Optimising the All-Weather portfolio [Mavens]
by The Accumulator
on July 23, 2024
The All-Weather portfolio is reputedly better than conventional portfolios at balancing our need to take investing risks while at the same time cushioning us from the worst of:
- Recessions
- Inflationary shocks
- Terrifying stock market crashes
We tested these claims in our All-Weather portfolio explainer post. We concluded that the strategy really has delivered over the long-term, reaching back to the 1930s.
This article can be read by selected Monevator members. Please see our membership plans and consider joining! Already a member? Sign in here.
Comments on this entry are closed.
Great article.
I assume that in this part:
“Here’s the sweet spot I found at the 60% equities mark”
Money Market should be 5% rather than 0.05%?
@TA: First an immense thank you both for Parts 1 and 2 of this superb series on the All Weather Portfolio (‘AWP’). It’s super appreciated, especially as I contemplate if, when and how to move away from having nearly 100% in global equities – as my 50s come into sight, and the equity sequence of return risk becomes ever less easily recoverable in the decreasing time left until planned retirement (at 60).
Secondly, do you have any thoughts about:
1). Replacing the money market funds allocation with one for trend following and/or global macro listed hedge funds instead?
2). Using global equities instead of UK?
3). Going with physical replication ETFs for gold and broad commodities in the 60% equity, 5% money market, 25% commodity and 10% gold variation on the AWP discussed above, but then using WTEF ETF (1.5x a 60% S&P 500 tracker and 40% US intermediate Treasuries – so effectively 90% physical replication of US large caps with 60% synthetic replication with futures for US Treasuries) for the 60% equity sleeve?
Thirdly, on the dangers of ‘torturing the data’, the Mathematical Investor blog has some excellent pieces illustrating the problem. The nearest thing to a solution is to discover in sample and then test the discovery out of sample. So if you had data, say, for US stocks from 1926 to 2023, one might discover using only the 1926 to 1994 data and then out of sample test that discovery separately with the 1995 to 2023 data.
[NB: tiny typo spotted – should be 5% allocated to the money market, and not 0.05% as shown in the chart for the ‘sweet spot’ at the 60% equities mark].
Thanks both for the heads-up on the typo. Fixed now!
@TA – great article. Thanks.
In my portfolio I’m holding 50% equities roughly half HSBC FTSE All World and half VHYL (cos I reckon it’ll be a little bit value-y and avoid those Magnificent 7 stocks. Probably a bit Doh!)
The other 50% is equal parts Gold, Commodities, Inflation-linked bond fund (GISG), Cash and BHMG. Brod’s Bit-Blustery? Do you think it’ll catch on?
I’ve got 8 years till I receive an indexed income floor so thinking I’ll cash in GISG and buy a short ladder to hold to maturity which will get me .
Btw, the cash is in Premium Bonds, the payout is about 4%. I reckon the downside is floored (admittedly at 0%) but the upside could be sensational. And all tax free! Anybody else?
A really enlightening article. Thank you.
@ Delta Hedge – By replacing money markets I take it you want to increase the growth side of the portfolio? You’re not suggesting those alternatives as an alternative to the function of cash in the mix?
Trend following does seem to be a viable strategy for boosting equity return albeit with the usual caveats that it may not work in the future / may not work in your lifetime, or for prolonged periods. How are you implementing it? Momentum is one of the factors in my multi-factor ETF but it’s pretty diluted.
As for the hedge funds and WTEF – how do you get around the problems of transparency? What are your criteria for assessing them? Will you cut them if they didn’t outpace a particular benchmark over a certain period of time? Or if they lag for a decade, will you keep them having decided they’re a good idea, well implemented, whose time has simply not yet arrived?
What I like about the all-weather portfolio is that we have 90 years worth of data showing that the recipe works.
That this combination of asset classes is about as diversified as you need to be, and can be tweaked to improve return at the expense of volatility (or volatility at the expense of return) without having to trust in any fancy investment house’s promises.
Re: UK equities – yes, absolutely I’d choose global equities. The only reason I’m using UK equity returns for the long-term view is because World index data doesn’t exist pre-1970 in the public domain. Although I think I may have just found a solution for that.
Thank you for the pointer towards the Mathematical Investor – that sounds interesting. A simple way to conduct out-of-sample checks is exactly what I need.
