Since January 2000, the correlation between annual changes in the All-Share index and in the euro/sterling rate has been 0.42. This implies that euros have given investors better protection against falling share prices than gold, whose correlation (in sterling terms) with All-Share returns during this time has only been 0.09. What's more, there are good reasons to fear that gilts will be more correlated with equities in future than they have been recently - because, say, fears about a tightening of global monetary policy might be bad for both. So it could be that euros will do a better job even than government bonds of diversifying equity risk.
This correlation doesn't just matter for investors looking to protect themselves against equity risk. It also warns bulls that it might not be a great idea to buy European equities. Because global stock markets have been highly correlated, there's also been a negative correlation between European stock markets and the euro, which implies that gains on European shares might be partly offset by losses on the euro. (I say partly because the correlation is less than one, and because the euro/sterling rate is less variable than share prices.) If this is the case, then stock market bulls might be better off sticking with UK or US equities.
This poses the question: why has sterling tended to fall, and the euro rise, when stock markets fall?
One reason is that for years the UK has run a current account deficit - we've been net borrowers from overseas - whereas the eurozone has generally been in surplus. This has tended to make sterling a risky currency - because lending to debtors is risky. As a result, sterling has tended to rise and fall as global investors' appetite for risk has risen and fallen. And this has caused it to move in the same direction as shares.
Herein, though, lies a problem. Economists have known ever since the work of Dick Meese and Ken Rogoff in the early 1980s that exchange rates are only loosely linked to macroeconomic developments. This has been corroborated by recent work at the European Central Bank, which has found that macroeconomic variables have "particularly low explanatory power" for changes in the euro. All this implies that we shouldn't rely upon the euro remaining negatively correlated with stock markets. Currencies are just too cussed and random to be relied upon as a means of diversifying portfolios.
There's an especial reason to believe this now. It's that one reason for the negative correlation between equities and the euro might no longer hold. For most of the 2000s, sterling - like equities - was a liquidity play. In times of abundant liquidity, traders placed 'carry trades', borrowing low-yielding currencies to buy sterling. This cause sterling to rise in good times when liquidity was cheap and plentiful, but to fall - most obviously in 2008 - when it dried up, forcing traders to close such positions. However, while equities still depend upon liquidity conditions, sterling doesn't so much, simply because interest rate differentials across currencies are now much smaller than they were before 2008, so the returns to carry trades are smaller with the result that there are far fewer of them.
The investor who uses euros to diversify equity risk is therefore vulnerable to Goodhart's law asserting itself: any statistical regularity breaks down once we rely upon it. It could be that this has already happened, in a benign way; in the last 12 months, the euro and equities have both risen.
There are two messages here, one general and one specific.
The general message is that correlations are not hard facts, to be relied upon unquestioningly. They can vary with macroeconomic conditions, so investors should ask before diversifying; is this correlation likely to remain favourable?
The specific message is that we can't rely upon euros being a great way of spreading equity risk - although if exchange rates are a random walk the correlation with equities should be zero, implying they might have some use. It could be simply that it is harder to spread equity risk these days than it used to be - which implies that, in a very important sense, equities are riskier than they used to be.
