XIRR vs Unitized total return = Ratio?

george4064

Been crunching some numbers and have been comparing my (100% equity) ISA's unitized total return vs Excel's XIRR using the respective cashflows since April 2019 to 27 Aug 2021. 

Unitized total return = 33.5%
XIRR = 22.2%

XIRR formula: =XIRR(B4:B16,C4:C16) or =XIRR(Cashflows,Dates). First figure is minus, subsequent purchases are minus, withdrawals are positive and final portfolio value is positive.

Vanguard LS 100% Acc total return = 31.7%
Vanguard FTSE GAC Acc total return = 38.1%

Would also be interested to hear what other people's XIRR vs respective fund return is? Should be easy to get for you Vanguard investors because they calculate your XIRR (aka personal rate of return) for you.

I'm not so much concerned about my unitised total return vs Vanguard, but I was hoping that my XIRR was a little bit higher compared to the unitized total return! My understanding is that the XIRR is a money-weighted return and by it being lower suggests that I have timed my cashflows poorly (both in and out, mostly in)?

Does anyone know of any ratio that takes your XIRR and divides by the Unitised price return? Does such a 'calculation' exist? If so, does it have a name?

For example mine would be 22.2/33.5 = 66%

Would appreciate any thoughts/comments from other wise members of the forum, and apologies if this is a bit technical.

Thanks in advance.

masonic

I don't think this should come as any surprise when the broader trend is rising markets. If you were investing money gradually over the period, your return will be lower than someone who put in a lump sum at the start of the period (which is what unitisation models). During a period where there is a crash, but little difference between the start and end points, then XIRR would give you the higher number as you'd have benefited from buying the dip. If you had very little invested in 2019-2020, then went all in at the bottom of the Covid crash, your XIRR would look very good indeed.

Prism

XIRR gives an annual return so understandable that its lower than total return.

dales1

As I understand it, the XIRR function returns an Annualised return (not a Total return).

So it would need grossing up to be comparable.

However (unless your cashflows have been very uneven) a 22.2% XIRR return looks too high for an annualised return, yet too low for a total return, over almost 2.5 years of strong growth.

(My own figures (XIRR v Unitised) are very close to each other in recent years, on my own cashflows that don't vary much at all).

Perhaps you could recheck your XIRR calcs eg using simple cashflows that can be approximated manually.

(Or post them up here, scaled for anonymity).

tebbins

XIRR is an annualised time-weighted (the return your portfolio generated as it was a fund), Vanguard for example show the annualised money-weighted return. I hadn't heard of unitised until this thread but looking up a definition it seems like another name time-weighted.
My Vanguard MWRR is 55.49% (I was *very* lucky with Covid timing), the luckiest holding was my trusty old VMID on 73.36%. I would guess the XIRR would be much more normal since I've had the account for a few years.
If your MWRR > XIRR over a period, it means you timed the market well, ie that during that period you bought at points below the average trend, and vice versa.
For example someone who invested £10,000 on 1/1/20 lost -4.6% over 2020, exactly the same as the index. Someone investing the same amount on 23/3/20 saw a 59% gain, 42% annualised. That's all it tells you.

However I'm sure a thread such as this wouldn't suggest we start market timing.

masonic

Prism said:

XIRR gives an annual return so understandable that its lower than total return.

Yes unitised returns should always be annualised before comparing with XIRR, but I assume the OP has done the opposite and then used the annualised XIRR to derive a total return figure over the period. The annualised Vanguard LS total return is 9.9% (the annualised unitised return will be very close to this figure based on comparing the total return over the period with VLS). An XIRR of 22.2% per year is unrealistic and is probably derived from an annualised figure of about 7%, which is more plausible. So it is 9.9% vs ~7% that would be typically compared. Comparing 22.2% with the 33.5% is a like for like comparison, but it magnifies the difference for longer time periods compared with an annualised comparison, which is standardised.

OP, it would have been helpful if your graph had a minimum of £0 on the y-axis (no need for labels), then we could have at least seen by what proportion the value had increased (currently it looks like the value has tripled based on the orange bars).

