UK government bonds have given investors a painful kick in the portfolio recently. Many of us found out bond funds are riskier than we realised. But with one simple(-ish) metric you can assess the riskiness of your bond assets ahead of any market crash. That metric is bond duration.
Quick note – Duration applies to bond funds, individual bonds, and portfolios of individual bonds. I’ll mostly just refer to ‘bonds’ throughout the article because it’s snappier. I’ll specifically call out bond funds when duration applies differently to them. Please check out our bond jargon buster for a brief refresher on confusing bond terminology.
What is bond duration?
Bond duration expresses a bond’s vulnerability to interest rate risk. The larger the bond duration number, the more reactive a bond’s price is to interest rate changes, as the bond’s yield adjusts to reflect those changes.
For example, if a bond’s duration number is 11, then it:
- Loses approximately 11% of its market value for every 1% rise in its yield1
- Gains approximately 11% for every 1% fall in its yield
Whatever your bond’s duration number2, that’s how big a gain or loss you can expect for every 1% move in its yield.
A duration three bond will rise or fall in value by approximately 3% if its yield moves by 1%.
In a rising interest rate environment? Shorter duration bonds will be less risky than longer duration equivalents. But they won’t do much for you when rates fall.
Conversely, long duration bonds are more comforting than your favourite teddy bear when interest rates fall. They go up in price!
But that would-be teddy bear is about as welcome as a grizzly at a picnic when interest rates rise.
What affects bond duration?
A bond’s time to maturity, yield, and coupon rate determine its duration:
Remaining time to maturity
The more coupon payments a bond has yet to make until it matures, the more price-sensitive it is to interest rate changes.
That’s because a long-dated bond is stuck with its fixed interest advantage or disadvantage for many years in the future. A short-dated bond has only a few more payments due.
- A distant maturity date implies a higher duration.
- A near-term maturity date implies a lower duration.
Yield and / or coupon payment
Bonds with lower yields / coupon payments are more price-sensitive than similar types with higher yields / coupons.
- A higher yield implies a shorter duration – because the bond returns your money at a faster rate.
- A lower yield implies a longer duration.
The diagram below shows the tug-of-war that resolves a bond’s duration:
- Higher durations are primarily a function of longer bond maturities. Low bond yields / coupons also contribute.
- Lower durations are primarily a function of shorter bond maturities. High bond yields / coupons also contribute.
Higher durations equate to a more volatile bond price (up or down) when interest rates change.
Lower durations mean smaller price swings.
All this helps explain why long duration bonds took horrible losses in 2022. As interest rates escalated, bonds trading in the market became less valuable.
Though it’s little comfort right now, duration also sheds light on why long bonds stepped up in value when interest rates plunged during the Global Financial Crisis.
The ups and downs of being a bond
Bear in mind that duration is an approximate measure. It makes various simplifying assumptions about the relationship between interest rates, bond prices, and yields.
But it helps to remember these opposing bond dynamics:
- When interest rates rise, bond prices fall.
- When interest rates fall, bond prices rise.
- When bond prices fall, yields rise.
- When bond prices rise, yields fall.
Long bonds react more violently to these forces, for good or for ill.
When interest rates rise, investors demand more compensation for tying up their money in bonds.
New bonds entering the market must have higher coupon payments to match the rate increase.
But longer bonds are saddled with their old, lower, coupon payments for years – even decades. So their price falls to reflect their less competitive fixed rates.
That price cut pushes the old bond’s yield up. It rises to the point where it’s just as attractive to a buyer as a new bond (of the same type) that waltzes in flashing its higher coupon payment.
An analogy with cash savings accounts might help.
Let’s say you’re in the market for a fixed-rate savings account. Now suppose that interest rates had just risen from 3% to 4%. There’s no way you’d pick the same 3% account you might have gone for yesterday. At least not without a hefty bribe cashback offer.
Bonds on sale
The discounted price of a less competitive bond is a bit like cashback given to new buyers to make it just as profitable as the new bonds they’d otherwise choose.
In the savings business, banks withdraw old, fixed-interest accounts from the market. Existing savers, however, are stuck with their outmoded choice. Curses.
With government bonds, debt obligations are seldom taken off the market. Instead they’re priced at a discount or premium to reflect their altered competitiveness, as interest rates yo-yo.
Naturally, the process works in reverse, too. You earn a premium on bonds boasting a yield higher than prevailing interest rates.
Bond duration captures the short-term capital gain (price premium) or loss (price discount) part of these moves in one simple number.
(Although even this this isn’t the end of the story. Counterintuitively, bond funds have higher expected returns after a price drop. That’s due to the impact of rising bond yields.)
How far do bond yields move?
