• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

Handling log returns

Kuroda

Kuroda

...
Joined
10/4/11
Messages
2
Points
11
Hello everybody, I would like to ask some question about how would you handle log returns, since I'm getting "trickied" by my results.
Suppose you got monthly log returns for, say 10 years for a total dataset of 120.
  • Trying to get an annualized return for the - say - industry: I took the exponential of the entire dataset, then computed the product out all of them to get cumulative result. The CumulRet was then adjusted: (CumulRet^(1/T))-1 to get the annualized result. T=10
  • Standard Dev: Adjusted the STDev multipling it by SQRT(T). T=10
Hoping the procedure is right, what do you think about setting T=years? Is it correct? I felt unconfortable because of some crazy results got during computations... I started wondering thereafter if I was doing good or not... Or need the monthly data to be turned first into yearly ones prior to every calculation?
What do you also think about computing other statistics, Kurt./Skew., would they need to be computed on specifically adjusted returns?
Thanks..
 
Please help me with this simple math problem:

A trading account has the following annual returns in a 3-year period:

50% first year, 100% second year and 100% in the final year.

What is the formula to calculate total percentage return?

total = A% * B% * C%

am I wrong ?
 
Kuroda:
I'm a little unclear as to what you're doing, but I will say this. Log returns generally refer to doing the calculation ln(S(i)/S(i-1)), where S(i) is the level of a given asset at time i, to represent the return of the asset between i-1 and i. It's used because of the nice properties of continuously compounded interest for options theory. To calculate the annualized volatility of your data set (again coming at this from an options point of view), you'd square all those log returns, sum them up, multiply by 252/(# of returns), then take the square root of the result.

ExSan:
It's total = (1+A%)*(1+B%)*(1+C%) - 1
 
Back
Top