I’ve written about this before, but it is getting close to doing it for me, hence a worked example. This is particularly useful for people with a legacy DB pension although the technique also works to smooth your income between retiring and getting the State pension, which is another defined benefit pension [ref]a company DB pension is defined after each year you work for the company – they can change the terms for future accrual but not retrospectively for defined benefits already accrued. Whereas the State pension is defined by government and can be redefined – as has been the case recently[/ref].
A DB pension is defined only at a particular retirement age, usually 60 or 65, or use this calculator in the case of the State Pension. With a company DB pension you can often retire earlier, but you will take an income hit called an actuarial reduction if you retire earlier than the scheme NRA. In my case the NRA for most of my DB pension is 60, I will eat the loss from ‘retiring five years early’ for the last three years accrued when it was shifted to 65.
When these schemes were designed in the 1970s and early 1980s there was more goodwill between companies and their workforce which was seen as more of an asset than now, the pension was part of aiming at staff retention, which is largely gone in a faster-moving, possibly more efficient and definitely more dog-eat-dog employer/employee relationship now[ref]until you get to the parasitic executive level, which seems to featherbed a ‘because we’re worth it’ layer of scum to loot shareholders more and more, because of course you have ‘pay the going rate’ to recruit top talent despite the fact that CEO pay used to be about 40 times that of the grunts (US study, Table 6), compared to over 200 times now and there’s been no notable increase in company profitability since then[/ref]. The actuarial reduction is rarely defined, and gives companies wiggle room to reduce their costs. As a result it’s usually best to take a DB pension at NRA, because it’s nailed down what you get. Individual circumstances can sometimes mean you’re better off to take it earlier, but that’s usually more to do with paying off debts with any tax-free lump sum.
What to do if you want to retire before NRA?
What you ideally want is a short pension to front-run the main pension until its NRA. This pays out between the date of your early retirement and the DB pension NRA. If you want to retire before 55 you also need to save enough money in an ISA or unwrapped to pay your way to the earliest date you can draw pension savings, currently 55 but scheduled to rise to 57 and further – a gotcha to watch.
When I retired in 2012 a short pension wasn’t an option – if I had used a SIPP I wouldn’t have been able to draw it down, but all this has changed now. So the question is now how much can I save into a SIPP such that I can run the SIPP flat in the (for me) 5 years between getting hold of it and the pension NRA, without paying any tax. There’s not much mileage[ref]but there is some – even if you pay 20% tax on all of your SIPP income the 25% pension commencement lump sum saves you a quarter of the BR tax you’d otherwise have paid[/ref] in a basic rate taxpayer saving tax on the way into a SIPP only to pay tax on the way out, although any sort of higher rate taxpayer will gain a useful amount drawing down a SIPP even above the tax-free personal allowance up to the 40% tax threshold.
Put another way, I want to know how much can I put into a SIPP, such that I can withdraw the 25% tax-free lump sum up front and then a personal allowance worth each year, for (in my case) five years, from 55 to 60. With a NRA of 65 that would be 10 years. It’s reasonable to hold a five-year amount in cash, a ten-year amount would need to have some investment component for inflation protection – either some exposure to equities or some fixed interest bond-like stuff.
Show me the money…
At first sight this is easy enough. Estimate the personal allowance at £11,000 (it is currently £10,600 and rising to £12500 by government manifesto diktat in 2020) and in five years I have £55,000 worth of personal allowance. I can add a third of that amount to the total, ie £55,000 × 4 ÷ 3 ≈ £73000. The first year I take a PCLS of £18,250 (£73,000 ÷ 4) plus £10,600 tax-free personal allowance, leaving £44150, which I then run out at £11,000 a year for the remaining four years.
