Telegraph reporting - pensions tax threat

Malthusian

mat1964 said:

I really hope they leave salary sacrifice alone - it is really helpful if you're earning above £100k as the loss of the personal allowance means paying in effect 60% tax.

Conventional "net pay" or "relief at source" pension contributions also get you 60% tax relief within the £100k - £125.14k band. It doesn't have to be salary sacrifice.

The £100,000 calculation is based on "adjusted net income" which means pension contributions are deducted regardless of method.

 So if you earn £120k and sacrifice £20k for pension contributions, you are getting 60% tax relief on that amount (or more if your employer passes on the NI savings). 

If you are salary sacrificing you always get the employee NI saving. The bit that your employer might or might not pass on is the 13.8% employer NI saving.

So the maximum tax relief within that band is 67% if your employer pays the full NI saving into the pension. (£23,202 in your pension vs £7,600 taxed income.)

OldScientist

Albermarle said:

As someone who has a fairly large DB scheme and wife likewise. I have to reluctantly agree that the 20x seems very low in the current environment. On the flip side annuity rates are also very low at the moment. If the factor was 40x and annuity rates rose, it would look too high.

I think there was speculation in the past of increasing the multiplier to  25X as at least a step towards a more realistic figure .

Does anyone know the reasoning (if any!) behind the current 20x factor? I note that, possibly by coincidence, the historical safe withdrawal rate for a mixed portfolio of stocks/bonds for a cohort of retirees (50/50 male/female) retiring in a single year (i.e. taking into account mortality rates) is approximately 5% (rather than the 3.5-4.0% for an individual planning for a 30 or so year retirement). The mortality credits available in collective defined contribution pensions (should any ever get off the ground now that the legislation has been passed) may further muddy this area.

hugheskevi

OldScientist said:

Does anyone know the reasoning (if any!) behind the current 20x factor?

Way back in December 2002 it was initially proposed that:

B39 The new pension rules will call for valuations of DB pension rights for two purposes: for comparison against the lifetime limit when benefits are drawn; and to check against the annual value inflow limit. There is a choice about how to approach these valuations: • either accurately reflecting market conditions; or • using simplified rules to make the calculations quick and easy.

B40 The Government proposes to set a standard actuarial table of age sensitive conversion factors to express DB pension rights as capital values. The table would be set by order and updated from time to time, perhaps at intervals of 3 to 5 years.

B41 This table of factors will make standard stylised assumptions, producing robust rounded factors which would not pretend to accuracy for all users of the test. For simplicity, the factors will be unisex and assume that pension rights include a survivor’s pension set at half of the member’s pension and rights to RPI indexation of the pension.

Then in 2003 following consultation it was decided instead:

A24 The standard factor for determining the value, for tax purposes, of a defined benefit pension promise is £20 for every £1 of pension promised. This factor assumes that the aggregate of any survivors’ pension payable is no more than the member’s pension and that the member’s pension is indexed by RPI. Where a pension scheme allows commutation of pension benefits in return for a lump sum, the dependants’ pension benefits must not exceed the member’s gross pension before commutation. The scheme may decide whether to commute the member’s pension, dependants’ pensions, or both. Schemes that increase pensions in payment by a fixed rate no greater than 5 per cent may use the same 20:1 factor. The factor should be applied to the net pension, after any lump sum commutation. The cash value of the lump sum should be added to the value of the net pension

I would assume the factor of 20 was based on actuarial advice at the time given the reasoning is not described in the documents linked above, similar to the single factor of 16 used for Annual Allowance, which was based on the value of pension based on the typical affected member at the time it was set.

Interestingly, the 2010 Annual Allowance report by the Government Actuary Department linked above notes:

Using the same assumptions as were used to derive the 16:1 factor recommended in this report, the equivalent factor on retirement at age 60 would be 23.6:1 and at 65 would be 22.2:1. That is, I am effectively assuming that pensions are more valuable than implied by the existing 20:1 Lifetime Allowance factor, since I am assuming higher life expectancy and lower discount rates. In my opinion, this seems appropriate compared to the position of 5 to 10 years ago when the previous factors were adopted. 

The discount rates used to set the Annual Allowance factor 16 were: 

• 3.25% (plus inflation) a year before benefits come into payment,
• 3.0% (plus inflation) a year after benefits come into payment but before age 70, and 
• 1.75% (plus inflation) a year after age 70.

OldScientist

hugheskevi: many thanks for the reading (I'll have a proper look later on). That last note you quoted is particularly interesting.

DT2001

The 20 x multiple is simple. It covers many scenarios and hopefully aids people in their  retirement planning.
If you have a GMP element in your pension and inflation at 5% the ‘value’ of your DB pension is considerably lower than one without GMP.
I think my pension, at 65, is unusual in that 1/3 rd will increase at RPI, 1/3rd at CPI and 1/3rd not at all. A system to cope ‘fairly’ with the variations would have to be quite complex. Personally I’m quite happy to take the ups and downs of a simpler system to help planning.

Notepad_Phil

DT2001 said:

The 20 x multiple is simple. It covers many scenarios and hopefully aids people in their  retirement planning.
If you have a GMP element in your pension and inflation at 5% the ‘value’ of your DB pension is considerably lower than one without GMP.

A 20 x multiplier does seem quite generous to DB pensions with good inflation proofing - but most pensions like that exist only in the public sector nowadays and so any increase to make it more 'realistic' would likely only cause grief to whichever government tried to bring that in.

Personally I'd like to see DB pensions with lower inflation proofing be offered lower multiples than that offered to gold (or silver) plated public sector versions, but that's never going to happen.

zagfles

Or as I've said before, if they go for flat rate relief they could abolish the LTA and use the higher rate threshold to impose a similar penalty (eg 25% relief then 40% tax in retirement is similar to 40% relief then 55% tax).

That would swing the balance back towards DC or those with worse DB.

xylophone

Mr S is already a State pensioner, 

And how much of the increase for "wealthy" pensioners will be repaid to HMRC in tax?  

OldScientist

Notepad_Phil said:

DT2001 said:

The 20 x multiple is simple. It covers many scenarios and hopefully aids people in their  retirement planning.
If you have a GMP element in your pension and inflation at 5% the ‘value’ of your DB pension is considerably lower than one without GMP.

A 20 x multiplier does seem quite generous to DB pensions with good inflation proofing - but most pensions like that exist only in the public sector nowadays and so any increase to make it more 'realistic' would likely only cause grief to whichever government tried to bring that in.

Personally I'd like to see DB pensions with lower inflation proofing be offered lower multiples than that offered to gold (or silver) plated public sector versions, but that's never going to happen.

Agree that the 20x is simple, but also agree that not every DB scheme is the same. Interesting that the 2010 government actuary linked and quoted by hugheskevi above suggests that a DB scheme with the same assumptions made for the annual allowance would be about 10-15% more valuable than 20x (those assumptions include a portfolio with two-thirds stocks, a risk-free rate that is much higher than today and the purchase of an annuity at 70 - all subject to some argument/discussion and changes in the assumptions would make large changes in the valuation factor).

The row over NHS doctors having their pensions hit the LTA gives some indication of what would happen in the broader world for any government (easiest to phase it in starting, say 5 years hence!).