Video transcript
Speaker
Robert Sears – Chief Investment Officer, Rathbones
Transcript
Introduction: looking beneath the headlines
Robert: Hello, I'm Robert Sears, Chief Investment Officer at Rathbones.
Politics matters, particularly when it changes fiscal, trade or regulatory policy. But one lesson of recent weeks is that politics hasn't been the main driver of markets.
Markets have been far more focussed on earnings, inflation, bond yields and, more and more, whether the huge investment in artificial intelligence (AI) will earn an adequate return.
Beneath the headline indices, something more interesting has been happening. Market leadership has shifted sharply. Some of the biggest winners of the AI boom have struggled, while other parts of the equity market have done much better. Questions about how AI investment is financed have become more important. And government bonds have started to behave more like the diversifier investors once expected them to be.
I think these developments tell us something important. The market is becoming more discerning, and that's not necessarily a bad thing.
So today I want to look at three signals beneath the headlines. First, the broadening of equity markets. Second, the shift in AI from a technology story to a story about financing and returns. And third, whether bonds are starting to regain their role as portfolio protection.
One index, two markets
Let's start with equities. If you only look at the major indices, you can get a misleading picture of what's actually been happening.
October was a good example. The S&P 500 was slightly weaker over the month, but underneath that headline number, performance varied enormously. The equal-weighted S&P 500 rose by around 5%. That's the version of the index where every company counts equally, rather than the largest technology companies dominating. Semiconductor shares, meanwhile, fell by around 15%. And the parts of the US market outside the major AI enablers did considerably better.
So this wasn't simply a market sell-off. It was a change in leadership, and that difference matters.
For much of the past few years, investors have been rewarded for owning a fairly small group of very large companies linked to technology and AI. Real earnings growth has supported that concentration. But markets rarely stand still. As expectations rise in one part of the market, opportunities can open up elsewhere.
We've seen better performance from areas including financials, healthcare and energy, and from companies that were simply overlooked during the excitement around AI. That could be healthy. A broader market is ultimately a more resilient market.
It also reinforces something we've been saying for some time: diversification isn't just about owning more securities. It's about owning different sources of return. So for investors, the opportunity may be less about deciding whether to be 'in' or 'out' of equities, and more about being increasingly selective about where in equities they take risk.
AI: from spending to financing
That brings me to artificial intelligence. In my last update, I argued that the AI story was moving from silicon to balance sheets. Over the past few weeks, we've started to see what that means in practice.
Alphabet is a useful example. The company reported strong underlying growth. But investors focussed on how much it's spending on investment, the pressure that spending puts on free cash flow, and how much more investment may still be needed. The sharp fall in the share price was striking, because the results themselves weren't obviously poor. In effect, the market was raising the bar.
That tells us a lot about where we are in the AI cycle. Investors no longer reward spending just because it's linked to AI. More and more, they want evidence that the spending will earn an adequate return.
So the question is no longer just how much money will be spent on AI. We know the answer to that: an enormous amount. The more important questions are who pays for that spending, how heavily the infrastructure will be used and, ultimately, whether the revenue will justify the money invested.
This matters because AI isn't one investment. The largest technology platforms are still highly profitable businesses, with substantial cash flows and strong balance sheets. But around them is a growing ecosystem of data centres, specialist cloud providers, semiconductor companies, power infrastructure and financing vehicles. The economics of those businesses can be very different.
We're also seeing more of the AI build-out paid for through bond markets and other forms of borrowing. That doesn't necessarily mean there's a problem. But it does mean the cost and availability of money matter much more than they did in the first phase of the AI boom.
A company paying for AI investment from large existing cash flows can afford to be patient. A heavily indebted business that depends on constantly raising new money has far less room for error.
So the real dividing line in AI may not be between believers and sceptics. It may be between businesses that can pay for the investment and earn an attractive return from it, and those that can't.
The technology can succeed while the investment disappoints
This is why I think the debate about whether AI is a bubble can sometimes be unhelpful. AI could be truly transformative. I believe it probably will be. But a transformative technology doesn't guarantee attractive returns from every investment made to build it.
History shows this again and again. Railways transformed economies. The internet transformed the world. But in both cases, huge amounts of money were invested, and not everyone who provided it earned an attractive return.
The same applies today. More and more, the key test is utilisation. Are all these data centres, chips and networks being used enough? Are customers getting enough economic value from AI to keep paying for more computing power? And will the revenue be enough to justify the money being invested?
