How will we know AI is delivering?
Welcome to Uncommon Truths, Paul Jackson and Andras Vig’s regular in-depth look at the big topics impacting markets.
As is usual when new technologies arrive, the benefits of AI have so far accrued largely to the enablers (model builders, data providers and semi-conductor manufacturers). Paul believes that at some stage we should see broader economic benefits (if this truly is a revolutionary technology).
Paul suggests that if AI does anything, it should boost productivity. However, productivity is cyclical, so it will be difficult to separate the underlying wheat from the cyclical chaff and we may not be aware of the gains until well after the fact. Paul believes that other effects would be disappointing job growth (jobless recoveries, say), lower inflation and wider profit margins. The period after the mid-1990s (technology boom) showed rising productivity growth and falling inflation and this sort of wedge in the future could be a sign that AI is working.
As for financial markets, the effect on bond yields could be mixed (higher growth could raise real yields, but lower inflation could reduce inflation breakevens). Paul suspects they may rise a bit. Equities should benefit from wider margins but the effect could be dampened if real bond yields rise. Finally, as the benefits of AI spread to users (rather than enablers), Paul expects the extreme concentration in US stock indices to wane.
Catch up on the last few editions:
FAQs
The optimal portfolios are theoretical and not real. We use optimisation processes to guide our allocations around “neutral” and within prescribed policy ranges based on our estimations of expected returns and using historical covariance information. This guides the allocation to global asset groups (equities, government bonds etc.), which is the most important level of decision. For Uncommon Truths, the optimal portfolios are constructed with a one-year horizon.
We’ve chosen to include equities, bonds (government, corporate investment grade and corporate high-yield), real estate investment trusts (REITs, to represent real estate), commodities and cash, on a global level. We use cross-asset correlations to decide which decisions are the most important.
Using a covariance matrix, based on monthly local currency total returns for the last five years, we run an optimisation process that maximises the Sharpe Ratio. Another version maximises Return subject to volatility not exceeding that of our Neutral Portfolio. The optimiser is based on the Markowitz model.
Related articles
Asset allocation Capital market assumptions | 2025
Invesco Solutions develops capital market assumptions (CMAs) that provide long-term estimates for the behaviour of major asset classes globally.
Asset allocation Applied philosophy – Are global equities in a typical mid-cycle phase?
Welcome to Applied philosophy, our view on global equity market model sector allocation.
Asset allocation Quarterly Global Asset Allocation Portfolio 2025 Outlook
Paul Jackson, Global Head of Asset Allocation Research for EMEA, discusses his insights on portfolio allocations and strategies for the 2025 outlook.
-
Investment risks
The value of investments and any income will fluctuate (this may partly be the result of exchange rate fluctuations) and investors may not get back the full amount invested.
Important information
Views and opinions are based on current market conditions and are subject to change.
This is marketing material and not financial advice. It is not intended as a recommendation to buy or sell any particular asset class, security or strategy. Regulatory requirements that require impartiality of investment/investment strategy recommendations are therefore not applicable nor are any prohibitions to trade before publication.