The relationship between oil and bond yields, and on to stock prices, is now well established. Oil up, bond yields up, stocks down-with the day-to-day trading relationship between 10-year yields and oil futures this month reaching its tightest on record.
This week followed the pattern pretty well. On Monday and Tuesday, bond yields fell a little as oil prices retreated; on Wednesday yields soared and oil rose; and on Thursday both rose again.
Here's the thing: What if the bond market's making a terrible mistake to pay so much attention to oil? If it is, yields might have risen too much, and all their knock on effects-on government debt, stock prices and the rest-could be overdone.
At one level, the link between higher oil and higher Treasury yields seems obvious. Higher oil prices are fueling inflation, and the Federal Reserve under new Chairman Kevin Warsh has concluded it has to respond. So higher oil means higher short-term interest rates.
What's less obvious is that higher interest rates should mean higher long-dated bond yields.
Despite that, the 10-year yield is at 5.2%, its highest since June 2007, just before the global financial crisis. "It makes no sense that what happens to oil today should affect the 10-year," says Vítor Constâncio, a former vice president of the European Central Bank.
The basic problem is that higher oil prices are a one-off. Quite apart from probably being temporary-no one believes the Persian Gulf will stay shut forever-higher prices only lift inflation once (or twice if you count knock-on costs).
Once oil stops going up, at whatever price level, its direct impact on inflation disappears. And if at that point it is so expensive it damages the economy, high oil prices would reduce non-oil inflation.
Yet, the link between oil and bonds, the correlation, has recently been extremely tight, and, perversely, tighter for long-dated than short-dated debt. The link remains tight even when near-term yields are stripped out to create a synthetic bond starting in five years.
Two-year Treasurys are strongly affected by what the Fed is expected to do at its next few meetings. So it isn't surprising that their yields should be affected by moves in oil, which have pushed up expectations for rate rises.
Futures now price a more than 50% chance of four more hikes by the end of next year. But the effects of near-term changes in interest rates should be much smaller for a 10-year, and negligible for a 30-year bond, unless you expect higher rates to last.
Some theories:
Expensive oil is forcing governments to respond, adding subsidies, tax cuts and export bans, either worsening government deficits or damaging trade and growth and tax revenue. Bond investors are already concerned about high levels of debt. If the result of higher oil is yet more debt, or less ability to service existing debt, they ought to demand a higher yield to compensate.
Final straw syndrome. High oil comes when the economy's already booming thanks to spending on artificial intelligence. The causation isn't only in one direction: A stronger economy can afford to pay up for oil, and it's plausible that faster expected growth is lifting both bond yields and oil prices, as it did after strong data on Wednesday. Maybe the coincidence of AI-driven growth and the supply shock from the U.S. war on Iran is what's making the correlation so high.
Investors don't treat long bonds properly. Alan Blinder, former Fed vice chairman, pointed out three decades ago that there's a strong correlation between daily prices of one-year debt and the implied one-year rate 29 years in the future, even though almost nothing that happens on any given day should matter in 29 years. He suggested that long yields are far more sensitive to short rates than they should be if pricing were rational.
The opposite: Sushil Wadhwani, a former hedge-fund manager and Bank of England policymaker, suggests markets don't know-because no one knows-how much of a rate rise is likely to be permanent and how much temporary. So markets assume at least some part is permanent unless rates are obviously already too high. That means changes in near-term rate expectations feed through into longer-dated bonds.
Constâncio, the former ECB official, dismisses the day-to-day correlation as meaning little, a mix of coincidences and co-movement rather than causation. Though he does argue the 10-year is rightly higher because real, after-inflation, interest rates need to be higher in the long run.
I think there's more to the high oil-bond correlation than chance. It's previously been very high in the first Gulf War of 1990-91, when the Gulf trade was threatened; in 2010 amid worries about growth (an example of both bonds and oil being driven by something else); in 2012 as the eurozone imploded (ditto); in 2016 when the oil industry was in deep recession, holding back the economy, and in 2020 when Covid drove prices of everything.
There's no science here. And, as I explained in my last column, bond gauges offer some confusing contradictions. But I can see why traders think oil and bonds should move together.
Depending on your theory, the resulting high yields might be very attractive (if there's too much focus on short-term effects) or still nowhere near enough to make up for the risk (if it's the result of yet more government debt).
I think there's probably a bit of all the theories at work. I like the yield from Treasurys (and even more from U.K. Gilts) mainly as a hedge against the AI trade going badly wrong for stocks-but I share everyone's worries about U.S. debt, a Fed that's been too slow to respond to rapid growth and, of course, oil.
Comments