Civil Service pension: alpha v. partnership & voluntary topup

chiefnoodle

Hi all. I've accepted a job offer which pays a Civil Service pension and am trying to get my head around the options: alpha and partnership. 

People say it is a no-brainer, alpha is far better. It is true that Alpha has an employer contribution of 28.97% - but this seems to be practically meaningless as it doesn’t go into your pot. All that matters is the defined benefit, i.e. how much it will actually pay after retirement. And I personally find the control of partnership appealing, and seeing the pot building up would be fun.

So I think the reasonable way to compare is to make some assumptions on how your investments will do (under partnership), and about annuity costs. Then, compare the annuity you can buy with the built up lump sum (under partnership), to what you would get under alpha.

I'm 38, single, and the job is a 3 year contract - not sure what will happen after the three years.

For partnership, with a 3% employee contribution and 14% employer contribution (11% + matched 3%), the contribution to the pot is 17% of your wage. Assuming investment return of 3% above inflation, after the 0.12% management fee, over 30 years until I'm 68, means at retirement the pot will more than double to 0.17 * 1.03^30, or 41% of my wage this year, adjusted for inflation. 

Looking at annuities best buy rates for Single life, 3% escalation, it looks like a lump sum can be converted to an annuity with a rate of about 5-6% of the value of the lump sum - call it 5.5%. So 0.41*0.055 comes to 2.27% - i.e. the annuity purchased with the lump sum will be worth 2.27% of my annual salary, adjusted for inflation.

Well with alpha, the defined benefit is 2.32% (also adjusted for inflation). So there isn't all that much difference. On balance I think I'll probably go for alpha as it gives you certainty. If I find a spouse, then the comparable Joint life 50%, 3% escalation annuity rate is 5%, so 2.06% instead of 2.27%. But I think that alpha isn't the incredible giveaway people think it is.

(And of course investments can often do much better than CPI + 3%/year. And the employee contribution is more expensive for alpha, at 5.45% instead of 3% in my example, so you are paying quite a lot more for a negligible difference.)


OK so if that all makes sense, that brings me to the second part of my question. Not having paid anything into my pension until just a few years ago, I'd like to make some additional contributions out of my monthly salary. There are 2 ways to do this: "Added pension", or "Civil Service Additional Voluntary Contribution Scheme" (CSAVCS).

As I understand it, the former is basically boosting the alpha, it increases the defined benefit; while the latter is defined contribution, it goes into an investment pot. So I'm trying to compare and make an informed decision which to go for. But the added pension page doesn't exactly make this easy as there is no indication of what the defined benefit is! Just a a link to a rather unintuitive excel spreadsheet calculator. How much will I get then, for every pound I pay in now? What basis for comparison is there between these two additional contribution methods?

With the first question, there are employer contributions which are twice as generous for alpha than they are for partnership, which is what tips the scales in that direction. However with voluntary top-ups, there are no employer contributions. So my guess is that there won't be any strong reason either way?

Thanks! 

DE_612183

The question about alpha and partnership very much depends as well on your age - I'm a bit older than you - just joined CS - so Alpha for me IS a no-brainer.

hugheskevi

chiefnoodle said:

Alpha has an employer contribution of 28.97% - but this seems to be practically meaningless as it doesn’t go into your pot. All that matters is the defined benefit, i.e. how much it will actually pay after retirement. 

Correct, the employer contribution is meaningless, as it contains adjustments for notional past service deficits which have no relevance to the value of new pension accrual.

chiefnoodle said:

So I think the reasonable way to compare is to make some assumptions on how your investments will do (under partnership), and about annuity costs. 

Although it is a common approach, I don't favour the annuity comparison. That is appropriate if in the counter-factual you would actually buy an annuity, but most don't, they use drawdown of some type. Using an annuity comparison therefore does not reflect what will happen, and significantly biases the outcome depending on the current level of gilt and corporate bond yields (which determine annuity pricing) and also builds in the profit margin and capital reserving of annuity providers. There is also an inflation mismatch between the comparisons which is likely to be significant (3% escalation is not a good comparator to uncapped CPI with a floor 0%). The reason for using annuity rates is the convenience of having a figure which can be used to compare to a similar figure, rather than considering the stream of cash-flows in retirement under the most likely withdrawal method, and the associated volatility which will need to be managed one way or another.

chiefnoodle said:

I think that alpha isn't the incredible giveaway people think it is.

