How long do bond funds take to recover after a sharp rise in yields?


The short answer is that after a step increase in yields, the value of a bond fund will reach the value it would have had if the step had not occurred after an interval roughly equal to the effective duration of the bond fund (see the next post for a more detailed example). Different bond funds have different effective durations, for those matching the UK all gilts index such as iShares UK Gilts All Stocks Index Fund, the effective duration is currently about 9 years, while funds with other remits will have different durations, e.g., iShares Over 15 Years Gilts Index Fund has an effective duration of 17 years, while the iShares 0-5 yr ETF (IGLS) has an effective duration of just over 2 years.
Of course, for those worrying about the drop in portfolio values with the recent relatively rapid increases in bond yields from their recent historically low values, the more pertinent question is likely to be ‘what to do now’? While the following observations might be a starting point for discussion, I’m sure others will have different views.
For those still accumulating, the increase in bond yields is an opportunity to buy at reduced prices and, in the absence of further large increases in rates, to get a better return. Might be worth shortening the duration of bond funds held as retirement is approached.
For those approaching retirement or retired and holding long duration bond funds, there are essentially two alternatives
a) Hold and wait – eventually (after a period roughly equal to the duration) the fund will recover. Rebalancing will take advantage of the drop in prices (although ‘rebalancing’ to a shorter duration fund might also be an option).
b) Sell, take the hit, and buy shorter duration funds or other investments
Comments
-
A slightly longer answer to the question posed in the thread is illustrated in the following figure, where the Net Asset Value (NAV) of three different bond funds (‘short’ with a duration of roughly 3 years, ‘intermediate’ with a duration of 7 years, and ‘long’ with a duration of approximately 12 years) for two yield environments (constant yields and yields with a single step increase after 5 years).
For constant yield, as expected, the NAV increases at a constant rate with, in this example, the longer duration funds doing a bit better. With the sudden increase in yields introduced in Year 5, the NAV undergoes a strong decrease which is larger for the longer duration funds with falls in NAV of approximately 7% for the short fund, 20% for the intermediate fund, and 35% for the long duration fund.
What is interesting, is that the NAV then rises at close to the new yield rate (remember this is modelled as a single step increase in yields rather than the drawn out increases that we are currently undergoing) and eventually crosses the value the fund would have had if the change in yields had not occurred (marked with the circles in each case). For the short duration bond fund, the recovery occurs about 3 years after the step, while for the intermediate and long duration funds, the recovery periods are about 7 years and 12.5 years, respectively – i.e., in each case, roughly equal to the duration of the funds. However, it is also clear that the NAV of the longer duration funds does not exceed that of the short duration fund for a lot longer than is shown (over 30 years for the intermediate duration fund, more than 45 years for the long duration).
0 -
For those interested in the details of the results. To model the effect of a yield step, I’ve used my own implementation of the bond fund simulator developed by the bogleheads (see https://www.bogleheads.org/forum/viewtopic.php?t=179425&sid=ad6185a3f51c89f8778143ac0c83b837 ). For the case without a step (and for the years prior to the step), I’ve used par yields to maturity of 0.2%, 1.0%, and 1.2% at maturities of 6 months, 10 years, and 30 years, respectively (this is very roughly equivalent to the UK yield curve in 2021), while the par yields after the step were 3.8%, 4.6%, and 4.8% for the same maturities (roughly equivalent to the UK yield curve in 2002 - this doesn't have the inverted yield characteristics currently seen in 2023). The yields have been linearly interpolated for maturities between those stated. I note that there are simpler ways to model the response to a change in yields, e.g., using the modified duration of the fund to estimate the fall in price and then plotting the NAV at the new yield rate.
The bond funds simulated are a 5-1 (this means bonds are bought with 5 year maturities and sold once they reach a maturity of a year) for the ‘short’ fund, 15-1 for the ‘intermediate’ fund, and 30-1 for the long fund.
