There’s a table you’ve probably seen on just about every investment platform known to humanity. It shows recent returns history and looks something like this:
Cumulative performance
| Investment | 1-year (%) | 5-year (%) | 10-year (%) |
| All Stocks gilts | 5 | -24 | -0.8 |
Nominal cumulative total returns 2015-2025. Data from JST Macrohistory 1, FTSE Russell, and British Government Securities Database 2. May 2026.
This kind of data is so ubiquitous it’s only natural to believe it must be helpful.
For example, it enables you to make quick fire comparisons over what seems like quite a long period of time:
| Investment | 1-year (%) | 5-year (%) | 10-year (%) |
| All Stocks gilts | 5 | -24 | -0.8 |
| Money market | 4.3 | 16.9 | 19.5 |
The conclusion looks obvious in this case. Gilts (UK government bonds) have been a disaster. Cash has quietly ticked along. If you want a defensive diversifier to offset equity risk, then the numbers speak for themselves.
Except they don’t. They’re only telling you something about the recent past.
Given the way we’re wired, though, it takes a hefty dollop of willpower not to extrapolate those numbers out into the future. Like an implicit join-the-dots exercise.
But on your guard or not, the table is still an attribute substitution honey trap! What we want to know is whether an investment will do well in the future.
That’s an impossible question to answer so the table plays to our cognitive biases, and invites us to subconsciously smuggle in an easier question, “What has the investment returned over the last 10 years?”
Beguiling figures like this fulfil our need for a quick resolution but deny us the full picture.
A postcard from the last war
The five and 10-year gilt returns in this table don’t tell us that bonds are broken or that cash is the superior investment over time.
Rather, it’s mostly a record of one seismic event: the interest rate shock of 2021 to 2023. When rates rise sharply, existing bond prices fall.
UK government bonds lost over 40% in real terms during that period. 3 And that asteroid strike is now baked into the return figures like the line of iridium which marks the end of the dinosaurs.
This isn’t evidence of a chronic problem with bonds. Gilts had one very bad year in the past half century. (Indeed 2022 was the second worst year on record in real terms.)
But that loss – when spread out across the ten-year return average in the table – makes bonds look like a long-term loser.
The nuance, the underlying cause and how it applies, the market awareness – the table skates past all of this.
Most importantly, it doesn’t show the silver lining. That the same rate rise which massacred bonds in 2022 simultaneously reset yields to the point where expected returns from bonds are considerably higher now than before.
It was all going so well
Let’s look at the same table as it appeared at the end of 2020:
| Investment | 1-year (%) | 5-year (%) | 10-year (%) |
| All Stocks gilts | 8.2 | 30.5 | 70.4 |
| Money market | 0.2 | 2.3 | 4.3 |
Bonds were crushing cash! Once again the conclusion is obvious. Only this time it pointed in the opposite direction.
The table flattered bonds in 2020 – pumped up as they were by falling interest rates in the wake of the Global Financial Crisis and the pandemic.
As Covid-19 vaccines were rolled out, many investors fretted that a similar shot in the arm of the economy could spell trouble for bonds as rates rose again.
But neither they, nor the table, could predict the scale and speed of the interest rate snapback. They couldn’t predict how fast the global economy would reopen, or the size of President Biden’s economic stimulus, or Vladimir Putin cutting gas supplies, or central bank dithering, or fire-starter Liz Truss as prime minister.
In retrospect the bond massacre looks inevitable. In reality, it was contingent.
So the table didn’t just fail to warn you. It actively pointed in the wrong direction relative to the risks, twice.
And these issues aren’t just a problem with bonds.
Hold my beer!
US equities and gold look amazing right now in similar tables due to their multi-year hot streaks.
- Does their run-up in value signal a tottering Jenga tower of risk piled upon risk?
- Or has the playing field fundamentally tilted in favour of these markets?
- Or are these cycles perfectly normal (if three-body problem unpredictable) when you examine the behaviour of risky assets?
Bet now!
A better picture
The next chart compares UK government bonds against the money markets over multiple periods from one to 50 years.
Orange means gilts won over a particular time-frame. Green means money market won:
- The numbers in the boxes show the winning asset’s lead in percentage points.
- The rows enable you to see which asset class led at the end of each year.
For example, money markets beat All Stock gilts by 1.8 percentage points annualised over the ten years up to 2025. Whereas, gilts beat money markets by 3% per year (on average) over the 10 years 2011 to 2021.
All numbers are inflation-adjusted.
As you can see, while the money markets score some wins, especially over shorter timeframes, gilts dominate overall.
And gilts maintain their edge over the very long term, too.