@ Brod – That’s a great name for a portfolio. Mine would be TA’s Turbulent Tonic 🙂
Do you ever see yourself adding a conventional bond component?
@TA – Yes, all the time 🙂
When I get brave enough to pull the plug, I’ll need to fund an 8 year gap before my floor kicks in. I’ve thought of moving a wodge from equities to 10 year nominal bond fund or something then consuming my index linked and nominal bond funds over the gap.
Or buying diversify the risk by buying one linker and one nominal bond to mature each year for the next 8 years. After which I won’t really need either (but might to maintain a nice low volatility portfolio to reliably bolster spending.)
Either option will cover me (probably). As will as-is.
Big fan of portfoliocharts.com, and results since 1970 suggest I’m counting angels on a pin head. As-is is probably more return at the cost slightly higher volatility. Not much in it.
Also, the problem of how to model BHG is portfoliocharts. Currently I regard it as catastrophe insurance and don’t include it in the AA or portfolio value. Any return is a bonus and who knows if it’s a waste.
Sounds like your rationalisation of BHMG is a bit like mine with my multi-factor allocation – it’s a bit of a punt, it may or may not work, there’s an underlying rationale for it that I can articulate, at the very least it’s additional diversification, mentally I’m prepared for disappointment.
+1 for Portfolio Charts.
Very interesting to hear about your approach. Fingers crossed that the premium bonds romp home for you. I have one worth £5 which was gifted to me when I was a baby. [Censored] decades later and I’m still waiting for my first win 🙂
@Brod. If you could model BHMG in terms of more basic asset classes, it would follow that you could replicate it with those assets. That replication would probably be cheaper, invalidating any reason to hold BHMG. portfoliocharts is going to be basically useless for BHMG since you are not buying assets, you are buying trading.
@TA – the human mind is a wonderful thing. It can rationalise anything 😀
Btw there is, or was, a great analysis of the maths of premium bonds on moneysavingexpert.com suggesting you really need to hold about £25K+ to receive any kind of win close to the average prize fund payout as a whole. So grandchildren given one premium bond 50 years ago are essentially subsidising the rest of us. But I think the minimum win now is £25 so if your £5 bond wins that, champagne all round!
Re: nominal bonds, I have been wondering if we’re at the start of a new secular period where equities underperform bonds. I can’t see interest raters rising any further, so is now the time to lock in those 10 years rates?
@ZX – yes, that makes sense and is maybe what I’m doing by excluding it altogether. Insurance which I hope not to need. But if I do, I hope it works, although I remember your points previously about maybe being too big to scale. Ho-hum.
Thanks @TA (#6) for your thoughtful response.
I’m a bit of an asset allocation obsessive, despite being in practice, by default so far, a nearly all equities person.
I’m also slightly haunted here by @ZX’s very perceptive and obviously correct observation (when I commented under the @Time Like Infinity /@TLI moniker) that I was overly equity centric, which I clearly am.
The problem is working out how to get from that position to where I’d like to be at with a diversified portfolio which delivers lower maximum drawdown, a better Ulcer Index, a better Sharpe and a better Sortino but still somehow manages to get close to equities in long term CAGR.
It’s like searching for the end of the rainbow.
There’s also bewildering choices of combinations, even with the much more limited pallet of asset class ETFs here in the UK as compared to the US.
But I have to do it, as decumulation is now only 11 years off, and if there’s another 2008-09, 2000-03, or 1972-74 inbetween, then there’s not necessarily enough years left for an all equities portfolio to recover.
I also don’t have the same relative scope which I used to to invest from income when prices fall (Buy The Dip) as, on the one hand, I’m fortunate that the combined ISA+SIPP+GIA have now hit seven figures but, on the other hand, my gross salary is still just under £80k p.a., so BTD with fresh money just doesn’t make much difference anymore.
Avoiding serious losses, however, does now make a big difference but it’s a bit of a departure from the mantra Peter Lynch reportedly started back in the 1990s that: “Far more money has been lost by investors in preparing for corrections, or anticipating corrections, than has been lost in the corrections themselves”.
Your above piece is especially excellent because you’ve got data spanning several market and economic regimes and, whilst there is an argument that more recent data is more relevant than older data; more and longer data is always better.
