AI, War, Debt and Disorder

Why the collision of an oil shock and an AI boom will push the US further along the road to fiscal dominance

Photo by UX Gun on Unsplash

Interested in the portfolio implications?

I will be delving into this topic and discussing the portfolio implications in a live webinar with Newstalk presenter Shane Coleman on Wednesday May 6th. Register here: Link

With equities bouncing sharply in April to hit all-time highs, many investors have been left wondering whether the market “knows more”, seeing through to an end of the conflict, or is simply ignoring the trouble ahead.

The real issue is that markets are being driven by two profound structural forces: the oil shock from the conflict in Iran and the productivity shock from AI.

February saw a widely read research report raise fears that AI would be transformative, particularly for white collar workers, pushing bond yields and software stocks down. March saw the conflict in Iran begin, pushing oil prices up, bond yields up and equities down. While the conflict rumbled on through April, hopes of a resolution coupled with a switch back to a focus on AI saw equities bounce back.

In simple terms the war in Iran represents an adverse supply shock, while AI represents a positive supply shock. But that oversimplifies the matter. Both shocks have significant impacts on demand, both create winners and losers, and both create dilemmas for policy makers.

What’s more, although we have seen both types of shocks in the past, they have rarely occurred concurrently. And this time they are unfolding in the midst of a changed macro regime, a regime already characterised by high and rising debt, supply constraints and a political bias towards more fiscal spending.

The big point for investors is not necessarily which shock dominates. But that both forces will likely push the US further along the path to fiscal dominance and the world more broadly towards greater macro volatility.

The Adverse Supply Shock

The oil shock from the conflict in Iran is the worst kind of shock for markets. In economic terms it is an adverse supply shock. The cost of production rises, translating into lower economic growth and higher prices - stagflation.

It creates a dilemma for policy makers, particularly the Fed, which is mandated to achieve maximum employment and price stability. Should it raise interest rates in response to higher inflation and risk accentuate a downturn, or cut rates in response to weaker growth and risk adding to inflationary pressures?

Central bankers have struggled with this dilemma before. In 2008 the ECB raised interest rates in May when oil prices spiked, only to have to cut rates later that year when the global financial crisis took hold.

The key risk for central bankers is inflation expectations. If expectations become de-anchored, the risk of higher inflation becoming embedded in the system becomes too great and they are forced to prioritise inflation over economic growth, regardless of the consequences for output.

AI: The Positive Supply Shock

AI, on the other hand, is a positive supply shock. As new Fed Chair Kevin Warsh said last year, “it makes everything cheaper”. In economic terms it represents an outward shift in the aggregate supply curve. Because it boosts productivity, overall output expands while unit labour costs decline, putting downward pressure on inflation. It produces disinflationary growth.

While stagflation is a nightmare for central bankers, disinflationary growth is a gift. It was this backdrop that allowed Alan Greenspan to hold off on raising interest rates in 1996 when the economy was gaining momentum but strong productivity growth was keeping inflation in check. Kevin Warsh and parts of the current administration have pointed to this precedent as supporting the case for lower rates today.

It Is Not That Simple

In theory the recent market gyrations reflect markets weighing the positive supply shock of AI, which dominated in February and April, against the adverse impact of the oil shock, which dominated in March.

But that oversimplifies matters considerably.

Both shocks have important demand dimensions that are easy to overlook. Higher oil prices raise gas and energy costs, reducing real disposable incomes and weighing on consumer spending. At the same time higher oil prices make a growing number of oil rigs commercially viable, potentially boosting spending on drilling and refining.

AI raises supply but also critically impacts demand. While there is enormous optimism about AI’s potential impact on productivity, to date it remains largely that: potential. Where the impact is already visible is in spending on chips, data centres and AI infrastructure. That has been a key driver of the equity rally in April.

But recently markets have taken a more questioning view on the likely return on investment from the large capital expenditure programmes at the likes of Meta in particular. Markets are assuming this capex splurge will last for years, but for that to materialise a material return on investment will need to be evident

Both shocks also have significant distributional consequences. Higher oil prices disproportionately hit lower income households, for whom spending on gas and energy represents a higher share of total expenditure. At the same time wealthy asset owners may benefit from exposure to high growth technology stocks. Both forces accentuate the K shaped nature of the US economy.

Debt, Disorder and the Macro Regime

Critically, both shocks are unfolding in a macro environment that is fundamentally different from the last time the world faced either.

The 1970s oil shock arrived before the era of inflation targeting and with debt levels a fraction of what they are today. The productivity boom of the 1990s unfolded against a backdrop of fiscal surpluses, contained geopolitical risk and a Fed whose independence was unquestioned.

None of those conditions apply now. The US is running a deficit of 6 to 7% of GDP with a debt to GDP ratio above 100%. The political zeitgeist, across the spectrum, is biased towards spending more not less.

Whether the priority is fiscal support for workers displaced by AI, military spending driven by geopolitical disorder, or energy investment prompted by supply insecurity, the pressure on the public finances points in one direction.

That matters because it constrains the policy response to whichever shock dominates. If the oil shock proves persistent and inflation expectations rise, the Fed faces the prospect of raising rates and debt servicing costs.

If AI proves as disinflationary as Warsh believes and the Fed cuts rates in response, it risks inflating an already elevated equity market and storing up a bust that the policy toolkit may no longer be able to contain. Household equity holdings now stand at around $40 trillion. The wealth effect works in both directions.

And it’s not just a matter of one shock dominating the other, they could conceivably interact in ways that could amplify the macroeconomic volatility.

For example, if the conflict in Iran rumbles on, and oil prices remain high, the immediate impact could be stagflationary. But if weaker economic growth then precipitates accelerated adoption of AI, lower hiring and job losses stagflation could quickly shift to being deflationary. That would likely necessitate an even more aggressive policy response both on the monetary and the fiscal side.

The result is a wider distribution of outcomes, a higher risk of policy error and a regime defined less by stable disinflation and more by macro volatility.

Some asset classes are already responding to this shift. Bond markets, commodities and gold have begun to reflect a world of higher macro uncertainty and fiscal pressure. Equities, for now, remain more anchored to the AI-driven growth narrative.

The key question is not whether AI offsets the oil shock or vice versa. It is whether policy can respond effectively in a world AI, of high debt, political pressure and competing shocks.

That is far less certain.

Interested in the portfolio implications?

I will be delving into this topic and discussing the portfolio implications in a live webinar with Newstalk presenter Shane Coleman on Wednesday May 6th.

Register here: Link