It’s all very well saying duration measures the price change sparked by a 1% yield move. But how far – and how fast – can bond yields bounce in the real world?
It’s the size and speed of your bond’s yield change that determines the scale of your capital loss or gain.
Below is a snapshot of UK gilt yields, with changes in yield over the course of a day, month, and year:
Source: Trading Economics.12 October 2022.
The daily, monthly, and annual shifts in yield shows you the impact of recent changes in market interest rates for each UK government bond in the table.
You can see, for instance, that the yield (note: not the price) on the UK’s benchmark ten-year gilt rose 1.4% in the last month. In the last year the yield is up 3.5%.
Indeed, every gilt with a maturity of three years or more saw its yield increase at least 1.35% in the last month, up to 3.8% over the past year.
If you multiply shifts of that size by duration then that’s going to hurt. As every bond fund owner knows all too well in 2022!
Moving too fast
The duration calculation assumes instantaneous moves. But the longer the change actually takes to unfold, the less violent the price swing. Reinvested cashflows mitigate the impact.
So it’s not quite right to multiply duration by real world movements that evolve over a year.
All the same, the size of the yield rises in the table above show us that just multiplying your bond’s duration by 1% doesn’t nearly capture the scale of the drama that can engulf us.
Which bond is my bond fund like? – Compare your bond fund to individual bonds of the same type. Look up your fund’s weighted average maturity. It’ll behave similarly to an equivalent individual bond with approximately the same maturity. The yield-to-maturity of the fund and the bond should be pretty close. Do check the dates though. Published bond fund yields can be quite stale.
Bond duration: making your money back
There’s another aspect of bond duration which is much more debatable.
This assertion is that your bond’s duration number tells you how many years it takes to recover from a capital loss after a yield rise – your breakeven point.
Or, to put it another way: how long it will take to make the annualised returns you expected before rising yields put a dent in your portfolio.
Let’s say you own a duration 11 bond fund, with a yield-to-maturity of 4%.
Interest rates go up, prices go down, and your bond’s value takes a hit.
However, your bond fund fully makes up that lost ground by the 11 year mark. At that point, you’ve now earned a 4% annualised return over the entire period going back 11 years. The scar of the price drop has healed. It’s as if the interest rate rise never happened.
Beyond 11 years, you’re up on the deal. That’s because your higher-yielding bonds pay you a better return than you would have received without the rate rise, when the yield would have remained lower.
All this assumes that your coupon payments and maturing bonds are reinvested.
The maths work the other way round, too.
If yields fall, then your bond return immediately jacks up (capital gain). But ultimately your returns soften like a tyre with a slow puncture. Beyond your duration number (expressed in years), you’re worse off over the whole period, because your cashflows are reinvested into lower-yielding bonds.
The downgrade in return happens to a duration 3 bond after three years. A duration 11 bond has more staying power. It wouldn’t show a worse annualised return until 11 years passed.
Here for the duration
All this is rule-of-thumb stuff. It works just fine for an individual bond that’s held until maturity, declining in duration as its coupons pay out.
However ‘holding for the duration’ is less applicable to bond funds operating in the real world.
In reality, bond funds turn over their holdings to keep the fund’s average maturity and duration relatively stable. The same goes for rolling portfolios of individual bonds.
Moreover, interest rates don’t change course only once, and then remain static. They weave around like a drunk at a wedding reception.
The traditional advice is to match your bond duration to your time horizon to ensure you get your money back.
But that is based on assumptions that are about as realistic as diesel emissions tests.
Indeed, there’s evidence to suggest you may have to wait for up to twice your bond fund’s initial duration in years to earn your initially expected yield-to-maturity.
The twice duration rule-of-thumb
This rule of thumb says that twice your bond fund’s initial duration is a better guide to your breakeven point.
Of course, you could earn your initially expected return faster if interest rates trend down and you enjoy a series of capital gain boosts.
But when your holding period is dominated by rising rates then twice duration is a more pragmatic time horizon.
This bracing finding comes from a research paper: Constant-Duration Bond Portfolios’ Initial (Rolling) Yield Forecasts Return Best at Twice Duration. The author is Gabriel A. Lozada, associate professor of economics at the University of Utah.
A hat tip to Occam Investing. Occam pointed to Lozada’s research as part of a very good piece on bond returns.
The ‘twice duration’ paper specifically investigates the returns of bond portfolios held at constant durations. It employs a more realistic model for fluctuating interest rates than allowed for above.
The author also empirically tested his model versus 60-years worth of historical returns.
Lozada’s conclusion is you’re more likely to earn your initial yield-to-maturity over a twice duration timespan in a world where interest rates can go for a random walk, or trend upward for decades.
