By: Trevor Llanwarne, ILC-UK Trustee

Following on from the launch of ‘Pension Pots and How to Survive Them’, ILC-UK Trustee Trevor Llanwarne responds to the report by Cass Business School.

• First, a big thank you to Les, David and Douglas for producing a very important contribution to the necessary discussions/debates on what people retiring will now need to consider under the new freedom and choice regime
• Let me start with a few words on why I was happy to accept the invitation to speak today on the panel. As well as my seven year stint as Government Actuary (and I can tell you that complaints about annuity rates for the GAD drawdown tables was the biggest element in my post-box as for many MPs), I can also draw on the work I did at PwC leading on big projects to help insurance clients sort out a lot of mis-selling in years gone by. We did thousands of 1:1 interviews across all socio-economic groups. Much was in precisely the sort of areas under discussion and I learnt a lot of practical issues in the process. So I hope I can add some practical and technical experience and input to this discussion.
• In reading today’s paper, what came across to me was how rarely an annuity purchase is the right answer. Maybe just at age 80 if I read correctly. Of course, in order to stick to broad principles, a number of more complex and hybrid options have rightly been omitted by the authors and, if added in, the conclusions would become less clear. But even allowing for this, my experience tells me to not be surprised by their conclusion – at least in today’s type of regulatory and economic conditions on yields, inflation etc. for the UK
• So what I would like to do is not to mention technical challenges (and there are some), but to offer a list of items which have a bearing on this expectation. I will leave others to judge on relative importance and how to put together – but I do hold to the position that they are all relevant

List of Items

• 1. There is a value on flexibility. The increasingly dominant so-called Bayesian thinking in analytical work would strongly support this both in theory and in practice. Yet, in the field of pensions, when did you last see any comparative analysis which puts relative values on flexibility before comparing monetary amounts and NPVs? But individuals do this subconsciously and, in concept, they are right.
• 2. State pension is DB and for a married worker on average weekly earnings of currently £494 (let’s say £25,000 per annum)[1], the State pension might take up 75-80% of the total pension. So what we have is a position where the percentage exposure from the private sector pension is small for perhaps well over half the retiring population.
• My point here is that every individual’s position will be unique. Most stylised examples ignore some crucial issues such as the contribution of State pensions. Indeed for many people retiring over the next 20 years the additional contribution of a Deferred DB pension, even if accruals stopped some years back, can also be a major element.
• 3. So being provocative to aid discussion, let me ask “is it the case that the people most vulnerable are perhaps those in the top one-third of the retiring population”. To what extent should we worry about them? Is it not this very group that is almost wholly in owner occupation with plenty of potential equity release “ballast”, if I can use that term, if things do go wrong. At least this seems likely for a decade or two to come. This aspect was excellently referred to in the paper under consideration today.
• 4. The evidence seems to suggest that people, on average, under freedom and choice will more likely cut expenditure than blow it all away – fully accepting that there will be a proportion that does the opposite. Although we know there have been many retirements since April with full cash-outs, I don’t think we know what has happened with the funds and it is clearly not the case that all has been spent or frittered away. We really do need some data on use of funds.
• 5. I have yet to meet someone in all the discussions I and the PwC mis-selling correction team had with retirees over many years who knew exactly, and could guarantee, at the point of retirement what they needed by way of income for ten years going forward let alone thirty. The point when a pension arrangement vests is invariably the wrong time to take a lifelong committed inflexible decision. Most need at least a year to see how their change in lifestyle pans out.
• 6. Solvency II makes cost of annuities in EU higher than in non-EU
• 7. Pointing in a different direction, many people understate how long they are going to live
• So what I now want to move on to is give some thoughts on current and future developments.

Some current and future developments

– (i) So-called “Robo-advice” for people at retirement. I do know a bit about this and can comment on some of the issues if anyone has questions

– (ii) The use of a temporary (holding) drawdown for say up to three years with a trigger for advice when much more appropriate

– (iii) The fact that it may be much easier to make investment selection brought in to the decision process

– (iv) I don’t agree with Les that the decision has, of necessity, become more complex at retirement. I could put forward a case, under a Bayesian concept, that a temporary drawdown at retirement with much better targeted decisions a few years later is actually simpler than the historical annuity regime where there is a real difficulty with …

….. rates of escalation, dependants’ pensions (to whom and what proportions), splitting between providers, guarantees, unit-linked and with profit options, payment frequency, mentioning of ill-health status all to be decided at the wrong time.

