When I began planning my financial future, I became obsessed with nailing a realistic rate of return. All of the investment calculators required one.
Plus, everything else flowed from that number – such as how much I needed to save, and how long it would be before I could declare financial independence.
It seemed important. Because if I highballed the number then I was telling myself a fairy story, wasn’t I?
Eventually I read enough fusty old PDFs and insomnia-curing books to convince myself I had an answer.
The average inflation-adjusted rate of return for a portfolio of global equites was about 5%. More than 100 years of returns data said so.
You could dig up a similar number for bonds, too, and all the rest.
Do the maths, and hey presto! One time-tested, personalised rate of return.
Data mining
Then you get down to the hard work. Years of hacking away at the FI coalface. Celebrating when you hit a seam of double-digit returns. Face blackened when you’re scorched by a fireball of negative numbers.
But it’s the damnedest thing. That oh-so-achievable looking positive average return hardly ever turns up. Because investment returns are rarely average:
Data from JST Macrohistory 1, The Big Bang 2, Before the Cult of Equity 3, A Century of UK Economic Trends 4, St. Petersburg Stock Exchange Project 5, World Financial Markets 6, and MSCI. February 2026.
No matter how many annual return charts I see, I never get used to how nuts the variance is. Yet this carnival of volatility is a far better portrayal of the actual investment experience.
In the chart above, the blue line is the average annualised return for World equities 1900 to 2025. It currently stands at 5.6%. (All returns in this post are inflation-adjusted, GBP total returns).
However you can count on your fingers the number of annual returns that remotely resembled that figure. Across 126 years!
Which is fine and dandy when returns come in over the blue line: “Yay, I’m above average – maybe I’ll get to retire early?”
But it’s super-bleak whenever the bad years roll in. Then, everyone wonders if they’ve been sold a pup.
Optimism biased
Luckily a string of defeats doesn’t happen very often, as you can see from the chart. We haven’t experienced more than a single negative year in a row since the Dotcom Bust of 2000 to 2002.
Since then though, interest in DIY investing has exploded. I can only imagine the fear and loathing that’ll reverberate through the community if (when…) we suffer a sequence more like the 2000s, the 1970s, or the 1930s.
There’s no cure for human nature I suppose. But the Pollyanna problem has been on my mind lately, given nerve-janglingly extreme US market valuations.
Gold fingered
The wide variation of returns we see with equities holds too for every other asset class you can plausibly take refuge in. Such as gold…
Data from The London Bullion Market Association. February 2026.
Gold won the past decade. It’s also having a great year (so far).Tempted? Beware that gold annual returns are certifiably insane.
The last 20 years have been amazing. But the 20 years between 1980 and the year 2000? Not so much.
Necessary historical footnote: The GBP gold price before 1975 was mostly either fixed or distorted by the impact of government regulation. Find out more in our deep dive into gold.
Show me the money
Data from JST Macrohistory 7, British Government Securities Database 8, and Millennium of Macroeconomic Data for the UK, 9. February 2026.
Cash operates in a narrower range, sure. Yet inflation and abrupt interest rate swings can send returns haywire.
I still wonder why everyone piled into money market funds when interest rates spiked in 2022. Had they forgotten the enormous cash bear market that raged from 2009?
Money markets lost over 27% from 2009 to 2023. Every year bar one was a loser. But it just didn’t feel like it because we don’t keep it real. (By which I mean inflation-adjusted!)
His skid mark materials
AQR 10, Summerhaven 11, and BCOM TR. February 2026.
Commodities are even scarier than equities. Some 42% of years are negative versus just 30% for World equities. You need a cast iron stomach to withstand that level of volatility.
But also look at the number of years commodities returned over 20% – and even 40% – in comparison to equities.
The penny finally drops when you discover that bonza commodities years often occur when equities are in the toilet.
Commodities’ average return looks pretty good, too: 4.3% annualised. Then again, this asset class is the epitome of ‘anything can happen and it probably will’.
Gilt complex
Data from JST Macrohistory 12, and FTSE Russell. February 2026.
Lastly, if not leastly, there’s government bonds – whose approval rating sank to Trumpian levels when gilts dished out their second-worst annual return on record in 2022.
All Stocks gilts (as featured in most UK government bond funds and ETFs) aren’t really much easier on the nerves than equities. Even worse, their average return is a miserable 0.76%.
The secret though is not to view bonds on their own. Bonds don’t make any sense in isolation. The magic happens when you throw them into a pot with other assets.
Kinda like how most people don’t eat raw chillies, but there’s widespread agreement that they add something to curries.
Enter the Pot-folio
Don’t even think about stealing my amazing new Pot-folioTM idea. I’ve trademarked the bejesus out of it. (What’s that? “Just stick to the charts, mate…?”)
The improvement wrought by sufficient diversification isn’t totally obvious in chart form. The down rods are definitely fewer and stumpier, though.
However looking at the raw numbers highlights the difference more clearly:
| World equities | The Pot-folio | |
| Annualised return | 5.6% | 5% |
| Deepest drawdown | -51.8% | -36.5% |
| Longest drawdown | 13 years | 10 years |
| % years -10% or worse | 15% | 9% |
| Volatility | 16.2% | 11.6% |
| Ulcer Index | 18.4 | 9.8 |
| Ulcer Performance Index | 0.28 | 0.47 |
In exchange for giving up a little return, you get fewer and less severe down years. That means:
- Shallower drawdowns
- Shorter drawdowns
- Less volatility
- Better risk-adjusted performance
The Ulcer Index is a measure of downside pain that translates drawdown depth and length into a single metric. A lower number is better.
Portfolio Charts introduced me to the Ulcer Index as devised by Peter Martin.
The Ulcer Performance Index is a risk-adjusted performance ratio that divides the excess annualised return by the Ulcer Index number. Here higher is better.
You say portfolio, I say Pot-folio, you say “Go do one”
I haven’t spent time optimising the Pot-folio. It’s just an equity-tilted variant of an All-Weather portfolio.
Essentially, you maintain positions in assets that when combined can cope with most people’s shopping list of worries:
- Growth – equities
- Inflation – commodities, index-linked gilts
- Recession / panic – government bonds, gold, cash
- Stability / liquidity – cash
However, as much as everyone buys into the concept of diversification, it’s fair to say investors spend more time thinking about how to satisfy their immediate desires. Such as making bank as quickly as possible, if not quicker. Right up to the point that the risk chickens come home to roost – and crap all over the place.
So if you’re nervous about AI bubbles or whatnot, be bolder with your diversification. By which I mean, consider investing in asset classes that look painful when viewed in a vacuum, but that can be blended together to smooth out your ride.
This way you can aspire to be a bit more average most years – and if that means the difference between you staying invested for the long run and bailing out at some market bottom, it’ll make all the difference.