So many asset allocations out there only back test 20-50 years. IMO it’s just not enough. So it’s great to see the data which you use going back to 1934.
And it’s no surprise that commodities seem to make a difference.
In the very few really long term (i.e. nearly century long) multi asset backtests out there (like yours) commodities feature prominently (e.g. Bogleheads has some AWP threads with this).
FWIW, IMHO the WTEF ETF is more or less straightforward and transparent.
You get 90% SPY ETF and a 60% exposure to Treasuries via rolling futures with the rest. It’s all automated, GBP denominated, and plain vanilla return stacking at pretty tame leverage (1.5x) with two assets which should have low or negative correlations most of the time, although this very obviously wasn’t the case in 2022.
One idea I’m quite enamoured of at the moment is Taleb’s barbell asset allocation of 90% low risk TIPS held to maturity with 10% maximally risky equities – which might be the traditional daily reset 3x equity index LETFs or option like stocks, such as microcap moonshots or top 20 positive optionality S&P 500 giants (0.5% each) by high balance sheet strength, high free cash flow margins, and high R&D spend as a percentage of revenues (e.g. Meta, CSCO, GOOG, MSFT, Nvidia, QCOM, ANSS, SNPS and ABNB).
In the worst case you lose 10% of the starting portfolio value. In the best case that 10% does multiples of the TIPS return in the remaining 90%.
In either eventuality (or any scenarios inbetween) the 90% in TIPS returns US CPI plus 2% or so p.a.
@ Brod #10 – this is also a good site for looking at Premium Bonds: https://premiumbondsprizes.com/detailed#50000 especially if you look at the ‘See more details’ page.
@ Delta Hedge – that’s an interesting summary of your conundrum and I think the crossroads that many Monevator readers arrive at.
I think part of the dilemma is that we think we’ve failed, or are somehow stupid, if we don’t pick the optimum path. Even speaking as a passive investor who didn’t try to beat the market, it’s still easy to beat *myself* up because my version of the market was trounced by another version i.e. the S&P 500.
But where we really fail, I think, is in giving sufficient credence to our competing personal objectives – like the need to avoid crippling losses – and accepting that means we should no longer be making big swings for the fences.
The temptation then is to trust in clever engineering which seemingly allows us to duck the moment of truth, or to accept a compromise. Most of these investing ‘technologies’ should probably be filed under “too good to be true”. Anything that promises we can “have it all” should be treated with extra scepticism and subjected to fierce scrutiny in my view.
A better mousetrap often turns out to be a mousetrap laden with superfluous bells, whistles, and glockenspiels for twice the price.
My instinct is that the Peter Lynch quote relates to chasing the market, as opposed to derisking a portfolio because that’s your best strategic choice when you’ve won the game and are on the home straight?
@TA: It’s sort of the Rorschach’s test of investing.
Look at it one way and it’s an opaque, hand wavy, word salad of inadequately IRL tested hopeium.
Then look at it from a slightly different angle and it looks like it could be the answer to so many investor problems.
Case in point, I give you as an example the Direxion HCM Tactical Enhanced US ETF, launched in the US last year and, who knows, maybe one that will become available too here eventually, if the FCA recognise US ETFs under the Overseas Funds Regime:
https://pictureperfectportfolios.com/hcmt-etf-review-direxion-hcm-tactical-enhanced-us-etf-review/
Is it a better mousetrap or an investor death trap?
Could it be both depending on the market regime?
In a persistent sideways market it could end up getting whipsaw slaughtered like a turkey before Thanksgiving.
But in a strongly (and enduring) trending up or down market it could end up either going up like Maverick on afterburners in Top Gun (by using 2x leverage), or give a fund saving parachute by moving into cash.
I don’t know the answers. Maybe simple is best. Maybe it isn’t.
But at least it’s automated and quantitative, rather than manual and subjective.
I’m trying not to have preconceptions and to keep an open mind, but not so open my brain falls out. It’s difficult.
Nicely put, Delta Hedge. Love the word “hopeium”.