It was only simpler before because a lot of complex analysis was bypassed (for simplicity) and we are now starting to see the problems emerge – e.g. with FCA reviews on annuity mis-selling.

• I shall provide a copy of my script for an ILC blog if they do one. I will add to that script some thoughts on further research and some technical questions for the authors of the paper.
• In conclusion, a very welcome paper to add its own very important perspective to help all of us start map out a way forward. Well done.

Trevor Llanwarne

12th November 2015

PS For those interested in my thoughts on how to look at the bigger picture with a focus on encouraging more people saving and people saving more, go to booklet https://blog.ilcuk.org.uk/2015/09/23/greater-savings-for-retirement-more-people-saving-people-saving-more-all-voluntarily/

Ideas for further research

1. Optimal ages for buying an annuity (ignoring other person-specific details) allowing for value of flexibility.
2. More work on my contention that the age of vesting is the wrong age for taking fixed long-term decisions.
3. Combination ideas on defaults such as a temporary drawdown with robo-advice every three years and some sort of staging into annuity purchase around age 75-80 as the default.
4. Consider different and varying economic conditions.
5. Where the cash-out money is going.
6. Data on current retiree pensions (between State, DB, DC) and links with home ownership.
7. Ideas on how to provide guarantees that are outside solvency II.
8. Try to put some indication of the numbers of people impacted in each of the stylised examples so as to get a feel for relative importance.

Technical questions for the authors

Note – none of the below are fundamental to the big picture messages of the paper

9. Did you consider any other categories in addition to your three of poor health, home owner and bequest motive? My work some years ago suggest that other critical categories would include silver divorcees and part time workers in retirement.
10. Although I stated above that you needed to keep the options simple to bring out the key messages, did you consider putting hybrids or equity-linked annuities into the range of options? Or indeed why does the paper give the impression that drawdown once started is fixed (whether in £ or on a standard formula) as it is for life since this is clearly not the case?
11. I think your table in 2.3 would warrant some checking and peer review. For example, person type 6 of poor health and bequest motive seems ideal to me for drawdown (the one option that is presented as a bad option). Why does drawdown work well for type 6? Because for someone in poor health, drawdown provides just as much living expense as full pot extraction but at a lower tax rate and if early death occurs, the funds get bequested free of IHT. Neither option A nor C deliver on this. C is only good if someone is in poor health and then lives a long time. A is only better than B if government changes the tax rules at a later date with no advance warning. B is also the better option for person type 4 on similar reasoning. Even for person type 1, drawdown is best in early years and/or in lots of sub-divisions (e.g. to give just one example, someone who has a partner in poor health but does not wish to leave partner without income if they die first – the norm). Whether or not you agree these comments, do you agree more work is needed on the categorisations and potential options for each?
12. There is a lot of focus on how unisex annuity rates create bias for the decision. This is true if a single life annuity is to be bought. But did you do any work to consider unisex impact on joint life annuities? A lot of work I have seen suggests that joint life rates are very similar between sexes and hardly impacted by unisex.
13. To what extent have you had input from tax experts in terms of how you have presented the relative merits of options where tax is mentioned? I suspect the tax position is more complex than suggested with more options to change the conclusions and you may wish to revisit/re-check some of the detail.
14. Have you looked at different economic conditions or variable conditions?
15. In practice, people on drawdown do not carry on at the same drawdown level without regular reviews. So when you apply the test e.g. in pages 15-17 that as long as drawdown does not run out of money then it must be better than an annuity, this is not the test that is actually applied in practice at least for all years past. Instead, the test is whether at each regular review, perhaps every three years, can drawdown continue at current levels without reducing security on expectations. A consequence of this is that reductions in drawdown amounts is quite common at some stage and then there is a “peace of mind” issue about whether people are happy to accept this potential volatility. Have you done any analysis to look at impact of regular reviews and do you have any thoughts about whether reviews have a role?

These questions should be seen as pointers for more follow-up analysis rather than items which undermine the conclusions to what is a very good paper

[1] in receipt of a pension totalling 50% of pre-retirement earnings