Real annualised returns to year-end 2025
| Investment | 75-year (%) | 100-year (%) | 126-year (%) |
| All Stocks gilts | 1.2 | 1.4 | 0.8 |
| Money market | 0.8 | 0.4 | 0.4 |
It’s so over for money market funds – they earn half the long-run average of gilts!
Actually, it’s so not over…
When money markets win
The mosaic chart above shows that 1981 was the last time money markets scored a 10-year victory over gilts before 2022.
That’s because interest rates and inflation were stratospheric in the 1970s.
These are the known failure conditions for nominal bonds: inflationary environments where spiralling prices wreck fixed income returns and central banks push rates higher to limit the damage.
To be fair, both asset classes are typically hit hard in these circumstances. But it’s better to be caught in a shorter duration interest-bearing asset like cash when inflation stalks the land.
So what happened when Britain last experienced a long period of rising rates?
Real annualised returns by decade: rising rate environment
| Investment | 1950s (%) | 1960s (%) | 1970s (%) |
| All Stocks gilts | -3.7 | -1.7 | -3.3 |
| Money market | -1.8 | 2.1 | -2.7 |
Good grief! The money markets did beat gilts for three decades (and change). Even though cash-like funds were clearly no picnic at the time either.
From my perspective, this reminds me that even a 126-year long-run return, shorn of context, doesn’t tell me everything I need to know about the relative merits of two asset classes.
During that particular period in history, successive British Governments stamped on the interest rate brakes to contain episodic inflationary surges – but they eased off again too soon as unemployment rose, setting the conditions for the next CPI pressure wave.
It was a terrible time for bonds but cash made a huge loss too.
Gonna need a bigger framework
I’ve come to the conclusion that return tables alone are a seductive but misleading tool. They compress a complex, time-dependent story into a single number that skips the ifs and buts.
I don’t believe that you, me, or anyone that we could hire can predict the future.
If it was so damn easy then why was anyone holding bonds in 2022?
And if bonds are doomed now, why is anyone still holding them?
It’s because bonds aren’t doomed. Their expected returns are better now than they were in 2020, as I’ve already mentioned.
Nominal government bonds also have a specific strategic role to play in portfolios as an:
- Equity diversifier
- Deflation / disinflation hedge
- Volatility dampener
- Refuge in a demand-led recession
So much for bonds. But people will dump gold and equities too next time they run into serious trouble. Mostly when it’s too late already.
We clearly need a better framework for deciding which assets to hold.
The minimum viable alternative to a quick returns comparison
I think a strategic investor should ask:
- What role does this asset play in my portfolio?
- Under what conditions does it work? When does it not?
- Why might it continue to work in the future?
- What’s my back-up if the asset fails for a protracted period?
There are various tools at our disposal to answer these interconnected questions.
Financial theory
This helps explain what assets are for, their sources of return, and so whether we have reasonable grounds to expect the investment to work in the future.
Expected returns
Enable you to take a view on the prospects for an asset class in the years ahead.
The advantage of expected returns is that they’re informed by current market conditions. Hence they can be a useful corrective for the very human tendency to project out recent trends.
The disadvantage is that market conditions can change quickly.
It’s important therefore not to take expected returns too seriously. They’re not forecasts. They’re formulas that are easily defeated by the unforeseen.
Long-term asset class history
The long term view reveals how each asset class performed during different economic regimes.
This enables you to understand:
- When it works
- When it doesn’t work
- How regularly an asset class experiences conditions that cause it to thrive or dive.
- How often does an asset class experience negative returns? How long and deep can those drawdowns be? Can you live with that?
If you hold an asset class as a diversifier, for example, then does it actually work? That is, does it have a track record of diversifying the appropriate risk?
For instance, if you hold an asset that’s reputed to rise when equities fall, how often does it do that? Once or twice in spectacular fashion? Or on a recurring basis?
Under what circumstances does the diversifier fail to respond? Does it actively flourish when equities drop, or just limit the damage by falling less far?
Ask whether an asset can behave the way you need it to, when you need it to. What are the chances?
Bear in mind that if an asset class wilts in unfavourable circumstances for such an investment, that’s evidence it’s behaving as expected, not that it’s useless.
Every asset can win big, drift sideways for years, dive underwater for a decade, behave unexpectedly… If you think you have found something that doesn’t, think again.
Ask how much of this asset should I hold given I know it can fail badly for extended periods?
Ten years worth of returns tells you next to nothing. A quarter of a century doesn’t really cut it.
Fifty years is okay and 100 years is good. Starting from 1900 is ideal.
Don’t rule out sepia-tinged events just because they happened a long time ago. I’m specifically thinking of the Great Depression or major wars.
Granted, the economy has changed. But the nature of risk has changed less so. Recall the dictum: history doesn’t repeat but it rhymes.
Predict the unpredictable
Most of all, stay lively to recency bias and resist plausible but simplistic theories. The world is rarely so neat. Bolts from the blue can upend current trends without warning.