@Delta Hedge — In our ongoing discussion across a few threads about the more complex products you’re currently into, thought you might find the following new paper interesting:
We examine factors that predict the success and failure of financial product innovations using a novel comprehensive database, which contains surviving and defunct commodity futures contracts traded on 36 exchanges between 1871 and 2022. New products are more likely to fail if they do not sufficiently compensate investors for risk, or if they experience extreme returns. Contracts are also less likely to survive if they face significant competitive pressure from other products or exchanges. Finally, innovations fail because they experience systemic shocks to the financial intermediary sector such as wars.
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4730746
Not making a comment on your choices per se! Just background reading.
Obviously nobody ever expects *their* innovative product to fail… 😉
Thanks @TI. That’s super kind and interesting. 🙂
At the moment I’ve only dipped a little toe in WTEF due to illiquidity concerns with the tiny (but slowly growing) AUM and large spread.
But it’s getting harder and harder to run with the all equities crowd when the stakes are so much higher than just ten years ago, and the potential for an irrecoverable life changing drawdown before starting decumulation is much more real than before. And I sense some recent returns have been pulled from the future. I’ll have to diversify. It’s the mechanics of doing so now, not the decision to.
I’ve realised that, even where money’s concerned, it’s actually time that’s the most precious commodity – time for markets to recover and losses to be made good, time to compound, time to earn in order to save to invest. And they’re not making any more time!
The end of the accumulation game comes into sight on the distant horizon so much more quickly than you think it will when you’re young. Hard to believe that we were all the future once.
From the Janardanan, Qiao, Rouwenhorst paper: “Our second hypothesis links the probability of failure to extreme returns. The experience of a large loss may cause investors to cast doubt about the fairness of a new product innovation. Particularly in a product’s early stages of gaining market acceptance, market participants may find it difficult to distinguish between losses that are due to bad luck or those attributable to poor contract design that favor the other side of the contract. Limited willingness to absorb losses will cause participants to withdraw from the market when returns to emerging contracts are poor”.
We’ve seen this play out time and again IRL with recency bias. New product or fund launched. Early returns are good. Investors extrapolate and pile in. Good returns tend to be followed by bad. Investor shocked and quit. Returns then improve. Investors pile back in again just as the returns deteriorate. Rinse and repeat.
Also the reverse pattern: The earliest returns are poor. No one is willing to touch the product/ fund. Returns then improve, as good returns tend to follow bad. Investors wait on sidelines as they still remember the bad returns. When the investors finally take a punt on the now not so new product / fund the returns are just cycling back into bad. Investors then take fright and abandon, ecetera.
@Delta Hedge — Indeed. Related: I read an interesting bit of research this week deconstructing ARKK’s returns and claiming all the alpha was due to fund flows and Cathie Wood’s buying actually driving up the prices of the underlying. It’ll be in the links Saturday. 🙂
As we’re dealing with improving the All Weather Portfolio (“AWP”) I thought I would share a DIY attempt to do so with a leveraged AWP:
https://www.reddit.com/r/LETFs/comments/rtxuv8/a_leveraged_allweathertype_portfolio_with/
This has to be one of, if not the, most complex and difficult to implement (in practice) examples out there of marrying Tactical Asset Allocation, Leverage and the AWP. Just thinking about what would be involved in executing this on an ongoing basis would give an aspirin a headache. I’m not a ready buyer for the one hour per month claim to maintain it 😉
@TA: if you were able to use your new 1919-2023 (or even 1900-2023) data for UK, US and global equity returns for investigating improving the classic AWP, would it then perhaps be possible for you, as part of that investigation, to also look into what this AWP allocation shows as regards real terms average CAGR, the Sharpe ratio and the Maximum Drawdown (both % wise and duration wise) over that extended period?:
40% US (large capitalisation) equities
30% US intermediate Treasury bonds
20% Broad Commodities
10% Gold
I ask as the above is the very approximate economically / functionally equivalent mix of the following:
a). 60% WTEF LETF: 1.5 x (60% S&P 500 + 40% intermediate Treasuries)
b). 25% broad commodities (UC15 ETF)
c). 10% gold (SGLN ETF)
d). 5% trend following (Winton)
As trend following/managed futures data only goes back (at most) several decades or so, for the approximation of the effective allocations above, I’ve just treated it as though it were part of the allocation to gold and then rounded that down to 10%.
Many many thanks in advance 🙂