There’s an old joke that Saudis have a lot of money not because of oil, but because they don’t let their wives spend it. With prices making 2023 highs this week, saying no must be getting harder, even if first-half oil revenues for the kingdom were a quarter down on last year.
Extended production cuts by the kingdom and Russia are supposedly to blame for the recent price surge. The two countries supply about 40 per cent of the world’s oil and have expensive footballers and drones to buy.
Cue the usual worries about growth and inflation. Financial markets were rattled. Should they be? It seems like this column is always taking news everyone is panicked about and saying it doesn’t matter. Perhaps I spend too much time at the beach.
But once again the popular narratives are not supported by data. To be sure higher oil prices hurt economic growth. The cost of producing goods and services goes up. Wealth is transferred from many consumers to a few producers.
There is plenty of research confirming this. The International Energy Agency’s rule of thumb is a $10 increase in a barrel of oil means 0.5 per cent off global output the following year. Even a $5 move makes a difference, calculates the IMF.
The effect of oil prices on asset values is far more complicated. Take bonds. You would think it a no brainer that higher energy costs mean higher inflation means higher interest rates — which is bad for fixed income securities.
And sometimes it is, as a recent paper by the University of Pretoria looking at 161 years of US data shows. The trouble is the causal relationship is bidirectional and time-varying, to use statistician speak. In other words, bonds can go up when oil prices rise, or they can fall.
Since the first world war the correlation has mostly been negative (that is, oil up, bonds down) but not always. It depends which period you look at. There is also no reason why the relationship shouldn’t turn positive again. Very helpful.
What explains the bi-directionality? It makes sense, really. You have the no-brainer causation above. But we also know that higher oil prices have a negative economic impact, which can lead to a demand for bonds due to them being haven assets.
Meanwhile, research on equity returns is more confusing still. Again, this seems counter-intuitive. Surely, for example, higher energy costs lead to lower future margins and free cash flows. Discounting them means lower valuations.
For some sectors, yes, a causal relationship can be shown. A 2015 paper by Bing Xu on UK industries shows that oil prices have strong predictive power when it comes to energy stocks (positively) as well as for sectors where oil is a major input, such as consumer goods (negatively).
Well, duh. However, at an aggregate level — the effect of oil prices on overall stock indices — academics haven’t a clue. Or rather too many different clues. Some papers find no relationship whatsoever, others a strong one.
A study of Japan found the correlation to be negative (oil up, stocks down). In other markets such as Norway, it is positive. And it is not only the make-up of the indices that explains the mixed results. The big problem is oil prices move for myriad reasons.
If they are rising because the global economy is booming, higher input costs should be offset by stronger revenue growth (as I’ve written previously with interest rates). Indeed, in such periods when all commodities are seeing spikes in demand, even Japanese stocks love higher oil prices.
But it is different when oil alone experiences a demand shock — perhaps when buyers bring forward purchases due to concerns about future production. In these cases the relationship turns negative once more.
Here you would expect a supply shock — or even a mild surprise as we’re seeing now — to be equally bad for equity markets. If producers simply reduce output, prices rise without a background of stronger demand.
Again, the data doesn’t play ball. Research by the likes of economists Lutz Kilian and Cheolbeom Park among others suggests the supply side is less important than global demand shocks or even industry-specific movements in oil prices.
So it’s not quite the simple oil story that headlines suggest, then. Indeed, it is harder still for investors. That’s because it is not enough to identify that two variables have a causal relationship. What matters is forecasting in advance that one of them is going to change.
And no one gets it wrong more than oil analysts. By far the most regularly embarrassing calls I’ve heard during my years in finance were predictions for the price of Brent or WTI. Given the importance of oil to human livelihoods — if not always markets — this is unfortunate.
So much so that in 2011 the US Federal Reserve commissioned an international discussion paper which should have had the title: “How the f*ck can we forecast the price of oil accurately?” One answer is by tracking US inflation, according to causality tests. So too narrow measures of money supply.
But interest rates and currency moves have no forecasting power, according to Fed number crunchers. Nor do changes in real GDP predict movements in real oil prices. More important, given how many central banks rely on them, futures prices are no more prescient than spot prices. Indeed over one to three months, they are “inferior to tossing a coin”.
All of this means investors need not fret over oil at $100. Likewise $50, or the IEA’s latest projection that fossil fuel demand, including for oil, will peak before the decade is out. Everyone will tell us we should, though.
Oil is at 2023 highs but, then again, so is my portfolio.
The author is a former portfolio manager. Email: stuart.kirk@ft.com; X: @stuartkirk__
Comments are closed, but trackbacks and pingbacks are open.