XIRR Problem

D

Dkline

I am trying to use XIRR to solve for the rate for a known schedule of dates
and value. That schedule is below. I have solved for the rate using IRR and
also a Goal Seek method to confirm the IRR. I have a spreadsheet to show all
this.

The answer per the IRR and the Goal Seek is -0.009270531%. I also confirmed
this on my HP-12C.

How does XIRR need to be set up to handle this sitation? This worksheet will
be updated monthly and I also need to calculate the ROR from inception, from
start of year, for the month.

Dates XIRR
12/31/2007 $61,445.04
1/7/2008 $4,062.21
2/6/2008 $4,062.02
3/6/2008 $3,921.48
3/31/2008 -$72,906.91
#NUM!
 
R

Ron Rosenfeld

I am trying to use XIRR to solve for the rate for a known schedule of dates
and value. That schedule is below. I have solved for the rate using IRR and
also a Goal Seek method to confirm the IRR. I have a spreadsheet to show all
this.

The answer per the IRR and the Goal Seek is -0.009270531%. I also confirmed
this on my HP-12C.

How does XIRR need to be set up to handle this sitation? This worksheet will
be updated monthly and I also need to calculate the ROR from inception, from
start of year, for the month.

Dates XIRR
12/31/2007 $61,445.04
1/7/2008 $4,062.21
2/6/2008 $4,062.02
3/6/2008 $3,921.48
3/31/2008 -$72,906.91
#NUM!

I don't understand your results or what "situation" you are trying to handle.

Also what "rate" are you trying to compute?

Applying Excel's IRR function to your series of cash flows (not really
appropriate since they are at irregular intervals), I get

=IRR({61445.04;4062.21;4062.02;3921.48;-72906.91}) --> -0.21689688%

very different from your reported result of -0.009270531%

I have no idea what parameters you set for Goal Seek so I can't comment on
that.

XIRR gives a result of -3.37% which is the Annual rate, based on a 365 day
year.

--ron
 
D

Dkline

My IRR function is =IRR(R2C2:R94C2,-0.009).

The Goal Seek proof is the next column running from Row 2 to Row 93 using
the same cash flow and accumulated at the IRR rate.

If it would be helpful, I could email the spreadsheet,
 
J

joeu2004

The answer per the IRR and the Goal Seek is -0.009270531%.

You cannot use IRR on the following example, since the periods are
irregular. Well, what I mean is: GIGO. IRR will give you a result,
but it assumes the cash flows are equally spaced. Yours are not.
How does XIRR need to be set up to handle this sitation?
[....]
Dates XIRR
12/31/2007 $61,445.04
1/7/2008 $4,062.21
2/6/2008 $4,062.02
3/6/2008 $3,921.48
3/31/2008 -$72,906.91

XIRR always provides an annualized rate of return. IRR always
provides a "periodic" rate of return. (Recall: IRR assumes equally
spaced periods.) Since your cash flows are __almost__ monthly, I
would expect (1+IRR(...))^12 - 1 to be very close to the XIRR result.

With your data, I compute the annualized IRR to be -0.025682814, and
XIRR computes -0.033609948. They are equal when rounded to 2 decimal
places. "Close enough for goverment work" ;-)


------ original posting -----
 
J

joeu2004

PS....

I wrote:"
Dates XIRR
12/31/2007 $61,445.04
1/7/2008 $4,062.21
2/6/2008 $4,062.02
3/6/2008 $3,921.48
3/31/2008 -$72,906.91
[....]
With your data, I compute the annualized IRR to be -0.025682814, and
XIRR computes -0.033609948. They are equal when rounded to 2 decimal
places. "Close enough for goverment work" ;-)

For an even better comparison, change the first date to 12/7/2007 and
the last date to 4/6/2008, so that the periods are now almost
monthly. The annualized IRR is -0.02568, and XIRR returns -0.02583 --
almost the same, as expected.

Note: There will always be at least a small difference between the
annualized IRR and XIRR because XIRR the varying number of days
between "monthly" dates (i.e. say day number each month).
 
