Understanding Black-Scholes Model Assumptions

The Black-Scholes Option Pricing Model is a fundamental concept in financial mathematics used to estimate the theoretical value of European-style options. Like any mathematical model, it relies on several key assumptions about the market and the behavior of the underlying asset. Understanding these assumptions is crucial for applying the model correctly and appreciating its limitations.

Key Assumption: Underlying Asset Price Distribution

One of the most critical assumptions in the Black-Scholes model concerns how the price of the underlying asset (like a stock) behaves over time. The model assumes a specific type of random movement for these prices.

Let's examine the options provided regarding the distribution of underlying asset prices:

• Equi-distributed: This term is not standard terminology used in finance to describe asset price movements in models like Black-Scholes. An equi-distribution implies all outcomes are equally likely, which doesn't reflect the dynamic and volatile nature of stock prices.
• Marginally distributed: Marginal distribution refers to the distribution of a subset of variables within a larger probability distribution. While relevant in multivariate statistics, it's not the term used to describe the fundamental process governing the underlying asset's price movement in Black-Scholes.
• Log-normally distributed: This is a key assumption of the Black-Scholes model. It assumes that the logarithm of the underlying asset's price follows a normal distribution. This implies that the asset's returns over continuous time intervals are normally distributed, and prices themselves are always positive, as the exponential of any real number is positive. This characteristic makes log-normal distribution suitable for modeling asset prices, which cannot fall below zero.
• Normally distributed: While the Black-Scholes model assumes that the continuous returns are normally distributed, it does *not* assume that the asset prices themselves are normally distributed. If prices were normally distributed, they could theoretically become negative, which is impossible for a stock price.

Therefore, the assumption in the Black-Scholes Option Pricing Model is that the underlying asset prices are log-normally distributed.

Why Log-Normal Distribution for Asset Prices?

The log-normal distribution is preferred over the normal distribution for modeling asset prices in the Black-Scholes model for several reasons:

• Positivity: Stock prices cannot be negative. A log-normal distribution naturally ensures that the price is always positive. A normal distribution, in contrast, allows for negative values.
• Compounding Returns: The model assumes continuous compounding. If continuous returns are normally distributed, the terminal price (which is the exponential of the sum of continuous returns) will be log-normally distributed.
• Matches Empirical Data: While not perfectly true in reality (e.g., asset price movements often exhibit "fat tails"), the log-normal distribution is a reasonable approximation for asset price behavior over moderate time horizons in financial modeling.

In summary, the Black-Scholes model models the continuous returns as normally distributed, which leads to the assumption that the underlying asset's price follows a log-normal distribution.

Revision Table: Black-Scholes Assumptions Overview

Additional Information: Log-Normal Distribution

A variable \(X\) is log-normally distributed if \(\ln(X)\) is normally distributed. If \(Y = \ln(X)\) and \(Y \sim N(\mu, \sigma^2)\), then \(X\) is log-normally distributed. The probability density function of a log-normal distribution is defined for \(x > 0\). This property of being strictly positive makes it suitable for modeling quantities that cannot be negative, such as asset prices.

In the context of Black-Scholes, if \(S_t\) is the asset price at time \(t\), the model assumes that \(\ln(S_t/S_0)\) (which represents the total continuous return from time 0 to \(t\)) is normally distributed. This is equivalent to saying that the price \(S_t\) is log-normally distributed, given the initial price \(S_0\).