An All-Weather portfolio

From this imperative came Dalio’s All Weather strategy, now one of the core products of Bridgewater Associates, the company Dalio founded in 1975 and which, from its base in the New England state of Connecticut, runs about $235bn (£205bn).

Taking the idea to inception involved much trial and error, and Bridgewater is coy about the strategy’s returns, as it is about disclosing the performance details of any of its funds. Back in 2009, when the All Weather plan had been active for 13 years, Bridgewater said it had generated annualised returns of 8.4 per cent with volatility – movement around its average – of about 11 per cent. From the perspective of the UK, such returns are, indeed, equity style, with the bonus that the comparative absence of volatility means that, statistically speaking, annualised losses would be worse than 2.6 per cent just one-sixth of the time. By contrast, a portfolio wholly exposed to London-listed shares, and based on returns over the same period, could expect to labour with annualised losses of over 8 per cent one month in six.

Since then, the simplicity of the All Weather approach means it has attracted many imitators. For example, the Lazy Portfolio ETF website, which constructs and back-tests many investing plans based on exchange-traded funds, runs a Ray Dalio All Weather Portfolio. It says that back testing from 1973 has produced returns of 5 per cent a year above inflation for its All Weather variant, with accompanying volatility of not quite 8 per cent.

Now Investors’ Chronicle has produced a UK-friendly version of the All Weather strategy using exchange traded funds (ETFs) that are listed on the London market and are eligible for both self-invested personal pensions (Sipps) and individual savings accounts (Isas). Back testing, mostly using the underlying indexes tracked by the ETFs in the Investors’ Chronicle UK All Weather Portfolio, produces equally impressive results. The IC All Weather Portfolio would have generated annualised returns averaging 6.5 per cent from December 1997 to June 2022. That is only 0.7 percentage points less than the total return for UK shares over the same period, as measured by the FTSE All-Share index, yet with much less than half the volatility. Simultaneously, the average real return – ie, in excess of inflation as measured by the retail price index – was 3.6 per cent a year (see Table 1 and Chart 1).        

That such smooth yet satisfactory returns can be generated from a no-effort portfolio containing just five or six holdings is a compelling recommendation; as is the point that the fund can be assembled at any time – whether bull market or bear market prevails – and held for as long as needed. In other words, timing barely matters.

The prime insight that drives the IC All Weather Portfolio’s make-up is that returns on almost any asset are driven by whether both economic growth and inflation are rising or falling relative to investors’ average expectations. While these may be the driving factors, different assets react to them in different ways. For equities, fast economic growth is good while slowing growth is bad. Yet for conventional fixed-interest bonds slowing growth is fine since that implies a drop in interest rates, which will drive up the value of their fixed dividends.

So mixing equities and fixed-interest bonds in a portfolio stabilises returns. But only to an extent. Come the current stagflationary conditions, which feature slowing growth and rising inflation, then both equities and conventional bonds suffer. What’s now needed are other assets to play the offsetting role. Commodities are useful. To an extent, they tolerate rising inflation but prefer rising growth. Yet gold, as the ultimate flight-to-safety asset, is quite happy when growth tanks and inflation surges.

Index-linked bonds, whose dividends and return on capital are both tied to a measure of consumer price inflation, also play a vital balancing role when inflation rises. Since their future cash flows are almost guaranteed to rise in line with higher inflation, their value is protected even when the value of most other assets is falling. That said, and as we will discuss shortly, even that logic comes under pressure if inflation rises too fast, too far as it is doing currently.

The second vital insight behind the All Weather approach is that the focus of the average investment portfolio is on allocating capital according to the expected returns of various asset classes rather than according to their risk (or volatility). As a result, portfolios tend to be too heavily weighted to equities.

True, that has brought the reward of higher returns over the long haul and – barring catastrophe on a vast sale – should continue to do so in the future. However, the outperformance comes at the cost of high volatility. Research by Bridgewater showed that a typical diversified portfolio allocated 60 per cent of its capital to equities. As a result, equities – a particularly volatile asset class – accounted for almost 90 per cent of the portfolio’s risk. Put another way, the overexposure to equities means that other assets with less volatility, especially bonds, can do little to reduce overall portfolio volatility. Either way, the effect of over-reliance on equities means portfolio value can be badly damaged just at the point where capital needs to be turned into cash.

So what’s needed is the correct spread of assets – equities, bonds and so on – but, crucially, they should be mixed in the right proportions. Hence Table 2, which shows the asset mix that comprises Bridgewater’s All Weather approach. Essentially, it shows that the portfolio is divided into four equal parts, each of which holds the assets that do best when inflation or growth rise or fall in relation to market expectations.