A not unreasonable conclusion, at least for a 38 year old. The CPI + 1.7% discount rate used to value alpha is fairly low, although it is arguable that a low discount rate is appropriate to value a very low risk pension. If you were investing in a DC pension, you would expect a better long-run return than CPI+1.7% net of charges.

Do remember that there are other benefits such as death-in-service, ill-health, redundancy and buy-out of actuarial reduction, death guarantee period, enhanced survivor benefits on death, injury benefit, spouse and dependent benefits, etc, which should all be taken into account (Partnership has some, but different to alpha).

Note that if you did a similar calculation but for a 65 year old you would find very different figures - alpha is far more generous for older employees than it is for younger employees.

chiefnoodle said:

So I'm trying to compare and make an informed decision which to go for. But the added pension page doesn't exactly make this easy as there is no indication of what the defined benefit is! Just a a link to a rather unintuitive excel spreadsheet calculator. How much will I get then, for every pound I pay in now? What basis for comparison is there between these two additional contribution methods?

The cost of Added Pension is calculated with a discount rate of CPI+1.7% and the cost varies by age (more expensive for older members, cheaper for younger members). If you are happy to accept a (nearly) guaranteed rate of return of CPI+1.7% and want an annual income, then Added Pension will be good. If you think you can do better with DC and would prefer a more flexible pot of money, then  AVCs will be better.

chiefnoodle

DE_612183 said:

The question about alpha and partnership very much depends as well on your age - I'm a bit older than you - just joined CS - so Alpha for me IS a no-brainer.

Right. Just to expand on your comment, how well partnership would do depends on your age for two reasons:

1) the greater the age, the less years the lump sum will be invested for, so less compounding
2) employer contribution increases at certain ages (link).

The two reasons point in different directions though. As a rule of thumb, the closer you are in age to one of the thresholds (36, 41, 46) when the employer contribution gets bumped up, the better partnership is. As you move from 46 to retirement, the worse partnership is (as @hugheskevi wrote, "alpha is far more generous for older employees than it is for younger employees").

A rough spreadsheet gives the following chart:


(note employer contributions include matching of the employee contribution up to 3%).  

The bottom two lines show that from about age 48 onwards, alpha starts getting better and better. But until then it isn't terribly clear cut.

Universidad

One thing you're not taking into account is that certainty is expensive.

Alpha offers a tremendous amount of certainty, and the reason comparison with annuities is a good one is that it shows how expensive that certainty is to buy privately.

r6mile

Do you have any other DB pension? Do you see yourself staying in the Civil Service for a while? I think this can shape which choice you go for. If you already have a lot of DB pension (I assume not, given your age), then I guess building a bit of a DC pot will give you extra flexibility. 

If not, and you are only likely to stay in the CS for a few years, then taking this opportunity (rare, given that DB pensions are basically non existent outside the public sector and academia) to build some DB pension via Alpha (plus potentially added pension) could be worthwhile.

I am also on my 30s, and having looked at Added Pension, I concluded it was not great value so chose the AVC scheme for my extra contributions, to complement my Alpha pension.

Another option for you might be to join Alpha, potentially transfer in any other DC pension you may have to top up your DB, and then separately join the AVC scheme. Lots of possibilities depending on what you are after.

chiefnoodle

hugheskevi said:

chiefnoodle said:

So I think the reasonable way to compare is to make some assumptions on how your investments will do (under partnership), and about annuity costs. 

Although it is a common approach, I don't favour the annuity comparison. That is appropriate if in the counter-factual you would actually buy an annuity, but most don't, they use drawdown of some type. Using an annuity comparison therefore does not reflect what will happen, and significantly biases the outcome depending on the current level of gilt and corporate bond yields (which determine annuity pricing) and also builds in the profit margin and capital reserving of annuity providers. There is also an inflation mismatch between the comparisons which is likely to be significant (3% escalation is not a good comparator to uncapped CPI with a floor 0%). The reason for using annuity rates is the convenience of having a figure which can be used to compare to a similar figure, rather than considering the stream of cash-flows in retirement under the most likely withdrawal method, and the associated volatility which will need to be managed one way or another.