A better, not perfect, guide
Here’s the key finding for ordinary investors:
In summary, almost all the time, initial yield was within a percent or two of average annual realized return with a horizon of twice initial duration.
Are you a fellow glass-half-empty type? Then know this rule-of-thumb looks more rigorous than the happy-clappy ‘just hold for the duration’ advice of old.
But the message isn’t that it will definitely take 22 years to earn say a 4% annualised return from a duration 11 bond fund. We’re not trapped by some boa constrictor of fate.
If interest rates stayed relatively flat for the next eleven years, your bond’s yield would be about what you could expect.
And if rates go down then you may earn more for a while. Though longer-term you’ll likely earn less.
But given that interest rates are inherently unpredictable – and could relentlessly trend up – estimating that it could take somewhere between your bond fund’s duration and twice its duration to earn its yield is the hardheaded approach.
Even then, this doesn’t tell you much about real returns.3
Other complicating factors
If inflation is higher than expected, nominal bonds do poorly. If inflation is lower than expected, nominal bonds do relatively well.
It’s true too that if you pound-cost average into your bond fund then you’ll:
- Shorten your path to higher overall returns in a rising rate environment
- Shorten your path to worse overall returns in a falling rate environment
Lozada says his findings apply to default-free bonds that aren’t callable (i.e. high-grade government bonds), but notes they aren’t a good fit for long bond portfolios.
Sometimes the best fit was 1.25 times duration or 1.75 times. Much depended on the type of bonds and the time period Lozada put under examination.
Lozada also didn’t look at what happens if you periodically rebalance, withdraw cash, or spend interest. Or any other of the common investor behaviours that influence your particular outcome.
Twice duration then is no more than a rule-of-thumb for short and intermediate government bond fund risk. Albeit a more steely-eyed (steely-thumbed?) one.
If that doesn’t sound especially reassuring then check out the Banker On Wheels bond ETF calculator.
This suggests the twice duration rule should be reserved for gloomy scenarios when rates rise constantly during your time horizon.
Duration and convexity
Duration simplifies the real world complexity of bond maths to broad strokes. Convexity fills in more of the detail.
Convexity provides more accurate insights into bond price sensitivity because it accounts for the fact that yield changes also alter a bond’s duration.
Picture the difference in price outcomes between the two measures like this:
The relationship between bond prices and yields is curved, whereas duration assumes it’s linear.
The practical outcome is:
- Duration (white line) tends to underestimate the bond price rise (green line) when yields fall. (Left-hand side of pic).
- Duration typically overestimates price drops when yields rise. (Right-hand side).
The difference between the green line and the white line reveals convexity at work. The convex curve of the bond price shows how it differs from the duration estimate as yields change.
When bonds exhibit positive convexity (as pictured above):
- Yield falls, price spikes, duration lengthens (duration underestimates actual price rise)
- Yields rise, price drops, duration shortens (duration overestimates actual price fall)
Essentially, the lower yields go, the faster bond prices accelerate versus duration’s estimate.
Meanwhile, the higher yields float, the slower bond prices decline vs duration’s readout.
Convexity amounts to a welcome tailwind. One that enhances your portfolio protection in a falling rate environment. And moderates expected bond damage in rising rate conditions.
The effect is barely noticeable for short bonds. But is pronounced at extreme ends of the yield curve, as bond maturities head over 15 years until maturity.
Portfolio Charts has produced some fantastic graphs that give you a proper feel for convexity.
And we demonstrate convexity’s effects in our bond prices post.
Incidentally, watch out for negative convexity. This occurs when bonds become less price sensitive as yields fall. (And vice versa). It’s the exact opposite of what you’d want to happen.
Negative convexity is not a concern for default-free, non-callable government bonds. It is a worry if you stray into corporate / municipal bond territory where call options rear their heads.
If convexity is more accurate then why does everyone use duration? Mainly because it’s simpler, but also because duration is good enough in most circumstances.
Where to find bond duration numbers
A bond fund’s home page should tell you its duration number.
Though as usual, providers love to shower us with a confusion of different terms.
Average duration – A bond fund’s duration is the weighted average of the individual bond durations that it contains. So no cause for alarm if you see this label.
You can flip Vanguard’s site to the financial advisor view (wee dropdown menu, top-right, on desktop) to see its duration figures. For some unearthly reason you can’t see them on the consumer site.
Modified duration – Strictly-speaking the correct term for the type of duration that measures price sensitivity to interest rate changes. Use this number where you see it.
Effective duration – Modified duration diluted by the effect of any bonds with call options in the portfolio. Effective duration trumps modified duration if a fund gives you the choice.
Use Trade Web to find out the modified duration for individual gilts.
If you’d like to calculate bond duration then check out this calculator.
