President Trump cannot be credited with the creation of economic bubbles, but he will certainly be debited with the fallout if they should burst on his watch. This needs qualification: Trump’s policies have achieved an extraordinary short-term economic equilibrium, but at a long-term strategic cost to the US of the loss of its alliances and the erosion of the quality of US democratic institutions, as well as rocketing levels of government debt. Trump has monetised trade, alliances, immigration, military protection and even coerced foreign investment from trading partners. The sums received will average about US$ 300bn for each of the next two years. That masks the fiscal damage of the Big Beautiful Budget. After that, the fiscal damage will be in evidence again. But what boosts bubbles and markets is the short term. And in that Trump has been successful.

How do bubbles work? First, we can quantify the bubbles. Then we must consider the policies to deal with the crises that emerge when they have burst. After that, you can make up your own mind about whether their collapse should be considered as a ‘capital market correction’ or an ‘economic catastrophe’. All bubbles have three parts: irrational exuberance and a credit bubble that finances an asset bubble. When I talk about bubbles, I include both the asset and liability side.

When a bubble bursts, its size in relation to GDP and its penetration of the productive economy will determine the damage to the economy. For a bursting bubble to be a ‘correction’, it must be confined to financial markets and their near relations (like housing). Credit transmission must be left intact. The bubble must also be relatively moderate in size (otherwise there will be spillover effects to wealth and risk appetite, which signal more lasting economic damage).

The ability of regulators to ‘deal with’ a bursting bubble varies inversely with its size, surprise and the suddenness of collapse. The successful resolution of bubbles requires a leadership that recognises it for what it is and has clear ideas of what needs to be done.

Bubbles can survive for a long time provided the capital that creates them remains underpriced (Japan’s bubble economy of 1986-91, for example). A long time is not, however, forever. The undoing of a bubble is based on price discovery, and that is rarely gradual for bubble assets. As a consequence, the collapse of the bubble is unlikely to be gradual. It can be argued that in an economy where trade and capital flows take place behind protectionist walls, bubbles may be more frequent and of more lasting duration. In such an economy, price discovery may be distorted, and financial repression may bias capital allocation.

The potential bubbles in the economy are as follows: Non-Bank Private Credit is now bigger in Europe (3.8 times GDP) than in the US (3.1 times GDP) but huge in both. So who said Europe was a laggard at financial innovation? Banks, never ones to miss out on a good party, have lent about 10 per cent of their loan book to the intermediaries that create Non-Bank Private Credit. If that were to be lost, the banks would also have to write down a substantial part of their Tier 1 capital, which stands on the other side of their balance sheet as a buffer against bad times. Tier 1 capital ratios for EU banks are 16 per cent, for their US peers 11.5 per cent.

Non-Bank Private Credit largely escapes regulators’ attention. Ironically, at the same time, the process of credit disintermediation – by which the creation of credit by banks becomes a lesser part of total credit, and that due to non-banks a greater part – also owes much to Basel-type regulation that forces banks to cut back lending where it absorbs a lot of bank capital as a buffer.

Non-Bank Private Credit is currently involved in a broad scope of credit to SMEs and lesser-quality borrowers by fintech and hedge funds. Insurance companies, in the hunt for yield, are major funders. Over 46 per cent of US Life insurers’ bond holdings are ‘private bonds’ ( that is, those not registered with the US Security and Exchange Commission, private placements, and section 144A bonds). The figures for European Lifers are unknown. In short, Non-Bank Credit portfolio pricing is opaque, but we know that shortfalls in debt service are often added to final maturity payments and not accounted for as Non-Performing Loans.

Sovereign debt is also back on the bubble menu because its level and trajectory raise the issue of sustainability. Sovereign debt to GDP for Advanced Economies will reach 95 per cent of GDP this year and could reach 123 per cent by the end of the decade. When the European Union had its Sovereign Debt Crisis back in the 2020s, EU debt was only 87 per cent of GDP. The US, Japan and China are now leading the way. The US, according to both the IMF and the CBO, will see its debt to GDP ratio rise by 20 percentage points to 143 per cent of GDP by the end of the decade. That is mainly down to the Big Beautiful Budget.

It is worth recalling that, if sovereign debt rises as a proportion of GDP, it is because the new debt is not being invested productively enough to produce the GDP to service it. Private equity is estimated to be capitalised at US$ 3.1 trn in Europe and the US. How well it is invested is obscure. Private equity cannot be marked to market because there is no market. So valuations are only policed by the auditors. The investment thesis is that, by sacrificing liquidity and transparency, private equity investors’ money can be managed to achieve superior returns. Why should that be? There is no evidence that equity investments that are shielded from scrutiny earn higher rates of return than equities on publicly quoted markets.

