Macro Reality Check: Markets, Rates and the ‘Unpriceable’ World

That world has gone

What we are witnessing is a return to something that economic historians would recognise as neo-mercantilism: the use of trade policy, industrial strategy, and geopolitical leverage to advance national interests at the expense of multilateral cooperation.

The numbers are stark.

US effective tariff rates have risen to their highest level since the 1930s, representing a structural change to the framework that governed global trade for nearly eighty years.

And in my view, this is not a one-administration phenomenon. Quite apart from US domestic political pressures, China’s ongoing balance sheet downturn is putting more and more emphasis on exports as the release for huge levels of industrial over-capacity. In other words, large scale changes to the global trading system are the result of both the US and China’s policies.

To put it in context, Chinese export prices have been cut by around 15ppts since 2022. The examples of Chinese exports undercutting western industries in recent years are there to see from the EV market to renewable energy to the chemicals industry.

This is something that will continue to stoke protectionist sentiment outside of the US and fragment the global trading system further.

The structural adjustment of global supply chains is still in its early stages. Reshoring, friendshoring, and strategic inventory-building are real trends. But they will eventually come at a cost. They sacrifice comparative advantage. They mean more domestic production of things that other countries could produce more cheaply. This is not a transient cost — it is a permanent, structural addition to the cost base of the global economy.

The macroeconomic consequence — and this is what matters most for asset allocators — is that we now live in a world of more frequent and more persistent adverse supply shocks. And adverse supply shocks are the most difficult environment for economic policy, because they force a brutal choice: accept higher inflation, or accept lower growth.

Look at recent history. The energy price shock of 2020 to 2022 added approximately 5 percentage points to core US inflation — far more than models predicted, because loose fiscal and monetary policy allowed firms to pass through cost increases rapidly.

The pandemic, the Russia-Ukraine conflict, the Middle East crisis: each has generated a new supply shock. And the geopolitical environment that produces these shocks has not improved. It has deteriorated.

This is likely to mean inflation will be more volatile, not just higher on average. The risk distribution is not symmetric: adverse supply shocks push prices up sharply; demand destruction is the only mechanism that brings them back down, and it is slow and painful.

That is likely to make it harder for central banks to credibly commit to stable inflation, which has direct implications for the risk premia demanded by bond investors.

Of course, supply chains are not the only reason we have seen such volatile inflation in recent years – macroeconomic policy has played a crucial role. Specifically, overly loose monetary and fiscal policy.

Which brings me to the second structural shift – fiscal activism.

For most of the post-GFC era, the dominant policy framework was one of fiscal restraint paired with monetary accommodation. Central banks were the first and last line of defence. Governments, constrained by austerity politics and debt sustainability concerns, largely kept their powder dry. Monetary policy was everything.

That framework has inverted

Fiscal policy is now the active policy tool. Governments across the developed world have rediscovered spending as a political lever, and the Middle East conflict has further loosened fiscal norms, generating a larger global fiscal impulse than many anticipated, with China’s fiscal stimulus roughly doubling relative to pre-conflict assumptions.

To give you a sense of how big an influence fiscal policy has become on the economy, unanticipated movements in fiscal policy – so changes to the budget that come in year – are responsible for about 40% of economists GDP growth forecast errors. That is huge.

In the United States, we expect a 2026 fiscal deficit of $2.8 trillion — the largest since 2021 — driven by the full impact of tax cuts, declining tariff receipts, and elevated spending.

At 8.6% of GDP, that is a deficit of extraordinary magnitude for an economy growing broadly in line with trend and an unemployment rate of 4.3%.

The question I consistently hear from our asset management clients across the world is whether the US is heading to a fiscal crisis. Their argument goes that if there are no “adults in the room”, then does fiscal discipline have to be enforced by the market?

I am relatively sanguine about US fiscal position over the next five years. The risk of a funding crisis is low thanks to the dollar’s reserve currency status and the insatiable appetite for US assets, especially as AI dominates investors’ thinking.

In any case we expect political constraints will likely force some consolidation as Social Security trust fund depletion approaches in the early 2030s.

But — and this is an important qualification — “not a solvency crisis” is not the same as “no consequences.”

This fiscal trajectory is likely to reinforce the inflation volatility I just mentioned. Fiscal activism in the face of adverse supply shocks is inflationary. When governments run large deficits in tight labour markets, it can add demand pressure that monetary policy must then counter. The result is a policy mix that is structurally less effective at stabilising inflation.

As this chart shows, the twenty years pre-pandemic were the aberration – with abnormally low and stable inflation compared to history. More active fiscal policy and weak global supply chains imply that the future may look more “normal”.

At the same time, fiscal activism contributes to a permanent repricing of fiscal risk in bond markets. Even absent a crisis, investors are rationally demanding greater compensation for holding duration in an environment of large and persistent deficits. This manifests as structurally higher term premia — a theme I will return to later.

The third structural shift we see is the one that generates the most excitement — and, in my view, the most analytical confusion. The United States is experiencing a genuine AI-driven investment boom but I think there are a few key elements that are misunderstood.

For example, the growth contribution of AI is real but remarkably narrow in its transmission mechanism.

Oxford Economics estimates AI-related investment added approximately 0.5 percentage points to US GDP growth in 2025. That much less than many assume due to the high import content of AI spending. It is not broad-based productivity improvement — not yet. It is running primarily through one channel: the wealth effect on top-end consumers.

