The skill of investing lies not merely in knowing things. Instead, we must know what is relevant. We know the oil price is close to a 17-year low. The question is: so what? Only if this fact tells us anything about future returns does it matter for investors. So does it?
The efficient market hypothesis says it doesn’t. It says that all current information – and especially something so salient as oil prices – should be already embedded in prices and so should not predict future returns.
But the efficient market hypothesis is called a hypothesis for a reason – because we must test it. And, to a limited extent, the hypothesis is wrong because the oil price has predicted some returns in the past.
For years, there has been a negative correlation between the price of Brent crude and the change in the All-Share index in the following six months. Low oil prices in 1989, 1993 and 1999 led to good returns, for example, while high prices in 2008 and 2012 led to equities falling. Overall, since 1988 the correlation has been minus 0.17. That’s not a lot, but it is statistically significantly different from the zero predicted by the efficient market hypothesis. Low oil prices, then, have tended to be mildly good for equities, and high prices bad. Each $10-per-barrel lower oil price has been associated, on average, with returns being 0.5 percentage points higher than average in the following six months.
This is not a new finding. In 2003 Tilburg University’s Ben Jacobsen and colleagues showed that between 1973 and 2003 equity returns around the world tended to be higher in the month after oil prices had fallen and lower in the month after oil had risen.
There are two FTSE sectors where this is especially the case: oil and gas, and mining. Since 1988 each $10-per-barrel lower oil price has led on average to returns on mining stocks being 1.4 percentage points above average in the following six months.
The story behind these numbers is perhaps a simple one: investors overreact. They interpret a drop in the oil price as a symptom of weak global demand and so sell equities, especially cyclicals such as miners. In doing so, though, they are more likely to sell too much than too little, perhaps because they underestimate the ratio of noise to signal in oil prices. The upshot is a tendency – only that – for shares to fall too far when oil prices fall and so recover in the following months.
There are, however, two huge caveats here. One is that for most FTSE sectors there is indeed no significant correlation between the oil price and subsequent returns. For most shares, the efficient market hypothesis is (in this context) correct. The oil price tells us nothing about the future and so should not influence our decisions either way.
Secondly, what is true in normal times is sometimes not true in crises (and vice versa). One reason to believe this is the case now is that the low oil price is not mere noise but a signal: global demand is significantly weaker and more uncertain – a fact that should justify lower equity prices perhaps for several months.
On the other hand, however, oil’s recent fall has attracted a lot more attention than its gyrations usually do. And when people pay a lot of attention to anything, they are more likely to overreact to it. This suggests equities will rise over the next few months.
Overall, the evidence points to a case for buying equities now, or at least oil and miners. But the evidence is not terribly strong – and perhaps certainly not strong enough to overcome reasonable levels of risk aversion to these sectors. Perhaps the best response to low oil prices, therefore, is to do nothing – a course of action that is very often wise.