Digital Assets, like Bitcoin, Ethereum et al., are capitalised at US$ 4-5 trn. Stablecoins are not included here because they represent the privatisation of money used for transactions. Bitcoin and its like are speculative assets whose volatility excludes them from the transaction role of money. Their intrinsic value – apart from mining costs – remain a mystery to me. When investing in a normal security you get a share of the assets owned by the company and the revenues produced by those assets. Bitcoin has neither.

Artificial Intelligence is an asset bubble. It is socially useful; I use it to build models and mine information. And it will change employment (I used to employ people to do these things). So, over time, it will boost productivity. But I don’t have to pay for it. I simply use open-source models. My argument is that AI, while being a social good (when not abused), will not generate the sort of returns to pay for the ego-driven, gobsmacking amounts of capital being poured into it. Annual investment in AI is forecast to reach US$ 1.5 trillion in 2025. That is nearly seven per cent of EU or US GDP.

Don’t take my word for it. AI investment accounted for 40 per cent of all US fixed asset (ex-construction) investment in the US last year. Yet it contributed only 0.5 per cent of GDP growth. That is forecast to fall to 0.2 per cent this year. Worse still, in 2025, US Fixed Investment (excluding AI and other IT) has shrunk by around 3-5 per cent (QoQ annualised). AI investment has continued to grow by double digit percentages. Will this be a case of delayed return or no return? And what happened to all the United States’ wonky bridges, potholed roads and decaying social infrastructure (railways, schools, hospitals etc.)? That is a reminder that there is an issue of social inequality inherent in the AI bubble that makes it even less sustainable by enriching the few and impoverishing the rest.

Credit and asset bubbles are all interlocked – excessive credit in one sector fuels an asset bubble in another. This is always the case, but perhaps never quite like now. Moreover, losses in one bubble sector, like non-bank credit, would be compounded by losses in other bubble areas, such as AI or Digital Assets.

The sums involved today dwarf most previous credit bubbles, and they penetrate the real economy deeply and ubiquitously, like arteries.

It is unlikely that a Non-Bank Credit failure could be limited to a correction by regulators. Even a 20 per cent loss on European Non-Bank Credit (with a loss given default of 30-40 per cent) would result in a net loss of US$ 5trn. That’s the size of the German economy. The figure for the US would be US$ 7trn, or 24 per cent of US GDP – the equivalent of French and Italian GDP.

What would cause such a collapse? It could be short term shocks like an adverse decision by the US Supreme Court on the use by the Trump Administration of the International Emergency Economic Powers Act (IEEPA), emergency powers to impose tariffs. This would cause the fiscal edifice of Trump’s policies to collapse. Or, it may be rising interest rates caused by the long-term costs of the Trump Administration policies. Bubbles only thrive on under-priced capital. Or, it might be the lesson we learned as kids on the beach when piling dry sand on itself until the heap collapsed, that the last grain was just too much.

What kind of policies can deal with the aftermath of bubbles? Here, I must rely on historical precedent.

No matter what the damage, a 1929-style Great Depression is very unlikely. The 1929 Crash was made much worse by policy. Initially, interest rates were raised. Monetary policy was constrained by the gold standard. Fiscal policy was tightened to balance budgets. There was widespread bank failure. The thinking of the day was non-interventionist. None of those policy mistakes would be made today.

What about the Dot.com crash? This destroyed equity and bond capital amounting to US$ 5-9 trn. That is about the low estimate for a bubble burst today of several of the candidate sectors we have identified. And yet, the recession it caused was very mild (about 1.3 per cent of US GDP) and it lasted just eight months (March to November 2001).

Policy reaction was swift. The Federal Reserve cut rates from 6.5 to 1 per cent. The Bush administration passed tax cuts in 2002 and 2003.

But the biggest determinant of damage limitation was the nature of the Dot.Com bubble itself. The banks were only marginally involved. The credit machine remained intact. Most of the damage was limited to the tech sector. The broader indices fared quite well. And the economic damage was also to tech, with traditional sectors weathering the storm.

Is this a model for the fallout if current bubbles burst? Probably not. The reason is that much of the current excesses are large components of the non-bank credit machine and therefore bigger than the banks, which are nevertheless involved, to the tune of 10 per cent of their assets, in lending to originators of non-bank credit.