1 -
One of my guiding principles has always been "the Goldie Locks approach". ie don't go to extremes. So I split my money between pensions, general investing and extra mortgage payments and I own a bond heavy, US Vanguard, multi-asset fund (Wellesley - VWINX) with an average duration of 7 years to mitigate interest rate sensitivity.And so we beat on, boats against the current, borne back ceaselessly into the past.1
-
I'd agree. The original Bengen safe withdrawal rate paper used intermediate term US government bonds (which I think are the 5 year duration ones in the SBBI dataset) which provide, as you say, a nice compromise between their returns being (usually) higher than cash, but are less affected by interest rate risk than longer duration bonds. Interestingly, the bond component of the Vanguard Lifestrategy funds appear to have a duration of about 8 years too.
0 -
I took the hit and sold my bond funds. They were a small percentage of my portfolio anyway.
This coincided with my pension fund introducing a global equity index tracker which I wish I had had access to originally, rather than having to have a mix of global-ex-UK and UK equities. One fund to rule them all! I just wanted to put money in my pension without wondering if I've got the right percentages or needing to rebalance or all that malarky.
So I did a switcheroo over a few days. I put £10k in the cash account (which is actually more than I had in bonds) earning BOE base rate in interest and the rest into the global equities tracker.
This is an extra DC pot. I have one deferred DB and a second DB I'm currently paying into (until employer scraps it, which I fear is coming).1 -
IN real terms, given bond yields are probably correlated with inflation, doesn't the step basically reflect a change in inflation expectations so the 'recovery in value is actually an illusory nominal recovery?I think....0
-
michaels said:IN real terms, given bond yields are probably correlated with inflation, doesn't the step basically reflect a change in inflation expectations so the 'recovery in value is actually an illusory nominal recovery?
I think a similar response to the one I've plotted would be seen with inflation linked gilts, except the real NAV would be declining without the step (since real rates were negative) and subsequent to the step, the real NAV would have dropped because of the rate rise, but would then increase, in real terms, because real yields are now positive.
1 -
I retired a couple of years ago and prior to that had been increasing the bond allocation in my portfolio for fear of sequence of returns. However I was becoming aware of the 'profitless risk' of the low yields, mainly from this forum. I didn't entirely shift away from bonds, but decided to split the defensive side of my portfolio between bonds, cash and a 'wealth preservation' trust.I was surprised how fast and deep my bonds crashed, so I'm still nursing a nasty bruise in my bond portfoilio, and this question has been on my mind, but my plan is to hold, take the additional bond ncome, and use the other parts of my portfolio for drawdown.I was also wondering what would happen during a severe recession, i.e. would my bond portfoilio recover at that point for it to be used to cover that period.Time will tell, but I'm happy to be well covered by my diversification and not feeling troubled.Retired 1st July 2021.
This is not investment advice.
Your money may go "down and up and down and up and down and up and down ... down and up and down and up and down and up and down ... I got all tricked up and came up to this thing, lookin' so fire hot, a twenty out of ten..."0 -
Since I was curious, I thought I'd plot the response of an intermediate duration (i.e., 15-1 as before, duration~7 years) fund consisting of Inflation linked gilts using the yield curves from June 2021 and July 2023.
While the basic form is the same (i.e., there is a large drop with the rise in rates, a real return of -27% in this example), in real terms, the crossover takes 2 years longer (~9 years) than it did in the nominal case (~7 years).
0 -
So just to pose a simple naive question, I think people are saying that following interest rises, bond funds will gradually return to normal behaviour if not to previous values, as their existing bonds mature and are replaced. If that's roughly correct, what will happen if interest rates start falling?0
Confirm your email address to Create Threads and Reply

Categories
- All Categories
- 348.4K Banking & Borrowing
- 252.1K Reduce Debt & Boost Income
- 452.4K Spending & Discounts
- 241K Work, Benefits & Business
- 617.3K Mortgages, Homes & Bills
- 175.7K Life & Family
- 254.1K Travel & Transport
- 1.5M Hobbies & Leisure
- 16K Discuss & Feedback
- 15.1K Coronavirus Support Boards