Take it steady,
The Accumulator
- 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.[↩]
- Kuvshinov D, Zimmermann K. 2021. “The
Big Bang: Stock Market Capitalization in the Long Run.” Journal of Financial Economics,
Forthcoming.[↩] - Campbell G, Grossman R, Turner JD. 2021. “Before the cult of equity: the British stock market, 1829–1929.” European Review of Economic History. 25. 10.1093/ereh/heab003.[↩]
- Chadha J, Rincon-Aznar A, Srinivasan S, Thomas R. “A Century of UK Economic Trends.” ESCoE, NIESR and Bank of England.[↩]
- Radchenko P. “St. Petersburg Stock Exchange Project.” Yale School of Management, International Center for Finance.[↩]
- Moore L. “World Financial Markets, 1900–25.” Working paper.[↩]
- 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.[↩]
- Thomas R, Dimsdale N. 2017. “A Millennium of Macroeconomic Data for the UK.” Bank of England.[↩]
- Levine, Ooi, Richardson, Sasseville. 2018. “Commodities for the Long Run.” FAJ.[↩]
- Bhardwaj, Janardanan G, Rajkumar, Geert Rouwenhorst K. 2020. “The First Commodity Futures Index of 1933.” Journal of Commodity Markets. 2020.[↩]
- 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.[↩]
@TA,
Nice post and nice approach.
A few questions to clarify please:
a) what deflator did you use
b) are returns total returns derived from the indices with dividends assumed to be fully re-invested
c) what are the source(s) of exchange rates
d) are any costs/frictions, such as fund costs and or platform fees (inc. any applicable to FX), accounted for
In short – are the returns calculated ever achievable by the man in the street with a so-called perfect portfolio (which AFAICT nobody holds either)?
If the answer is no, could you indicate the likely gap practical to theoretical?
Thanks very much in advance for any additional info.
@Al Cam – isn’t the point of the post to illustrate the potential benefits of diversification rather than as a how-soon-can-I-hit-my-number post?
So that Pot-folio(R) maybe won’t give you 5% in the long run, but it will give you similar return to 100% equities and similar reduced drawdowns, volatility, Ulcer Index, etc, etc, etc.
A nice bit of analysis which reminds us of the variability of real returns.
IMV for planning purposes a range of ‘likely’ returns (which will depend on assets and timescales) may be more useful than a single ‘average’
For example, using the macrohistory database, over rolling 40 year periods (a period that might be useful at the start of accumulation or retirement)
UK equities had real annualised returns ranging from 1.7% (1st percentile), 5.0% (median) to 9.1% (99th percentile)
For UK gilts* these values were -3.6%, 0.7%, and 6.0%, respectively
And for 60/40 combo, 0.3%, 3.2%, and 8.0%
For planning purposes, the 1st percentile might be used as a ‘worst case’ while the median might be used as a more benevolent example (or anywhere in between to taste).
For shorter periods, the extreme values become more extreme (e.g., for 10 years, the range of returns for equities were -7.4%, 5.7%, and 16.4% at 1st, 50th, and 99th percentiles).
* On a technical note the gilts in macrohistory.com (R6) were consols before about 1960, 20 year gilts before 1990, and 15 year gilts after that with a transition to 10 year gilts from 2016.
@Al Cam (#1)
I can’t answer for all the databases, but macrohistory.com quotes total returns for equities, bonds, and cash.
Subtract 20-30bp from the returns for fees (historical fees would, of course, been much, much higher).
@Brod,
Agree with your conclusion.
However, and fwiw IMO costs/frictions are not inconsiderate and I think they should, at least, be hinted at. And whilst @Alan S provides IMO a good pointer, I think his numbers might exclude FX fees, which, depending on your Pot/approach, may well be non-trivial. Being a generally conservative person, I would probably opt for a somewhat higher value for total costs, etc.
The figures given AIUI are gross returns and that opens up a whole load of other issues including wrappers, etc. That topic, however, is covered in some details elsewhere @Monevator.
Lastly, @Alan S’s suggestion of a range of returns certainly has merit, albeit somewhat more complicated to grasp. At its simplest interpretation I wonder if the median returns would be a better indicator than the (assumed geometrical) average? Just a thought, really.
@Al Cam (#4)
FWIW, personally for quick calculations I assume 0% real for two reasons
1) The calculations become relatively easy and, for those of us who like mental arithmetic, largely doable ‘in head’ (or at least on paper).
2) For longer planning periods (upwards of about 20 years), 0% is lower than the worst case, but includes a bit of fudge for sequence of returns (since simple constant return calculators do not account for this*).
However, I note that for shorter periods, 0% lies somewhere between the worst and median cases so while probably useful enough doesn’t represent the worst case.
* When using a spreadsheet calculator and median returns (i.e., the 5% or so in the article), modelling a 50% drop in portfolio value at the end of an accumulation period or at the start of a retirement period (with a recovery time to suit) forms a useful test of the effect of a bad sequence of returns.
Finally, while more complicated, running calculations over different historical rolling periods illustrates how big a part luck plays in the outcome. For example, a historical accumulator investing one inflation adjusted pound per year in 100% UK equities over a 40 year period with no costs, would have ended up with, in inflation adjusted terms, between £40 (worst case), £128 (median case) and £371 (best case). In other words the lucky accumulator would have ended up with about 9 times the amount in their pension pot compared to the unlucky accumulator. Also note that the worst case essentially had a 0% real return.
Have not dug deep on the topic yet and they get mentioned here at times but if somebody decides the safest way to get a spread of assets in retirement is to buy an annuity what are the costs? In the sense i know there are financial advisers or platform fees to the likes of HL & Fidelity etc. and probably commission but i can’t see the likes of Royal London etc swapping a pension plan for an annuity and then losing a management & fund fee so is it like a HL or AJ Bell monthly fee within the annuity?
Love this article and this discussion of asset allocation very helpful. I’d like to know what the long term safe withdrawal rate (over say 30 years) for the pot folio would be? This is the most important metric to look at for me heading into decumulation phase, and doesn’t get talked about enough in UK forums. I want to compare portfolios and pick the one that’s most robust and allows me to sustainably draw out as much cash as I can. That’s what I am focussed on and I’m sure I’m not alone.
The more experience I have, the less confident I am about any investment strategy for the long term.
Bonds and money market funds are the conventional way to reduce portfolio volatility but so too would be to keep half your money in a box under the bed.
The idea of negative correlation doesn’t seem to work. I have believed that a principle behind a bond/equity portfolio is that, as one outperforms the other, rebalancing has the effect of overcoming our inclination to hang onto the winner for too long and ensure we, effectively, sell high and buy cheap. If both follow the same trend up or down I question the merit of holding bonds and money market funds.
Some have talked of total return funds as an alternative but I am unconvinced. The cost of derivative-based insurance seems too much of a drag on performance. Like any insurance product, it is cheap when risk is low and expensive when risk is high.
The so-called evidence-based investment approach is compelling but, other than diversifying with smaller cap tracker funds, I am less convinced about other specialised factors such as momentum etc. where I feel you are simply betting on both sides of a tug of war.
Commodity funds have performed well but my understanding is that, whilst long-term, they have been sightly negatively correlated, since 2005 they have been positively correlated. I feel to have missed the boat on the more recent rise but am reluctant to buy into these funds now at such elevated prices especially in areas such as gold. True, they may still be in the foothills for other commodities, but the differences suggest either targeting more specific commodity sectors with fairer value or relying on active managers to do so on my behalf.
I have been led astray by the tax-benefits of VCT and EIS investments but, fortunately, only with a small proportion of my funds. Their risk is too high for mere mortals like me.
I do have some individual shares but am not convinced of my ability to know when to sell them. Returns have been ok but nothing special compared to global tracker funds.
