Bond fund duration

BobR64

I have been dithering for too long and need to get my portfolio sorted, but find myself blocked by one main question that I am hoping I might be able to get some help with here: the question of what is an appropriate effective duration for a bond fund.

The portfolio is actually my SIPP but I am trying this sub-forum as don't think the question I have is pension-specific except in as much as we are looking at a fairly long time horizon from just before and throughout retirement.

I have been trying to understand bond funds for some time and I think I have arrived at some conclusions. This was after, among other things such as lurking on this forum, reading Lars Kroijer's and Tim Bale's books. Without going into too much detail, I want to go for a simple index fund plus bond fund approach, in proportions to be decided but probably around 70/30, and my rationale for holding bonds is to reduce volatility. I have been looking at UK gilt ETFs and have identified the following two:

• IGL5: iShares UK Gilts 0-5yr UCITS ETF
• IGLT: iShares Core UK Gilts UCITS ETF

IGL5 has an effective duration of 2.32 and IGLT and effective duration of 7.58.

My issue boils down to the fact that I have seen more than one person make each of two different recommendations:

• Time Hale (and the PensionCraft guy in a recent one of his videos) recommended short duration funds. That would favour IGL5.
• Lars Kroijer (and someone on this forum recently - unfortunately I can't find it now) recommended intermediate duration funds. That would tend to favour IGLT.

The rationale for favouring shorter duration as I understand it is that it is less subject to interest rate risk.

I am not really sure what the rationale is for intermediate duration. (Maybe a hand-wavy rationale is that a longer duration is more aligned with the time horizon over which they are held and that things tend to come out in the wash over a longer period.)

I would be grateful for any discussion or insights that might help me to understand how to reconcile the differing advice and help me come to my own conclusion.

A secondary question is that the IGLT fund is perhaps at the longer end of what would be considered intermediate, so, even if I do end up targeting intermediate duration, would it make sense to go for a mix of the two funds? 

aroominyork

I can think of two main reasons for favouring longer duration funds. The first is that, unless the yield curve is inverted, you get a higher yield; as with savings accounts, the longer you tie up your money for, the greater the return. Visit this BoE site https://www.bankofengland.co.uk/statistics/yield-curves and read the section about "Yield curve terminology and concepts". The second reason is if you believe interest rates will fall more/faster than expected and you think you can make a capital gain by buying longer duration. 

If you are buying individual gilts rather than a fund, then you might want a duration that equates with when you will need the cash; that way you know the exact income flows and that you will get £100 redeemed with no nasty surprises. 

Regarding IGLT being at the longer end of intermediate, by international standards that is true - a global government fund will have a year or two's shorter duration than a UK gilt index fund. What you do with that information is a matter of choice.

InvesterJones

I'd class 5-10yrs as intermediate, so IGLT is pretty much in the middle. Bonds can go out to nearly 49 years  

I don't know kroijer's rational for intermediate over short - it could be there's more of an inverse correlation with equities at the intermediate end so there's more smoothing effect. Or they're just targeting different points in the economic cycle - the short duration recommendation is particularly good when interest rates are being used to bring inflation down, but once that's done there's probably more yield gradient at the intermediate range

masonic

You haven't stated the role of these bonds in your portfolio, which is the most important determinant of what you should hold.

The rationale for intermediate duration is that the return is higher when the yield curve is normal, and they have the potential to rise more in recessions (when interest rates generally fall) to counterbalance equities.

Short dated bond funds behave like cash and generally should be expected to deliver below 1% real return. Consumer savings accounts provide an attractive alternative with rates often in excess of returns from short bonds.

I think you need to watch some more recent PensionCraft videos to get the latest view on bond duration, as it has changed in recent months.

In my case, the majority of my bond holdings are in Index Linked Gilts with a maturity of greater than 25 years, because I my objective is to have this capital available to me in mid-late retirement with a high degree of certainty that it will have greater spending power than it does today. It may be used to purchase an index linked annuity depending on circumstances then, or it could be drawn down as cash. I do not want to put this money at risk, including inflation risk.