Beware that duration doesn’t capture every dimension of bond risk. Credit quality is another major factor – and duration does not address this at all.
Bond risk: higher or lower?
As a Brucie Bonus, bond funds actually become less risky after the yield rises and the price falls.
I appreciate that’s in complete contrast to our instincts after big capital gains and losses. But the eagle-eyed might have noticed their own bond fund’s duration shorten following the recent falls.
For instance, here’s how the key numbers have moved for the SPDR’s intermediate gilt ETF (ticker: GLTY):
On 30 April 2020:
- Duration: 13.85
- Yield-to-maturity: 0.35%
On 30 September 2022:
- Duration: 10
- Yield-to-maturity: 4.09%
The fund’s yield is vastly improved while its lower duration number shows its price sensitivity is less volatile than a couple of years ago.
The fund is now a better investment prospect than it was in 2020! But as ever after a big investment shock, some people will be too bruised to go back for more.
Investing often defies our human intuitions. And bond investing perhaps most of all.
Take it steady,
The Accumulator
PS – When we mention ‘interest rates’ in this post we’re referring to bond market interest rates, not central bank interest rates. References to ‘yield’ mean yield-to-maturity. Please see our bond jargon buster for more.
- Yield to maturity. You can think of it like the interest rate you’ll get if you hold the bond to maturity. [↩]
- Technically, it’s called ‘modified duration’. [↩]
- Real returns subtract inflation from your investment results. In other words, they’re a more accurate portrayal of your capital growth in relation to purchasing power than standard nominal returns. [↩]
What’s that I hear banging in the wind? 😉
Thanks TA, especially useful explanation about Convexity. I read on Bogleheads a good estimate for breakeven is 2D-1. But it’s much the same I guess.
Does this apply to inflation linked bond (funds) too? Assuming inflation constant, I guess so?
Btw, can anyone help me with the duration of GIST/GISG? If you average the holdings of the fund, it looks about 5 years. I just suspect that there’s something sneaky where the longer duration bond holdings can’t simply be averaged with the shorter ones to come up with a mean. TIA.
Thanks TA, that’s another article that merits some sort of public service award – not the easiest subject to summarise, industry jargon included.
For people already familiar with NPV calculations via other topics, is there any disadvantage to using the same framing here? Future cashflows for an existing bond are fixed/known. But the choice of discount rate can be thought of in opportunity-cost terms, with fresh/rival bonds as the relevant opportunity. So rising yields (in the market for new issues) drive down the valuation of existing bonds/funds.
Only relevant to a niche audience, of course. But I wonder whether a small minority of people encounter NPV calcs in general before they think about bond funds for isas/pensions.
@ Brod – yes, I saw a few sources talk about twice duration -1 and Lozada talks about that too but he says twice duration is the best fit. However, as you say, it’s neither here nor there. I tried to quote elements from the paper that show this stuff is an approximation at best. Certainly, the rule isn’t as clean as once duration but neither can twice duration or 2D-1 be relied upon with precision either. It depends on the bond and the path of future interest rates. Best to think of it as a potential range between once duration and twice duration.
Re: index-linked bonds, Lozada has a cautious paragraph saying that the paper’s theory implies the same outcome for linkers but he hasn’t tested it using historical data.
GIST/GISG – Lyxor publish the modified duration as 4.85
@ Ballard – cheers! Trying to make bonds digestible is the stuff of nightmares. I’d say I’ve failed by some margin here. The piece is more than double the length I hoped for 🙂
@TA – thanks for the duration. Never thought of going to the Lyxor site. Doh!
So my method of saying they’ve got equal-ish percentages in 1 to 3 / 3 to 5 / 5 to 7 and 7 to 10 Years so it’ll be 5 years cos that’s in the middle was good enough. I rock!
Interesting article particularly the link to the paper suggesting duration x2 to minimise risk. As an ex mathematician I will study that with interest.
There is one point that has been omitted from your description is the effect of the shape of the yield curve, that changes the numbers considerably.
The yield curve is presently slightly inverted but broadly flat, making descriptions of the change in maturity much simpler, however with a steep yield curve, which is not uncommon it affects the numbers a lot. I appreciate you don’t want to over complicate the description of the mechanism of bonds but it can have quite an effect, might have been worth a mention, particularly if the yield curve changes during the holding period considered….
Thank you for this, but my head still hurts. I have just lost a packet in INXG (or rather I haven’t because I haven’t sold them), and my question to the panel is: what have bonds got that 1 year deposits with banks and BSs paying 4.5%, have not?