I can do a little bit better by using the non-earner’s annual pension contribution allowance of £3600, costing me £2880 p.a, and tossing that into the pension for each for each of the 5 years. I only need ~ £49k of savings as opposed to £73k up front. I save less tax with the short SIPP, because the PCLS is smaller. However, I leave a lot more of my AVC fund intact, and get that tax-free later, so I have shown the difference as a lump sum PCLS coming out in 2020, computing the tax saving on that to add in. I need to take a 360 degree view of all pension savings rather than just compute it for the SIPP in isolation.
Now it happens that I already have £10,800 in a HL SIPP as cash, because I acted as soon as I heard of Osborne’s pension reforms which was in March 2014, so I was able to use the 2013/4, 14/15 and now the 15/16 tax years non-earner’s £3600 pension savings allowance. That’s a bit of a way short of £73,000 but I saved 1/4 of my DB pension capital in AVCs, which I have the right to transfer to a SIPP, so I can get a hold of the residual amount from there, leaving the rest as AVCs. That AVC fund was originally targeted at getting 1/4 of my pension as a tax-free lump sum, by pre-saving that rather than commuting DB pension. With a short pension I can get hold of some of that earlier, still tax-free, and be able to defer my DB pension to NRA. Since I lose ~5% of DB pension per year drawn early, I am effectively investing this capital in more pension income. Although this exposes me a bit more to the risk of the DB pension failing, it reduces the risk of me screwing up the investment strategy. On balance the risk of the latter is probably worse than the pension failing, but this is a risk/reward tradeoff. Of course I don’t get ¼ of my DB pension capital as a tax-free lump sum because I’ve already transferred some of it out, but since I don’t have any mortgage to pay off with the PCLS that’s no big deal – getting some of it it tax-free earlier and the rest tax-free later is better than getting all of it tax-free later. The £12250 SIPP PCLS isn’t a bad match for next year’s ISA allowance, I only need top it up £3000-ish.
The same principle of running a SIPP flat or drawing at a higher rate pre SP could be used to match the State pension for early retirement ten years before taking it, smoothing the income profile without taking a suckout.
Breaking your pension savings into silos (ISA to 55/57, short SIPP 55 to 65, DB pension)has problems of its own. Any situation where you have to divide your lifetime savings into different silos with different conditions is always worse than if you didn’t have the silos. Running an ISA or SIPP flat comes with uncertainty if it’s exposed to the stock market. I can qualify this because my investment horizon is 5 years max and less than 2.5 years average[ref]more than half the money is out by 2018, the midpoint, because the PCLS front-loads it[/ref], so cash is a perfectly decent asset class to use. For a ten-year SIPP the investment horizon is on average a shade under 5 years, cash isn’t necessarily right for that. With an ISA you can aim to over-save – if you don’t run it down to 0 you can still get a tax-free income from what’s left. If the aim is to run the SIPP to 0 before drawing the DB pension then adverse investment performance could mean you flatten it before your DB NRA. However, you’d still gain by not drawing it as early as you’d have to otherwise – before Osborne’s changes I was going to draw my pension at 55 and invest the large AVC PCLS in ISAs over several years to compensate for the actuarial reduction. You don’t have the option to draw the State pension early, so you need to have some flexibility in your SIPP drawdown before 67 or whatever it is for you to not end up with an income hole.
The theory of this comes with a massive DYOR warning – it’s possible I’ve screwed up somewhere in the calculations[ref]original spreadsheet at google docs[/ref] The trouble with pension generalisations is that by definition people coming up to pension age have had a lifetime of complexity accrued in their financial story so no two people are even remotely similar. Saving your children from the requirement to earn their way by leaving an inheritance also complicates the issues.
Osborne’s changes have created an opportunity to front-run a pension which was previously only possible for people with £20,000 p.a. of other guaranteed pension income. That’s useful to some people, but like anything to do with pensions the numbers are big for many people – the average DC pension savings at retirement is £25,000. And readers who don’t find the numbers big will probably wonder what’s the point of sweating to save a trivial amount like £14k over five years 😉
|total savings grs||67000|
|no annual contrib|