InvesterJones

True about the latest view at pensioncraft - the recent many happy returns podcast asked which part of the bond yield would you take in a fund for a hedge to equities, and it was medium duration.

BobR64

Thanks all for the comments. As always some good insights and some things to ponder.

Regarding @masonic's question:

masonic said:

You haven't stated the role of these bonds in your portfolio, which is the most important determinant of what you should hold.

The rationale for intermediate duration is that the return is higher when the yield curve is normal, and they have the potential to rise more in recessions (when interest rates generally fall) to counterbalance equities.

The role is intended to be what I understand to be the "usual" one of being a way to reduce volatility and help me sleep at night. My understanding is that we hope with bonds that there is going to be some inverse correlation with equities and it is this that helps with volatility. I think what you are saying here is that the rationale for intermediate duration is consistent with what I am trying to achieve.

Regarding PensionCraft as mentioned by a couple of people: I have obviously missed some of the more recent ones, so I will catch up with those.

masonic

BobR64 said:

Thanks all for the comments. As always some good insights and some things to ponder.

Regarding @masonic's question:

masonic said:

You haven't stated the role of these bonds in your portfolio, which is the most important determinant of what you should hold.

The rationale for intermediate duration is that the return is higher when the yield curve is normal, and they have the potential to rise more in recessions (when interest rates generally fall) to counterbalance equities.

The role is intended to be what I understand to be the "usual" one of being a way to reduce volatility and help me sleep at night. My understanding is that we hope with bonds that there is going to be some inverse correlation with equities and it is this that helps with volatility. I think what you are saying here is that the rationale for intermediate duration is consistent with what I am trying to achieve.

Regarding PensionCraft as mentioned by a couple of people: I have obviously missed some of the more recent ones, so I will catch up with those.

Yes, it's the interest rate sensitivity of the intermediate duration that gives you the inverse correlation. Since interest rates tend to be cut during recessions to stimulate growth.

OldScientist

Where yields tended to increase (e.g., from 1940 to 1980 in the US and UK and possibly some other countries), shorter duration fixed income (i.e., MMF and short duration bonds) did better than longer duration gilts.

Where yields tended to decrease (e.g., from 1980 to 2020 or so), then longer duration bonds did better than cash or shorter duration bonds.

Therefore, from a performance point of view, intermediate gilts would lie somewhere between the extremes and represent a neither worst nor best performance and is not making a bet on future results.

While the nominal volatility of short bonds will be less than intermediate and less than long, real volatility also depends on inflation (which itself can be quite variable).

In other words, a case can be made for holding either short or intermediate bond funds (or both), but there is little case for longer bonds..

Until fairly recently, available datasets tended to only have returns for cash and longish bonds (e.g., in the macrohistory.net US treasuries are nominally 10 years, while, IIRC, UK gilts are consols through to the 1960s, 15-20 year maturities until the 1980s/90s and the 'all stocks' gilt index after that) although the freely available simba spreadsheet (https://www.bogleheads.org/wiki/Simba's_backtesting_spreadsheet ) does contain estimates of US bond returns for funds at a range of maturities and some analysis of the effect of bond maturity on SWR is presented at https://www.bogleheads.org/forum/viewtopic.php?t=412851 .

In passing, I note that 'intermediate' for the US appears to be shorter than 'intermediate' for the UK, but practically anything under 2-3 years can probably count as short, anything over 10 years can probably count as long and anything in between shades from one to the other.

FWIW, we have ended up with fixed income of an average duration of between 1 and 2 years (a mix of fixed term cash accounts, short global fund and an 'all bonds' global fund since my OH likes cash and it reduces volatility on our variable portfolio withdrawals (probably at the expense of long term returns). Fixed income is the only area where I allow myself a bit of active investing - I tend to decrease duration when I think yields will fall and vice versa. Overall it will have little effect on the outcomes but, for me, scratches an itch to tinker!