I also found this post helpful… it’s all very well explained. But I’m wondering how one would act on this in practice. The implication seems to be that the bond fund duration should not be matched to the time horizon, but actually rather shorter, if not half of time horizon. It makes a big difference in terms of investment choices (I think I remember Lars Kroijer issuing the “match to time horizon” advice”). So it’s very useful to have this discussion.
This also leaves me wondering how investors should react given that time horizon reduces as the years go by. For example, if an investor has a 20 year time horizon, and invests in a bond fund with average duration of 10 years, the time horizon is going to reduce to 19 years, 18 years… should the investor be gradually transferring into a shorter-duration fund? Or simply selling all units in one fund and switching to another? What criterion would trigger the switch?
@B. Lackdown
You generally can’t earn 4-5% on cash if it’s in your pension (although there are SIPPs which do allow you to deposit money into interest bearing accounts (I transferred money into one just before interest rates plummettd and ended up transferring it out again)). In addition, you only get £85,000 of FSCS protection, so you’d potentially have to spread your money around a bit.
Suppose:
in 2010 we buy 100 into BonFun with duration 10 yield 1%
in 2020 yield rises to 2%, cap val falls 10% wiping out the 2010’s distribs,
in 2030 we are up to about 110 with the 2% distribs, so back to 2010
wealth after 2*duration. But we lost 10 years.
… by 2040 we have about 130 which means we have got the principal and all 30 years of the 1% yield that seemed such a great deal in 2010 back.
Ah! so the deal is you get the yield you bought at after 2*duration from the
rate rise, not from 2*duration from when you bought. So anyone buying 10 years before the rise hoping to retire 10 years after rise has to adjust their plan
sorry folks, that maths is wrong.
should be: you have 120 by 2030, so you still get the 1%yield after 20 years that
you origianally wanted. except now inflation has been 2% instead of 1% through the 2020’s, so you still find yourself wanting 10% more money, which your BonFun agreed back in 2010 cannot provide.
@ B. Lackdown – I would hold some cash, but it doesn’t offer you the opportunity to make a big capital gain in the event of an equity market downturn. This piece shows how that worked during some of the 21st Century’s biggest drawdowns: https://monevator.com/are-bonds-a-good-investment/
Index-linked gilts offer inflation protection but INXG embeds too much interest rate risk:
https://monevator.com/why-uk-inflation-linked-funds-may-not-protect-you-against-inflation/
GISG as mentioned by Brod above is an option as is Royal London’s Short Duration Global Index Linked fund (GBP hedged).
Both these funds have much shorter durations than INXG, so you take less risk but get reasonable inflation protection.
There are more ideas here:
https://monevator.com/60-40-portfolio/
@ Haphazard – I’m going to write a piece on duration matching – but as I’ve researched it I find myself wondering who actually does this with funds and why?
The idea of duration matching is that you earn a certain yield to maturity over a particular timeframe which enables you to match bond investments to your liabilities. But what is your time horizon? Say you retire at 60. Is your time horizon 20 years? 30 years? 40 years? Life expectancy? And how do you account for inflation?
Some commentators used to describe duration matching as a method to avoid losing money. I don’t think that holds up at all.
Your questions are spot on. How does it work exactly? The more I look into it, the less viable it seems for funds. Duration matching with individual bonds looks doable. With funds, possible in theory but very difficult in practice.
Anyway, I haven’t finished my research yet so maybe I’ll stumble on some fresh insight before I’m done.
@TA (#12):
According to LDI Misapplied, Idzorek & Blanchett, 2017:
“Common LDI approaches include the ultra-conservative approach of cash flow matching (matching the timing and size of cash flows from the assets with the required cash flows of the liability), followed by duration matching (matching the interest rate sensitivity of the asset portfolio to that of the liability), and liability-relative optimization, also called surplus optimization.”
My understanding is that for an individual cash flow matching LDI (which is what I think you are describing rather than duration matching) can only ever work with non-callable bonds held to maturity. And inflation protection is either built-in before the fact using an estimate and/or you use linkers.
However, even using this approach, there are issues around max duration of bonds and the availability of bonds for each year required. Most such schemes seem to assume bonds can be rolled at their duration with no change in characteristics – which of course may well not be true – if coverage is needed beyond that provided by the longest duration bond that can be purchased, and the next nearest available bond will suffice for “missing” years.
Someone once said: do not invest your money in something you do not understand.
To me it’s bond funds.
Just take Vanguard inflation linked bond fund. Everage Joe like me would most likely think that, when inflation goes up then such fund also goes up. In reality, the complette oposite has happened.
Keeping money in simple savings account solves the safe part of my portfolio, the rest is shares.
@Accumulator – on the what is your time horizon question: I suppose it depends what you’re investing for. I can see that for someone investing a pension in drawdown, it’s not very helpful. But for those of us saving for a thing with a date (retirement, or something else), I suppose it’s the time until that date.