Take the top right quadrant, which shows what’s needed if, as is currently the case, inflation is rising relative to expectations. Then an equal mix of index-linked bonds, commodities and – perhaps surprising – emerging market debt is required. Similarly, if growth is falling (bottom left quadrant), then an equal proportion of fixed-interest and index-linked bonds are needed.

For retail investors, the shortcoming of this portfolio make-up is getting exposure to emerging market debt in particular, and perhaps to corporate debt. Happily, the Lazy Portfolio approach is simpler and more accessible. It ditches those two requirements and, as Table 3 shows, holds just five asset classes – equities, fixed-interest bonds, index-linked bonds, gold and diversified commodities. It is this asset split on which Investors’ Chronicle's All Weather approach is based (see Table 4), although the back testing employed a further tweak.

Intuitively, the weightings of the portfolio look odd; at least, they are not what investors are used to seeing. That’s because the weightings take to heart Ray Dalio’s point that conventional portfolios put too much emphasis on equities. Peg that back and, with the right mix of bonds and commodities, acceptable returns have been there for the taking with much less volatility.

As to proof of that via back testing, the detailed results, as mentioned, are shown in Table 1 where the comparison is with the total return (ie, including dividends received) for the FTSE All-Share index, the broadest index of London-listed shares, and with the retail price index, the original measure of UK inflation which, handily for our purposes, produces higher rates of inflation than the officially-favoured consumer price index.

The data are based on year-on-year returns for the 295 months running from December 1997 to June 2022. The key figures are for the average 12-month return and the standard deviation, a measure of the variation around that average. In practical terms, statistics tell us that two-thirds of the time data will fall within plus and minus one standard deviation of the average and, of that remaining one-third of the time, for only half of it would the data fall on the downside. That is especially significant for the UK All Weather result. It means that for five-sixths of the past nigh-on 25 years it would be in profit. By contrast, the All-Share index’s total return would be running losses of up to 8 per cent. In addition, the worst 12-month loss that the All Weather portfolio would have to bear was 15 per cent; for the All-Share, the worst was a 34 per cent loss. For All Weather investors, running losses worse than 10 per cent came round at the rate of once every three years; for All-Share investors, they occurred almost once every six months. True, there was a trade-off of sorts. The best 12-month result that All Weather investors saw was 18 per cent. For All-Share investors, the best was 52 per cent. Predictably enough, however, that was 12 months on from when they were having to cope with big losses – about 30 per cent in March 2009.

That the weighting of the holdings is crucial is shown in Table 5. This contrasts the actual returns from UK All Weather, as shown in Table 1, with what the All Weather approach would have produced had the weighting of the assets been equal or if they had been set to intuitive levels that put a little less emphasis on fixed-interest bonds and a little more on all the other asset classes.

Set the weightings equally and volatility rises by more than a half with just a small rise in average 12-month returns (6.8 per cent versus 6.5 per cent). The number of falls over 10 per cent almost doubles, as does the rate at which such losses crop up. True, the volatility is not quite equity-style, but the very small extra return is not worth the extra risk. The intuitive version of UK All Weather involves no more than fine tuning and it produces nothing other than a fine-tuned difference in returns. The best return is a bit better; the worse is a bit worse. The average is a bit higher, but the volatility is too. And so on. In short, it does not seem worth the bother to search for better weightings than those offered by Lazy Portfolio’s take on the Ray Dalio original.

Dig down to the granular level and Table 6 shows how the component parts of the UK version performed. Experience bore out intuition. The riskier assets produced higher average returns at the cost of higher volatility. The safer ones did the opposite. Returns for equities, commodities and gold produced data that were in the same ball park – for each, the best, worst and standard deviation were all close, although average returns for commodities were distinctly lower.

The big test is the extent to which returns from these assets offset each other. That is, how much volatility is reduced without – hopefully – affecting average returns over the long haul. This is quantified in Table 7 which shows the correlation co-efficient in price changes between each pair of assets; the lower the value (including minus figures), the more that changes move in opposite directions. So, for example, equities and conventional gilts tended to move in opposite directions (a correlation of -0.4), as did gold and equities. For the All Weather portfolio as a whole, its highest correlation was with equities. However, their unconventionally low weighting in the portfolio meant the volatility to returns that equities might usually bring was nullified by the bigger presence of all the other assets, all of which had a comparatively low correlation co-efficient with the portfolio.

