How not to lose money

My first table shows the performance of a very simple portfolio – one with 50 per cent invested in MSCI’s world index, 20 per cent in gilts and 10 per cent each in gold, sterling cash and US dollar cash; I chose these proportions as reasonable round numbers rather than because they would have optimised our portfolio. Such a portfolio would have returned only slightly less than global equities since 1990, but with much lower volatility. It has had a much higher Sharpe ratio than the MSCI, and a lower worst-case annual loss.

We can improve on this, though. My second column shows what happens if we add the Halloween rule. This keeps the same portfolio between November and April, but ditches equities on 30 April in favour of a portfolio split equally between gold, gilts, UK and US dollar cash. This would have given us higher returns than equities at much less risk.

My third column adds a different rule to our basic portfolio – the 10-month average rule. This keeps the basic portfolio in equities after the MSCI index has been above its 10-month average, but ditches equities altogether when the index drops below that average in favour of equal weights between gold, gilts, UK and US cash. This has a slightly lower Sharpe ratio than the one using the Halloween rule, but it has the advantage of a small maximum drawdown. Its worst annual loss (in the 12 months to January 1995) was a mere 5.9 per cent.

All three of these portfolios did well when equities did badly. The worst 12 months for global equities in sterling terms since 1990 came in the 12 months to May 2003 when MSCI’s index lost 31.7 per cent. But our balanced portfolio with the 10-month rule lost only 0.8 per cent then because it got us out of the bear market well in advance.

During the financial crisis this portfolio did even better. It gained 21.8 per cent in the 12 months to February 2009 because it was out of equities but into gold and US cash, both of which did well as sterling slumped.

What’s more, these portfolios have done as well or better than many fund managers recently. My second table shows how the basic balanced portfolio has performed in each of the past five years compared with the 172 funds in Trustnet’s database of mixed investment funds with 20-60 per cent in equities. Only in 2017 did it come outside the top third of such funds. And in two years it was in the top 10.

I wouldn’t set much store by this, because the last five years have seen an atypically strong bull market in global equities. But it does suggest something important – that perhaps you don’t need to pay hefty fees to fund managers to achieve reasonably stable returns. You can do it yourself.

Or can you?

The problem is that the past 30 years might not be a guide to the future.

For one thing they’ve seen a fantastic bull market in bonds, and therefore in gold too, which (to put it mildly) might not continue. And it’s unlikely that lower returns on gold and gilts will be offset by higher ones on equities. It’s reasonable to suppose, therefore, that returns on balanced portfolios won’t be as good in future as they have been in the past.

Worse still, there might also be less room for diversification. Since the 1990s, we’ve been able to offset losses on equities by holding bonds. But this need not remain the case. If real interest rates rise – say because inflation finally rises – then we could see gilts and equities fall together. This was quite common before the 1990s.

There’s not much we can do to protect ourselves from the danger that previously uncorrelated assets will fall at the same time, except to hold cash which means foregoing returns.

Running a balanced portfolio in recent years has been easy. But it might not continue to be so – although whether this means fund managers will actually earn their money remains to be seen.