Perhaps the “target date” idea with retirement has become a bit cloudier with the popularity of drawdown over annuities. I know a lot of pension funds and target date funds now assume their beneficiaries will want to go into drawdown. But for those of us with a more conservative mindset, who want to annuitise some or all of what we have, there really is a “date”.
The same would apply to someone saving for a child’s university fees, etc.
Perhaps we could have Lars Kroijer back on and ask him what he thinks about all this?!
@TA Id have thought the chance of decent capital gains haven’t been there over the past couple of years, with interest rates near zero and more likely to rise if anything? From here it appears short duration bonds do have something to offer over cash with known capital at term. Assuming UK plc remains solvent of course! I’m still questioning the merits of long duration bonds as a reserve asset (which might be underwater for decades).
@ Peter – that’s fair comment. What’s happened with UK index linked funds has been horrendous.
@ Calculus – I understand your view and I wouldn’t personally invest in a long bond fund. At the same time, bond yields are up and they still have the potential to shield a portfolio during a demand slump. Imagine something happens that throws everything into reverse. Say, a new variant of COVID, pandemic is back on, but this time we dan’t money print our way out of it. Bonds will be a lifesaver.
I’m extremely wary of the ‘conventional bonds underwater for decades’ scenario you flag, too.
We live in uncertain times when – in just the last 2 years – we’ve witnessed huge lurches from near deflation to rapid inflation. I say let’s be ready for anything. Hold a slug of complementary assets so something will always perform.
The downside of that diversification is we’ll inevitably be left cursing one asset class or other. But I’d rather that than guess and be completely wrong.
@ Haphazard – I agree, if you stay in drawdown then duration matching with funds is of very limited use. Wanting to have a set amount of cash available to invest in an annuity does make sense as a target. An annuity is actually a much better way of duration matching in retirement i.e. it’ll produce a fixed income to meet my liabilities every year I’m alive.
Although as Al Cam points out, it needs to be index-linked to protect you against inflation.
@ Al Cam – I agree with you. Duration matching makes more sense with individual bonds. If you needed coverage beyond the longest dated bond then hopefully you’d still have some equities to sell. In reality, I’d solve this problem with an annuity.
@TA: (#17):
In some ways an ‘ideal’ solution is to build your ladder to age X and also purchase a deferred annuity (DA) that kicks in from age X. In a very real sense such a DA is purely longevity insurance; and the younger you are when you purchase it the cheaper it should be! However, I am not sure if such a DA product is even available!
Inflation protection should be a consideration, but the precise level of protection required may be less, or indeed more, than what is provided by index linking.
Thank you very much for another helpful & informative post
@ Al Cam – I don’t think deferred annuities are available in the UK.
@TA (20) With the exception of deferred care needs annuities, which I have personal experience of! Regarding deferred annuities for retirement income, for example to extend a bond ladder, I believe even US writers concede that the main problem with them is that, in their market, they are not available with any inflation linking.
Given current high yields, I’m considering a gilt ladder in my SIPP from age 55 (next year) to age 67 when state pension kicks in. Inflation aside, are there any other considerations or reasons to go for a shorter ladder, say 6 years, then repeat? Or any compelling reasons to use bond funds as part of the solution?
I’ll have no other meaningful income so thinking it sensible to take the tax-free 25% PCLS at 55, empty the remainder of the SIPP over 12 years to make use of my personal allowance, then use the taxable state pension and ISA after age 67 (potentially deferring state pension if need be). ISA would start as 100% equities with some phased into less riskier assets as 67 draws closer.
I like the idea of certainty of income with 3-4% gilt yields until the floor of the state pension arrives, whilst allowing a global tracker to (hopefully) do some heavy lifting over those dozen years or more.
Any pointers very much welcomed.
@DavidV (#21):
I believe it is not possible to currently buy any annuity (including a ‘normal’ lifetime income annuity) in the US market that is inflation linked; and that this has been the case for some time now. Probably a coincidence but, if you like a a conspiracy theory, might just make you wonder what US life insurers knew back then!
@Norman (#22):
Your plan sounds rather familiar to me.