The world wasn’t preparing for a pandemic in 2019. People weren’t talking about AI before Chat GPT3 launched in 2022. (Zuckerberg was betting on the metaverse at the time, for goodness sake).
Remember that everything you know is already priced in. For example, demographic decline and the size of government debt.
Embrace uncertainty and risk. That’s the source of your excess returns over those who go nowhere in cash.
Take it steady,
The Accumulator
Bonus appendix – even more gilts vs money market tables
I wrote up these tables then cut them from the main article. I’ll leave them here in case anyone finds them useful.
Nominal annualised returns to year-end 2025
| Investment | 1-year | 5-year | 10-year | 15-year |
| All Stocks gilts | 5 | -5.3 | -0.07 | 1.8 |
| Money market | 4.3 | 3.2 | 1.8 | 1.3 |
Real annualised returns to year-end 2025
| Investment | 1-year | 5-year | 10-year | 15-year |
| All Stocks gilts | 1.6 | -9.8 | -3.3 | -1.2 |
| Money market | 0.9 | -1.7 | -1.5 | -1.6 |
Gilts only achieve a real positive return on a 22-year view. Money market on a 28-year view.
Nominal annualised returns to year-end 2020
| Investment | 1-year | 5-year | 10-year | 15-year |
| All Stocks gilts | 8.2 | 5.5 | 5.5 | 5.3 |
| Money market | 0.2 | 0.5 | 0.4 | 1.3 |
Real annualised returns to year-end 2020
| Investment | 1-year | 5-year | 10-year | 15-year |
| All Stocks gilts | 7.3 | 3.7 | 3.4 | 3 |
| Money market | -0.7 | -1.2 | -1.5 | -0.9 |
Long-term real annualised returns
All Stocks gilts
- 1.2% (1900-2020)
- 0.8% (1900-2025)
Money market
- 0.49% (1900-2020)
- 0.4% (1900-2025)
- Jordà O, Knoll K, Kuvshinov D, Schularick M, Taylor AM. 2019. “The Rate of Return on Everything, 1870–2015.” Quarterly Journal of Economics, 134(3), 1225-1298.[↩]
- Cairns A, Wilkie D, ESCoE Historical Data Repository. “Heriot-Watt / Institute and Faculty of Actuaries / ESCoE British Government Securities Database.” ESCoE.[↩]
- Most UK government bond funds follow the All Stocks gilts index.[↩]
Thanks for continuing to provide a constant stream of thought provoking articles.
A couple of thoughts on expected returns – if you hold gilts to maturity then the returns will be pretty much defined by the yield. The thing you can’t predict if you buy conventional gilts is inflation. So agree real returns on conventional gilts are unpredictable. Real returns on inflation gilts much more predictable (again if held to maturity)
(Geek corner) I wonder how much gilts over-performance over money market funds could be anticipated through he shape of the yield curve. Yields on gilts of different terms (in fact come to think of it there was a Mavens article looking at whether those differences might be exploited ). Now yields on short-dated gilts tend to mirror MMF returns quite closely. But there is often a “term premium” where longer term gilts have higher yields. So my question if we wind back the clock 30-years to 1994 and see what the term premium on a 30-year gilt was – how does this compare the actual outperformance of gilts over MMFs over the 30 years to 2024 – 1.7 if I’m not mistaken.
I picked 1994 because that’s when I first started a proper job, and looking at gilt yields curves was part of the job. I’d guess the term premium then was about 2.5% (though it was a long time ago). So relative to that gilts underperformed 1% compared to what you might expect based on the what
And after 30 years of saving hard and investing wisely (and no small degree of luck accumulating during one of the greatest bull markets of all time) , I’ve now decided to hit the RE bit of FIRE!
Hit submit before proof reading, for some reason didn’t get the option to edit my first post but could edit the second! Typos:
A) Yields on gilts of different terms…are not the same.
B) gilts over-performance over money market funds could be anticipated through The shape of the yield curve
C) So relative to that gilts underperformed 1% compared to what you might expect based on the…Term Premium
And a couple of follow up thoughts. The term premium comparison would only be valid if comparing with an investment in 30-year gilts. In practice I expect the term of the gilts used to determine the performance in the tables above is much less than this.
And where does the line between passive and active end when investing in gilts? If I pick the gilt fund with a certain term or but individual gilts because they match my expected spending is that active antics?
What if I’m influenced by the current shape of the yield curve…?
Those tables are like “we had to put something” – graphs tell more of the story, but really if it gave you the peace to stay the course then it did it’s job regardless
Not sure it does still give the peace though as we see rates can and do go up
Also everything that’s known may be priced in, but each person’s tax situation is different, so it’s value will be different to someone who doesn’t have sufficient cash ISA options
Are the money market returns in the first table in the appendix really annualised? Even for nominal returns, 21.8% over the last 15 years seems a lot.