What has worked for me, despite the best efforts of successive governments, has been a split between global equity tracker funds and a portfolio of rental properties. With the latter, either value is static and rents increase or rents are static and value increases. Property is, I feel, a terrible diversifier – apart from all the other options!
That leaves the non-property part of my portfolio. I agree with the more aggressive diversification need as a way of diluting the heady level of some larger stocks currently, but any move away from a global tracker involves a more active choice on my part of how to allocate between markets, sectors and cap levels which, like buying individual stocks, returns to the impossible decisions of when to buy, when to sell and when to rebalance.
With a portfolio of over £10m now, I also have to consider IHT. I am starting to realise that my strategy of living forever might not work!
@Al Cam – Cheers. Annual UK inflation is from Bank of England’s A Millenium of Macroeconomic Data. ONS annual CPI from 2017 (when MOMED ends).
Yes, total returns i.e. with dividends reinvested.
Exchange rates are from A Millenium of Macroeconomic Data, A Century Of UK Economic Trends, and directly from Bank of England after those datasets conclude.
Investment costs and taxes are not accounted for as is standard practice. I could in the future include a simulated OCF aligned to leading ETFs – I’d be interested in your views on this. It wouldn’t be conventional but would be interesting to see.
Maybe a starting point is just to write an article showing what difference it makes?
As for whether it’s achievable. You mean a return of 5% real on a diversified portfolio? Sure, in principle we’re just talking about investing in suitable low-cost ETFs at a low cost broker. However, if some significant chunk of your asset allocation is violently sub-average then probably not.
But to me it’s the behaviour of the diversified portfolio that’s interesting like @Brod says. How significantly it can reduce drawdown pain for a relatively acceptable reduction in performance.
The ‘Potfolio’ is not a recommendation. And I spent zero time optimising it. I’m just trying to show that a sufficiently diversified portfolio – featuring assets with positive long-term returns and relatively low correlations – has a good chance of reducing pain without hurting gain (much).
But our actual future returns are largely out of our hands.
@ Alan S – Yes, I think you’re right. I’ve just upped my chart game of late so I’ll put your thoughts on my ideas list and will rustle something up to try and illustrate 🙂
@KLJ – I previously bought annuities for my mum – it was a while ago, so the details are a bit hazy. I think I started with a comparison site then eventually spoke to a few different sales agents directly. I can’t remember if they worked for the annuity companies or not. They offered me a price which included their commission. I negotiated over a couple of rounds and then eventually bought from two different companies. There aren’t ongoing costs. I assume that’s something they account for in their initial offer.
This site was useful to me as a starting point: http://www.sharingpensions.co.uk/annuity_rates.htm
@Iain Coward – Cheers! I need to go back to SWRs. I previously worked on calculating SWRs for different portfolio allocations – exactly as you suggest. I could do it but my method was labour intensive, and I wanted a better version of a World index to use as a baseline. I’ve got that better World index now and I’ve been improving my spreadsheet behind-the-scenes to make it easier to output results for whatever portfolios readers are interested in. Nearly there now 🙂
I’d also like to add that I’ve been greatly helped and encouraged by @Alan S’s endeavours in producing a database of returns for various gilt indices back to 1900 (which he made available in the public domain). So now it’s possible to separate out the effect of say short bonds from long bonds from the perspective of a UK investor.
If you’d like to cut to the chase, I think this approach to settling on a SWR still works:
https://monevator.com/what-is-a-sustainable-withdrawal-rate-for-a-world-portfolio/
https://monevator.com/how-to-improve-your-sustainable-withdrawal-rate/
Thanks for the excellent piece.
It is indeed a “carnival of volatility”. SoR risk rules.
Would you consider a Maven/Mogul update to your July 2024 piece here?:
https://monevator.com/all-weather-portfolio-optimisation/
Especially on “the sweet spots” and anything you’ve discovered in your researching AWP optimisation in the past 19 months.
I think a lot of us equity heads get freaked out by the % gilts needed in the AWP to get the Sharpe up (and Ulcer index down).
One way round that on the run in from accumulation to deaccumulation (in fifties) is to switch into an optimised AWP variant from a 100% global equities position but perhaps keeping when doing so both a global SCV tilt (from AVSV ETF or likewise) and a multi factor element (from FSWD ETF or similar) and using capital efficient products (like WGEC ETF) for some of the bond exposure (i e. global HQ gov bonds hedged to Stg in place or some of the gilts in the AWP) and with a trend following overlay on the DM cap weight equity element using Winton Enhanced Global Equity.
I have a (slightly?) mad ‘double’ DCA idea, which has been knocking around my head recently, and which won’t go away, which is too risky/ arguably bonkers for most folk to spell out the actual strategy here (but happy to DM you or @TI) for putting 5-10% into a MMF and then using that as a funding source to double DCA drip feed (at a 0.5x rate when within 20% off of 52 week high and at a 1.5x rate when below 20% off of 52 week high) into a maximally risk/ return strategy.
That preserves 90-95% in an orthodox (lower risk) approach (optimised AWP), but then harvests volatility on the 5-10% max asymmetric strategy with the baseline DCA (1x amount) being 0.5%-1% of the MMF starting amount per month (given an objective to drip feed the DCA amounts in over 10 to 15 to 20 years).
@The Accumulator – Thanks for the feedback based on your dealings and also the link – must admit apart from the front end charge i did wonder if there is/was a hidden fee paid ongoing much like the old (in)famous initial units on pensions going back about 30 years ago that may not be obvious at first look.
@TA: I enjoyed “His skid mark materials” a bit too much – chapeau!
This topic was covered extensively on US financial websites long before U.K. financial thoughts on this item appeared
It was all I had many (30+) years ago when I started out
6% for equities and 3% for bonds were common prognostications on many sensible websites-Vanguard Diehards (now Bogleheads) etc and were all that was needed for an amateur investor like me to set sensible and achievable saving parameters ie portfolio size required for retirement,sustainable withdrawal rate and Asset Allocation construction etc
Any U.K. thoughts from U.K. financial sources (often insurance companies) were plainly wishful thinking on growth rates and even more so on withdrawal rates-ridiculously optimistic and actually down right dangerous for an amateur U.K. investor setting out on his stockmarket adventure
I was saved by the Americans -once again!
xxd09
Yet more thought provoking analysis from Monevator, thanks @TA (and thanks @Alan S for sharing your data).
When I get asked about equity returns I like to say if the real return from equities was a steady amount pa the price would be bid up and the prospective returns would fall. No risk no reward.
It’s theoretical but of interest to think what would is the risk free real asset? Short dated inflation-linked gilts? These have only been around since the 80s so unfortunately can’t be compared over the time horizons in @TAs charts.
Longer dated linkers lock into a real yield for longer time horizons, but you need to hold them to maturity to get that yield, so doesn’t a like for like comparison when looking at one-year returns.
@TA, @Alan S:
SWR (or sequence of returns risk (SORR)) is largely not a feature of accumulation. IIRC, years ago Pfau had a go at estimating SORR as another reduction to expected returns for people in de-accumulation. I do get the sense of Alans 50% drawdown approach too. Not everybody is/will be a slave to SORR, just like it is not true to generalise and say that “The No. 1 risk in retirement, hands down, is longevity risk”; see e.g. https://howmuchcaniaffordtospendinretirement.blogspot.com/2025/08/measuring-retirement-risks.html
Little people rarely (if ever) in my experience get those BoE exchange rates, and, depending on how you go about things, could be quite significantly penalised.
