For years there has been a large and partly tedious debate about how useful economic theory is. Sometimes, though, we should question theory even when it works. This is true for the dollar/sterling rate.
A year ago, I wrote that “sterling is cheap and will bounce back”. It did. Sterling has risen more than 10 per cent against the US dollar since then. This wasn’t because I was clever. It’s simply because I applied an old piece of economic theory, albeit one with more recent evidence – the idea that exchange rates tend to overshoot and so bounce back after big falls.
But there’s a problem here. From one perspective, this should not have happened.
To see it, consider the real exchange rate. This is simply the plain nominal rate, multiplied by the ratio of UK to US prices: in my chart, I’m using consumer price indices*. The idea here is that the UK can lose competitiveness in one of two ways: either by the nominal exchange rate rising, or by prices rising faster here than elsewhere, causing us to price ourselves out of some overseas markets.
My forecast that sterling would rise was based upon the idea that real exchange rates should be more or less constant over time. This means they should tend to mean-revert so that a low real rate leads to a rising one, with the rise occurring via a rise in the nominal rate.
But here’s the thing. There’s a good reason why this should not have happened. Most economists believe that Brexit will depress the UK’s long-run growth rate, other things being equal. To mitigate this, we need a lower real exchange rate – so that greater price competitiveness offsets the loss of frictionless trade with the EU. To put this another way, weaker trend growth should mean a lower real interest rates in future, which should depress the real exchange rate now.
Brexit, though, isn’t the only issue here. The real dollar/sterling rate was low months before the EU referendum. In itself, this might not mean much: exchange rates go all over the place. If we consider it alongside the underperformance of UK equities (even in local currency terms), however, it has a troubling message. It suggests markets were pessimistic about the UK’s growth potential (relative to the US) even before the Brexit vote. This is consistent with the UK’s poor productivity performance and low investment in new technologies; the OECD recently pointed out that the UK has fewer robots per worker than other developed nations, for example.
This leaves investors with a dilemma.
On the one hand, simple overshooting points to sterling continuing to rise; even now, the real dollar/sterling exchange rate is 13 per cent below the post-1990 average.
On the other hand, though, there’s a good reason to suspect the rate should stay below average: weaker trend growth should mean a lower real exchange rate.
We might get a solution to this dilemma. If we get good news about medium-term economic growth, sterling’s real exchange rate should rise even without the help of mean-reversion.
And we have had such good news recently: labour productivity rose sharply in the second half of last year. If this can be sustained confidence should increase about medium-term growth.
Whether it can be sustained is, however, doubtful. On the one hand, it’s possible that tight labour markets are forcing companies to find better ways of using their existing workforces. On the other, though, it could be that Brexit-related uncertainty has caused some companies to temporarily limit hiring, in which case the rise in productivity might be only temporary: its rise last year was due mostly to a drop in hours worked.
For this reason, I’m no longer confident that sterling will rise.
Bland as it seems, this is only to be expected. Even the strongest sceptic of the efficient markets theory must believe that assets must at least sometimes be fairly priced and difficult to predict.
*In recent years the nominal and real dollar/sterling rates have been highly correlated. This is because relative consumer prices in the two countries haven’t changed much – a fact that is significant in itself.