Excellent piece.
Bonds will be back. Or not…In any event, time will tell, as it does always.
The lesson for me is multi strategy as well as multi asset, multi market, and multi sectors
Diversification really is the only free lunch out there, but, IMHO, you really have to take the implications of it seriously, and to recognise that it means much more than just holding different shares, indices and assets. It also and perhaps most valuably means adopting different approaches and styles of investing, because every dog has its day, and all systems and strategies are somewhat regime dependent.
Excellent article once again – illustrating, amongst other things, the critical effect start and end points have on comparisons.
FWIW, my favourite tool for comparing returns is Bogle’s TellTale Chart (there’s a nice article from 2002 that can be searched for) which is “devised simply by dividing the cumulative returns of one data series into another, year after year”. IMV, the advantage of the TellTale chart over simple ‘growth of $1’ graphs is that the periods when one asset was beating another become so much clearer.
@prospector (#1, #2)
According to the data at the BoE, the 20 year spot yield at the end of May 1994 was 8.2%, while the 6 month spot yield was 5.0% and 4.8% at 3 months so the term premium was just over 3 percentage points. AFAIK, there were no 30 year gilts in issue at the time (the longest issued in 1994 were two maturing in 2010 – both of which only had cover ratios of 1.2 at auction).
Excellent article, with bonds we can almost ignore previous returns, apart from the long term trends.
What matters most is the present yield to maturity , the shape of the yield curve , inflation outlook , credit quality looking forward, when do you need the money ?
Long duration bonds can provide equity like returns ( and losses) anyone who loaded up with long bonds 20 to 30 years ago did very well but when the yield to maturity dropped to very low levels , that was the signal to massively shorten duration , same with linkers , buying on 2.5% real makes more sense than holding a bond on -3% !
We live in interesting times ….
PP for the win!
Set it and forget it!
NFA, DYOR
@Reactive – Thank you for pointing that out! You are right those numbers are way too big. I pasted the money market nominal cumulative returns into the appendix tables instead of annualised. Sorted now, thank you.
@Alan S – I haven’t heard of Bogle’s TellTale Chart before. I’ll check it out. The Bogleheads have a post on it here: https://www.bogleheads.org/wiki/Telltale_chart
@DH – I think that’s right. One of the things that worries me is that it’s extremely hard to understand what loss feels like unless you’ve experienced it personally. (Like feeling another’s pain… we can empathise but we can’t really understand how godawful it is until it’s shot through our own nervous system.) One of the possible outcomes of that is when major league investing losses do show up, they’re so shocking that we panic, that we never want to experience it again, that we display all the symptoms of shell-shock that yesteryear’s investors exhibited.
For example, I’ve published all these charts about 70% losses in the bond markets or equities etc but I’ve not experienced anything like that myself. I really don’t know how I’d deal with it in actuality.
In the meantime it seems like the best I can do is to diversify across different sources of return, not think I can avoid risk, try to be as indifferent as I can be my own opinions about the direction of travel, look at more charts showing giant losses 🙂
@PPYS – LOL
As a decumulator, my focus is on cashflows rather than asset/portfolio values/performance. As such, I don’t hold nominal bonds at all – linkers to maturity and a large slug of MM to cover my drawdown needs until pensions arrive.
And I find Harry Browne’s PP underpinnings a useful framework. Linkers will cover my cashflow needs in two of his four states (inflation and stagflation); stocks in a growth scenario and deflation… um, with fiat currencies I think this the least likely scenario and I’ll muddle through with the reduced value of maturing linkers topped up with MM fund.
Also, a flexible ISA is very handy. I can withdraw money in early April; purchase a linker; sell in late March; receive the inflation protection for that slug of money and a bit of yield; and pop the money back into the ISA; rinse and repeat. Kind of like a inflation linked cash account. Unless events move massively against me over the turn of the financial year, all should be good?
Or have I missed something obvious?
Many great points here. I think the lessons from historical data are limited. The sample is quite small, the world changes a lot over decades, and we may be stuck with unhelpful data. E.g. nominal returns instead of real; or asset class correlations based on short-term volatility (the usual way this is calculated) instead of long-term which matters more.
So, for FIRE the goal is to build real wealth over the long term, with the main problems/risks: long-term stock market crashes, inflation and longevity.
I agree with @Brod that disinflation is less likely because politics is pushing the other way. And less of a problem (if we manage stock market risk) because it keeps our liabilities constant too.
> What role does this asset play in my portfolio?
This has been the key question for me. Each asset should have a clear “job” towards the goal and/or the main risks. This helps remove needless complexity and to hold on to the important stuff through the inevitable bad periods.