I like your OCF idea – but think it would vary depending on the portfolio choices and definitely not be a one size fits all number. Seems that @Alan S is even more conservative than me. FWIW, I have tended to use 6% nominals [as achievable] for equities with other assets lower to suit.
@Delta Hedge:
Elegance and simplicity both have a lot going for them. Alan S points out very nicely just “how big a part luck plays in the outcome”.
@xxd09:
Agree that back in the day US FIRE websites were way ahead of anything available in the UK. Not sure I would more generally rely on the Americans right now though!
Very interesting, we all want to put a number out there and optimise the portfolio.
After many decades I conclude it’s unknowable.
If you’re at the beginning of the accumulation stage then it does not matter , just keep throwing in what you can at equities, if markets are bad then it’s Buffets do you want to pay more or less for Hamburgers if you anticipate a lifetime need for them. It’s probably more worrying if the good times seem to roll too long and too high..
At the decumulation end of the timescale, then enough cash/bonds /state pension / pension income for 10 years or so and you can probably throw the rest into equities.
When you’re approaching retirement then start building the cash/bond pot to reach that 10 year reserve.
Many of want to optimise and plan but just doing something that will be satisfactory probably requires only 5% of the effort and may well be 100% as satisfactory in providing for a good retirement… we don’t know when the end date is and thus satisfactory is good enough.
Hariseldon – Unknowable is right but………
You do have to have an investment plan in place with reasonable parameters as an aim
Another fact I forgot to mention that was brought home to me so long ago by those American financial blogs was the very large amount of savings required for a “satisfactory “ retirement income
(£100000 in a 60/40 portfolio used to bring in a safe £4000 pa before tax-nearer £3000 pa now)
Most of my peers at the time were focused on property-their houses and then BTLs for pension income etc
I never had enough money left over for more than a cottage but alternatively had accumulated enough money (in stocks and bonds) to retire at 57 and travel the world while wife and I were both fit
Brits seemed to do property (too illiquid and too much hassle for me)
Americans did investing in the stockmarket as their priority-everything else then followed
I followed the American example and still do
xxd09
@Stephen Francis – “The more experience I have, the less confident I am about any investment strategy for the long term.”
I feel the same way. I remember reading about how US equities have never experienced a loss over a holding period of more than 20 years. That was very comforting when I started out. Now I know it’s perfectly possible for that to happen.
As human beings we crave certainty and security but the truth is those two things don’t exist in absolute terms. We can’t predict the future and we can’t hedge out every last vulnerability.
My solution is to spread my bets among assets that exhibit a low degree of correlation in a crisis and not worry about outperforming an alternate version of me that got every decision right. All I need do is ensure there’s enough for me and mine to enjoy a reasonable quality of life.
Negative correlation wouldn’t work if you owned two assets that cancelled each out. Happily that’s not the deal on offer. Government bonds are sometimes negatively correlated with equities during recessions. But they don’t always work – especially when the recession is inflationary in nature.
Commodities typically do work during highly inflationary periods but are terrible during demand-side recessions – when govies tend to work.
Check out the performance of commodities in 2022 when inflation kicked off and global equities fell: https://monevator.com/asset-allocation-quilt/
They were negatively correlated right on cue.
Commodities spend a lot of their time being godawful though. Then they shoot up over relatively short periods to record a positive overall return. On the one hand, they’re an excellent diversifier for equities and bonds. On the other, they’re extremely hard to live with.
Gold and cash fill in the diversification gaps.
Maybe the gold price is elevated but nobody knows. It’s impossible to accurately value. It’s helped me to understand that mean reversion doesn’t really exist. It’s an artefact we can tease out of the data, it may even be a widely held belief that affects prices, but it’s not a law of nature.
The gold price declined for 19 years before turning around. Now it’s been rising for the past 25 years minus a lost decade. Yet if the rising gold price is fuelled by dedollarisation, asset financialisation, instability in the global financial system and increasing geopolitical tension – well, those trends that could run for decades yet. Or they might not. Or the gold price could be affected by something else.
I do wonder what will happen to faith in global equities next time they go nowhere for ten or twenty years. I hope it never happens but I own the diversifiers (unsatisfactory though they are) because I know that one day equities will let me down.
@Delta Hedge – Thank you as always. I’m beavering away behind the scenes on improving my ability to delve into the historical record for the main asset classes using GBP returns.
I should be able to rustle up a good trend investing backtest soon – I think you’ve encouraged me to do this before IIRC, along with Algernond and others. Apologies if I’m misremembering.
I love your ideas as always. Feel free to post them on the site somewhere. I like the way you use old threads to log your findings. I read many of them but time often prevents me from replying. Re: Accidentally posting the formula for investing anthrax 😉 TI would just delete anything he thought liable to destroy the universe but he’s a reasonably tolerant fella 🙂
@Jiffy – Haha. Cheers! I’m glad someone liked it. There’s nothing like a terrible pun to bring out the schoolboy in me.
@xxdo9 – I totally agree, the work of an amazing generation of publicly spirited Americans completely changed the game.
At the same time, I think a few things didn’t translate as well over here.
Most famously the 4% rule.
But also the 60/40 portfolio’s over reliance on bonds as a diversifier. 40% in bonds makes less sense in the UK where historically higher inflation and longer duration government bonds have proved less protective than the US equivalent. (At the same time, I’m reassured by how well it’s gone for you.)
All the same, I don’t think Americans should be comfortable relying exclusively on domestic stocks and bonds and I’m certain citizens of not-so-superpowers should not.
@Prospector – I suppose the very long-term risk-free asset has to be the 0.4% inflation-adjusted result for the money market. One way to think of linkers is as a bet on the market under-estimating future inflation rates, while nominal bonds represent the opposite.
I agree though – from the perspective of my portfolio, index-linked gilts is the closest thing I can get to a risk-free real return asset – so long as negative yields aren’t a thing. I don’t really think of ‘em as risk-free but it’s the best I can do.
@Al Cam – Re: exchange rates: We’re not talking about the exchange rates you and I can get. For investing purposes, we’re talking about the exchange rates fund managers obtain and whether that’s a source of profit for them.
I’ve just quickly compared MSCI World Index GBP returns with the GBP return of SWLD – SPDR’s MSCI World ETF. These figures are after costs:
2025 – SWLD outperforms the index by 7bp.
2024 – Index outperforms by 8bp.
2023 – ETF wins by 39bp.
2022 – Index wins by 6bp.
I’ve stopped there because those are the figures I can get in a jiffy. Still, as a quick test, it indicates that the market is so competitive these days, you don’t have to worry too much about leeching return to costs if you purchase a decent tracker. Bogle won.
Regardless, historical backtests aren’t predicated on a promised level of return. At best, they can tell us something about the relationships between different assets and how they behaved during past economic circumstances. Our future returns are subject to a high degree of variance as per @Alan S. And there’s a futility to over-optimisation as per @Hariseldon.
All the same, I think you’re wise to counsel the use of conservative benchmarks when considering future returns.
@Hariseldon – Amen.
@TA:
Re exchange rates – I did say “depending on how you go about things” and if you want to diversify currencies then, in my experience, “the exchange rates fund managers obtain” are not applicable.