D

Dkline

The cash flow is the actual history of this investment, When I set up the IRR
function, I have every date, (4/28/2006 through 3/31/2008) amount invested,
surrender value, and account value from start to finish in a column from
start to end in from R2C1:R705, Any cells without values are set to 0 so the
IRR will function properly with the calculation being done for the entire
history of the investment i.e. inception to date.

My proof also matches.

My intent is to able to get the same answer using XIRR instead of using this
brute force method. I'm certain the end result is valid. I'm trying to learn
how to shape the XIRR request to match what I know is right.

joeu2004 said:
PS....

I wrote:"
Dates XIRR
12/31/2007 $61,445.04
1/7/2008 $4,062.21
2/6/2008 $4,062.02
3/6/2008 $3,921.48
3/31/2008 -$72,906.91
[....]
With your data, I compute the annualized IRR to be -0.025682814, and
XIRR computes -0.033609948. They are equal when rounded to 2 decimal
places. "Close enough for goverment work" ;-)

For an even better comparison, change the first date to 12/7/2007 and
the last date to 4/6/2008, so that the periods are now almost
monthly. The annualized IRR is -0.02568, and XIRR returns -0.02583 --
almost the same, as expected.

Note: There will always be at least a small difference between the
annualized IRR and XIRR because XIRR the varying number of days
between "monthly" dates (i.e. say day number each month).
 
J

joeu2004

The cash flow is the actual history of this investment,
When I set up the IRR function, I have every date,
(4/28/2006 through 3/31/2008) [...].
Any cells without values are set to 0 so the

If you are saying that you have an entry for every date (including Sat
and Sun), then IRR will compute a daily rate of return. In that case,
you would annualize it by the formula: (1 + IRR(range,guess))^365 -
1.

You will probably need the "guess" parameter in this case, as I
believe you are doing, based on one of your follow-up postings.

If you do not include Sat and Sun, or if you do not otherwise have
equal cash flows, the IRR result will be misleading.

amount invested, surrender value, and account value
from start to finish

IRR and XIRR deal with cash flows, not level amounts. But it would be
equally misleading to treat the difference in investment value from
time to time as a cash flow. The cash flows are: amounts that you
invest (inflow); interest earned or dividend payments (inflow if
reinvested; outflow otherwise); and ending value.

My intent is to able to get the same answer using XIRR
instead of using this brute force method.

As I explained previously, IRR computes a __periodic__ rate, assuming
that each cash flow represents an equally-space period. To compare
its result with XIRR, you must annualize the IRR result based on the
size of the period. In general:

daily IRR: =(1 + IRR(range,guess))^365 - 1
weekly IRR: =(1 + IRR(range,guess))^52 - 1
multiple months IRR: =(1 + IRR(range,guess))^(12/m) - 1

where "m" is the number of months per period.

I'm trying to learn how to shape the XIRR request
to match what I know is right.

The "shape" of the XIRR model is the same as the IRR model, with the
additional detail of specific dates. You use XIRR only if the cash
flows are irregularly spaced (or you want to bother with annualizing
the IRR result); otherwise, use IRR.

If the XIRR result does not closely match the annualized IRR, the
problem is likely in your model -- the values that you are using as
"cash flows" -- which would adversely effect the IRR result as well as
the XIRR result.

Let me know if this is helpful. Otherwise, I will stop wasting my
time beating a dead horse.


----- original posting -----

The cash flow is the actual history of this investment, When I set up the IRR
function, I have every date,  (4/28/2006 through 3/31/2008) amount invested,
surrender value, and account value from start to finish in a column from
start to end in from R2C1:R705, Any cells without values are set to 0 so the
IRR will function properly with the calculation being done for the entire
history of the investment i.e. inception to date.

My proof also matches.