The PP framework is very useful but I wouldn’t recommend the PP. Not diversified enough, too much in inflation-prone assets, not enough in stocks.
@Brod – linker prices move quite a lot, especially towards the end of their maturity. I think there was a Monevator piece about this.
And the spreads are high for retail investors.
@ Sparschwein 11
Yes the PP has issues but for risk-adjusted returns on a long-only portfolio, it is hard to beat!
Plus if you are very averse to drawdowns it is also very hard to beat!
https://www.optimizedportfolio.com/wp-content/uploads/2026/04/permanent-portfolio-performance.png
When I ask myself what’s missing from our savings & investments the answer is Foreign Currencies. Suppose I fancy Singapore dollars, Swiss Francs and Norwegian krone. What’s a good way to hold them?
(True, fixed interest bonds are missing too but my policy is that I’ll buy those only when the circs are so favourable that even a coward should jump in – such as in late ’99 when it was obviously the time to sell our equities.)
@Sparcschwein – thank you for your response.
If “linker prices move quite a lot” at the end of their maturity, then I can buy onbe further out? I’m only seeking to capture the inflation adjustment for the 11 and a half months I’m holding it.
Spreads may be a problem, but don’t they tend to collapse from the quoted spreads?
@dearieme – if you hold e.g. an All World equity fund, you’ll get plenty of currency exposure there.
@Brod but if you buy and sell further out, you may not get an inflation linked return. It would depend on any change in the real yield while you held it.
Maybe I’ve misunderstood your aim, but if you aren’t going to hold all the way to maturity you might consider buying and holding something like this fund in your flexible ISA:
https://markets.ft.com/data/funds/tearsheet/summary?s=GB00BN091H11:GBX
Again, there’s no guarantee you get a return that exactly matches inflation, but the duration is short enough that you wouldn’t be far off most of the time. You’d lose about 5% if real yields rose by 1%pt (and would have lost a bit more than that in 2022).
@Tetronimo – thanks for your help.
To be honest, this was just a thought experiment that seems to have got a little out of hand. I am building a linker ladder in my GIA by withdrawing monthly from my SIPP and buying a linker rung every month or so, and I wondered if I could accelerate the process by using the flexible ISA to take out the lump sum and purchase all today (rather than monthly over this financial year) and replace the money in my flexible ISA before April 5th. This then somehow morphed into the above.
Sorry for wasting everybody’s time 🙂
@Brod, no, you’re not wasting anyone’s time. Always interested in new ideas or conundrums.
I don’t see any problem with your latest description if it’s a one off, i.e. you use your flexible (cash?) ISA to fund the GIA, plan to hold those linkers to maturity in the GIA, and you know you can replenish the flexible ISA from the SIPP before tax year end.
I had misunderstood your original outline. I thought you were trying to work around the lack of flexible ISAs at good platforms for individual linkers. And my earlier suggestion definitely wouldn’t work, seeing as tax treatment means you don’t want to hold that fund in a GIA.
@Brod – a very short duration gilt (linker or not) sold with a couple of months to maturity – the price should be pretty close to par. So you could make a small gain or loss relative to holding to maturity but it shouldn’t matter much. I don’t think it’s correct that linker prices move around a great deal towards the end. They should move relatively smoothly towards par.
You can see this if you look at the clean price of TR26 (GB00BYY5F144) before it matured in March 22nd this year. It scarcely moves because the bond is so short in the last few months. (See prices on Tradeweb).
The yield goes nuts. But that’s just an artefact of small changes in price looking artificially large when expressed as an annualised metric. The changes are minuscule if calculated based on time to maturity.
Inflation indexation of bond lags the RPI index by 3 months. For example, May RPI informs the gilt’s indexation from 1 August. That could throw you if there were big changes in inflation just before you sold which you were expecting to show up in the dirty price.
The price will drop after the last coupon payment. Again, the adjustment may look artificially large if you weren’t expecting it.
You’ll pay a transaction fee you didn’t need to but I guess that’s neither here nor there if you’re doing it online with a broker like AJ Bell and not over the telephone.
You’d be selling T27 4 months early. (Maturity date of Nov relative to the following March). 7 months early for T28 and TR31 which seems like a lot. and TR31. T3oI is an 8-month indexation lag linker and would be sold 8 months early.
Back to 4 months early for T32 and T33.
Looking at the maturity dates, I think I’d let them mature and put the money in a money market fund for the remaining months until you need the dosh.
Buying a linker out further along the curve and then selling means a stronger likelihood of capital loss (or gain).
I have a feeling there’s something about your plan I’m missing entirely but hope this helps 🙂
So what happens with bond funds if the gov switches to mostly issuing short debt?