I’m not really sure what your additional figures are telling me. Is this principally a FOREX issue or a tracking error issue or some combination of them? AIUI, trackers can only ever beat the index if they are using full replication with a lot of securities lending or alternative (sampling, synthetic, swap-based, etc) approaches. That is, taking some form of additional risk(s). However, I am far from expert in this area. FWIW, I think there are a whole host of issues “under the hood” with trackers.
Fully agree with everything about: learn more to seemingly only know less, etc
Thanks for this.
I’m increasingly seeing 7% real/10% nominal quoted as the long term return form equities, which always seems over optimistic to me. Any idea where this number is from? Is it US inflation and USD investor? In global equities or US equities?
I’ve decided that whatever numbers you put in your projections, anything you attempt to predict more than 3-5 years hence is fantasy land.
@Vanguardfan #20: Answer: US large caps only, and in USD (S&P 500 1957 onwards, S&P 90 1928-57): ~10% pa nominal, ~7% pa real. Total return. Going back to 1790s for US possible (but much less reliable), and brings it down to a bit under 6% pa real total return. Global markets, capitalisation weighted, equities total real returns since 1900 are as per @TA’s piece above, although I’m not sure if the 5.6% CAGR is pre or post adjustment for the disappearance of the Russian stock exchange from 1917-92 and China’s from 1949-90 (and, indeed, other such instances).
For the future, who knows?
If we were to assume a little under 2% pa productivity gain per capita and around 1% pa population growth both since 1800 (consistent with the historical record); then a future no productivity growth and a minus 1% pa shrinking population (roughly consistent with 1 TFR) globally (say by the 2050s) would knock 4% pa off of that 5% to 6% pa real return (an equilibrium state scenario, albeit very pessimistic I hope).
Now that it’s probably too late to matter I think I’ve got an investment policy. Start with our three biggest assets in increasing order. (i) State Retirement Pensions: these are rather like index-linked gilts. (ii) DB pensions. I’ve decided that the part covered by the Pension Protection Fund can be likened to ILGs too, but the capitalised value of the other bit is best viewed as equity-like.
(iii) House – it has two parts. (a) The house we would accept living in – fewer rooms, smaller garden, address less tied to work – though we won’t trade down to that because of hassle and stamp duty. That throws off the imputed rent i.e. the rent we don’t pay because we own it. Again ILG-like. (b) The rest of the house – say 25%. That’s more equity-like.
Given that we already have two lumps of equity-like investments, and heaps of ILG-like, what should we invest our rather modest “portfolio” in? Probably in stuff that the DB pensions don’t own much or any of – gold, commodities, cash. Cash is particularly good because it can absorb all my limited investment obsession while I lever it (stoozing) and search for best interest rates. That means I’m more likely to be a sensible chap and check the rest of my investments no more often than annually.
What is the portfolio allocation for the potfolio? What percentage in which assets?
@Al Cam (#15)
Re: SORR in accumulation
This does exist – imagine a case where over 40 years the real returns were 0% in every year except one where the return +50%. For an accumulator investing a real pound each year, if the outsized return occurred in the year after their first investment, they would, at the end of 40 years have £40.5. On the other hand, if the outsized return occurred after their 39th investment, they would have £59.5 after 40 years. The order in which returns occurs does matter.
The graph in this link* (https://postimg.cc/7bBBLCym) shows the final portfolio value (FPV) after 40 years (expressed in terms of real GBP per real GPB invested per year) for a historical UK accumulator holding 100% UK equities as function of real annualised returns over the 40 year period. Each dot represents a single historical 40-year accumulation period. The blue line is the FPV value assuming the real return was constant over the period.
The scatter about the blue line indicates the effect of sequence of returns on the final portfolio value.
* The link works for me – hopefully it will work for others too.
@DH – I finally found a credible source of returns for the major markets pre-WW1 – Russia, Austria and South Africa – that were missing from the original version of the index: https://monevator.com/world-index/
Monevator’s World index real annualised return is now pretty close to the last figures I can find for the DMS global index.
1900-2020
DMS / Monevator
5.3% / 5.49%
1900-2024
DMS / Monevator
5.2% / 5.53%
DMS is now a global index (i.e. emerging markets included), whereas Monevator’s open-source version remains a developed world index so I’m happy with the differential. I use MSCI World returns since 1970.
FWIW, I did find some good stuff on the pre-war Shanghai stock exchange but the gist was it was insignificant in global terms.
@Potfolio: You’re sailing very close to IP infringement with that name of yours. I hope you can prove you were christened Potfolio by your parents or my lawyers are gonna have a field day 😉
In other news, the Potfolio is 60% equities, 10% All Stocks gilts, 10% money market and 20% commodities. I sneaked the deets into the strapline of the chart but even I missed it when I was rereading yesterday so maybe that wasn’t my brightest idea 🙂
@Alan S – I can see your chart. Very useful, cheers!
@Alan S (#24):
You are quite right – my bad.
Interesting graph.
Is the annual addition/subtraction of new money key, ie no new money no difference. That is, the order in which the returns occur makes no difference to a one off lump sums end value.
Link does work – with “interesting” ads too?
Also, in accumulation the effect is seemingly reversed (vs deaccumulation) ie ideally you want your best returns towards the end of the accumulation period. Hence, why folks like Pfau etc talk about the “danger zone” as being plus and minus N years around your date of retirement.
@Al Cam (#26/27)
Yes, the addition of new money is key since the returns each instalment ‘sees’ are different to the return over, in the example, 40 years (the second instalment sees a return over 39 years, the third 38 years, etc.).
Sorry about the ads – none show up for me using safari or firefox (previously, I’d only ever used postimage for direct embedded links at bogleheads).
Yes, the requirements are reversed (after all, accumulators want to buy cheap investments) and the danger zone is definitely around the period close to retirement (which is why traditionally derisking into bond/cash occurs at that time, or more recently, potentially securing income flooring using a bond ladder).
Alan S (#28):
Thanks. No worries about the adds. FWIW, I use Edge.
Re: “which is why traditionally derisking into bond/cash occurs at that time, or more recently, potentially securing income flooring using a bond ladder”
AIUI, a lot of lifecycle economists would actually posit this is because you are coming to the end of your human capital, and the transition this implies and your accumulation period holdings should reflect your type of work (and not just your age), see e,g, Are you a Stock or a Bond by Milevsky*. The aim being to balance your overall asset (inc. human capital) allocation overall throughout the lifecycle to your risk/age profile. This is why most of these guys say they can see no point in pre-programmed glidepaths which increase bonds much beyond your date of retirement, such as say 100-age in stocks, etc. Any subsequent pensions should also play a role too.
*in a nutshell, folks in safe steady jobs (aka low risk or according to M being a bond) – e.g. fully tenured university lecturer** should go full in to stocks during accumulation, whereas folks in risky jobs should probably take a more conservative investment approach during accumulation.
**not sure if such roles still actually exist in the UK, but I trust you get the point
Definitely a stock and not a bond at heart here @Al Cam #29, albeit that a generous DB to look forward to in the next 5 to 10 years (depending on retirement dates) objectively makes me more bond like.
Superb point @Alan S #24. Yes. There is no definitive cure for SoRR, DCA or otherwise, only mitigations.
