Michael Roberts: Iran and the US economy

Iran and the US economy

by Michael Roberts

The Iran war rages on.  Having failed to repeat his Venezuela option’’ ie decapitating the Iranian leaders and then getting Iran to surrender, US President Trump has been dragged into a long war.  So far, he has opted for escalation, cajoled by his advisers and forced on by unbridled attacks on Iran and Lebanon by Israel.  Strikes by both sides on so-called upstream gas production facilities in recent days are a significant escalation, with potentially long-term consequences. The latest strikes were the first time facilities associated with the production of fossil fuel energy had been hit in the conflict, rather than sites associated more generally with the oil and gas industry.

As I said in my first post at the outbreak of the attack by the US and Israel, that “two things need to happen before oil prices shoot up to $100/b or above. First, there must be significant and prolonged disruption of all traffic through the Strait of Hormuz, given that the Strait carries about one in five barrels of oil in the world. Second, the missile and drone attacks must start hitting oil production installations. If those two factors come into play, then the oil price per barrel could be in triple figures.” 

This has come to pass.  Today, crude oil prices hit $116/b (before falling back to $110) and even worse, natural gas prices in Europe exploded to over €68 per MWh, thus reaching their highest levels in over three years.

The International Energy Agency (IEA) now reckons that the war in the Middle East is @creating the largest supply disruption in the history of the global oil market@.  “With crude and oil product flows through the Strait of Hormuz plunging from around 20 mb/d before the war to a trickle currently, limited capacity available to bypass the crucial waterway, and storage filling up, Gulf countries have cut total oil production by at least 10 mb/d. In the absence of a rapid resumption of shipping flows, supply losses are set to increase.”

Global oil supply is projected to plunge by 8 mb/d in March, with curtailments in the Middle East partly offset by higher output from non-OPEC+ producers, Kazakhstan and Russia following disruptions at the start of the year. The loss of energy imports and the rise in prices hits some countries more than others. Asia in particular, suffers, followed by Europe, while, at least for energy, the US economy is relatively least affected.

Indeed, one part of the US economy is benefiting, namely US oil companies.  They stand to receive a windfall of more than $60bn this year if crude prices maintain the levels they have hit since the start of the Iran war. Modelling by investment bank Jefferies estimates American producers will generate an extra $5bn cash flow this month alone following a roughly 47 per cent rise in oil prices since the conflict began.  The capitulation of Venezuela to US control is also enabling US energy companies to raise production and increase sharply revenues from the now highly priced Venezuelan oil exports.

But for the rest of the US economy, the sharp rise in energy prices, whether at the gas stations or in home heating and industry, is already starting to feed through to prices overall.   Even before the war started, US producer prices (ie the prices that manufacturers sell their goods to wholesalers and retailers) were on the rise.  The producer price index (PPI) rose 0.7% in February with fuel and related products up 1.1%.  That meant PPI inflation was up 3.4% from February last year. Inflation in the US was not heading towards the Federal Reserve target of 2% a year, but instead heading back up. 

As for economic growth, the second estimate of US real GDP growth in Q4 2025 was revised down sharply to an annualised 0.7% qoq, well below 1.4% in the advance estimate. The new estimate reflected downward revisions in every component of GDP: exports, consumer spending, government spending, and investment. US real GDP growth for 2025 is now estimated at 2%, down from 2.4% in 2024 and 3.4% in 2023, while per capita real income rose only 1.1% in 2025 and fell in the last quarter of that year.

Slowing growth in national output is now also accompanied by falling employment growth.  In January, the US economy lost 92,000 jobs. Job openings in the professional and business services sector have fallen to just 4.0 per 100 employees, the lowest since the 2020 pandemic slump and down nearly 60% since the peak of white-collar employment in 2022. When white-collar hiring slows this sharply, the rest of the job market usually follows. 

Now the Iran war will widen the scissors between slowing economic growth and employment and rising inflation – in other words, stagflation is the order of the day. Donald Trump’s one-time pick to lead the Bureau of Labor Statistics said the US economy is too weak to handle oil at over $100 per barrel: “I don’t think this is an economy that is going to be able to handle $100 a barrel for oil, it’s just not,” EJ Antoni told the FT. “The economy is weaker than we thought it was, and inflation is worse than we thought it was.” New home sales plunged by 17.6% in January, the sharpest decline since 2013. 

