Navigating global uncertainty: AI risks, the US Dollar pivot, and geopolitical conflicts in 2027

You commented last year that the fixed income landscape is more uncertain than at any time in your career as much of the post-WWII global architecture is being revised, do you still feel that way or have things become clearer?

Have things gotten any less uncertain in the past year? I would have to say no. Let's think about it.

We have had an on-and-off-again trade war with various countries, and we are apparently on again with Canada just this week. We have had a war with Iran begin six months ago, and we do not seem to have a plan for how to end it. Oil prices are up 40% around the world from the beginning of the year. Finally, US Treasury yields at the long end are now at their highest levels in about 19 years.

Furthermore, most of the energy in the economy seems to be involved in AI investment, and nobody knows how that is going to end. So, I would have to say no, things are not more certain.

You have said that a hiccup in the AI narrative is the biggest risk to the US dollar. Can you take us through the potential consequences to the US and global economy and fixed income, if this occurs and what might cause it?

What could go wrong with the AI investment boom?

I think the thing to watch is revenue for the specific AI products. We are watching two segments. We are watching retail, where we think it is going to be low-end and commoditised, but very broadly employed. We are also watching enterprise uses. These tend to be higher value-added, higher-end, and much more specific applications.

Our concerns are at the retail level: we do not see a moat, and we see it as very competitive, so we are not sure about the profitability story there. We are more optimistic about agent AI in enterprises, but again, it is going to depend on return on investment. If there are disappointments, then some of the first movers may not see the profitability that the market currently anticipates.

What would it look like and feel like?

Well, it would be double trouble. Number one, much of the CapEx right now is into data centre warehouse investments. Those might stop on a dime if people revisit their thesis.

At the same time, given how important AI and large-cap tech generally are now to investors' portfolios, we could see a bear market that would give us a negative wealth event measured in the trillions of dollars, possibly tens of trillions of dollars. That could cause a sudden stop in parts of the consumer sector.

On balance, what we are talking about is an event where both companies and individuals get very cautious at exactly the same time. That is a recession risk. In that world, it is going to be very interesting to see how not just the bond markets, but the currency markets react.

Our general view is that it would be very bearish for the US dollar, and non-dollar bonds would be a particularly useful hedge and diversification alternative.

You also noted the recent positive correlation between the US dollar and US equities. For wealth managers and financial advisors managing balanced portfolios, the historical negative correlation between the USD and risk assets has been a portfolio stabiliser. If this correlation is changing, how do you think portfolio allocators should react?

To continue with our "what happens if AI goes wrong" thesis, I would argue that for much of the past 40 years, the US dollar has been a safety asset. It is a place where people go when things get uncertain.

But when we look at how much foreign investment has gone into the US AI story and into US equities, we are impressed. Foreign capital flows have been bigger into equities than into bonds. That means, in many ways, the US is now the risk asset, not the safety asset.

In that case, an equity bear market, which would certainly be a risk event, would drive the dollar lower, not higher. There is a lot of foreign capital that would either need to sell or hedge. I think if you are losing money in a foreign stock market, the natural tendency is to either bring money home or to take out the currency risk of that position. Both of those would be dollar-bearish. It does not really matter which one they do; either one would tend to strengthen foreign currencies and weaken the US dollar.

The Iran conflict rumbles on without a definitive end, how destructive has this conflict been for the global economy and how have your asset allocations shifted from start to now?

Collectively, we have been surprised at the resilience of the global economy and how little oil prices have moved. They have moved—international oil prices are up 40% from the beginning of the year—but the doomsday scenarios have not taken place, and the world seems to be coping with oil in the $90-a-barrel range. They are coping better than we would have expected, so that has been a surprise.

What we did in the portfolios was to immediately de-risk on the currencies of the oil importers, and we have carried more dollar exposure than we might have, given our scepticism about the ending of the AI story. But the US is an oil exporter, so as long as these prices hold, it actually tends to hold the dollar up rather than push it down.

Clients have been asking us when they should consider adding duration to their portfolios. Are you adding duration in any country bonds right now, and if not what might prompt you to do so?

I have mentioned that interest rates at 10, 20, and 30-year maturities are now at 18, 19, and 20-year highs. In Japan, they are at more like 40-year highs. This has reached our valuation levels and we have begun to add duration, which has been modest and gradual.

I freely confess we probably added Japanese duration too soon. But in the US market, now that 10-year Treasuries are above 4.7% on any given day, we think that is fair value. So, we are perfectly willing to hold US duration, Euro duration, and Yen duration. We also have added duration to much smaller markets where yields are particularly attractive, ranging from very high-quality New Zealand debt, through to Brazil and certain other selected emerging markets.

What would prompt us to add more duration? I can think of two things.

Number one is valuation. If yields continue to rise, we will probably continue to get longer because that adds coupon income to the portfolio. Ultimately, it also adds exposure to capital gains if we in fact get an AI disappointment or any other kind of growth disappointment, which would tend to rally global bonds generally.

What is the one thing you think investors should be keeping in mind going into 2027?

One thing I think investors should keep in mind going into 2027 is that the future is not predictable. It may be less predictable than ever.

In a world of unpredictability, diversification is your friend. If you do not have high conviction in any particular theme or factor, it means you want to own a little bit of everything. That absolutely includes fixed income, it includes foreign country exposures, and it absolutely includes bonds from other countries and currencies.