How is AI transforming the macro outlook? Three potential scenarios

Potential paths ahead

Assessing the macro implications of AI is exceptionally challenging, given the theme’s many moving parts and uncertain destination. What we know for certain is that AI has the potential to shift the macro regime because it reshapes the outlook for fundamental concepts typically assumed to be fixed, such as productivity, the capital-labour ratio, saving behaviour and government tax policies. These are all core, long-term determinants of asset valuations.

Given the number of moving parts and the rapid pace of technological development, several potential scenarios are plausible.

At this stage, however, one of three broad paths appears likely: a capital supercycle, a boom-bust cycle or a messy transition, with the latter seemingly the path markets are currently following.

Scenario 1: a capital supercycle
If AI drives up aggregate productivity without universally compressing margins, we are likely to see the economy grow in aggregate. This higher growth should offset the job displacement caused by AI, keeping structural unemployment largely stable.

The inflation implications of AI remain uncertain. The price of some services should fall, but infrastructure bottlenecks — for instance, in commodities — could generate inflation. It is unclear how this relative price shock between different sectors and between goods and services will ultimately play out, but if monetary conditions remain broadly accommodative, it could result in a more favourable inflation-growth trade-off.

In such a supercycle, capital would be put to work through increased investment rather than being channelled into savings, pushing up real rates. Higher productivity, stronger demand for capital goods and higher returns on the invested capital would also point to higher real yields. The extent of any increase would depend on a range of unknowns. Not least is the degree to which AI lifts productivity and potential growth, as well as how far the associated investment proves genuinely additive.

Scenario 2: a boom-bust cycle
Alternatively, the AI-fuelled transition could be very positive for the near-term cycle as capex booms, but the destination could be strongly disinflationary, even recessionary, because of capital over-expenditure — a classic boom-bust cycle.

Central banks and markets understandably find it difficult to price supply-side shocks accurately. If AI proves to be transformative and productivity-enhancing for the economy overall, as appears likely, policy will need to be tightened to account for that growing productivity boost; otherwise, policy could become increasingly stimulative. In the absence of such adjustment, the risk of excess capacity would rise, potentially culminating in a recession or even deflation.

Scenario 3: a messy transition
A messy transition marked by a high level of uncertainty is another plausible scenario. Markets currently seem inclined to favour this scenario as concerns grow about how AI could upend many business models — and even entire sectors — that were previously viewed as relatively insulated. If this uncertainty persists, it could undermine assumptions about the viability of businesses and jobs and significantly reduce visibility about future cash flows, returns and even corporate lifespans.

This is an environment in which companies may freeze most new activity, with the possible exception of AI-related investment. The speed of associated job displacement could also outpace the ability of workers and organisations to adapt, pushing the unemployment rate higher. In that environment, precautionary savings may rise faster than net investment, driving real yields lower. Ironically, this would feel more like a risk-off environment, even if the underlying shock ultimately raises productivity.

Given the starting point of relative healthy corporate and household balance sheets, at least in aggregate, this path would most likely entail a shallow recession followed by a rebound that would resemble a boom.

Countervailing forces

The AI revolution is not happening in isolation. Several other important forces should be considered in any macro framework, notably:

1. Demographics and deglobalisation are rapidly lowering productivity and supply potential. It remains unclear whether the benefits of AI will accrue sufficiently quickly to offset their negative impact.
2. Income inequality and the concentration of wealth are already major political issues in many countries. If AI technology exacerbates these trends, there could be a policy response in the form of taxing the gains and raising general income thresholds, which would offset some of the productivity gains.

Which indicators to watch?

Markets may oscillate between the three paths outlined above, making it difficult to determine the ultimate direction. With macro data often viewed as less reliable by policymakers during periods of transition, signals may instead be found primarily in equity and credit markets. Specifically, we are monitoring:

• The relative performance of AI infrastructure versus AI application stocks, to help assess whether value is moving up the chain or is being compressed more broadly. Rising AI infrastructure valuations would point to a capital-deepening cycle. Conversely, a collapse in the valuations of AI application stocks would be a highly defensive signal.
• High-yield credit spreads to gauge whether the real economy is under stress.
• The performance of cyclicals relative to defensives to evaluate whether the market believes growth and capital spending will broaden across the real economy.
• Earnings revisions to assess whether AI is boosting profits and cash-flow expectations.
• Real yields as an indicator of whether the market believes AI will raise overall productivity or compress margins and be negative for the cycle. On balance, we think real yields will take their lead from the indicators above: yields are unlikely to rise if those signals are moving in the opposite direction.

Bringing it all together

Ultimately, we think that AI will be positive for the economy but, as we transition to a new macro regime, investors will face a period of near-permanent uncertainty on the eventual direction that the economy will follow. In particular, we have identified three potential paths with very different investment implications. Rather than making a definitive call on the end point of that journey, we believe it is more effective to track market indicators. These provide a clearer read on which macro regime the economy is heading towards and what that may mean for rates and risk assets.