I apologize for not being very clear. Here is my example that you might be
able to help me out.

Q1 return - 3%
Q2 return - 2%
Q3 return - (1.5%)
Q4 return - 1%

Simply adding the returns for each quarter together doesn't provide an
accurate yearly return because it doesn't take into consideration the
quarterly compounding as in this example. Without more data than this, is
there way to calculate the compound rate of return for the year.

"joeu2004" wrote:

On Jul 26, 2:42 pm, Biff wrote:
I am trying to take returns for each quarter of a year and calculate the
compounded return for the year. I tried using the XIRR function, but that
isn't returning me a correct value.

What do you mean by "correct value"? That is, what are your data;
what is the "correct value" you expect; and where did the "correct
value" come from?
Finally, what are your parameters to the XIRR() function?

If you provide the specifics, we can provide more specific answers.
But perhaps the following generalizations will be of some use.

First, there are many definitions for (or views on) how to compute
annual return rates. Second, there are several views on how to
annualize a sub-annual rate. From your ealier post, it is clear that
you want to deal with the geometric mean; but the source of the
"correct value" might not have.

Finally, even taking all that into account, I have never been able to
duplicate the annual rates reported in prospectuses and stock quotes
-- although sometimes I come close. One explanation might be: there
are at least two definitions of "returns", namely FMV and "total
return", not to mention different defintions of return rates (simple,
arithmetic mean and geometric mean). Often we do not have access to
the historical "total returns" on a per-period bases.

Okay, let's assume that you have quarter-to-quarter return rates; i.e.
r1 = q2/q1-1, r2 = q3/q2-1, etc, for consecutive quarters q1, q2, q3,
etc.

In some respects, XIRR is the more accurate computation of the
annualized compounded return rate. But it is usually more accurate
that the common method of annualizing a periodic IRR by compounding;
and many people to do not compound the periodic IRR at all (sigh).

By now, you know how to compute the geometric mean of consecutive
quarterly rates. One method: g = GEOMEAN(r1+1, r2+1,...) - 1. The
compounded annualized rate is typically (1+g)^4 - 1, or GEOMEAN(...)^4
- 1. The non-compounded annualize rate is 4*g or 4*(GEOMEAN(...)-1).

The compounded annualized rate computed in the manner above should be
comparable to the XIRR(), though not exact. But the arguments to the
XIRR() function would not be the return rates (r1, r2, etc).

Hope this helps get a leg-up on the problem. Again, if you provide
specific numbers for the questions above, we can better address your
question.