DireEmblem

I always unitise my performance(time weighted), that way its easier to compare like for like for like basis with index/mutual funds.

george4064

Prism said:

XIRR gives an annual return so understandable that its lower than total return.

My blunder, I did not annualise the unitised performance!

masonic said:

Prism said:

XIRR gives an annual return so understandable that its lower than total return.

Yes unitised returns should always be annualised before comparing with XIRR, but I assume the OP has done the opposite and then used the annualised XIRR to derive a total return figure over the period. The annualised Vanguard LS total return is 9.9% (the annualised unitised return will be very close to this figure based on comparing the total return over the period with VLS). An XIRR of 22.2% per year is unrealistic and is probably derived from an annualised figure of about 7%, which is more plausible. So it is 9.9% vs ~7% that would be typically compared. Comparing 22.2% with the 33.5% is a like for like comparison, but it magnifies the difference for longer time periods compared with an annualised comparison, which is standardised.

OP, it would have been helpful if your graph had a minimum of £0 on the y-axis (no need for labels), then we could have at least seen by what proportion the value had increased (currently it looks like the value has tripled based on the orange bars).

Minimum on Y-Axis is £0. Your assumption about value roughly tripling is correct (more like 278% according to my rough calculations), thats from a mixture of new money and positive performance.

@masonic@Prism@dales1@DireEmblem

Thank you all for your responses, you have pointed out my blunder that I had not annualised my unitised performance. Now having done that calculation the figures look much better!

XIRR = 22.2%
Unitised performance (p.a. %) = 13.2%

XIRR (total) = 59.8%
Unitised performance (total) = 33.5%

Finally, does any Vanguard investors here know if Vanguard's 'personal rate of return' gives the annualised IRR figure or do they calculate the total IRR over the period? My OH is with Vanguard so I like to compare.

Aceace

Vanguard's "Personal rate of return" is a total return since one started investing, so not annualised. 

masonic

george4064 said:

Thank you all for your responses, you have pointed out my blunder that I had not annualised my unitised performance. Now having done that calculation the figures look much better!

XIRR = 22.2%
Unitised performance (p.a. %) = 13.2%

XIRR (total) = 59.8%
Unitised performance (total) = 33.5%

Finally, does any Vanguard investors here know if Vanguard's 'personal rate of return' gives the annualised IRR figure or do they calculate the total IRR over the period? My OH is with Vanguard so I like to compare.

That's a very good result, although the figures look a bit out to me. For the annualised unitised figure, I make this 12.7% (Period is equal to 2.408 years; annualised value = 1.335 ^ (1/2.408) = 1.127). For the total XIRR figure I make it 62.1% ( = 1.222 ^ 2.408 = 1.621).

I don't know, but think it is very unlikely that Vanguard does different to virtually every other platform, which just calculates returns based on the current holdings and their average acquisition price. Looks like Aceace has a better answer for this one, perhaps it does include disposals, unlike many providers.

george4064

masonic said:

george4064 said:

Thank you all for your responses, you have pointed out my blunder that I had not annualised my unitised performance. Now having done that calculation the figures look much better!

XIRR = 22.2%
Unitised performance (p.a. %) = 13.2%

XIRR (total) = 59.8%
Unitised performance (total) = 33.5%

Finally, does any Vanguard investors here know if Vanguard's 'personal rate of return' gives the annualised IRR figure or do they calculate the total IRR over the period? My OH is with Vanguard so I like to compare.

That's a very good result, although the figures look a bit out to me. For the annualised unitised figure, I make this 12.7% (Period is equal to 2.408 years; annualised value = 1.335 ^ (1/2.408) = 1.127). For the total XIRR figure I make it 62.1% ( = 1.222 ^ 2.408 = 1.621).

I don't know, but think it is very unlikely that Vanguard does different to virtually every other platform, which just calculates returns based on the current holdings and their average acquisition price. Looks like Aceace has a better answer for this one, perhaps it does include disposals, unlike many providers.

My annualised calculations were as follows:

30 April 2019 to 31 August 2021 = 28 months

Annualised unitised performance:
1.335^(12/28) - 1 = 13.2%

And XIRR total return:
1.222^(28/12) - 1 = 59.6%