The mechanism is straightforward. AI investment drives technology stock valuations higher. Higher equity prices increase the net worth of households with significant financial assets — overwhelmingly higher-income households.

We estimate that every sustained dollar change in stock market wealth generates approximately a five-cent change in consumer spending. That sounds modest, but across tens of trillions of dollars of equity market capitalisation, it is a very large number.

The implication is uncomfortable: US growth is, to a significant degree, a function of the AI equity story holding together. Strip out the wealth effect on the top quartile of earners, and aggregate US consumer spending growth would have been around a percentage point weaker last year.

At the same time, lower-income households face the cumulative real income squeeze from years of high inflation as well as a softer labour market. The aggregate growth numbers mask this divergence.

As a result, the bubble question deserves to be taken seriously. My Alpine Macro colleagues do not currently characterise AI equity valuations as a classic bubble – today’s market, unlike the late 1990s, remains sensitive to profitability and capex plans. Furthermore, elevated inflation and bond yields have so far prevented valuations from getting out of control. But the concentration of market gains in a small number of mega-cap technology names creates a specific and significant tail risk.

Oxford Economics has modelled the scenario explicitly: a 25% fall in US technology stocks would be sufficient to drag global growth down to 2.1% in 2027 — 0.9 percentage points below the baseline.

That is a meaningful global slowdown.

But perhaps the most interesting misconception is around the inflationary impacts of AI. In the near term, AI investment is inflationary. The construction of data centres at scale is tightening labour markets and pushing up producer prices for electronic components in ways that are reminiscent of peak pandemic supply disruptions. The demand for power, cooling infrastructure, and specialist engineering skills is running well ahead of available supply.

But in the long run, AI could be profoundly disinflationary — and this is where the analytical picture becomes genuinely complex. If AI delivers on its productivity promise, it allows the economy to grow faster without generating excess inflation. It could dampen unit labour cost growth and promote disinflation in service prices, which have been the most stubborn component of recent inflation.

Oxford Economics estimates potential US productivity gains of around 3% over a decade from AI adoption — meaningful, but not transformative on a short horizon, and critically dependent on assumptions about adoption rates and diffusion that are highly speculative.

The historical analogy is instructive: productivity gains from previous general-purpose technologies – electricity, computing – took decades to fully materialise. We are, at best, in the early innings. For the next several years, the inflationary pressures from AI investment are likely to dominate. The disinflationary dividend, if it comes at all at scale, is more of a 2030s story.

For asset allocators, this creates a genuinely difficult positioning problem. You are being asked to hold duration in an environment where near-term AI investment is inflationary, while also being told that long-run AI productivity could justify lower nominal rates eventually. The honest answer is that the uncertainty band around that long-run outcome is very wide — and you should not be positioned for the optimistic scenario alone.

So what does all this mean for asset allocation? Let me draw out some implications that I believe are most directly relevant to this room.

First, term premia. Term premia are structurally higher across global bond markets than at any point in the last decade. Why? The era of secular stagnation and close to zero interest rates is over. Fiscal deterioration, inflation expectations that never fully re-anchored, higher frequency of supply shocks, and ongoing QT are keeping them elevated. These aren’t transient pressures. Duration risk has been permanently repriced and treating long-end sovereign exposure as a reliable hedge to risky asset returns is no longer defensible, at least not under non-recession scenarios.

Second, equity markets. The AI boom is real – we think we are past the point of questioning whether A.I. spend will generate a positive return on investment – it’s apparent the need for compute is structural, and the benefits of the infrastructure build-out will accrue to a broader set of companies. Physical infrastructure constraints – power, chips, data centres – represent a genuine bottleneck that markets aren’t fully pricing. At the same time, it isn’t a story of broad market strength. A market-cap-weighted allocation simply doesn’t provide meaningful diversification when concentration is this extreme.

Third – and perhaps most consequential for this room – the stock-bond correlation has flipped. For most of the last decade, the negative correlation between equities and bonds was the foundation of the balanced portfolio. That relationship has broken down. In a supply-shock world, both assets come under pressure simultaneously, and the diversification mechanism fails to deliver. Gold and energy stand out as genuine alternatives — but diversification is episodic, and they lack the absorption capacity institutional investors demand.

So let me pull this together.

My thesis this morning is that markets have, to their credit, adapted to the surface-level volatility of recent years remarkably well. Equities have absorbed multiple shocks and still sit near record highs. Credit markets have remained functional. The financial system has proved considerably more resilient than many feared in the darkest moments of the pandemic and the inflation surge.

But beneath the surface, three structural shifts are reshaping the macro regime in ways that I think are still incompletely priced. Neo-mercantilism means more frequent and more persistent supply shocks. Fiscal activism means a more volatile inflation environment and structurally higher risk premia on government bonds. And the AI boom is both a cyclical phenomenon and a structural shift.

Together, these shifts define a world that is genuinely harder to model, harder to hedge, and harder to price. The uncertainty is not merely cyclical. It is structural.

For insurers, I think the honest response is not to pretend that traditional frameworks still work without modification, and not to be paralysed by what we cannot know. It is to take seriously the possibility that the regime has changed — and to build portfolios that are resilient to a wider range of outcomes than a single-scenario base case would imply.

The world has become, in important respects, less predictable. Our job is not to eliminate that uncertainty — it cannot be eliminated. Our job is to be appropriately humble about what we know, appropriately rigorous about what the data tells us, and appropriately nimble when the regime shifts again.

Thank you.