FWIW, I would note the following:
a) there are costs and time associated with buying gilts and are you sure you can/should hold what you need in your SIPP;
b) have you considered using fixed rate savings bonds (or CD’s as our US cousins like to call them) in lieu of bonds, which will be tax free if held in ISA(s); I do not know of any CD’s that last for 12 years but IIRC 7-year terms exist and CD’s could be stacked and rolled as required too; this is what I did and as long as you are not too fussy about the precise spacing of the steps of your ladder IMO it can work well, albeit that it requires some work
c) plans change – often due to circumstances outside of your control – like you I originally planned a twelve year ‘gap’ to SP with a ten-year gap to my DB coming on stream. The ten-year gap will probably end up being around seven years now – largely, but not entirely, due to HMG changes;
d) Any decision to delay taking current SP may not be VFM – and I think it is even less compelling if the triple lock remains and high inflation persists;
e) flexibility has real value; e.g. with the benefit of hindsight I bought too much income for my gap but at least I was able to access and use the excess if I wanted – which would not be the case with annuities;
f) it is not unusual for pre-retirees to over-estimate their needs post jumping ship; but then again shxt happens so I would also hold an instant access emergency fund too;
g) to date (some six years after jumping ship) my assumed built-in inflation rate has exceeded reality albeit that the gap is rapidly closing and my interest received is somewhat higher than forecast too.
h) using the gap for tax arbitrage to at least your personal allowance and/or the annual ISA limit can work, but as usual all tax calcs are very situational (family, etc) and the scenario/circumstances could change too
Hope the above is of some use to you. Please remember this is just my experience and YMMV. Please DYOR.
@ Norman – Wade Pfau is very good on bond ladders. Here’s some links:
https://retirementresearcher.com/laddering-with-individual-bonds/
https://retirementresearcher.com/category/assets/bond-ladders/
I’d construct your rungs from index-linked bonds if possible so you’ve got built in inflation protection.
This report from the Debt Management Office shows you the gilts available to build your ladder from:
https://www.dmo.gov.uk/data/pdfdatareport?reportCode=D1A
Most of the major brokers buy/sell gilts but I’m not sure they cover the whole market so this link may also prove useful:
https://www.dmo.gov.uk/responsibilities/gilt-market/buying-selling/purchase-sale-service/
@Al Cam – Many thanks for taking the time to post a detailed reply. That’s extremely useful. A few quick answers in response:
a) I can’t buy gilts via my current SIPP provider, so it would require a transfer. Interactive Investor is looking the best bet for my situation. With no other income I’ll need cash from age 55 so it seemed prudent from a tax perspective to withdraw from a taxable investment vehicle now rather than my ISA.
b) I hadn’t considered CDs and would need to check their availability in a SIPP. As noted above, it seems sensible to use my SIPP now to make use of personal allowances. Will check out CDs though, thanks.
c/d) I have no DB pension, so really only the one gap to plan for from age 55-67. I’d make a call on deferring SP at the time but any other planning at this stage would be based on taking it at 67. I’ll work (with some contingency) on the basis that the government don’t shift the SP age goalposts again.
e/f) My thinking is to withdraw more than required annually from my SIPP and put any excess into my S&S ISA and/or add to an emergency cash ISA fund.
g) Index-linked products might be the best way to go but I’m heavily swayed by the current high yields on nominal bonds. Despite the recent dip I’m assuming yields will rise again if the BofE keeps hiking interest rates. But I need to look closer at index-linked gilts and crank some numbers.
h) Agreed. For example it’s not impossible that I’d have taxable earnings in the 12-year gap. However I think unlikely it would push me into a higher tax bracket when combined with SIPP withdrawal and that would probably be my main consideration.
@TA – thanks for the useful links and the tip re index-linked bonds (and of course this excellent series of bond articles). As mentioned above, I need to look closer at linkers and not be seduced by healthy-looking nominal yields.
Index-linked gilts are causing my head to explode. Can someone please put me out of my misery:
1) If I calculate yield to maturity in the same way as for nominal bonds, would an ILG with 0% YTM generally imply I will receive a return at the rate of inflation?
2) If inflation were to return to the BoE target level of 2% in the near future, would I be quids-in buying nominal bonds with YTMs higher than 2%, or is there a cumulative aspect to the ILG RPI uplift? I got lost at Base RPI numbers.
3) Is it fair to say that prices of nominal bonds already reflect expected inflation, therefore returns should work out the same as ILG? So the main benefit of ILG is protecting against unexpected inflation?
Thanks in advance.
@Norman (#26):
Perhaps it would have been clearer had I originally used “reversibility” rather than “flexibility” in e).
f) Ideally, I would want my emergency fund available from day one rather than being built over time.
h) As TA has previously described, high inflation paired with frozen allowances/bands is a pretty corrosive mix. I suspect such tactics are HMG’s favoured method for [surreptitiously] raising the tax burden!
@Norman (27) yes to all 3
A bond fund is not all bad, it’s effectively a bond ladder that is constantly refreshed to retain the initial duration characteristics.
Individual bonds allow you to tie down what you get and when, from a behavioural perspective you can ignore the intermediate value of bonds, however the changes in value and the market interest rates are merely disguised not eliminated….
