Gilt ladder in SIPP - some not availiable gilts in platform

coyrls

With conventional gilts you know the nominal return but not the real return but with index linked gilts you know the real return but not the nominal return.  Given the difference, I think it's inevitable that they have to be modelled differently.  The confusion is caused by adding in the nominal returns from a savings account, where assumed inflation has to be used to convert the nominal return into a real return but it's important to note that the assumed inflation figure is not applied to the index linked gilt returns, where the real return is known.

incus432

A last question on the topic of gilt ladder software.  As well as the streamlit software discussed above and that everyone is familiar with I recently came across the Ladder creator at 

https://giltsyield.com/ladder/income/?csrf_token=Ijk3ZGY3ZmZjYWY3NTFmYWQ2NjgzN2Y2Y2NhMGYyYTA1Zjk3NTk4MTEi.aXOnRg.MFLaAlDwQXarKQDllIF72XoK0-k&nb_years=20&target_income=14000&inflation=0&tax_rate=0&submit=Go 

This example shows a flat 14k for 20 yr 

This one seems to give a significantly lower overall cost (3 or 4%) for the same income compared to streamlit.  It uses more gilts with more bias to longer duration, so that may be why as they have higher YTM.  You cant allow for cash interest rate or set a start date, but you can set what escalation rate you want (as assumed inflation) .

Any thoughts? 

OldScientist

incus432 said:

A last question on the topic of gilt ladder software.  As well as the streamlit software discussed above and that everyone is familiar with I recently came across the Ladder creator at 

https://giltsyield.com/ladder/income/?csrf_token=Ijk3ZGY3ZmZjYWY3NTFmYWQ2NjgzN2Y2Y2NhMGYyYTA1Zjk3NTk4MTEi.aXOnRg.MFLaAlDwQXarKQDllIF72XoK0-k&nb_years=20&target_income=14000&inflation=0&tax_rate=0&submit=Go 

This example shows a flat 14k for 20 yr 

This one seems to give a significantly lower overall cost (3 or 4%) for the same income compared to streamlit.  It uses more gilts with more bias to longer duration, so that may be why as they have higher YTM.  You cant allow for cash interest rate or set a start date, but you can set what escalation rate you want (as assumed inflation) .

Any thoughts? 

On a quick glance it looks like the giltsyield tool can current only model a ladder with nominal gilts with an escalation.

A 10 year nominal ladder with no escalation providing £10k per year currently costs £81.4k with the lategenxer tool and £82k with the giltsyield tool. Close enough (I note that the lategenxer tool uses close of day prices from yesterday while I think giltsyield uses up to date prices).

DRS1

I didn't get beyond noticing that the tax rates in the drop down list were 0, 20 and 40.  So not intended for very high earners or Scots.

incus432

OldScientist said:

A 10 year nominal ladder with no escalation providing £10k per year currently costs £81.4k with the lategenxer tool and £82k with the giltsyield tool. Close enough (I note that the lategenxer tool uses close of day prices from yesterday while I think giltsyield uses up to date prices).

A £14k pa ladder for 20 years using nominal gilts costs £189.2k on lategenxer tool and £181.4k on giltsyeild. So larger difference on longer ladders. 

PS for 10k/10yr I am getting 84.87k for lategenxer (not 81.4k) cf 82k for giltsyield

MK62

incus432 said:

OldScientist said:

A 10 year nominal ladder with no escalation providing £10k per year currently costs £81.4k with the lategenxer tool and £82k with the giltsyield tool. Close enough (I note that the lategenxer tool uses close of day prices from yesterday while I think giltsyield uses up to date prices).

A £14k pa ladder for 20 years using nominal gilts costs £189.2k on lategenxer tool and £181.4k on giltsyeild. So larger difference on longer ladders. 

PS for 10k/10yr I am getting 84.87k for lategenxer (not 81.4k) cf 82k for giltsyield

For £14k pa over 20 years, I get £181436 on giltsyield, and £179970 on lategenxer's tool......pretty close, but they are not using the same gilts......perhaps they assume a different annual withdrawal date. Setting to monthly withdrawals on lategenxr pushes the price up (as expected) to £183044.

PS for 10k/10yr I am getting £81267 on lategenxer and  £82075 on giltsyield.

