Nouriel Roubini offers an explanation for the relatively benign global economic outlook, despite obvious risks

I have followed Nouriel Roubini for years and have considerable respect for some of his earlier work, particularly his warnings around the GFC and his subsequent criticism of the too-big-to-fail banking model. Which is why I find this latest piece all the more extraordinary.

Roubini manages to identify the Iran war, oil above US$100, disrupted shipping through Hormuz, enormous fiscal deficits, rapidly accumulating sovereign debt, rising bond yields, the possibility of credit crises, an increasingly questionable AI boom, stagflationary pressures and even the eventual prospect of debt monetisation. Having assembled that rather formidable collection of flashing lights across the global economic dashboard, he somehow arrives at a "relatively benign" global economic outlook.

For what it is worth, I arrive at the opposite conclusion - in fact, I increasingly believe we are watching the ingredients assemble for a financial perfect storm which could eventually combine elements of 1929, the dot-com collapse and the GFC - except this time they are sitting on top of sovereign balance sheets already carrying extraordinary levels of debt.

Perhaps the most interesting place to begin is Roubini's assertion that "the world is in the middle of a massive, positive long-term aggregate supply shock, owing to the AI investment boom."

I have read that sentence several times and I am still none the wiser as what he is claiming - there are two quite different propositions buried inside it.

If he means the present AI investment boom is itself already some enormous positive aggregate supply shock, I find that difficult to reconcile with what we can actually observe. The AI build-out is presently consuming extraordinary quantities of capital, electricity, gas generation, grid capacity, transformers, copper, semiconductors, cooling infrastructure and skilled labour. In several of those areas, AI isn't expanding available supply at all - it is competing aggressively for resources that are already scarce, and in some cases driving their cost considerably higher.

If, on the other hand, Roubini means that all this infrastructure will eventually enable AI to produce an enormous economy-wide productivity dividend, that is certainly possible. But it remains a forecast rather than an established economic fact, and personally I am considerably more pessimistic about the American side of that bet than Roubini appears to be.

The United States for the tome being still possess some technological advantages, particularly around frontier semiconductors and parts of the software ecosystem, but I think it would be an enormous mistake to assume those advantages are permanent. China is attacking those remaining weaknesses aggressively while already possessing extraordinary advantages elsewhere in the industrial system required to deploy AI at scale.

And that is where I think much of the Western discussion is looking at the wrong race.

The AI race will not necessarily be won by whoever possesses the world's fastest individual GPU. Surely, it will ultimately be won by whoever can produce the greatest quantity of economically useful compute at the lowest total system cost.

That means electricity, generation capacity, transmission, construction, financing, cooling, critical minerals, equipment, manufacturing depth, supply-chain resilience and eventually the cost of replacing enormous quantities of rapidly aging hardware.

Electricity alone should have alarm bells ringing.

China has spent decades building an immense electricity system and industrial supply chain, while the United States is now discovering that announcing another multi-gigawatt data centre is considerably easier than producing the electricity, transmission infrastructure, turbines, transformers and pipelines required to run it.

Oracle's Project Jupiter is already providing a fascinating glimpse of what happens when the AI dream collides with the physical economy. Anthropic is another extraordinary example, with hundreds of billions of dollars of future infrastructure commitments resting upon assumptions of equally extraordinary future demand and revenues.

None of this proves that American AI will fail. But I increasingly believe the underlying economics point toward something much more dangerous than Roubini's benign productivity revolution.

I think there is a serious possibility that America is constructing one of the greatest episodes of capital misallocation in modern economic history, while simultaneously competing against a Chinese industrial system capable of delivering much of the underlying physical infrastructure at substantially lower total cost.

If China also succeeds in progressively closing its remaining semiconductor and compute disadvantages, that equation could become brutal. And the numbers being committed are now so enormous that failure would no longer remain neatly contained inside Silicon Valley.

Capital expenditure and productive investment are not necessarily the same thing. Spending another trillion dollars unquestionably creates economic activity while the money is being spent, but whether it creates sufficient additional productive output and future cash flow to justify that trillion dollars is another matter entirely.

This is where I think Roubini's optimism becomes particularly dangerous. We already know the investment is massive. My concern is that the eventual productive return will prove nowhere near massive enough.

And where that investment has been financed through bonds, loans, leases, private credit and increasingly elaborate financial structures, there is a particularly uncomfortable asymmetry sitting underneath the entire exercise.

The hardware can become obsolete, but the debt doesn't.

This brings me to another familiar part of Roubini's argument, his warning that enormous fiscal deficits and public debt may "crowd out" private production, capital expenditure and consumption. There is certainly a legitimate argument about governments competing with the private economy for real resources and balance-sheet capacity, particularly when economies are running into physical constraints, but the conventional explanation invariably seems to drift back toward the same old bucket-of-money model.

Government wants more money, private enterprise wants more money, AI suddenly wants mountains of it, everybody competes for a supposedly finite pool of savings, and naturally its price rises.

But where is the banking system in this story?

Commercial banks don't simply wait for savers to deposit money and then lend those savings onward. When banks extend loans they create new deposit money, and once that mechanism is understood, the much more important economic question becomes where that newly created credit is directed and what subsequently happens to it.

Credit created to finance genuinely productive new capacity capable of generating sufficient additional output to service the resulting debt is one thing. Credit directed into existing property, financial assets, speculation, leveraged acquisition or investment incapable of producing an adequate return is something else entirely.