@TA #25: excellent work.
Removing disappearing markets makes less difference than I’d feared (20 to 30 bps off of CAGR, albeit averaged out globally and spread over a century and a quarter).
I’ll email you over the mad idea I had. It’s too risky (volatility and maybe permanent loss of capital) for a passive forum.
Maybe @Finumus will be interested.
I suspect that it’s going to be too systematically based for @TI’s idiosyncratic tastes and company first, build up from the fundamentals, approach.
In comment #22 above I wrote a little homily to myself explaining how I can diversify.
Separately I’ve been thinking of buying an annuity (100% for widow) shortly before the IHT rule change in April ’27.
It now occurs to me that I/we have no fixed interest investments so maybe that annuity should be a level annuity, contrary to my instinct to buy an inflation-linked one. Harry Browne would approve, I suppose?
@dearieme (31)
I thought your motivation for considering an annuity was that your wife has little interest in investment. Therefore, if you die before her, she will not appreciate the niceties of your careful asset allocation, but would appreciate a continuing no-effort income, preferably increasing with inflation.
Incidentally, I don’t understand the rationale for waiting until just before the IHT rule change. Don’t assets pass between spouses free of IHT?
@dearieme (#22, #31)
Re: Annuity. One thought is whether a single SP and any remaining DB pensions would support your OH’s income requirements or not. If the answer is ‘yes they would’ you may not need an annuity at all while if the answer is ‘no they wouldn’t’ then an RPI annuity would be ‘guaranteed’ to support the required real income long term whereas a nominal annuity would not.
@Al Cam (#29) UK tenure was removed under Thatcher (late 80s?). ‘Permanent’ contracts in the university system are permanent in the sense they can be dissolved by either party with appropriate notice and cause (usually redundancies or disciplinary).
@Alan S (#33):
Thanks. Thought it had gone. Are you familiar with M’s stocks/bonds idea?
Re your response to @dearieme: whilst costs for one are clearly more than half the costs for two, do you have a suggestion as to what they might be. There are several models e.g. [modified] OECD equivalence scales which suggests two thirds. I ask as I did some work a few years back that looked at a five year period* and noted the following things for our situation:
a) the value is not constant year-to-year
b) two thirds might be too little; and
c) the value may depend to some extent on which partner goes first.
Having said that, I do seem to have a tendency to be rather conservative with such budgeting/estimating things**.
*probably, on reflection, too short and possibly an unrepresentative sample***
**although, in part, this might be explained by a period (of around six years) of possibly anomalously low spending immediately following retirement and, of course, circumstances (including scope of annual spending) evolve too
***and coincident with the six years of possibly anomalously low spending noted at **
@ DavidV: “I thought your motivation for considering an annuity was that your wife has little interest in investment.” That’s certainly one of my motives but not the only one.
“Incidentally, I don’t understand the rationale for waiting until just before the IHT rule change. Don’t assets pass between spouses free of IHT?”
Because if I die before then the SIPP money can pass down the generations free of IHT. If I don’t, the money will pass IHT-free to my widow but might well be caught later for IHT when she dies.
@ Al Cam: “b) two thirds might be too little”. I assume she’d need 80% if she’s to feel secure and comfortable. I used my digital computer i.e. spat on my finger and held it up to the wind. In my experience, people with a feel for numbers can often get quite good estimates this way. I cite my 2020 predictions for inflation rates over the next few years which were much better than the Bank of England’s.
I accept that doing better than the BoE may be a rather undemanding test.
@dearieme (#35):
Ah good old “mucous digitus in aero” as nobody in particular ever said!
OOI my findings ranged from 68% to 81% with +/-2% for who first. There is some research that suggest the less you spend the higher the %need.
And, FWIW, 75% is sometimes written about as a pretty comfortable value.
Agree about BoE.
@Accumulator, regarding over reliance on bonds in a 60/40 portfolio and even worse in the uk.Could you expand please?( VGLS40/60 here). Thanks
Nice article.
I do think it would be great if soon you could include some the ‘Alt’ strategies in these portfolio performance illustrations. There is plenty of evidence out there that including strategies that can go long & short depending on various price signals (e.g. Trend Following, Equity Mkt Neutral, etc…) have meaningful impact on risk adjusted returns – hedge funds include such strategies for a reason right?
I know it’s quite a challenge for such illustrations due to data quality & availability, but there are now various Alt strategies available to us retail investors (e.g. Trend Following from Winton, Dunn, AQR) that we can incorporate into our portfolios.
I think I mentioned before the significant improvements that both Chat GPT & Grok have shown me (backtesting to ~ 25 yrs) for a Permanent Portfolio that replaces Bonds with Managed Futures (Trend Following).
@David – 60/40 portfolio of equities / bonds is susceptible to bouts of high inflation because equities and nominal bonds can both suffer during extreme episodes e.g. 1970s and WWI.
https://monevator.com/best-inflation-hedge-uk/
That’s fine if you’re happy to wait for the real average return of both assets – especially equities – to reassert itself. Younger investors especially can afford to do that. But it can leave you with an unguarded flank against inflation if you’re near or in retirement.
Better diversified portfolios like the All-Weather more fully account for the inflation problem: https://monevator.com/asset-allocation-for-all-weathers/
The historical record of UK nominal government bonds is worse than the US.
Three main reasons:
– Worse inflation
– Greater vulnerability to extreme geopolitical and financial events
– The historical indices for UK bonds typically exhibit longer durations than commonly used US equivalent e.g. All Stocks gilt index vs 10yr US Treasuries.
The first two challenges don’t seem likely to go away given the size of our economy and position in the world.
The third issue is a technicality in one sense – longer duration bond index looks worse than shorter duration one during periods of rising interest rates, better when rates fall.
But it has a real outcome in that UK government bond trackers tend to follow the All Stocks gilt index – historically quite a long duration index. (Though it looks OK at the mo.)
Hence UK DIY investors brought up on US financial education were painted a rosy picture of bonds that didn’t apply quite so well outside of the States. Then we tended to default into a riskier product too.
Very few sources pointed at unfortunate events such as this gilt bear market: https://monevator.com/bond-market-crash/
If they did then bonds would be a pretty hard sell. FWIW, I do think nominal government bonds have a role. They also outperform cash in the long term.
But I think people could do with being better diversified (or at least better informed) unless they have a fair chunk of inflation-protected income e.g. DB pensions + State Pension.
These pieces may help if you’d like to explore further:
https://monevator.com/diversified-portfolio/
https://monevator.com/60-40-portfolio/
https://monevator.com/defensive-asset-allocation/
@Algernond – I’m wiring up a spreadsheet at the moment to help me backtest trend strategies due to your previous prompting 🙂
That said, my intention is to stick to long only.
Monevator has two very able writers capable of discussing long/short strategies in Finumus and TI.
My remit is to explore the realm of the possible for people who want to keep things relatively simple (or at least as simple as possible but not simpler.)
For me, that doesn’t really include long-short strategies that are quite likely to blow up in your face – sooner or later.
By which I mean, even if some of the firms are brilliant (and I’m sure some are), every strategy has its shortcomings:
– It doesn’t work in the way it’s supposed to – often due to some unique combination of circumstances that didn’t show up in the backtest.