A better fit for today’s circumstances would be the 2008 Housing Crisis and ensuing Great Recession. The housing crisis was the result of a nationwide mania. The banking sector was front and centre in exposure to housing and to Mortgage-Backed Securities (MBS) and Collateralised Debt Obligations (CDOs). There was widespread fraud – another factor that makes bubbles burst upon discovery. Banks were extremely leveraged. Plummeting mortgage values wiped out capital (completely in two major banks). The credit system collapsed with banks afraid to lend to each other or to third parties.

Policy-makers reacted fast and in extremis. There were fiscal bailouts (TARP), Fiscal stimulus (the American Recovery and Reinvestment Act – a US$ 830bn fiscal stimulus package). The Fed cut rates to zero per cent and undertook Quantitative Easing (QE).

Yet the recession lasted 18 months (Dec 2007 to June 2009), not just eight months like the Dot.Com recession.  Job losses (8.7mn) spanned the entire economy. Unemployment rose to 10 per cent. And, of course, there was global contagion.

What made the 2008 Housing Collapse so bad compared to the 2001 Dot.Com bubble burst? Policy can’t be blamed. It was credit: 2008 broke the credit machine; 2001 did not. In addition, the excesses of the Dot.com bubble were far more limited in economic scope than the housing mania of 2008.

Consider today’s bubbles. At their core is non-bank credit, not banks. But non-bank credit has replaced banks as the financiers of excesses. Not only that, but the size of non-bank credit in relation to both US and European GDP (3.1- and 3.8-times GDP, respectively) indicates that this form of lending is ubiquitous in the economy, though we have few details of where the liabilities are located. That means it is the economic, and not just the financial, credit machine that will be in the cross hairs of any crisis. That could be the determining factor of the Correction Vs Catastrophe outcome.

How could leadership shape the next crisis? The two most recent crises showed timely, decisive leadership with a clear understanding of what needs to be done. As we have seen, the non-intervention of 1929 stands in marked contrast to 2001 and 2008. So, leadership is key to the outcome. In the next crisis, will we get it?

The ECB would likely plod along, but the current US administration may go into denial. The Trump administration has an entrenched policy commitment to digital assets (see Executive Order 14178 of 23 January 2025) and to Stablecoin (the Genius Act) as well as, reputedly, making US$1 bn from tokenised assets in personal gains. Failure of digital assets would also hurt a lot of Trump supporters. Finding culprits rather than solutions might be prioritised. Poor leadership from the top can magnify rather than resolve the fallout of bursting crypto bubbles.

The scenario of bursting bubbles could have an enormous impact on geopolitics. The end of Trump’s brand of populism, for one thing, and a long, very gradual return to more conventional national and international policies. But there are brakes on the wheels of such normalisation. Among them: a lack of international trust – US populism could regain power later (‘the issue is not Trump but a polarised US society’); resistance to free trade by lobby groups that benefited from Trump’s protectionism could be spurred; and continued US isolationism could have a serious effect on spending on military alliances.

In Europe, the loss of wealth equal to France’s GDP would leave Europe with a gaping hole in its ability to finance its own defences against Russia. A poorer EU is not going to be generous about the integration of Ukraine and other candidate countries. And a poorer Europe is likely to be a more populist one. Putin, or his successors, will exploit this opportunity to achieve its goals of subjugating and demilitarising the EU. The Kremlin is likely to find support among European populist governments.

In such a scenario, the credibility of the US deterrent will suffer among both friend and foe in Asia. For allies like Japan, Korea, Taiwan and Philippines, there’ll be doubts that a volatile US, also suffering from the economic costs of burst bubbles, is a reliable defence partner. President Xi Jinping may decide that it isn’t, and move on Taiwan kinetically (potentially with a trade embargo rather than D-day style landings).

The developing world, as always a victim of bursting Western bubbles through trade and credit channels, will become more convinced that the American Way is not their way and align increasingly with China – at least economically.

Throughout this essay, I have asked questions about economic bubbles and set up a framework for answering them. The questions are as follows: are there significant bubbles around? If so, what would make them burst? And if they burst, will the result be a market correction or an economic catastrophe? What makes the difference? What I cannot do is tell you what to think – I can only provide a framework to inform your judgement.

I will, however, tell you what I think. As things stand, there are enough bubbles that are set to burst within a year. When they do burst, they will cause significant economic disruption. The fallout won’t lead to a 1929-style crisis, and it won’t resemble the 2001 Dot.com burst either. Rather, it will be more like the Housing Crisis in 2008. I don’t think the equity market (overvalued as it is) will be the cause of the burst, but it will be the victim of it. The cause, as so often, lies elsewhere.