Thanks @hugheskevi. I agree with all your points, however they generally point in the direction that my comparison unfairly favours alpha - for example, by penalising partnership by including the profit margin of the annuity provider. The point I was working towards is that partnership isn't necessarily worse than alpha - so the very valid differences you point out strengthen that point. 

I imagine the other way they could be compared would be comparing lump sums - i.e. what lump sum can I exchange my alpha pension for at retirement? (I think you can only actually exchange up to 25%, but perhaps using that rate would be some basis for comparison). Doing it that way would however unfairly favour partnership, by including the cost of the exchange in the alpha calculation...

Happy to be educated about any other ways to compare!

chiefnoodle

hugheskevi said:

chiefnoodle said:

So I'm trying to compare and make an informed decision which to go for. But the added pension page doesn't exactly make this easy as there is no indication of what the defined benefit is! Just a a link to a rather unintuitive excel spreadsheet calculator. How much will I get then, for every pound I pay in now? What basis for comparison is there between these two additional contribution methods?

The cost of Added Pension is calculated with a discount rate of CPI+1.7% and the cost varies by age (more expensive for older members, cheaper for younger members). If you are happy to accept a (nearly) guaranteed rate of return of CPI+1.7% and want an annual income, then Added Pension will be good. If you think you can do better with DC and would prefer a more flexible pot of money, then  AVCs will be better.

@hugheskevi it sounds like you are saying that with AVC it goes into a pot and final value depends on investment performance. Whereas with Added Pension, while they obviously don't know how many payments are going to be made between retirement and death, they have some expected value for this; and they charge you that much, discounted by CPI+1.7% / year. And that is why cost varies by age, as for older people there is less discount. Is that right?

Not sure I understand this. So you pay a certain amount now and are immediately guaranteed a corresponding income post-retirement. Given that CPI for the years between now and then is unknown, how can it depend on CPI?

Anyway if I've understood you correctly, what it boils down to is simply: how much do I think inflation will be over the coming years, and how well do I think my investments will perform. So will my investment strategy outperform a rate of CPI+1.7%, in which case AVC is better - or will it not, in which case AP is better. 

Thanks a lot!

hugheskevi

chiefnoodle said:

hugheskevi said:

chiefnoodle said:

So I think the reasonable way to compare is to make some assumptions on how your investments will do (under partnership), and about annuity costs. 

Although it is a common approach, I don't favour the annuity comparison. That is appropriate if in the counter-factual you would actually buy an annuity, but most don't, they use drawdown of some type. Using an annuity comparison therefore does not reflect what will happen, and significantly biases the outcome depending on the current level of gilt and corporate bond yields (which determine annuity pricing) and also builds in the profit margin and capital reserving of annuity providers. There is also an inflation mismatch between the comparisons which is likely to be significant (3% escalation is not a good comparator to uncapped CPI with a floor 0%). The reason for using annuity rates is the convenience of having a figure which can be used to compare to a similar figure, rather than considering the stream of cash-flows in retirement under the most likely withdrawal method, and the associated volatility which will need to be managed one way or another.

Thanks @hugheskevi. I agree with all your points, however they generally point in the direction that my comparison unfairly favours alpha - for example, by penalising partnership by including the profit margin of the annuity provider. The point I was working towards is that partnership isn't necessarily worse than alpha - so the very valid differences you point out strengthen that point. 

I imagine the other way they could be compared would be comparing lump sums - i.e. what lump sum can I exchange my alpha pension for at retirement? (I think you can only actually exchange up to 25%, but perhaps using that rate would be some basis for comparison). Doing it that way would however unfairly favour partnership, by including the cost of the exchange in the alpha calculation...

Happy to be educated about any other ways to compare!

Also worth noting that with Partnership you get a 25% tax free lump sum, whereas due to the commutation factors in alpha there probably is little or even a negative overall value from commuting pension into lump sum.

Personally I would do a cashflow comparison from the current time to expected date of death, comparing alpha with whatever alternative I was planning, eg, drawdown, and comparing contributions and benefits received on a year-by-year basis. I don't see much point in artificially changing one thing or another into something it isn't, nor ever will be, rather than analysing the actual cashflows which would be expected to materialise.

chiefnoodle said:

hugheskevi said:

chiefnoodle said:

So I'm trying to compare and make an informed decision which to go for. But the added pension page doesn't exactly make this easy as there is no indication of what the defined benefit is! Just a a link to a rather unintuitive excel spreadsheet calculator. How much will I get then, for every pound I pay in now? What basis for comparison is there between these two additional contribution methods?

