{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

lec07B

# lec07B - Lecture 7 Building the Tree This lecture shows...

This preview shows pages 1–7. Sign up to view the full content.

Lecture 7: Building the Tree This lecture shows that the model is a reasonably accurate approximation of more realistic dynamics of the underlying security. We also derive expressions for u , d , and r , and consequently q , for a binomial tree which matches the statistical properties of the underlying security. I. Why is the Binomial Model Realistic? A. Log-Normal Model B. Log-Normal Approximation II. Choosing Binomial Model Inputs A. Stock Price Parameters u , d , and q B. Number of Periods n III. Choosing the Number of Periods

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Using the Binomial Model I. Why is the Binomial Model Realistic? SI At first glance, the binomial model seems to do a terrible job describing reality. Stock don’t either return u or d , They can take on any value. SI We’re now going to derive a more realistic model for the stock price dynamics The log-normal model. SI Then we’ll show that the binomial model can approximate the log-normal model arbitrarily well. It’s just a matter of picking the periods short “enough” SI We’ll mention some of the problems with the log- normal model. Bus 35100 Page 2 Robert Novy-Marx
Using the Binomial Model A. Log-Normal Model Denote the stock prices at the end of each year by S t , for t D 0, 1, 2, 3, ... The simple annual returns on the stock is the gross return minus one: r t D R t NUL 1 where R t D DLE S t S t NUL 1 DC1 . The log return (or continuously compounded return) is given by ln R t D ln DC2 S t S t NUL 1 DC3 D ln S t NUL ln S t NUL 1 Bus 35100 Page 3 Robert Novy-Marx

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Using the Binomial Model Example– Stock price last year: \$100; now: \$10. SI What are the simple and log returns? The simple return is 10 NUL 100 100 D NUL 0.9 D NUL 90 % . The continuously compounded return is ln DC2 10 100 DC3 D NUL 2.30 D NUL 230 % . SI Note: with log returns we’re not limited to 100% losses. What is the log return when a company goes bankrupt? The log return over several years is the sum of the annual returns ln S T NUL ln S 0 D ln R T C ... C ln R 1 D T X t D 1 ln R t . Bus 35100 Page 4 Robert Novy-Marx
Using the Binomial Model Log-normal Model Assumptions: These assumptions are approximately consistent with historical stock price data: 1. Log returns are independently distributed. SI I.e., ln R t is independent of ln R s , for t 6D s . 2. Log returns are identically distributed. SI I.e., the probability distribution of ln R t is the same as that of ln R s for all t and s . 3. The stock price evolves continuously. SI I.e., the stock price doesn’t jump. Remark: Assumptions 1 and 2 jointly are often abbreviated by saying that “log-returns are i.i.d.” SI This stands for “independently and identically distributed.” Q: any problems with these assumptions? We’ll spend some time on these problems (and some “fixes”) later in the quarter Bus 35100 Page 5 Robert Novy-Marx

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Using the Binomial Model Mathematical Fact: these assumptions imply that log returns are normally distributed SI I.e., they follow a “Brownian motion”.
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}