– The product was been designed around backtests that look good but were over-optimised to create a good marketing story. The product then disappoints in the future because its simulated returns were an artefact in the data that doesn’t repeat.
– The product is widely misunderstood.
– The product underperforms in certain conditions – those conditions prevail, people lose faith.
I could imagine being interested in a product that is a long/short version of a risk factor strategy. Because that’s how they’re meant to operate in purest form and I’d understand the risks. I appreciate that a long/short trend investing fund could operate along these lines. Are AQR offering anything like that?
Re: backtests. The clean data problem is paramount. If you can point me towards some high-quality independent sources then great. We have discussed that issue before but I don’t seem to have any good links recorded anywhere. My apologies if you’ve posted me some good stuff before and I’ve missed it.
Re: Chat and Grok. Are you sure they’re using quality sources? I don’t use AI to backtest but I have used it like a friendly search engine to try and track down useful sources I might have missed. Mostly it’s pointed me to garbage.
@TA (#39)
One of the motivations I had to calculate the historical returns for gilt funds of various maturities was that US research indicated that using intermediate maturities (5 years in the SBBI year book) had a better SWR than either cash or longer bonds (IIRC, 20 years in the SBBI year book). A similar result (i.e., intermediate bonds were better) also came out of the bogleheads simba spreadsheet. Before that, all freely available sources of UK fixed income return data before about 1990 included returns for only cash or long bonds (typically consols or 15-20 year maturities).
@TA (#40)
Some while ago I backtested simple (moving average) trend following using the monthly Shiller US dataset. There is no doubt that it backtests quite well over long periods. However, the problem comes down to the details since it can perform poorly compared to buy and hold over periods of a decade or more particularly when the buy signal following a sell trigger comes too late and the buy price is higher than the sell price. There can be quite a few of these before profit is potentially made on a bigger event. In other words implementation would require rigorous discipline. IIRC, Andrew Craig’s fund (VT PEF Global Multi-asset Fund A GBP?) uses this approach across a very wide range of assets.
@TA (#39):
A fine answer – amongst your best IMO.
Some more details/thoughts that I hope may be useful about [primarily UK vs US] inflation:
a) According to google AI the average inflation for the last century in the US is 3% to 3.3% and in the UK it is 3% to 4%;
b) For a decadal view, compare and contrast [for UK inflation] the second chart in this post: https://simplelivingsomerset.wordpress.com/2025/04/03/inflation-in-the-uk-a-deep-dive-part-2/ with the second chart/table in this post: https://www.theretirementmanifesto.com/inflation-the-silent-killer-of-retirement/ for US inflation;
c) some of this delta (UK > US) is down to differences in methods and approach – but not all of it; the UK generally has higher inflation than US;
d) the first post linked above mentions another interesting factor, namely that: “Up until a couple of years ago, some people (including at least one very well-known FIRE luminary) truly believed that the inflation monster had been fully tamed and that the historical record was interesting but of no real further relevance. ” IMO, this view was wrong then, is wrong now, and will always be wrong. Risks may be dormant for long periods, but rarely, if ever, are they destroyed*. IIRC, an appropriate technical term might be recency bias;
e) like the more familiar SWR, the sequence of inflation matters, this is particularly obvious in the case of a DB pension with capped inflation indexing;
f) inflation (as measured and reported) is a macro-economics concept whose utility (pun intended) is well understood; no more, and no less – and its applicability at a household level is limited**; by and large, individual household consumption choices/idiosyncrasies are generally far more influential on your financial well-being than household inflation, howsoever measured. Having said that, IMO in the UK, CPIH probably provides the best available household guidance as all attempts to measure the seemingly attractive personal inflation rate are, at best, a folly;
As usual, just my take and somewhat off the top of my head too!
*I still like Fritz’s closing thought: “The risk of inflation is not too big to keep you from retiring. It is, however, too big to ignore.”
**hence, I often question the real worth (another bad pun!) of an individual using ever more elaborate schemes to track official inflation; although, as Fritz said, ignoring inflation is not an option! Finding what works for you and yours is situational (and often not just purely numerical) and, no doubt, tricky; as somebody commented recently: “We aren’t all playing the same game”.
From Alan S
“Andrew Craig’s fund (VT PEF Global Multi-asset Fund A GBP?) uses this approach across a very wide range of assets.”
Thanks for this – I will look in to it. I am always looking for multi asset, absolute return funds, whether long/ short, long only, trend – following etc.
Particularly those that can zig when equities zag or can perform in any type of market environment.
Maybe a comprehensive article on these types of ” absolute return” funds would be something Monevator would consider?
@Alan S – When I first started, I read a Larry Swedroe book that said intermediate bonds were the sweet spot and long bonds were too risky. So I found the UK equivalent which was All Stocks. I could see they weren’t the same thing as US intermediates but there wasn’t really much choice at the time.
More recently I think I read in one of your papers that the gilt sweet spot was further out the yield curve than the US. So it seems like the exact position is something of an artefact in the data? Though your work still supports the thesis that long bonds aren’t worth the risk, and your 0-10 gilt index appears to be the sweet spot?
I also got a shock when I read the first edition of Tim Hale’s Smarter Investing. It had an usually extensive bond section that featured a chart of long-term gilt returns including the 40-year gilt bear market. The bond chapter was much smaller in the 2nd edition and that chart wasn’t reprinted.
@Al Cam – I agree about your comments on personal inflation. Even I don’t bother to measure it.
I’m glad I learned about the idea though, it is something I bear in mind. Useful as a vague yardstick rather than a precise metric, p’raps. Same with ZX’s warning about the earnings growth. Really useful to know about, and I factor it in to my thinking thanks to ZX, but it doesn’t keep me awake at night.
Yes, madness to think that inflation was dead. Same with deflation too, I think. Not as big a concern but I don’t rule it out.
@TA:
Thanks for coming back.
Most of my learning about the personal inflation rate is given in: https://simplelivingsomerset.wordpress.com/2025/07/17/how-to-determine-your-personal-rate-of-inflation/ There are also some additional details provided in the 6th part of that series of posts and in the comments to part VI (called: UK Inflation – anatomy of the Covid crisis) I highlight the impact of varying just two (albeit key) assumptions.
I am still running my related experiment, but strongly suspect that 2026 will be the last year. This is primarily because there are just a few residual itches I want to scratch. And, it does take some effort/time too!
It is a nice idea – but really it is fundamentally flawed!
@TA (#44)
“More recently I think I read in one of your papers that the gilt sweet spot was further out the yield curve than the US. So it seems like the exact position is something of an artefact in the data?”
An artefact of reality since the behaviour of the yield curve in the US and UK was different and hence the sweet spot was further out for gilts. Interestingly, prior to the 1970s, the US data I’ve used (the bogleheads simba spreadsheet) is currently derived from yield curves rather than the prices of individual securities. For the period 1941 to 1981 I generated an open source database of individual US treasuries (monthly prices) and hence returns of various treasury indices (see https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5422914) and, in places, these are different to those generated from yield curves (I also explored this difference for gilts in https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5192452 ).
What the data show (I need to see if I can find returns by maturity for other countries) is that ‘intermediate’ maturities may not be well defined (except broadly as not long and not short). A rule of thumb (that may be broken globally) is that a weighted fund maturity (or duration) over 15 years is definitely long, less than 1 year definitely short, and somewhere around the 3-8 year mark is probably intermediate enough! I note than in one of McQuarrie’s papers, he divided US treasuries into ‘short intermediate’ and ‘long intermediate’.