The cost of Added Pension is calculated with a discount rate of CPI+1.7% and the cost varies by age (more expensive for older members, cheaper for younger members). If you are happy to accept a (nearly) guaranteed rate of return of CPI+1.7% and want an annual income, then Added Pension will be good. If you think you can do better with DC and would prefer a more flexible pot of money, then  AVCs will be better.

@hugheskevi it sounds like you are saying that with AVC it goes into a pot and final value depends on investment performance. Whereas with Added Pension, while they obviously don't know how many payments are going to be made between retirement and death, they have some expected value for this; and they charge you that much, discounted by CPI+1.7% / year. And that is why cost varies by age, as for older people there is less discount. Is that right?

Yes.

Not sure I understand this. So you pay a certain amount now and are immediately guaranteed a corresponding income post-retirement. Given that CPI for the years between now and then is unknown, how can it depend on CPI?

You purchase a set amount of annual pension in the current year, which will be payable without reduction from State Pension age. Each year between now and State Pension age the value of the amount you purchased increases in line with CPI, whatever that might be in the future.

chiefnoodle said:

hugheskevi said:

chiefnoodle said:

So I'm trying to compare and make an informed decision which to go for. But the added pension page doesn't exactly make this easy as there is no indication of what the defined benefit is! Just a a link to a rather unintuitive excel spreadsheet calculator. How much will I get then, for every pound I pay in now? What basis for comparison is there between these two additional contribution methods?

The cost of Added Pension is calculated with a discount rate of CPI+1.7% and the cost varies by age (more expensive for older members, cheaper for younger members). If you are happy to accept a (nearly) guaranteed rate of return of CPI+1.7% and want an annual income, then Added Pension will be good. If you think you can do better with DC and would prefer a more flexible pot of money, then  AVCs will be better.

Anyway if I've understood you correctly, what it boils down to is simply: how much do I think inflation will be over the coming years, and how well do I think my investments will perform. So will my investment strategy outperform a rate of CPI+1.7%, in which case AVC is better - or will it not, in which case AP is better. 

Pretty much, with a bit of judgement about the benefits of certainty vs flexibility, value of income over lump sum, how it fits into your wider financial position, etc.

chiefnoodle

Building on the very helpful comments by @hugheskevi and @DE_612183, I've modelled Partnership in a spreadsheet to compare it with alpha under a range of assumptions, to show how age is the most significant factor here.

What I've found has seriously undermined my faith in the superiority of Alpha. A career starter at 21 would be foolish to opt for Alpha. In fact for all ages until about 37 it is a no-brainer to go for Partnership. Equally, for all ages above 50 it is a no-brainer to go for Alpha. It is only those years in between where there is any question.

Here is a plot - note that the modelling may be imperfect, but in general it is biased in favour of Alpha (see comments above).



Note: the employee contribution rate is set to 5.45%. This is to make it comparable to Alpha, where employees are charged that much if they earn between £34,200 and £56,000. The different rates are available here, if anyone wants a plot comparing for a different rate then LMK. This affects which ages should choose which plan.

The policy is to auto-enroll all employees on Alpha, unless they choose Partnership. Based on the above, I question whether that is generally in the best interest of employees. It looks like the younger employees in Alpha are effectively subsidising the pensions of the older employees, by paying in more than their fair share to allow the scheme to balance out overall. Those younger employees probably don't think too much about their options for retirement, and I would hazard a guess that they are not well represented amongst the senior decision makers who decide whose payments should subsidise whose pension... but that is enough cynicism for one day.

DE_612183

On that basis I think I made the right choice - I'm 59 and started with the CS this year after 20 odd years of contracting.

I've got a Final Salary Pension that pays about £9k per year now + about 3 stakeholder pensions that have built up pots, but a re now only being contributed to for a small amount.

I estimate that at the end of 3 years I'll have built up a pension payout of about £320 per month ( minimum ) for an outlay of about £8,500 - for a stakeholder pension I'd need about 70k ( even if 20% comes as tax relief ).

That's why for me I see it as a no brainer - but I'm happy to see if my figures work ( my salary is £60k ).