This stagflationary environment has thrown the US Federal Reserve into a quandary.  Should the Fed raise its policy interest rate in attempt to curb inflation; or should it lower the rate to support employment and growth?  Yesterday, the Fed decided by a majority on the monetary policy committee to do nothing. The Fed increased its forecast for inflation this year and indicated that only one rate cut was likely in 2026, if at all. Far from heading towards the Fed’s target for inflation of 2%, inflation is now heading back towards 3% or above.  Today, the UK’s Bank of England and the ECB also held their policy rates.

Mainstream economists reckon that what primarily causes inflation is a rise in ‘inflation expectations’, a behavioural theory that this blog has refuted several times.  Five-year, five-year forward inflation expectations have not moved much in the last five years. Rising inflation had mainly a supply-side cause in the post-pandemic period and it will be the same this time.

The impact of the war is intensifying the widening gap between the rich elite in the US and the rest of American households – a gap that mainstream economists have called a ‘K-shaped’’ economy.  Spending growth has been notably faster at the top end of the income spectrum, while consumers at the bottom, who saw a brief burst of high wage growth post-pandemic, are now seeing wage growth slow.

Forbes magazine just released its latest annual ranking of global billionaires. The pace at which extreme wealth is rising is simply staggering.  According to inequality expert, Gabriel Zucman, the wealth of global billionaires has now reached the equivalent of 17% of world GDP.

The Iran war is also exposing new risks to the US economy that could trigger a financial crash. The 2008 global financial crash was not caused by high public debt, as many mainstream economists continually argue.  On the contrary, it was the meltdown in private sector debt that led to bailouts by government and then the rise in public debt followed.  In 2026, a private debt meltdown is again the danger. The recent report by an obscure financial analyst group, Citrini Research, on the future impact of AI caused a stock market sell-off in software companies before financial investors decided that there would be no crash in that sector.

However, what has become an issue is the possibility of defaults and bankruptcies in firms that have borrowed money, not from the traditional commercial banks, but instead from what are called private credit sources. In the past two decades, direct lending by private funds has become a crucial strand of the US financial system, providing credit to start-ups and other companies that would struggle to get bank loans or sell bonds. A classic private credit fund takes money from pension funds and endowments and locks it up for five years or more. That allows these private funds to make long-term loans to companies without fear that their investors will want their money back.  

But some of the big boys in private finance decided to attract pension funds and others to invest by offering “semi-liquid” funds, which promised investors quarterly access to their money, with the caveat that withdrawals could be capped at 5 per cent of fund assets to avoid fire sales.  These ‘financial products’ were a hit, attracting nearly $200bn investment and growing 60 per cent annually between 2021 and last year.

But these private credit funds are not regulated like commercial bank lending, so there is an inherent risk involved, just as there was with sub-prime mortgage lending in the financial crash of 2007-8.  It’s true that the size of the private credit market is relatively small compared to the total US loans market. Also, private credit funds are highly capitalised, with equity typically accounting for 65-80% of total assets more than six times the capitalisation of the banks, where equity represents about 10%.  As a result, the Fed’s 2025 stress tests found that even under severe recession scenarios, private credit would not threaten financial stability. Across the full spectrum of US credit markets, private credit has only a modest share of total credit outstanding. 

So there’s nothing to worry about? That is what they said about the mortgage companies that lent wildly in 2008.  Small cogs that clog up can also cause blockages for large cogs. As the US economy slowed, the private credit default rate (ie companies borrowing from private credit funds) has hit 9.2%. That’s higher than the 2008 bank loan default rate.

UBS says private credit defaults could hit 15%. That’s three times the peak bank loan default rate in 2008.

As a result, investors in private credit funds are trying to get out. And while most private credit funds have rules that limit quarterly redemptions to 5 per cent of assets — enabling them to “gate” (ie prevent) excess outflows — the exodus already echoes 2008.