As no information is supplied about the giltsyield tool, we can only guess at the differences between the tools (personally I suspect it might be the assumed annual withdrawal dates), but in my view they are close enough for it not to be a big issue. OldScientist's observation about using different prices might also be in the mix (even though it's now a Saturday morning - I guess Sunday afternoon would give the best chance of the tools using the same prices)

Another possibility is how the tools treat coupon income......on lategenxr it's assumed they are banked at the interest rate you set in the tool......there's no information on how coupons are treated on giltsyield though.

OldScientist

I've just had another look - I think there is another difference. As far as I can tell from looking at the cashflow tab, the lategenxer tool assumes a withdrawal at the end of each time period (i.e., at the end of the year or end of the month) whereas I think (and there isn't enough detail to tell) the giltsyield tool is assuming a withdrawal at the start*.

For example, currently for a 20 year ladder, £10k per year (0% tax rate) I have

giltsyield: 129,596.85
lategenxer (annual):128,550.40
lategenxer (monthly): 130,745.70

Consequently, there are other differences in the construction probably related to the assumptions of when the withdrawal takes place. For example, in 2027, lategenxer uses TR27 (matures in December 2027) while giltsyield uses T27A (matures in January 2027).

* A quick play with the excel pmt function also shows how the payment rate will change depending on whether the withdrawal is at the beginning or end of the payment period, e.g.

End of period:
=PMT(4.6%,20,-100,0,0) gives 7.75% rate
Beginning of period:
=PMT(4.6%,20,-100,0,1) gives 7.41% rate

MK62

.....OldScientist said:

I've just had another look - I think there is another difference. As far as I can tell from looking at the cashflow tab, the lategenxer tool assumes a withdrawal at the end of each time period (i.e., at the end of the year or end of the month) whereas I think (and there isn't enough detail to tell) the giltsyield tool is assuming a withdrawal at the start*.

For example, currently for a 20 year ladder, £10k per year (0% tax rate) I have

giltsyield: 129,596.85
lategenxer (annual):128,550.40
lategenxer (monthly): 130,745.70

Consequently, there are other differences in the construction probably related to the assumptions of when the withdrawal takes place. For example, in 2027, lategenxer uses TR27 (matures in December 2027) while giltsyield uses T27A (matures in January 2027).

* A quick play with the excel pmt function also shows how the payment rate will change depending on whether the withdrawal is at the beginning or end of the payment period, e.g.

End of period:
=PMT(4.6%,20,-100,0,0) gives 7.75% rate
Beginning of period:
=PMT(4.6%,20,-100,0,1) gives 7.41% rate

......yes, looks like the difference might be in the withdrawal dates........on giltsyield, a 10yr ladder uses T26, which matures next weekend......though that suggests the first withdrawal date must be at least a week away (though when is not specified). If using the giltsyield ladder tool, it might be wise to check the maturity of the first gilt - there's little point paying the spread and dealing fee for a gilt which matures very soon after the start of your ladder, you might as well keep the cash.

incus432

For example, currently for a 20 year ladder, £10k per year (0% tax rate) I have

giltsyield: 129,596.85
lategenxer (annual):128,550.40
lategenxer (monthly): 130,745.70

EDITED

Just had another look. Latengenxer is specific on when the withdrawal is made and you can set it to when you want  (and you can check on the cashflow tab) but the giltsyield tool is not and the graphic just shows the amount for each calendar year.  Playing around it appears to assume starting payouts (annual) in 12m time.  The two tools then give the cost figures above

In this instance the implementation for giltsyield is more even over 20 gilts, while lategenxer uses 19, ignoring TR43 and puts double in T45.  Based on total cost i would award Lategenxer the prize for more efficient algorithm. And more flexible on date and cash interest options

There is clearly an optimum date for withdrawal for different assumptions of cash interest rate, and this influences which gilts are chosen in the tool. - eg if you don't earn any interest it makes no sense to have 6-8k sitting in cash for 10 months. 

MK62

Looking at the giltyield tool, it seems rather crude compared with the lategenxr tool.........for a ladder with a first payout a year away, suggesting you buy a gilt which matures in 7 days seems somewhat ridiculous and a quick way to lose some money. Even if the ladder's first payout was in 8 days, it would still be daft.