How to construct an all-weather ETF portfolio

The long-standing rule of thumb for an investor who wants to balance the opportunity for growth with protection against the worst falls in the market is a 60/40 split between riskier assets and safer ones. This isn’t a panacea for all crises and there have been times when other combinations would have delivered a better trade-off between risk and return – but it’s worked well on average over many decades and it’s simple and intuitive, hence its continued popularity.

For much of that time, that 60/40 split for a typical individual investor would have meant 60% in domestic shares and 40% in domestic bonds. But the rapid growth of exchange traded funds (ETFs) and index funds over the past decade has made it possible for any investor to build a much more diverse portfolio. You can hold all of the major asset classes – cash, conventional bonds, inflation-linked bonds, precious metals, real estate and shares. You can diversify internationally and you can fine-tune holdings within each asset class to reflect what you think offers the best opportunities. And you can do all of this very easily and cheaply, probably with fewer than ten funds.

A baptism of fire

At MoneyWeek, we’ve set out diversified portfolios using these principles that reflect our overall views on global markets several times over the years in articles and newsletters. The ETF portfolio that we’ve tracked most closely over time is one that I last updated about three years ago. At the time, the safer assets were cash (10%), UK conventional government bonds (10%), UK inflation-linked government bonds (5%), US inflation-linked government bonds (10%) and gold (5%). The riskier assets were UK and Europe large-cap shares (15%), UK mid-cap shares (10%), US shares (5%), Japan shares (10%), emerging market shares (10%) and property shares (10%).  

The coronavirus crisis marks the first real test of this asset allocation. So it’s reassuring to see that this portfolio would have done relatively well in such an unprecedented situation – results would depend slightly on exactly when you choose to rebalance the portfolio back to its target weights (we suggest once per year), but it would probably be up around 9% over three years (including dividends), compared to a fall of around 4% in the FTSE 100. Perhaps more significantly, so far this year it would be down by around 1%, against a fall of about 15% in the FTSE 100. When the crisis peaked on 23 March, the FTSE 100 was down by around 33%, while the model portfolio would have been down by around 13%.

That’s not because the portfolio got anything brilliantly right (indeed, a couple of things were definitely wrong in hindsight), but because diversifying internationally and across a range of assets helped insulate it from any single event going wrong. A portfolio like this deliberately doesn’t make big bets on any particular outcome – instead, it tries to reflect a balanced view of what is likely to lie ahead. Most of the assets are chosen to do well with adequate growth and modest inflation. However, there are assets that should do particularly well in times of higher inflation (gold, inflation-linked bonds, real estate). Others would be especially helpful in severe deflation (cash, conventional government bonds). If a more extreme environment becomes much more likely – eg, growing risks of high inflation or severe deflation – you would adjust the amount allocated to certain assets to reflect the changing threats.  

Give me inflation – but not yet

Many things have changed in the world since we first set out this portfolio, but up until this spring surprisingly little had altered the economic outlook or the prospects for markets. We still believed that inflation was the likely outcome of the policies that governments were following, but there was no sign of it gaining traction.

The coronavirus pandemic has now upended the world to a far greater extent and made the future much more uncertain. It seems unlikely that we will emerge from this and return to the same economic environment that we had before – we are likely to end up moving one way or another towards deflationary depression or high inflation (possibly paired with stagnant growth – stagflation).

It seems deflation must reign in the short term – too much demand has been destroyed by the global shutdown and the resulting surge in unemployment. That argues against loading up on inflation protection right now. However, in the medium term governments will try everything to restart their economies through both monetary and fiscal policy, with government budget deficits underwritten by central banks buying bonds. We have to assume that they will keep doing this until they get results – they fear a depression much more than inflation.

So there is finally an obvious path by which inflation will take off – and so I suspect a portfolio like this might look significantly different in five years’ time. But it’s not a done deal and it’s certainly not happening tomorrow – and thus the changes I’d make at this stage are still limited.

