Small-caps, oil, and the AI elephant in the room: Asset allocation insights

The Asset Allocation Committee remains predisposed to buy the next equity drawdown.

Our four main active risk contributors—equity overweight, short duration posture, tilt to small-caps over large-caps, and real assets exposure—have all added value over the last year. We are now assessing how much of that risk we want to keep.

Small-caps generated the most debate at our latest meeting. The overweight has worked, but earnings support is fading.

There was no consensus on whether to trim the position. One member framed three scenarios:

• A recession would be difficult for small-caps.

• A continued unwind in AI-related mega-cap stocks could favor small-caps and value stocks.

• A benign growth environment could leave relative performance closer to even.

Another member challenged the idea that smaller companies provide defense: “If we want to play defense, small-caps are a poor place to hide.”

A few members think large-cap growth is becoming more interesting. Several mega-cap AI companies still have strong earnings growth, albeit at decelerating rates, while their multiples have compressed.

A member described the AI build-out as a regime shift.

“The post-GFC secular stagnation regime is over. AI is a game changer. It’s going to be a long cycle, with ups and downs, but ultimately, we’ll have higher real growth and higher interest rates,” he argued.

The (existential?) elephant in the room

Of course, we couldn’t ignore the recent public discourse¹ around AI’s societal impacts.

Discussions are growing around the pace of frontier model development, competition from open-weight developers, and data center moratoriums. But AI-related capex continues to move full steam ahead, with increasingly advanced agentic AI systems providing another source of support.²

Bottom-up insights from our research platform show a widening range of conditions across AI infrastructure. Some providers are ramping up production. Others still can’t meet order volumes. Bottlenecks are migrating along the supply chain, and not all will prove durable.

The implications for pricing power and margins are significant, arguing for greater selectivity within AI-related areas—not a risk-off approach.

The oil debate

Energy is the other big wild card. One committee member outlined two binary scenarios for oil.

Continued geopolitical disruption and refinery outages could keep oil and diesel prices high, sustaining inflation pressure and pushing rates higher. Conversely, a de-escalation scenario could bring supply back quickly.

Oil markets could emerge from the Iran war with a larger surplus than before the conflict. OPEC has weakened considerably,³ Canada has actively increased production in recent years,⁴ there’s a concerted effort to boost oil output in Venezuela,⁵ and U.S. production shows no signs of slowing down.⁶

There are several “ifs” embedded in that scenario, including whether lower demand in China reflects genuine demand destruction rather than inventory drawdowns. But if the pieces fall into place, oil could eventually fall sharply and improve the inflation and interest rate backdrop.

The relationship between oil prices and the 10-year Treasury yield is particularly striking. For roughly the past month, the correlation between the two has been around 80%, placing it in the 99th percentile on a historical basis.⁷

Markets may be underappreciating this dynamic.

What we’re watching

The committee reviewed analysis from the firm’s Multi-Asset team showing that buying 10%, 15%, and 20% declines has typically paid off,⁸ but not when the sell-offs developed into extended bear markets.

It’s often said that stocks take the stairs up and the elevator down. But our analysis suggests that a “stairs down” scenario may be more worrisome.

In addition to the speed of the drawdown, the team identified several other conditions that can help distinguish a contained sell-off from a deeper bear market, including:

• The trailing 12-month market return

• Credit conditions

• Inflation surprises

Fast sell-offs that occur after a rising market, with credit still available and monetary policy able to respond, have tended to create better buying opportunities. Conventional bear markets, on the other hand, have usually unfolded more slowly, alongside tightening financial conditions or inflation that limits the policy response.

The bottom line

We left the meeting with our core positions intact and greater clarity on how we would respond to a sell-off.

The committee remains constructive on equities and is prepared to add on weakness if the drawdown arrives with the right signals. AI could keep the growth cycle running longer than expected, while oil could either reinforce the higher-for-longer interest rate regime or relieve it quickly.

For now, the positions stay on, and the buy-the-dip checklist stays close.

One more thing…

At the end of the meeting, as we were closing our laptops, I asked if anyone had anything they wanted to bring up.

“The U.S. midterm elections,”⁹ a member said.

The meeting ran another 15 minutes.