War-driven energy price spike increases risk of central bank policy errors

Entering the fourth week of conflict in the Middle East as I write, missiles are still flying, the Strait of Hormuz remains closed, and oil and gas facilities across the region are under threat. Brent crude oil has spent much time above USD 100 a barrel, while European natural gas is double its price on the eve of the war. News flow changes from negative to positive and back by the hour, with resulting energy price shifts.

What was the economic environment just before the war?

With the war ongoing, it can be easy to lose sight of how the world economy looked just a few weeks ago. But it is important to recall that baseline. When the war ends, there will undoubtedly be extra factors to evaluate—but these will build on the prevailing trends in place before the war. A few reminders:

1. Higher inflation was already a risk. At a recent portfolio strategy meeting, one of my colleagues remarked that “Xi Jinping has been the most successful Federal Reserve governor of recent times.” The point being that Chinese domestic policy to address a weak housing market and slow economic growth has been a significant driver of lower inflation around the world in recent years. But recent “anti‑involution” measures (which try to stop industry from producing goods at a loss) have turned the tide. Exported global disinflationary factors have essentially ended; they are now potentially inflationary.
   
   Inflation risk was on the table even before the war, with inflation already above target in many countries.
   
   
2. The effects of AI on inflation were unclear. Arguments against the case for higher inflation were typically about either technical factors involved in the housing components of the consumer price index (CPI) or, more persuasively, how AI’s impending impact on the economy would work to restrain inflation. In the short term, spending on data center buildouts is likely to be inflationary—certainly, the cost of firing up those data centers is now much higher. But the core of the argument is that AI would enhance productivity, reduce labor demand, and therefore be disinflationary over the long term.
   
   I certainly find that rationale logical. However, regulation and other factors make the time horizon for this argument fuzzy, although I suspect the theory is anchoring long‑term inflation expectations.
   
   
3. U.S. growth was primed for an upswing. The One Big Beautiful Bill Act, Fed easing, looser financial conditions, and near‑record expected tax refunds had set the stage for a bumper summer ahead of the 250th anniversary of U.S. independence on July 4 and the midterm elections in November. Even beyond the U.S., monetary easing and expansionary fiscal policy (particularly focused on defense) made for a positive global growth backdrop.

What has changed?

Clearly, the cost of energy is higher. Much higher. There are shortages in some parts of the world, and damage to infrastructure in the Middle East means supply constraints—mainly the blockage of the Strait of Hormuz—are likely to persist for some time beyond the end of the conflict.

While we cannot be certain how long the war may persist, the incentives for all parties favor a conclusion sooner rather than later, in my view. Whether compelled by concerns around logistics, human, economic, or political costs, the odds are stacked toward de‑escalation. A simple rule in a crisis is to panic early and panic big! While headlines can feel bad, I wonder if the time for panic is now well past.

Away from the war, there has been a different type of destruction in markets. Much of the price action has simply been a “seek and destroy” of crowded positions. The forced unwind of consensus trades explains a substantial amount of the most extreme market moves, particularly in currencies and rates markets. 

Do not underestimate the degree of leveraged unwinds in certain markets. Some large rates markets have been less liquid than in 2020 at the onset of the COVID pandemic, and the combination of low liquidity and crowded positioning can lead to outsize market moves. UK government front‑end yields moving 100 basis points in only a few days had much to do with market technicals as opposed to fundamentals, in my view.

Looking ahead: actions in fixed income

Given that inflation was rising ahead of the war, I believe adding more inflation protection is the most obvious portfolio action to consider. With breakevens hardly changed and CPI prints likely to be very high in the coming months, inflation‑linked securities look cheap.

A second area to position for will arise from the way central banks react to the spike in energy prices. Two types of policy error are possible. Some central banks will likely hike when they shouldn’t (stifling their economies in the process); others will look through the energy spike and could sit on their hands (leading to higher inflation expectations). Since many central banks have chairs nearing the end of their tenure who may be motivated to take a monetary policy stance that burnishes their legacy, I suspect each type of policy error will occur frequently. 

There are two ways to position for policy error risks:

First, in currencies, I’d expect heightened volatility and carry profiles, leading to higher potential alpha. The U.S. dollar rally during the war has not been the result of its use as a risk hedge. Rather, I believe it has rallied because short positions were being unwound and because oil is priced in U.S. dollars. The upshot—the weaker U.S. dollar trend has not ended.

Second, an even more interesting way to take advantage of central bank policy errors is in the global rates market. Cross‑market relative value and dispersion are as broad as we have experienced since before the global financial crisis of 2008–2009. Carefully leveraging front‑end rate differentials could provide high‑quality returns with relatively low volatility. While the last four years have been about getting the direction of yields and the curve right, now I see the best opportunities in cross‑market positions.