Among this riskier assets, the current balance still reflects our views well. We think emerging markets are probably the cheapest stockmarkets overall (a suitable ETF for investing in them would be iShares Core MSCI Emerging Markets (LSE: EMIM) – there are plenty of choices, but to be simple I’m going to list cheap ETFs from iShares or Vanguard here, although other firms such as State Street or HSBC offer equally good alternatives). We also think Japan offers good value (Vanguard FTSE Japan (LSE: VJPN)), while Europe is not too dear (Vanguard FTSE Developed Europe (LSE: VEUR)). The allocation to mid-sized UK companies (Vanguard FTSE 250 (LSE: VMID)) provides more exposure to the domestic British economy (around 50% of FTSE 250 revenues come from the UK, compared with about a quarter of FTSE 100 revenues), although it hasn’t made much difference lately: FTSE 250 returns are only slightly ahead of the FTSE 100 over three years and five years (note that the FTSE Developed Europe index is about 25% UK, so we are not ignoring large UK companies altogether, simply tilting the portfolio towards mid-sized ones). 

There isn’t much in the US (Vanguard S&P 500 (LSE: VUSA)) because we’ve long considered the US expensive. That hasn’t worked very well – the US has continued to beat most other markets, largely due to the strong performance of large-cap tech stocks. But the S&P 500 still looks pricey by global standards, making it hard to justify raising this. A further mistake was to hold UK property stocks through the iShares UK Property ETF rather than a global fund, in part because the only global ETF (iShares Developed Markets Property Yield (LSE: IWDP)) has more than 50% of its holdings in the US, which seemed to offer worse value at the time. That hasn’t made much difference over three years, but is worse over five, largely due to the impact of the Brexit vote in June 2016. Given that all global property stocks have sold off hard, it seems a good time to switch into the global version instead. This indirectly increases US exposure by around five percentage points by buying one of the hardest-hit parts of the market – which appeals to me more on a value basis than upping exposure to the S&P 500.

Hold on to gold

Among the safer assets, there is one obvious change to suggest, largely due to a very specific technical risk. The inflation rate used for calculating payments on UK inflation-linked bonds is the old retail price index (RPI) rather than the newer consumer price index (CPI) – but the bonds are generally priced to reflect expectations for CPI. The government is consulting on whether the RPI formula should be changed to bring it into line with the CPI series. Inflation on CPI is usually lower, so that would reduce the expected long-term value of the payments on these bonds and so prices would be expected to fall. There is no certainty that the government will change the rules this time and any change would not be immediate (a date of 2025 is proposed as the earliest possible), but it’s increasingly likely it will happen eventually. So we’d drop the 5% holding in UK inflation-linked bonds.

US inflation-linked bonds (iShares $ TIPS (LSE: ITPS)) don’t have this problem. Unlike UK inflation-linked bonds, they also have a deflation floor, which means that the principal will not be reduced below its initial value even if inflation over the life of the bond is negative. It is unlikely that this would come into play even in a deflationary scenario except for very recently issued bonds (most have too much accumulated inflation), but it’s a tiny extra bit of insurance against the most extreme scenarios. So we’d retain a 10% holding here, as well as 10% in UK conventional government bonds (Vanguard UK Gilt (LSE: VGOV)). Yields on the latter are almost non-existent, but they should do well in deflation and help to stabilise the portfolio during a major sell-off. If inflation surges, it might be time to ditch them altogether, but for now they still justify a place in the portfolio.

The remaining 5% taken out of UK inflation-linked bonds can be shifted to gold (iShares Physical Gold (LSE: SGLN)). MoneyWeek’s long-standing view is to have 5%-10% of your portfolio in the metal; this portfolio has held the lower limit until now, but with uncertainty much higher than before and the risks much greater, it seems prudent to increase insurance against extreme events. 

• For an update to the portfolio, see here