These questions matter more as financing moves beyond the balance sheets of the biggest technology companies. That's why we're watching credit markets as well as equity markets. Equity markets tell us what investors expect companies to earn. Credit markets tell us whether investors are still willing to lend against those expectations.
What could turn the AI cycle?
This also changes how we think about risk. The obvious risk is that the technology disappoints. I don't think that's necessarily the most important one.
The AI investment cycle could slow even if the technology keeps improving. Past investment booms have often turned not because the technology stopped working, but because investors started questioning the economics of the build-out. That's why we're watching how fast AI investment is growing, not just how big it is.
If expected returns start to disappoint, companies may slow investment. Credit spreads could widen, borrowing could become more expensive and suppliers would then see fewer orders. Because AI investment now contributes a growing share of economic activity and market confidence, that could start to feed into the wider economy.
There's also a wealth effect. If a big fall in technology shares hit household wealth and confidence, spending could weaken just as companies were cutting investment.
So the risk isn't only that a recession causes markets to fall. A large enough market correction could itself help cause an economic slowdown. That's one of the signals we're watching most closely.
Bonds are auditioning for their old job
That brings us to bonds. The past five years have been very difficult for bond investors. The move from near-zero interest rates to much higher yields led to heavy losses, particularly in longer-dated government bonds.
But there's an important irony here. The same change that made bonds so painful has also made them more useful. Higher yields mean investors are once again being paid meaningful income to own high-quality government bonds. And during some recent bouts of worry about economic growth and AI investment, government bonds have started to rise when equities have fallen.
That's important. It suggests bonds may once again be able to provide some of the portfolio protection investors have traditionally expected from them.
But I'd add an important caveat. Bonds are conditional insurance, and how they behave depends on the type of shock. If the economy weakens, inflation falls and central banks can cut interest rates, government bonds should offer useful protection. But if the shock comes from inflation, energy prices or worries about government finances, bonds and equities can still fall together. We saw that very clearly in 2022.
So the lesson isn't that the traditional equity and bond portfolio has suddenly returned to normal. It's that higher starting yields have brought back an important potential source of diversification, one that was largely missing when yields were close to zero.
Different shocks need different protection
That leads to a broader point about portfolios. No single asset protects against every possible outcome.
Government bonds can protect against a conventional recession. Inflation-linked bonds and real assets can help when inflation is the problem. Commodities can protect against certain supply shocks. And within equities, broader exposure can reduce reliance on a small group of companies or a single investment theme.
The aim isn't to predict exactly which shock comes next. It's to build portfolios that don't depend on one particular outcome. That's especially important today, because the range of possible outcomes is still unusually wide.
AI could deliver big productivity gains. The investment cycle could spread into the rest of the economy. Or financing conditions could tighten and expose weaker parts of the AI ecosystem. Inflation could keep easing. Or pressure from energy prices and government finances could keep interest rates higher for longer. We should be prepared for more than one of these outcomes.
What does this mean for portfolios?
I'd draw three conclusions.
First, don't mistake a change in market leadership for the end of the equity cycle. The recent broadening of returns may create opportunities beyond the companies that have dominated markets in recent years. That strengthens the case for active management, discipline on valuations and broader equity exposure.
Second, stay invested in AI, but be more selective. We still believe AI is a powerful long-term investment theme. But we increasingly want to separate companies with strong balance sheets, real competitive advantages and clear cash generation from businesses whose economics depend on endless cheap money and ever-rising investment.
Third, bonds deserve another look. After one of the toughest periods for fixed income in decades, higher yields have brought back meaningful income and some potential portfolio protection. But bonds should be one part of a broadly diversified portfolio. We shouldn't expect them to protect against every possible shock.
Closing remarks: the market beneath the market
In investing, it's always tempting to focus on the headline. Who won the election? Is the S&P 500 up or down? Is AI a bubble or a revolution? Are interest rates going up or down?
But over the past few weeks, the more important signals have been beneath those headlines. Politics has mattered, but earnings and the cost of money have mattered more. Equity leadership has broadened. The AI debate is moving from spending to financing, utilisation and returns. And higher yields are starting to bring back some of the defensive qualities of government bonds.
None of this tells us exactly where markets will go next. But it does suggest the investment environment is becoming more discerning. That should reward portfolios that are diversified, careful about valuations and selective about where they take risk.
As I said in my last update: the forecast was never the plan. The plan is to build portfolios that can take part when markets rise, but are resilient enough to cope when leadership changes or the economy turns.
That means looking beyond the index, beyond the political headline, and focussing on the market beneath the market. Thank you.