So there would be a period where the bond fund duration shrank until it eventually turns into a money market fund – then there could be no argument about which was better!
And, because the central bank controls the short rate, the gov could pay 1% again while inflation runs at whatever it likes?
I suppose linkers would need to be nobbled somehow.
@Brod – you’re right about the linker spread. The quoted spread from AJB is huge (1-2%), but what I paid recently was actually quite close to the mid-point from AJB’s bid/ask, and that was quite close to the LSE price around that time.
I haven’t sold any linkers though. Presumably you’d get a similar deal.
@PPYS – PP stability is indeed impressive. Looks like the comparison starts in the early 70s when gold had the one-off boost from the gold standard? And it had another huge run recently.
I’d come back to my earlier point about limitations of historical data vs thinking about allocation from first principles today…
I’d absolutely not do the PP with nominal bonds, potentially half the portfolio ruined by inflation.
With the 25% bonds in a linker ladder, it may be quite good as a conservative portfolio for wealth preservation or drawdown… something to ponder/discuss.
@Meany – interesting question.
I suppose if central bankers were to act as govt lackeys and set the rate artificially low, then inflation would run hot, capital flees the country, currency drops. It would probably only work with tight capital controls. We’ve been there before, after WWII. Or the whole world may do zero-interest & QE again, then there’s nowhere to go for real returns.
Linkers have already been nobbled with the RPI-to-CPIH change. Another index change?
@Meany (#21)
It depends on the bond fund, the upper limit of maturity for new issues, and the proportion of already existing bonds at long maturities. But, yes, the weighted maturity or duration of something like All Stocks (which holds all nominal gilts in issue) has been highly variable. For example, the duration of the All Stocks index reached a minimum of 4.4 years at the end of 1974 and experienced a local maximum value of just under 13 years at the end of 2020 (low yields and coupons also contributed to the long duration). Since the issue of significant quantities of dated gilts in 1915, the peak duration for all stocks was nearly 25 years at the end of 1934 (at the time the largest issue matured in or after 1952, the next largest in 1962, and the third largest group of issues were perpetuals – in other words, the index was heavily weighted to long maturities).
The effect of issue weighting on funds with defined maturity ranges (e.g., under 5 years, under 10 years, etc.) is less marked, but yields and coupons also play a part in determining duration.
@sparschwein 23
Well that is kind of the point of the PP – something always comes along and saves you – that is how it is designed. I believe Monevator has done an article on the PP and highlighted this.
233 years of returns can’t be wrong!
https://www.lazyportfolioetf.com/allocation/harry-browne-permanent/
Not sure how the 233 years worth of returns are measuring gold. The price is essentially flat before 1968 due to various iterations of the gold standard.
Set that aside, I agree with you, PPYS, that the Permanent Portfolio is hard to beat on a risk-adjusted basis.
But we can only fairly measure the PP back to 1968 and probably shouldn’t start the clock until about 1975 due to the one-off gold price discovery issues noted by @Sparschwein.
That’s still 50 years worth of returns but essentially misses the bulk of the rising rate era from 1946.
It’s deeply unfair to the portfolio but if you measured say the 30 year period from 1946-1975, the PP is terrible.
The PP looks good again if you sub out gold for commodities during that era.
Plus the PP is formulated before the advent index-linked bonds.
So it seems to me there’s a contemporary version of the PP that includes allocations to linkers and commodities alongside the rest.
Then it features the two assets best fitted to deal with stagflationary regimes.
Guys,
See: https://retireearlyhomepage.com/reallife26.html for real life* ability of Harry Browne (aka PP) and other portfolios to support 4% SWR from 1994 to 2025. JPG, an early FIRE pioneer, is IMO a star!
*inc. fees and other frictions, etc
@Alan S – thanks for those numbers.
Right, so a bond fund in the mid-70’s had got much lower duration hence much safer to buy in then, even though the 70’s inflation was ongoing.
But then, presumably, when rates finally top out, being in longer bonds would be a good move but an all-bonds fund would be concentrated in shorter gilts and not get so much of a bump.
This seems another worry with big bond funds – they put you in slightly the wrong place.
i.e. if you were investing using bonds in 1974, was your best strategy:
– get the top paying gilt below time horizon, or
– use MMF until good leg up in rates then a time-apt gilt, or
– get the bond fund: lose to inflation for the rest of the up-leg then very slowly dictated by the gov issue duration shift up into higher payers. hmm..
Didn’t mean to end so conspiratorial above btw, it may be even more interesting q why we need duration at all, and how an honest market operating freely would change without it. – is duration just a convenience from old tymes because they didn’t have the tech to reissue/reprice everything 100 times a second?