My intent is to able to get the same answer using XIRR instead of using this
brute force method. I'm certain the end result is valid. I'm trying to learn
how to shape the XIRR request to match what I know is right.



joeu2004 said:
I wrote:"
Dates                   XIRR
12/31/2007      $61,445.04
1/7/2008                          $4,062.21
2/6/2008                          $4,062.02
3/6/2008                          $3,921.48
3/31/2008                       -$72,906.91
[....]
With your data, I compute the annualized IRR to be -0.025682814, and
XIRR computes -0.033609948.  They are equal when rounded to 2 decimal
places.  "Close enough for goverment work" ;-)
For an even better comparison, change the first date to 12/7/2007 and
the last date to 4/6/2008, so that the periods are now almost
monthly.  The annualized IRR is -0.02568, and XIRR returns -0.02583 --
almost the same, as expected.
Note:  There will always be at least a small difference between the
annualized IRR and XIRR because XIRR the varying number of days
between "monthly" dates (i.e. say day number each month).
 
D

Dkline

I appreciate you sticking with me. I'm looking at your reply now - 3:55 PM EDT.

joeu2004 said:
The cash flow is the actual history of this investment,
When I set up the IRR function, I have every date,
(4/28/2006 through 3/31/2008) [...].
Any cells without values are set to 0 so the

If you are saying that you have an entry for every date (including Sat
and Sun), then IRR will compute a daily rate of return. In that case,
you would annualize it by the formula: (1 + IRR(range,guess))^365 -
1.

You will probably need the "guess" parameter in this case, as I
believe you are doing, based on one of your follow-up postings.

If you do not include Sat and Sun, or if you do not otherwise have
equal cash flows, the IRR result will be misleading.

amount invested, surrender value, and account value
from start to finish

IRR and XIRR deal with cash flows, not level amounts. But it would be
equally misleading to treat the difference in investment value from
time to time as a cash flow. The cash flows are: amounts that you
invest (inflow); interest earned or dividend payments (inflow if
reinvested; outflow otherwise); and ending value.

My intent is to able to get the same answer using XIRR
instead of using this brute force method.

As I explained previously, IRR computes a __periodic__ rate, assuming
that each cash flow represents an equally-space period. To compare
its result with XIRR, you must annualize the IRR result based on the
size of the period. In general:

daily IRR: =(1 + IRR(range,guess))^365 - 1
weekly IRR: =(1 + IRR(range,guess))^52 - 1
multiple months IRR: =(1 + IRR(range,guess))^(12/m) - 1

where "m" is the number of months per period.

I'm trying to learn how to shape the XIRR request
to match what I know is right.

The "shape" of the XIRR model is the same as the IRR model, with the
additional detail of specific dates. You use XIRR only if the cash
flows are irregularly spaced (or you want to bother with annualizing
the IRR result); otherwise, use IRR.

If the XIRR result does not closely match the annualized IRR, the
problem is likely in your model -- the values that you are using as
"cash flows" -- which would adversely effect the IRR result as well as
the XIRR result.

Let me know if this is helpful. Otherwise, I will stop wasting my
time beating a dead horse.


----- original posting -----

The cash flow is the actual history of this investment, When I set up the IRR
function, I have every date, (4/28/2006 through 3/31/2008) amount invested,
surrender value, and account value from start to finish in a column from
start to end in from R2C1:R705, Any cells without values are set to 0 so the
IRR will function properly with the calculation being done for the entire
history of the investment i.e. inception to date.

My proof also matches.

My intent is to able to get the same answer using XIRR instead of using this
brute force method. I'm certain the end result is valid. I'm trying to learn
how to shape the XIRR request to match what I know is right.



joeu2004 said:
I wrote:"
Dates XIRR
12/31/2007 $61,445.04
1/7/2008 $4,062.21
2/6/2008 $4,062.02
3/6/2008 $3,921.48
3/31/2008 -$72,906.91
[....]
With your data, I compute the annualized IRR to be -0.025682814, and
XIRR computes -0.033609948. They are equal when rounded to 2 decimal
places. "Close enough for goverment work" ;-)
For an even better comparison, change the first date to 12/7/2007 and
the last date to 4/6/2008, so that the periods are now almost
monthly. The annualized IRR is -0.02568, and XIRR returns -0.02583 --
almost the same, as expected.
Note: There will always be at least a small difference between the
annualized IRR and XIRR because XIRR the varying number of days
between "monthly" dates (i.e. say day number each month).
 
D

Dkline

I finally got it to work. Thanks for your help.