Various of these characteristics are also shown in the charts. Chart 2 illustrates the way in which annual returns from the UK All Weather portfolio sit in between the bouncy returns from global equities and the calmer returns from index-linked gilts. Charts 3 and 4 do a similar job. Rather than show the 12-month percentage changes, they index the price of the assets from the start of available data in December 1996. Against comparatively volatile assets – equities, energy and gold – UK All Weather tracks a middle course in Chart 3.

In Chart 4, plotted against calmer assets – although it is debatable whether commodities should be included there – All Weather is led only by index-linked gilts, which is also the only asset whose standard deviation (ie, volatility) is less than its average returns. Simultaneously, however, index-linked gilts have a higher volatility than conventional fixed-interest gilts. This seems odd, especially as index-linked bonds, whose real returns are pretty well guaranteed, should be the ultimate low-risk asset and therefore especially low on volatility.

An explanatory clue lies in Chart 4 where the line for index-linked gilts drops precipitously in the few months leading up to the present. In fact, the underlying index, Bloomberg UK Government index-linked bonds, fell 27 per cent in the seven months to June, a decline twice as rapid as that of global equities as measured by the MSCI World total return index. In a sense, this shouldn’t happen, so what’s going on?

The likely explanation is that markets are having to recalibrate. For years, index-linked gilts were priced on the presumption of low UK inflation where ‘low’ meant anything from, say, zero to 3 per cent. Against that backdrop, real returns (ie, those in excess of inflation) could be low too. Suddenly, however, the paradigm has shifted. In this new era real returns must be higher to compensate for the extra risk of persistently high and/or volatile inflation. Prices of index-linked stock are adjusting in what, hopefully, will be a fairly short-term lurch downwards.

Time will tell whether this is an adequate explanation. Meanwhile, there is longer-run data, stretching back to December 1987, which offers some comfort. True, it would be nice to back-test UK All Weather to, say, 1970 and thus track it against the UK’s high inflation and economic malaise of that period. Miserably, the data are not there, chiefly because index-linked gilts, which were created in response to 1970s’ inflation, were not launched until 1981. As it is, the data going back to 1987 have the index for conventional gilts spliced together from Bank of England data, and some of the earlier data for index-linked returns are guesstimates based on subsequent actual index-linked returns.

Hopefully, that hasn’t distorted the results shown in Table 8 and Chart 5 too much. Certainly they look consistent with the good quality data of 1996 to 2002 and the higher equity market returns of the 1990s. Sure, there is extra volatility and the number of months that the portfolio showed a 10 per cent-plus year-on-year loss more than doubled even though the period was only a third longer. Probably most reassuring, though, is that the worst 12-month running loss an investor would be looking at – 15.7 per cent – was barely more than in the shorter period even though the best running profit was almost three-quarters higher.

The absence of a major asset class – property – from the UK All Weather portfolio is worth a mention. We did construct All Weather variants with a property index added. However, since our property index was based on the S&P Global Reit index it was really just adding more equities exposure to the portfolio. After adjusting for that, the All Weather with property variant produced returns that were much the same as the base level UK All Weather mix. Besides, most investors should not want to add a property dimension to their All Weather portfolio for the simple reason that, via their home, they are likely to have more than enough of their wealth tied up in property anyway.

This does, however, prompt the final thought – what sort of investor would want to sink capital into an Investors’ Chronicle UK All Weather portfolio? Clearly it is easy to construct, requiring just five (or perhaps six) exchange-traded funds (see Table 4) and cheap to run. The obvious investor would be one who wants to preserve the real value of capital already built up while extracting some income. Assume that returns remain around 6.5 per cent then that’s sufficient to take 3 per cent for income and leave the remainder to preserve the capital’s real value.

In addition, for those with sufficient capital and know-how there is the potential to use the All Weather approach as a means to lever up returns, which is how Ray Dalio’s Bridgewater Associates uses it. With that 6.5 per cent average return underpinning it, the prospect of borrowing at, say, 4 per cent, while using the portfolio as collateral, means that levered returns could be limited chiefly by tolerance of risk.

Yet let’s not exclude someone wanting to build a big pot of capital from a small pot over a working lifetime’s saving. Again, assume that 6.5 per cent return and, for argument’s sake, inflation averaging 3 per cent. That leaves a 3.5 per cent real return. Take that return, compound it for 35 years and every £1,000 of spending power becomes £3,300. Not to be scoffed at. Seems like the UK All Weather approach is all right for all investors who prefer a hands-off approach.