Moreover, private credit and commercial banks are closely connected. “Banks are lenders, counter parties, service providers and, at times, backstops to non-bank entities,” observes Hernández de Cos, lamenting the “complex ecosystems of leverage, liquidity transformations and duration risk” beyond regulators’ control, which makes private credit a potential channel of systemic risk. US banks have $300 billion in exposure to private credit: Wells Fargo leads with $60 billion in loans to private credit funds. JPMorgan, which recently marked down software-linked loans and curbed lending, has $22 billion in exposure.

Goldman Sachs estimates that up to $70bn could flow out of private credit funds in the next two years and force the worst-hit managers to sell loans to meet redemption requests. And the longer the Middle East turmoil grinds on, the more risks will rise. Or to put it another way: the Iran war-private credit combination may not seem damaging enough to cause a global recession, but it could certainly spark a financial crash.

But maybe the AI technology boom will come to the aid of the US economy.  Some are arguing that already the US productivity of labour is rising faster as a result of the adoption of AI models and AI agents in companies.  In 2025, US labour productivity rose 2.8% compared to 2.3% in 2024, above the long-term historic average and above consensus forecasts.

Labour productivity is calculated as real GDP divided by hours worked. This can vary with changes in technology and in the amount of capital per worker.  But mainstream economists also look at total factor productivity (TFP), which is a measure of the growth in productivity not accounted for by increased capital investment or labour intensity. That is also picking up.

Everything depends on how quickly companies and their employees adopt AI models in their work and how far that spreads through the economy.  The St Louis Fed economists reckon that workers who used AI models could save 5.4% of their work hours, or 2.2 hours a week. But a 2024 working paper by Kathryn Bonney and others found that only 5.4% of firms had formally adopted generative AI as of February 2024.  That suggests that worker adoption remains mostly informal and won’t appear in productivity statistics.  

In a paper, Jed Kolko reviewed recent research on AI and its impact on the US labour market.  He concluded that @early research findings on AI’s impact on the labor market are inconclusive, weak signals about the future, and only one part of the AI research landscape.  And that @the commercial diffusion of the current generation of large language models (LLMs) is so recent that any lasting economic impact would likely take years to show up in employment, output, or productivity data.@

Current data from the Census Bureau’s Business Trends and Outlook Survey shows that fewer than one-fifth of firms are using AI in any capacity, and even fewer are using AI directly for producing goods and services. Indeed, the @transitional disruption from AI to date is not outpacing recent technological changes. The occupational mix has changed over the past three years at a similar pace to the years after the start of the commercial computer era (1984) and the commercial Internet era (1996) and has not accelerated since the release of ChatGPT.

So the hoped for productivity gains from shedding human labour and replacing it with AI agents still seem some time away.  Meanwhile, the huge investment bubble in AI could soon burst.  Take the leader in AI, OpenAI. It is a $730 billion company in invested assets, but last year it generated just $13.1 billion in revenue, losing $8 billion in doing so. This year, the losses could hit $14 billion, with cumulative losses reaching $143 billion by 2029!  These projected losses are five times greater than Uber accumulated before making a profit. OpenAI claims it will be profitable by 2029, but its AI model ChatGPT’s web traffic share has dropped from 86.7% to 64.5% in the last 12 months as Google’s Gemini eats into its market share. And the cheap Chinese DeepSeek can match the performance of ChatGPT at just 1/30th of the cost.

OpenAI needs 1.2bn paying subscribers to make a profit by 2029. That does not seem likely. OpenAI hopes to keep the loans and equity investments coming because it claims it can soon achieve an AI model that is super-intelligent, reasoning on its own at a superior level to the human brain. This is the AI companies’ ‘holy grail’, the moment of total enlightenment.  But the holy grail was just 19th century fiction. 

As Ruchir Sharma put it back last October, @America is now one big bet on AI@. It’s seen as the magic fix for every threat to the US economy.  But can it deliver?  More likely, there will be an AI financial bust first and possibly a recession before that question will be answered.  So AI as the saviour for Trump and the US economy remains an each-way bet.