That distinction has occupied much of my own research for years because I increasingly believe it sits near the heart of what Western economics repeatedly gets wrong. We obsess endlessly over the price of credit while paying remarkably little attention to its quantity and destination.

Which brings us neatly to interest rates.

Roubini suggests that persistent inflation may require central banks to raise rates further, yet much of the inflationary danger he has just described arises from war, oil above US$100, disrupted shipping, energy constraints and other supply problems.

Increasing the cost of credit cannot produce another barrel of oil, reopen the Strait of Hormuz, manufacture another transformer or add another gigawatt of electricity to the grid. It can certainly destroy demand, but it also raises the financing cost of the productive businesses attempting to expand supply in the first place.

And then we arrive at the part of this perfect storm which I believe could ultimately prove far more dangerous than the dot-com collapse or even the GFC - the sovereign debt sitting underneath it.

Higher sovereign yields progressively increase government refinancing costs. Greater interest expense worsens fiscal deficits, everything else being equal. Larger deficits require additional borrowing, while that additional issuance must then be absorbed by a bond market already demanding greater compensation for inflation, duration, geopolitical uncertainty and fiscal risk.

If yields rise further, another turn is simply fed into the same wheel.

Higher yields feed higher interest expense, which feeds larger deficits, which require still more borrowing, which can place still more pressure on yields.

The US 10-year Treasury has already touched around 5.34%, its highest level since 2002, while inflation concerns, oil, enormous government borrowing and AI-related corporate issuance are all colliding in the same bond market.

This is where comparisons with earlier financial collapses become particularly uncomfortable.

The dot-com bubble involved spectacular technology valuations and enormous over investment, but it did not sit on top of today's sovereign debt structure.

The GFC involved catastrophic leverage and misallocated credit throughout the banking and property systems, but governments and central banks entered that crisis with considerably more room to absorb private-sector losses and suppress borrowing costs.

And 1929 demonstrated what can happen when excessive speculation, leverage and confidence finally reverse together.

My fear is that the next major crisis could contain elements of all three simultaneously - an historic technology investment bubble, enormous quantities of increasingly interconnected credit, grotesquely inflated financial assets, a bond market already repricing sovereign risk, an energy and geopolitical shock, and governments themselves entering the storm carrying balance sheets that severely restrict their ability to rescue everybody else.

That is why I don't see a relatively benign outlook.

I see the possibility of 1929, dot-com and the GFC arriving at the same party - and this time the sovereign is already drunk.

There is another reason I find Roubini's analysis particularly frustrating. He has previously been highly critical of too-big-to-fail banking, attacked the conflicts inherent in giant financial conglomerates and argued for separating ordinary commercial banking from the Wall Street casino. I therefore wouldn't describe him as some mindless cheerleader for the private banking establishment - his own record doesn't support that accusation.

But I do think he stops one enormous step too early.

He has been willing to question what the banks have done without seriously questioning the extraordinary monetary privilege sitting underneath the banking system itself - the private creation and allocation of most of the money circulating through our economies as interest-bearing debt.

Once you begin examining the system from that direction, the landscape changes dramatically. Who creates the money becomes important. Who receives the newly created credit becomes important. Whether that credit finances productive capacity or existing assets becomes enormously important, and so does the wisdom of attempting to control this vast credit-creation machine primarily by repeatedly changing one blunt instrument - the price of debt.

These aren't obscure monetary questions sitting somewhere on the fringes of economics. They sit directly underneath the inflation, asset bubbles, banking instability, sovereign debt, rising interest expense and increasingly fragile bond markets that Roubini himself is describing.

That is what I find so utterly frustrating about this article. He has identified many of the symptoms, and in several cases identified them very specifically, but once again the discussion screeches to a halt before any serious examination whatsoever of the monetary architecture sitting underneath this Western-centric financial debacle.

This is a classic case of the eCONomic mysticism that has pervaded global financial architecture for more than a century - endlessly analysing interest rates, inflation, debt, deficits, bond yields and banking crises while somehow avoiding the most fundamental questions of all - who creates the money, where the resulting credit goes, who receives the interest, and whether the debt created alongside it finances sufficient productive output to service it.

Roubini might have predicted the GFC, and he deserves credit for doing so, but today he appears remarkably relaxed as an historic AI investment bubble collides with grotesquely indebted sovereign balance sheets, rapidly rising borrowing costs, geopolitical and energy shocks, increasingly fragile bond markets and a fiat monetary architecture whose ultimate backstop is precisely those same indebted sovereigns and their central banks.

I see nothing remotely benign about that convergence.

Rather, I see the bond market and the Western fiat experiment approaching their ultimate reckoning - and when this perfect storm finally comes together, I believe 1929, the dot-com collapse, and the GFC, will look like a picnic in Central Park compared to what is about to unfold.

Is it resilience or scrambling to put more fingers in the leaking dikes?

Asset valuations are only notional until reverted to cash or other fungible tokens.

"which also increasingly includes higher defense spending". And this is a good thing? Here, we seem quite prepared to starve our health services of money for staff, while contemplating a much higher spend on military hardware like warships. To be used against? Alice in Blunderland.

Roubini has made an account of factors effecting world economies but omits the natural world geophysical and climate forces and their effect on us present and future. He hasn't picked any trends from what I read. Dr Doom has taken to to his armchair and is watching the news.

We continue to swim in a sea of uncertainty, an uncertainty mixed with fear which is not unreasonable because international leadership is to put it mildly a mixed bunch.

Perhaps our situation can be framed as like an overpopulation of deer on an over browsed island, with more than a few competing wolf packs?