Talking of conspiracy nuts, my reading of Harry Browne’s PP was that he thought any or even up to 3, but very unlikely all 4 of the quadrants could go to zero. So the PP idea is really to get 4 times as much as you ought to need with a quarter in each then you should make it through almost anything. FIRE people by contrast try to get 30+ years use out of 25 years’ money with investment and the 4% rule.
@Meany – your best strategy depended on the future path of interest rates which was unknowable.
All Stocks gilts were bad in the 1970s. Third worst decade since 1900.
Best decade ever was the 1980s because yields had been driven so high in the 1970s but then inflation was brought down. It wasn’t a done deal. Bond holders suffered a 30+ year bear market before that point because inflation wasn’t contained.
Alan’s the expert on the composition of the All Stocks gilts index but duration is mediated by yield AIUI. So I think the reduction in duration was, to some extent, caused by high yields. Would be great to hear what Alan has to say on that 🙂
Very interesting re: Harry Browne. One question I have about the Permanent Portfolio is why did Browne allocate so much to gold given that – at the time he was writing – gold only had 10yrs history as an investible asset, and for most of Harry’s lifetime it was strictly controlled by the government? Did he caution that private gold could be made unlawful again in the US? Or did he believe that era was firmly consigned to history?
@Brod “@dearieme – if you hold e.g. an All World equity fund, you’ll get plenty of currency exposure there.”
Yeah, but I’ll also get equity exposure that I don’t want. The same argument applies against an international fixed interest bond fund. How to hold just the currencies, that’s my question.
Maybe an international index-linked bond fund might be a least bad possibility. Any other suggestions?
@Accumulator 30
Harry Browne seems to have been a wise man. I think it was JP Morgan who said “Gold is Money, all else is credit”.
Maybe Harry had a worst case scenario in mind – if the financial world collapses, Gold is all you have left.
It is interesting to note that many older civilisations which have been around the block a few times (China, India and Asia in general) value Gold in a way that Westerners cannot understand. It is understood as portable wealth. These civilisations have seen governments, stockmarkets, money, come and go. Only Gold endures. And family! Which is why they also have very strong family ties.
@TA Based on the short history in the Rowland book on the Permanent Portfolio, it seems to me that Browne would have seen his 25% gold allocation as small because of the path he’d taken.
By which I mean he made good returns from gold, silver and the Swiss franc by investing in them before Nixon broke the gold standard, but then he and his team wanted to diversify further to protect what they had made.
The book doesn’t quote an allocation but I take it to mean he would have been reducing his gold allocation from some very high percentage down to 25%.
Incidentally, looping back to your earlier comment, the book also says the original plan for the Permanent Portfolio was more diverse, and may have covered your point on commodities. It included: stocks, bonds, cash, gold, silver, Swiss francs, and natural resources. Then later on it was simplified to the first four.
@TA / @PPYS
>for most of Harry’s lifetime it was strictly controlled by the government?
if you want a quick overview of Harry’s attitude to gov & gold,
page 132 of his book
https://dn711000.ca.archive.org/0/items/banks-for-nutin/Harry%20Browne%20-%20How%20I%20Found%20Freedom%20in%20an%20Unfree%20World.pdf
– it’s what PPYS suggests, Harry wasn’t just insuring against finance collapsing, the whole US government might go down and he would be in with a fighting chance. It’s basically the Prepper Bible, as y’all probably know.
@dearieme – I think one gets a Spreadex account and learns fx options
@dearieme as you suggest, any of the unhedged, shorter duration fixed income ETFs could do that job for you. There are unhedged dollar versions of ERNS (ultrashort bond), and TI5G (0-5 year TIPS). There are also unhedged 1-3 yr nominal US treasury ETFs.
(I’m quoting the hedged versions here just because I’m familiar with them, but you can find the unhedged versions easily enough on the iShares site – take a look at ETFs like ERNA and TIP5).
Thanks, Meany, tetromino.
The foreign currency I really don’t want to be invested in is USD. It faces a dramatically eventful future (in my guess).
Regarding ways to speculate on currencies, there are plenty of currency etp/etfs listed in the uk out there which one can buy in a sipp, isa etc
Wisdomtree does a few
Ta, PPYS. I’ll take a gander.
@PPYS, Meany, tetromino – thanks for that. That makes sense. I guess Browne wasn’t thinking about UCITS funds and had a very impressive vault!
@tetromino –
“the book also says the original plan for the Permanent Portfolio was more diverse, and may have covered your point on commodities. It included: stocks, bonds, cash, gold, silver, Swiss francs, and natural resources.”
There is a Permanent Portfolio branded fund that contains exactly that combo:
https://permanentportfoliofunds.com/permanent-portfolio.html
I guess they took inspiration from the book.
