Tightening Into a Fuel Shortage: Global Macro, September 2026

TWO DATASETS, ONE ECONOMY

The International Monetary Fund has cut its 2026 global growth forecast three times, from 3.3% in January to 3.1% in April and 3.0% in July, citing the energy shock from the Middle East war, partly offset by AI investment. Yet the J.P. Morgan Global Composite PMI rose to 53.5 in August, a 27-month high, underpinned by a surge in new orders and employment alongside the first increase in international trade volumes in six months.

The two measure different things. Surveys ask businesses how this month compares with the last, so they capture direction and mood. Forecasts estimate output over a full year. A sharp rebound from a weak spring can lift surveys to multi-year highs while annual growth still falls short.

That distinction would be academic if the survey signal were internally consistent, but it is not. The S&P Global US Composite PMI hit 56.0 in August, a 52-month high, implying Q3 growth near 3.0% annualized against 1.5% in Q2, with the Atlanta Fed's GDPNow at 4.4% on September 10. Over the same period, University of Michigan consumer sentiment fell to 47.8 from 51.7, its second-lowest reading on record and 13.2% below a year ago, with the expectations component down 11.1% and year-ahead inflation expectations jumping to 4.6% from 4.0% on fuel costs. In China, the private RatingDog manufacturing PMI rose to 51.5 while the official NBS index stayed in contraction at 49.8. In the eurozone, the composite held at 52.0 while France, the bloc's second-largest economy, contracted for an eighth month. Every strong reading has a weaker one beside it.

Central banks are acting on the stronger half. Robust PMIs describe an economy that can absorb higher rates, which is part of why the Fed is weighing a hike and the ECB has already delivered two. If the optimistic readings prove right, the IMF's 3.0% forecast is too low and today's tightening looks well timed. If the weaker readings prove right, policy is tightening into an economy that is already losing momentum, with rate rises biting hardest just as growth fades on its own. The next two to three months of output and spending data should settle which.

ENERGY IS DOING THE HEAVY LIFTING

US headline CPI held at 3.4% in August but accelerated to 0.4% MoM from 0.1%. Core eased to 2.4% YoY from 2.5%, though its 0.3% monthly gain beat the 0.2% consensus. Gasoline rose 3.9% MoM, more than a third of the total increase, and is up 16.3% over the year. PCE ran hotter still, with headline at 3.7% and core at 3.3% in July, but trimmed-mean PCE, which ignores whichever prices moved most that month in either direction, came in at 2.3%.

That 1.4 point gap is the evidence. If prices were rising broadly across a multitude of categories, removing the extremes would barely change the number. Instead it drops sharply, which means the extremes are carrying the headline. Core CPI at 2.4% and trimmed-mean PCE at 2.3% arrive at nearly the same answer by different routes.

The same pattern holds abroad. Eurozone CPI rose to 3.3% from 2.9% on energy accelerating to 14.3% from 10.3%, while core eased to 2.4% and services cooled to 3.0% from 3.3%, meaning the ECB hiked on September 10 into a headline moving the wrong way even as the underlying trend improved. China's CPI rose to 0.8% from 0.5% with producer prices up 3.8%, a fragile stabilization built on input costs rather than a revival in demand. US producer prices rose 5.4% YoY, this year's highest reading, while services producer prices rose just 0.1%, hinting that firms are absorbing costs rather than passing them through.

The labor data does not look like demand-driven inflation either. August payrolls added 162,000 jobs with unemployment steady at 4.1%, but over 60% of the gain came from food services and local government education, both seasonally sensitive. Real average hourly earnings fell 0.1%, the second straight month of declines. The quits rate, the closest proxy for worker bargaining power, stood at 1.9% in July against a peak above 3.0% in 2021 and 2022. Higher rates cool spending, but they cannot fix a fuel shortage. Tightening therefore addresses second-round effects rather than the cause, and that distinction defines what success looks like: if it works, it shows up in expectations and wage settlements, not in the energy component.

WHERE THE SHOCK LANDS

The US is moving at two speeds. Retail sales fell 0.6% in July, the steepest monthly drop since May 2025, landing in the same window as the drop in sentiment. Our view is that business-led growth holds but slows, and that real wages are the variable to track rather than sentiment, which has been weak for months without spending collapsing. With 162,000 jobs a month and 4.1% unemployment, income keeps flowing into households even when each paycheck buys less. That is a squeeze rather than a break. The specific risk is sequencing: if a September hike slows hiring while real wages are still negative, both supports weaken at once. The trigger to watch is not another weak sentiment print but the first payroll report that comes in soft while real earnings are still falling.

The eurozone is not stagnant, but its growth is narrow and defensive. GDP rose 0.6% QoQ in Q2, though Germany, France and Italy each managed only 0.2% while Spain grew 0.7% and Ireland's volatile figures flattered the headline. More telling is how unevenly the energy shock is landing. Spain's renewable capacity and fiscal support have cushioned consumers, and employment there grew 2.3% YoY. Germany's energy-intensive auto complex has not been cushioned: employment fell 0.5% YoY, industrial production dropped 1.1% in July, and 72% of automotive suppliers surveyed plan to cut investment in Germany. The ECB is setting one rate for economies absorbing the same shock very differently, too tight for Germany and arguably too loose for Spain. The risk is not a bloc-wide downturn but a German industrial slowdown spreading through the supply chain that depends on it, given that Slovakia's car sector alone generates 9.2% of GDP and close to half of its industrial production.

