Jessica Wachter Notes for Finance 604
How is it that the arithmetic average is 25% and the geometric is 0%? Although we know
geometric average is the correct HPR measure, it is useful to provide a numerical example
showing why the return over 2 years is 0%. Let’s see what happens to $100 in this fund:
•1st year (you earn 100%): 100 →200.
•2nd year (you lose 50%): 200 →100.
•Thus, you earn a return of zero over the two- year period.
Does it make any difference whether we experienced the 100% first or second? No:
•1st year (you lose 50%): 100 →50.
•2nd year (you earn 100%): 50 →100.
•Again, you earn a return of zero over the two- year period.
In the first period, you halve your wealth, and in the second you double it. You are back
where you started at the end of the second period. Once again, the arithmetic average
gives you 25%, but the geometric average gives you zero.
It can make a big difference whether you look at geometric or arithmetic average. If you
want to know what return investors in the mutual fund actually earned, you should pay
attention to the geometric average.
When might you be interested in arithmetic average?
There are two cases where it is correct to use the arithmetic mean (or average):
1. Simple interest: if you don’t reinvest your earnings in the fund.
•Thus, you start each year with the same amount of money, e.g. $100.
•In the previous example:
interest = 100 1st period + −50 2nd period.
•Over two periods, interest was $50 ⇒equivalent to earning 25% each period.
•Generally, we prefer geometric average because it tells us how an initial sum
grows ’untouched by human hands’. But if you like to add and subtract at the
end of each year to maintain the same dollar investment (you probably won’t
like the adding part), then arithmetic mean tells the truth.
2. As a statistician: if you are calculating the average return on this fund using
historical data, then you would most likely look at monthly returns over five or ten
years. In this case, the arithmetic mean would be your estimate of the average return.
3