One problem with the ‘All stocks’ gilt index is that the weighted maturity of the index varies with time given a) prevailing yields, and b) the maturities and coupons of gilts in issue so while currently the all stocks index has weighted maturity of over 10 years (and duration of about 8 years) putting it on the high side of intermediate, it was even higher in the run up to 2020. A constrained index such as the ‘under 10 year index’ cannot have a weighted maturity of more than 10 years and is more likely to be around the 5 year mark with durations a bit below this.
@Al Cam – I think what discussions of personal inflation and average wage growth did was nudge me to allow more room for manoeuvre as I plotted my escape route. In other words, I incorporated those concepts into my understanding of “what could go wrong” and made my FI number more baggy to accommodate.
You’ve described your journey from ‘conservative planner’ to someone who’s now comfortable living within their retirement means – if I’ve understood correctly. If anything I think you’ve concluded you may have overdone your conservatism on the planning side? 🙂
Partially it’s just really hard to plan for the great unknown years in advance. But then, once you start living it, after a time I think people just acclimatise to their new lifestyle. Most retirees I know – if not all – have gone through something like this. Me too. It’s quite heartening, I think. At the same time, everyone knows they can’t close out all the risks, but ain’t that life?
@Alan S – One thing I took from your work – apart from choice of bond portfolio being an interest-rate related crapshoot #243 – was that – on balance – I should probably switch to a 0-10 year gilt tracker.
@TA (#47):
Re: para that begins: “You’ve described your journey from …”
Yup, a good summary. There are a few bits & bobs here and there* that I would like to bottom out and understand a bit better – and, I may or may not succeed – but nothing that is overly worrying me.
It is heartening and the value, I suspect, for people yet to travel that road is in knowing that in most cases it seems to work out fine; albeit not perhaps as originally envisaged, but so be it. Long may the changes keep rolling!
*which may well be idiosyncratic, or not
I have felt more and more sure over the last year that something very bad is coming. I don’t mean a drop and bounceback but a real crash where the effects last for a very long time. So my portfolio is more hunkered down than ever before. While keeping a chunk of equities to help keep up with inflation, albeit they are mostly stoplossed, I have all manner of other stuff – straight bonds, inflation-linked bonds, infrastructure funds, precious metals, soft commodities, cash and near cash. I do hope I am wrong but I do sense that winter is coming.
You could be right @Hospitaller, and my own US exposure has been whittled down from (IIRC) ~64% (market cap weight in MSCI AC) a year to eighteen months ago, first down to 50% then down to 40%, then to 35%, and now to barely 30%; with positions taken in trend following, gold, gold miners, junior gold miners, silver miners, broad commodities, EM energy stocks (EC and PBR), uranium and copper miners, general miners (BRWM) etc to try and diversify sources of return (albeit, at the inevitable cost of expanding the vectors for risk).
However, whilst that’s the call which I’ve made (so far, and I may change my mind in a heartbeat), the mere fact that we’re both worried here about ‘the big one’ is a contrarian sign of sorts that there’s potentially fuel enough for a rally in equities, especially risk-on growthy US ones.
After all, the Bull market run of 2009-21 was hated the whole way up.
When the last Bear throws in the towel, and when Michael Burry and Nouriel Roubini start saying that, actually, there is a rational case for valuations; then there’s probably going to be no new buyers left.
But, until then, who knows?
Preparation not prediction.
Probabilities not forecasts.
On any day the S&P has had (since 1928) a 54% chance of being up.
Over a year it’s been around 75%.
Over ten years it’s 95%.
And over 20 years 100%.
It’s the antithesis of a casino. The longer you stay, then the higher the odds of making money.
@TA #40
Thanks for your reply.
Both Chat GPT & Grok told me they were using the SocGen CTA Index for the Trend Following part of the data, and this why they could only do the back testing of 25 years (since that index is only published from Jan 2000).
I did give a couple of links before, but I really can’t remember under which post it was (it wasn’t the @finumus one on TF).
I listen TTU (Top Traders Unplugged) podcast regularly, where they discuss TF every weekend, since that is the focus of the podcast and most of their guests. They discuss papers, but I usually don’t go and follow-up afterwards myself (I really should).
AQR (+Winton, Dunn) Trend Follow across commodities (agriculture, energies, metals), stock indices, bonds, interest rates, currencies, and volatility using medium to long term signals. And they position size, scaled by instrument volatility. They may trim or add to their positions as instrument volatility changes. Overall, the UCITS funds are targeting somewhere between 10-15% annualised portfolio volatility. UCITS doesn’t allow them to go higher than that I think.
(the AQR Managed Futures fund is as above, but their Alternative Trends fund has some added strategies and also some ‘more difficult to access’ commodities).
I know Finumus & TI are the people for discussing long/short strategies, but you’re the portfolio man !
And I appreciate it is difficult due to only having a 25-year dataset for Trend Following. Tyler at Portfolio Charts has the same issue when he’s showing the ‘Golden Ratio Portfolio’. He replaces the Managed Futures (Trend Following) part of the portfolio with Commodities, which is not the same thing at all!
I am much more comfortable ‘following the trend’ as part of my portfolio, than putting my faith into long-only government IOUs… It really does make me sleep better at night.
“It really does make me sleep better at night.”
Well, that’s great, you probably don’t want nor need me sticking my neb in 😉
Even if I can get SocGen CTA Index data, the fundamental problem is that it only goes back 25 years. It can only ever shed light on a relatively limited set of conditions. No 1970s, no World Wars or Great Depression.
It’s great that the index covers the Noughties and 2022 but I’d like to see it pitted against a greater range of financial foes.
Now I do think there’s a way. I remember the authors of the research paper, The Best Strategies for Inflationary Times, compared a trend strategy against historical high inflation regimes. I think it was a simulated long-short strategy. Here are the references:
“We follow the methodology of Hamill, Rattray, and Van Hemert (2016) and Harvey, Rattray, and Van Hemert (2021) and construct a time-series momentum (trend) strategy applied to liquid futures and forwards (or proxies) across assets, but extend the data back further than their original work. For equities, we have data for Japan, the UK, the US, Italy, Australia, and France, from the 1926 start of our sample period. Other markets enter as they become available. For bonds we have US bonds (different tenors) available since 1926, while most European bonds are included from 1950, some years after the dust of the Second World War had settled. In commodities, we have soybeans, corn, and wheat starting between 1940 and 1950. Currencies still only start after the end of Bretton Woods in 1973. We assume annual transaction costs of 0.8%.”
“We follow the methodology of Hamill, Rattray, and VanHemert (2016) and Harvey, Rattray, and Van Hemert (2021) and construct a time-series momentum (trend) strategy applied to liquid futures and forwards. The strategy has a 10% ex-ante annualized volatility target, and the weights to historical lags in the trend definition is chosen such that it best approximates the BTOP50 trend-following index returns.”
Cited: Harvey, C.R., Rattray, S., and Van Hemert, O. “Strategic Risk Management.” Wiley Finance. 2021.
So it seems like there is a way. Whether I’m capable of something like this, even with an AI-assist, I don’t know. Perhaps one day 🙂