@TA (#30)
Gilts: at the end of 1974, par yields were over 15% for virtually all gilts with maturities of over 10 years (the exceptions were some low coupon gilts that had lower yields because of their tax advantages). Consequently, the Macaulay duration, and modified duration were low. For example, the 7.75% gilt 2012-2015 (this gilt was actually called in 2012) had a yield of 17.3% and a MacDur of 6.3 years and ModDur of about 5.8. It is also worth noting that by issue weight, just over 60% of gilts were under 10 years.
@Alan S #40: I believe that for a UK investor 1974-76 was (with perfect foresight) the best time in the post 1945 era (and possibly ever) to put money into both Gilts and UK equities.
I don’t unfortunately have the actual concrete figures to hand for shares; but IIRC (based on what I recall reading somewhere some time ago) some blue chip companies in the FTSE Actuaries 30 Index (as it was then) were, at the very lowest point of the market despair in those years, priced on TTM PEs of 2 x to 4 x, and we’re giving investors dividend yields as high as 10% to 12% p.a.
When you get in the market at that level then, short of full scale expropriation, a recent times Lebanon or Sri Lanka scale financial and economic meltdown or an outright end of the civilised world scenario (aka Threads in 1984); you more or less guaranteed a good outcome over a 30 year or longer view.
Anything short of such calamities is already fully in the price at those levels.
Given that, and your deep research in this area of Gilts (and comparative) returns, would a perfect foresight investor at the absolute market nadir in 1974, 1975 and 1976 have fared better over the next 30 to 40 years in long Gilts or UK large cap equities? I’m curious to know.
Also, it would be interesting to have your take on whether there was then in 1974-76 an ex ante efficient frontier between Gilts and UK blue chip shares on a really long term 30 or 40 forward view?
Obviously this is all very much hindsight investing here, but it would still be very interesting to know the answers.
That should be “30 or 40 year forward view”. Sorry.
@DH (41)
Investing at the start of 1975 (during which year, the FTSE All-share had a nominal return of a shade under 150% as it recovered from nominal returns of -28% and -50% in 1973 and 1974), the annualised real returns over the next 30 years were 10.2% for FTSE All-share and 6.2% for the ‘over 15 year’ gilt index.
For someone who invested through the crash from the start of 1972 onwards for 33 years (i.e., to the same end point), had real annualised returns of 5.5% and 3.5% for all-share and long gilts, respectively.
Anyone who timed that particular crash would have done rather well!
What a time to be alive! Which I was. Just more interested in the contents of my milk bottle than my portfolio.
I wonder how many people thought 1975 was the time to go “all-in” on Britain? 🙂
@Alan S – Thank you for that insight. Is the composition of the All Stocks gilts index strikingly different now vs the 1970s? I just happened to notice at the weekend that the duration of an All Stocks Gilts index fund is around 7 now vs 12 several years ago. I’m quite happy with a duration of 7. My guess is that recent shift is a function of yield rather than index composition.
Currently (as of this morning), roughly 32.8% of nominal gilts (i.e., by issue weight) are under 5 years maturity, 25% between 5 and 10 years, and 11.3% between 10 and 15 years for a total of 69.1% under 15 years.
In 1974, the equivalent weightings were 43.7%, 18.5%, and 9.3% for a total of 71.6% under 15 years.
In other words, there was not a lot of difference between the amounts in issue under and over 15 years, but it was more heavily weighted at under 5 years in 1974.
A significant difference in duration comes with yield. To take an example, the longest duration gilt at the end of 1974 had a 3% coupon, a time to maturity of 21.77 years, a price of 26.25 and a yield of 13.71 and the modified duration was 9.1.
Imagine such a gilt had existed 5 or 6 years ago, when yields were as low as 1% or now with yields of about 5%, the modified duration would have been 16.9 and 14.6, respectively.
The gilt currently with the largest duration (26.8 according to giltsyield.com, although I make it 27.1 using the clean price) is TG61 (0.5% coupon, matures in October 2061, i.e., roughly 35.4 years time, and a yield ~5.2%). At 1% yield, the duration would be 32.0, while with a yield of 13.7%, the duration would be 12.3.
So, the duration of an individual gilt depends on both the yield and the coupon with lower values of either leading to higher durations. Of course, as yields change, the coupons of newly issued gilts change with them so when the yield goes high and stays high (as it did in the 1970s), for a given maturity and yield the durations gradually get smaller, while when the yield goes low and stays low (as it did post-GFC), then durations get longer.
I think your guess is correct, the duration of the All-Stocks gilt index depends on both the yields and coupons of the individual gilts and the maturity profile of the gilts in issue. While the yields can change quickly (as we have seen!), the coupons and maturity profile change relatively slowly as old gilts age and new ones are issued except, perhaps, in emergencies (e.g., covid, war, etc.) where a lot of debt is issued in a short time.
Thank you for taking the time, that’s super informative.