China's stabilization comes almost entirely from exports, which rose 25% YoY in August for a record monthly trade surplus of roughly $119 billion, with semiconductors up nearly 130% and autos 43%. Two caveats cut against reading that as strength. Export volumes actually fell 7.9%, implying prices rose roughly 150%, so the revenue gain reflects a global chip shortage more than expanded output. And the growth was redirected rather than expanded: exports to ASEAN, South Korea, Taiwan and Russia are all up 25-35% YTD while exports to the EU rose just 6.6%, with a significant share of the ASEAN and Hong Kong volume being transshipment. Domestic demand tells the opposite story, with imports missing expectations for the sixth time in eight months, property development investment down 19.2% YoY over the first seven months, and home prices still falling. The PBOC has held its main lending rates for 15 straight months at 3.0% and 3.5%.

The commodity picture makes the regional point most cleanly. Global inventories fell 95 million barrels in August, bringing the drawdown since February past 500 million, but non-OECD stocks fell 52 million led by China while OECD stocks actually rose 23 million. The acute shortage has moved from crude to refined products, particularly diesel: combined exports from the Gulf and Russia, regions supplying nearly 45% of global seaborne trade in that fuel, were 1.6 million barrels per day lower in August than in February. Pakistan, the Philippines and Sri Lanka have moved to four-day working weeks to conserve fuel. In the same month, Atlantic Basin refiners posted record profit margins on diesel demand they are positioned to supply. The damage and the benefit are both regional, and they do not overlap.

FOUR EVENTS IN FIVE DAYS

Iran meets the Gulf Cooperation Council, foreign ministers and Iraq in Oman on September 14 to discuss the Strait of Hormuz. The Senate votes on CLARITY Act cloture on September 15. The Fed decides on September 16, the same day the ECB's latest hike takes effect. The BOJ follows on September 17-18 with policy at 1.0% and a hike expected.

Oman is the most underpriced of the four. Every inflation print in this report traces back to energy, and energy traces back to Hormuz, which carries 20% of global oil and gas supply. Iran and Oman have already agreed a temporary corridor and a shipping map, and a wider regional framework would unwind the cost shock faster than most forecasts assume, making this month's tightening look mistimed within a quarter. Arguing against that expectation: the US is not party to the Iran-Oman track and is unlikely to accept terms stopping short of free passage.

Escalation is the larger risk in the other direction. Inventories have absorbed more than 500 million barrels of drawdown since February, so the buffer that cushioned the first shock no longer exists. A second disruption would transmit further and faster than the first.

On the rate decision itself, the July 29 hold at 3.50% to 3.75% was not unanimous. Three regional Fed presidents wanted to hike immediately, the biggest split on the committee since 2016. Hike odds then swung from about 70% after Chair Warsh's hawkish Jackson Hole comments to about 48% after Governor Waller signalled he would hold if inflation kept cooling, back above 70% on a hot producer price report, and to almost 90% after the September 11 CPI print. Meanwhile the Treasury moved the other way, tripling its bond buyback to $6 billion for the September 10 operation with $4 billion set as the new floor after 30-year yields hit their highest level since 2007. The Treasury calls that a liquidity measure; others call it stimulus arriving just as the Fed may be tightening.

WHAT IT MEANS FOR CRYPTO

Stablecoin supply shrank through June and July, then held steady in August. The standard reading treats that as risk appetite returning, which deserves qualification, because it applies a trading lens to a number that is mostly not about trading. Roughly 66% of global stablecoin supply sits in emerging markets per Goldman Sachs estimates, where these tokens function as payment and savings infrastructure rather than cash waiting to buy other crypto. USDT dominates that segment with around 74% of on-chain trading volume concentrated in Asia, Latin America and Africa, while USDC leads institutional flows in developed markets. A contraction in the global figure can therefore reflect remittance activity in Lagos or Manila as easily as positioning in New York.

That same regional split determines who the September 15 CLARITY Act vote actually affects. US markets hold the institutional infrastructure, meaning custody, regulated products and ETFs, so a failed vote matters most there. Europe already operates under MiCA, and Asian grassroots activity, which leads global adoption, runs largely outside US rules either way. The vote is a significant US institutional catalyst rather than a global one.

The two September events are independent and should not be read as one. CLARITY affects US institutional infrastructure; the rate path affects every risk asset. Of the two, the rate path is the larger variable.

THE BOTTOM LINE

The inflation now driving policy is a supply shock rather than a demand problem, with core and trimmed-mean measures near target while energy carries the headline in the US, the eurozone and China alike. The counterargument carries real weight: central banks are not targeting the shock itself but its second-round effects, and with year-ahead expectations at 4.6%, that concern is not hypothetical.

Regional differentiation matters more this quarter than aggregate calls. The same shock has cushioned Spain and squeezed Germany, delivered record margins to Atlantic Basin refiners and forced four-day working weeks in Pakistan and the Philippines. The more useful question is not whether the world economy is expanding, but where the shock is being absorbed and where it is being passed on.

Few dispute that policy should respond to inflation at these levels. What is disputed is whether now is the moment, given that central banks are acting on survey data while hard data points the other way. The near-term risk is timing rather than direction, and it resolves over the next two to three months in output, spending and payroll figures, not in any single decision this month.

Click here to read the full research paper.