I tried various guesses to see the effect. It takes a few digits before you
see a variance.

XIRR Dates
(61,445.04) 12/31/2007
(4,062.21) 1/7/2008
(4,062.02) 2/6/2008
(3,921.48) 3/6/2008
72,906.91 3/31/2008

ROR Guess
-0.03365756 -0.09
-0.03365756 -0.10
-0.03365756 -0.11
-0.03365756 -0.12
-0.03365756 -0.13
-0.03365756 -0.14
-0.03365756 -0.15
-0.03365757 -0.16
-0.03365757 -0.17
-0.03365756 -0.18
-0.03365756 -0.19
-0.03365756 -0.20


Dkline said:
I appreciate you sticking with me. I'm looking at your reply now - 3:55 PM EDT.

joeu2004 said:
The cash flow is the actual history of this investment,
When I set up the IRR function, I have every date,
(4/28/2006 through 3/31/2008) [...].
Any cells without values are set to 0 so the

If you are saying that you have an entry for every date (including Sat
and Sun), then IRR will compute a daily rate of return. In that case,
you would annualize it by the formula: (1 + IRR(range,guess))^365 -
1.

You will probably need the "guess" parameter in this case, as I
believe you are doing, based on one of your follow-up postings.

If you do not include Sat and Sun, or if you do not otherwise have
equal cash flows, the IRR result will be misleading.

amount invested, surrender value, and account value
from start to finish

IRR and XIRR deal with cash flows, not level amounts. But it would be
equally misleading to treat the difference in investment value from
time to time as a cash flow. The cash flows are: amounts that you
invest (inflow); interest earned or dividend payments (inflow if
reinvested; outflow otherwise); and ending value.

My intent is to able to get the same answer using XIRR
instead of using this brute force method.

As I explained previously, IRR computes a __periodic__ rate, assuming
that each cash flow represents an equally-space period. To compare
its result with XIRR, you must annualize the IRR result based on the
size of the period. In general:

daily IRR: =(1 + IRR(range,guess))^365 - 1
weekly IRR: =(1 + IRR(range,guess))^52 - 1
multiple months IRR: =(1 + IRR(range,guess))^(12/m) - 1

where "m" is the number of months per period.

I'm trying to learn how to shape the XIRR request
to match what I know is right.

The "shape" of the XIRR model is the same as the IRR model, with the
additional detail of specific dates. You use XIRR only if the cash
flows are irregularly spaced (or you want to bother with annualizing
the IRR result); otherwise, use IRR.

If the XIRR result does not closely match the annualized IRR, the
problem is likely in your model -- the values that you are using as
"cash flows" -- which would adversely effect the IRR result as well as
the XIRR result.

Let me know if this is helpful. Otherwise, I will stop wasting my
time beating a dead horse.


----- original posting -----

The cash flow is the actual history of this investment, When I set up the IRR
function, I have every date, (4/28/2006 through 3/31/2008) amount invested,
surrender value, and account value from start to finish in a column from
start to end in from R2C1:R705, Any cells without values are set to 0 so the
IRR will function properly with the calculation being done for the entire
history of the investment i.e. inception to date.

My proof also matches.

My intent is to able to get the same answer using XIRR instead of using this
brute force method. I'm certain the end result is valid. I'm trying to learn
how to shape the XIRR request to match what I know is right.



:
PS....

I wrote:"
Dates XIRR
12/31/2007 $61,445.04
1/7/2008 $4,062.21
2/6/2008 $4,062.02
3/6/2008 $3,921.48
3/31/2008 -$72,906.91
[....]
With your data, I compute the annualized IRR to be -0.025682814, and
XIRR computes -0.033609948. They are equal when rounded to 2 decimal
places. "Close enough for goverment work" ;-)

For an even better comparison, change the first date to 12/7/2007 and
the last date to 4/6/2008, so that the periods are now almost
monthly. The annualized IRR is -0.02568, and XIRR returns -0.02583 --
almost the same, as expected.

Note: There will always be at least a small difference between the
annualized IRR and XIRR because XIRR the varying number of days
between "monthly" dates (i.e. say day number each month).
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Similar Threads


Top