In general terms if you anticipate holding bonds for a long time then rising interest rates are good news, even if bond fund prices fall in the interim.
If your need is to finance consumption at particular points in time then individual bonds can provide certainty, logically this must be the objective, not maximising returns if there is a chance of failure to maintain living standards.
Inflation linked bonds can only protect against inflation that is not already priced in, the assumptions behind that “pricing in” may be wrong… Perhaps protecting the ‘essential’ part of your living costs is prudent.
@Al Cam – thanks again for the reply.
e) I think for my purposes over that particular timeframe I’m favouring certainty over anything else. Certainty of SIPP returns at those specific points in time and no tinkering temptations. I’ll probably leave a bit of flexibility in the ISA though.
f) Agreed. I should’ve said that part of my 25% tax-free PCLS would form the basis of an emergency fund.
@Hariseldon – thanks for confirming those 3 points. A bond fund or ETF could perhaps form part of my longer term ISA holdings beside the equities component. That’s something I’ll consider down the road when rebalancing.
Overall I think linkers would do the job required, so today’s job will be to try and profile a gilt ladder over that 12-year period. There are currently only 8 index-linked gilts in issue over that timeframe so I’ll need to see what the spacing would look like. I believe the DMO are issuing more gilts this week but I haven’t read any details and have no idea if these become instantly tradable on the secondary market.
Choice of platform is the other issue to resolve. Interactive Investor’s fixed fee SIPP is a decent choice for the size of my pension and it offers online gilt dealing (which I like), but not the full choice of gilts. The AJ Bell SIPP seems to be the other alternative from a charges perspective – percentage fees but capped (for gilts) at much the same level as II. I think AJ Bell does offer the full range of gilts. You can’t view them online but I’ve confirmed they offer all linkers at least and waiting on a reply to hear about nominal bond offerings. So it’s phone dealing only, albeit with online charges. Neither platform charges for pension withdrawals, so they would appear to be the frontrunners.
@Hariseldon (#29):
Re: “Perhaps protecting the ‘essential’ part of your living costs is prudent.”
Whilst I now agree that flooring your whole lifestyle is probably unnecessary [and may be prohibitively expensive] it took me several years and lived experience to reach this conclusion. A problem however remains – how much floor is sensible? Personally, I have always struggled with essentials vs discretionary splitting of expenses and have not found any other convenient rule of thumb/guidance in the literature. One thing that did occur to me is that your Covid level expenses may offer some practical clues. What do you think?
@Norman (#30):
If “over that particular timeframe I’m favouring certainty over anything else” then a fixed term annuity might also be worth a look. IIRC these are available.
Remember though that annuities are irreversible whereas bond ladders are reversible, albeit possibly with some dealing costs and exposure to the vagaries of intermediate price, etc gyrations. Were you to ladder CD’s it is worth noting that early surrender penalties have significantly worsened over recent years – meaning that reversing a CD ladder may have issues too.
@Al Cam – Thank you again.
I can’t find much to go on regarding CDs but they seem to be minimum purchase of £100k which would be far too high a denomination for my purposes. From what I’ve read the maximum duration is 5 years (although I also read there are some longer) but the majority are apparently 6 months and under. Coupled with the fact I don’t think I can buy them via a SIPP probably rules them out.
I had a go with a fixed term annuity quote and was surprised with the results. A £200k pension pot would pay out £26,187 annually (in arrears) for a 10-year term with no balance at end of term, so £261,870 in total. Paid in advance the amount reduces to £24,810 annually, so £248,100 in total.
Doing a very rough bonds calculation based on purchases of £200k in gilts, using a YTM of 3.5% as an annual coupon rate (coupons not reinvested), gilts being redeeemed at £20k per year over 10 years, would permit £23,850 to be withdrawn per year, so £238,500 in total. Don’t know if that’s in any way a decent approximation.
Does that comparison sound right to anyone? Can anybody advise as to how these should compare?
@Norman:
Most banks/building societies offer fixed rate savings accounts or bonds (a daft name but that is what they call them) with t’s & c’s not dissimilar to normal savings a/c’s – or at least they did the last time I looked. See for example https://www.skipton.co.uk/savings/fixed-rate-bonds
CD’s are a US product of the same generic type – ie a guaranteed interest rate for a fixed period with the benefit of [limited] deposit insurance (FSCS in the UK). Fixed rate ISA’s also exist, see e.g. https://www.skipton.co.uk/savings/fixed-rate-isas
@Al Cam – Yes, but I know of no way of holding these in a pension wrapper. I can’t get past thinking that using personal allowances is the best approach when I have no other income, so I’m really only looking at my SIPP to fund the 12-year gap period. Unless anyone can suggest a compelling reason to use my ISA instead.