!

Let's see this in action with a practical example. Suppose we have a stock price following the stochastic differential equation:

$$dS_t = \mu S_t dt + \sigma S_t dW_t$$

where $\mu$ is the expected return (say 8%) and $\sigma$ is the volatility (say 20%). Under the physical measure, this stock has drift $\mu$.

Now, we want to price an option on this stock. Using the Girsanov Theorem, we can transform to a risk-neutral measure where the stock's expected return equals the risk-free rate $r$ (say 3%). We set $\theta(t) = \frac{\mu - r}{\sigma}$, and under the new measure $Q$:

$$dS_t = r S_t dt + \sigma S_t d\tilde{W}_t$$

The stock now has the risk-free return as its drift, making option pricing much more tractable!

Risk-Neutral Measure Transformations in Practice

The transformation to risk-neutral measures is fundamental to modern derivative pricing. In the famous Black-Scholes model, this transformation allows us to price options without knowing the actual expected return of the underlying stock - pretty amazing, right? 🤯

Real-world applications include:

Options Pricing: Every time you see an option price quoted in the market, it's been calculated using risk-neutral measures. The Chicago Board Options Exchange processes millions of option contracts daily, all priced using these principles.

Interest Rate Derivatives: Banks use Girsanov transformations to price complex interest rate products like caps, floors, and swaptions. The global derivatives market, worth over $600 trillion notional, relies heavily on these mathematical tools.

Credit Risk Models: When banks assess the probability of default for loans, they often need to transform between real-world probabilities (for risk management) and risk-neutral probabilities (for pricing).

The beauty of the Girsanov Theorem is that it preserves the volatility structure of the original process. This means that while we change the expected direction of movement, the randomness and uncertainty remain exactly the same.

Technical Conditions and Practical Considerations

For the Girsanov Theorem to work properly, we need certain technical conditions to be satisfied. The most important is the Novikov condition, which ensures that our Radon-Nikodym derivative is well-behaved:

$$E^P\left[\exp\left(\frac{1}{2}\int_0^T \theta(s)^2 ds\right)\right] < \infty$$

This condition prevents our transformation from creating mathematical monsters! In practical terms, it means the drift change can't be too extreme.

Another important aspect is that the Girsanov transformation is reversible. If we can go from measure $P$ to measure $Q$, we can also go back from $Q$ to $P$ using the inverse transformation. This symmetry is crucial for consistency in financial modeling.

Conclusion

The Girsanov Theorem is truly one of the crown jewels of mathematical finance! 💎 We've seen how it allows us to change probability measures by transforming the drift of stochastic processes while preserving their volatility structure. Through the Radon-Nikodym derivative, we can move between the real-world measure (where assets have risk premiums) and the risk-neutral measure (where all assets earn the risk-free rate). This transformation is the mathematical foundation that makes modern derivative pricing possible, from simple stock options to complex structured products. The theorem's elegance lies in its ability to simplify complex pricing problems by changing our perspective on randomness itself.

Study Notes

• Girsanov Theorem Purpose: Allows changing probability measures for stochastic processes, particularly transforming drift while preserving volatility

• Radon-Nikodym Derivative: $\frac{dQ}{dP} = \exp\left(-\int_0^T \theta(s) dW_s - \frac{1}{2}\int_0^T \theta(s)^2 ds\right)$ - the "exchange rate" between probability measures

• Measure Transformation: Under new measure $Q$, $\tilde{W}_t = W_t + \int_0^t \theta(s) ds$ becomes a Brownian motion

• Risk-Neutral Transformation: Change from physical measure (real-world returns) to risk-neutral measure (risk-free returns) for derivative pricing

• Key Formula: $E^Q[X] = E^P\left[X \cdot \frac{dQ}{dP}\right]$ - how to compute expectations under the new measure

• Novikov Condition: $E^P\left[\exp\left(\frac{1}{2}\int_0^T \theta(s)^2 ds\right)\right] < \infty$ - ensures the transformation is mathematically valid

• Practical Application: Stock price $dS_t = \mu S_t dt + \sigma S_t dW_t$ becomes $dS_t = r S_t dt + \sigma S_t d\tilde{W}_t$ under risk-neutral measure

• Volatility Preservation: The Girsanov transformation changes drift but keeps the volatility structure unchanged

• Reversibility: Transformations can go both ways - from $P$ to $Q$ and back from $Q$ to $P$

• Market Applications: Used in options pricing, interest rate derivatives, credit risk models, and all modern derivative pricing frameworks