Is society being 'ghosted' by GDP?

Is an AI-led doomsday nigh?!

US Treasury yields were little changed over the past week and continue to trade in a tight range. Last week’s SCOTUS ruling against Trump’s IEEPA tariffs didn’t have a material market impact, with this outcome having been widely expected. Moreover, the US administration has wasted no time in implementing new tariffs, in the wake of those which have been taken away.

From this perspective, there may not be much to see that materially impacts the broader economic environment. However, we feel that there is a risk that realised tariff revenue may fall somewhat short of what had been planned, and at a national level there could be winners and losers from these changes.

Meanwhile, it has done nothing for Trump’s presidential approval rating, which has fallen to its lowest since last year’s inauguration. With this standing at just 40% and with the mid-terms ahead, it does leave the administration looking for some wins to turn the tide on momentum.

In this respect, economic data has been broadly constructive. However, inasmuch as economic activity is being spurred by huge AI investment, this has led to the popularisation of a phrase – that we are witnessing ‘Ghost GDP’.  That is to say that growth stats look good, but this ‘GDP’ doesn’t seem to show up in anyone’s pocket these days. Moreover, insofar as this investment represents a real risk to jobs and incomes in the future, this means that any AI gains are as much to be feared as celebrated, within broader society.

Meanwhile, a positive outcome in foreign policy appears to be a stretch. A conflict with Iran could lead to further pain on US consumers via oil prices.

Meanwhile, there are few signs of breakthrough with respect to Russia/Ukraine, with the conflict now into a sad fifth year, as young men continue to sacrifice their lives on battlefields tragically reminiscent of the Somme and the First World War.

An article from Citrini Research added to market angst with respect to the prospects for AI disruption. In 2026, whilst we remain more in the ‘infrastructure phase’ with respect to the build out of datacentres and the infrastructure needed to service these, the AI thematic continues to be growth additive.

However, as we transition into the ‘adoption phase’, there are growing fears that the ensuing disruption could lead to a more rapid downshift in the labour market. This will dent consumption as firms losing pricing power and market share will be forced to intensify their investment in technology and prompt further job losses.

Although the economy may be able to adjust and recycle workers should such a shift occur gradually over a 10-year period, such is the speed of advance in recent AI tools that there seems a growing risk that this could happen more suddenly, in a manner that creates material disruption in its wake.

It is also interesting to reflect on past periods of technological advance, which meant that consumers were able to purchase more goods at a higher quality than had been the case before. Yet in the context of professional services such as law, insurance or wealth management, it is not clear that technological investment will lead to increased consumption in these areas. If anything, it is likely to drive prices lower.

Although this could leave more income that can be spent elsewhere, within professional services themselves, there is a sense of an inelastic demand curve. In this case much of the current investment may deliver a zero-sum game, whereby the only way that investment returns are justified is at the expense of losses to other established operators.

Such an outcome could suggest a more materially disinflationary outcome ahead and this thematic has continued to help support longer-dated Treasuries over the past week. This line of thinking continues to drive a search for AI winners and losers at a national level as well as a sectoral level.

From our point of view, a relative loser at a country level could be India. In many respects, we feel that it is jobs that have been outsourced that will be among the first to be replaced by AI, and this could represent a material downside for the Indian economy. By contrast, Korea is more likely to be an AI beneficiary and in this respect an FX trade long won, short rupee, is a way of playing this dispersion.

On a sectoral basis, we see European banks as better defended from AI disruption given modest multiples, steep yield curves, and regulatory protection. Meanwhile, we remain concerned with respect to large parts of the software market and continue to anticipate further trouble for BDCs as pain in private markets continues to build.

In the UK, the Labour Party suffered a humiliating defeat at the Gorton by-election. This outcome was largely expected, but the margin of defeat continues to heap pressure on Prime Minister Starmer. Nevertheless, gilts have continued their recent outperformance, helped by an improving inflation narrative, which may also, in turn, benefit the outlook for UK government finances.

With a large share of the UK debt inflation linked, UK borrowing costs are particularly sensitive to the inflation trajectory. Additionally, the UK fiscal picture was boosted by strong self-assessment tax receipts this January, which has also helped mitigate the deficit. In part these tax receipts could reflect asset disposals, triggering capital gains taxes in anticipation that these tax rates may continue to rise under Labour, and so they may be a one-off.

However, any crumbs of good news are welcome in the UK these days, and with UK yields remaining well above levels in other European markets, there is a sense that valuations remain attractive.

In Japan, the latest Tokyo CPI printed 1.6% year-over-year and will likely remain below the BoJ’s 2% policy target over the coming months. We continue to see disinflation tailwinds in place in Japan, on the reversal of base effects linked to rice prices, which have normalised following a strong harvest and the reduced impact of past yen depreciation. However, we expect wages to remain firm in this year’s Shunto and expect the BoJ to gradually normalise interest rate policy, with rates reaching a neutral policy stance at 1.5% in 2027.

This week also saw the appointment of two new BoJ members largely seen as ‘reflationists’ in terms of their policy biases. Broadly speaking, we would not read too much into whoever Takaichi is picking for the BoJ, in the same way we think markets were overly concerned with rest of Trump’s imprint on the Fed. We know Takaichi wants to maximise growth, and this makes her dovish on monetary policy.

However, the BoJ can be expected to react to data and incoming newsflow. We remain constructive on the Japanese economy, and this is why we think conditions will warrant higher interest rates. We also think there is reluctance towards further yen weakness beyond 160 versus the US dollar. We retain a neutral stance on the yen, bullish duration on long-dated JGBs. and we continue to be upbeat on Japanese stocks.

Credit markets remain focused on deciphering the impacts of AI, which has led to increased dispersion and some healthy spread widening in areas vulnerable to AI disruption, including software, business services, private credit, and insurers with private credit exposure.

In some areas, the market has behaved in a ‘shoot first, ask questions later’ manner, beginning to create interesting dislocations. Overall, market spreads remain close to their tights, and taking on significant beta risk still makes limited sense. However, there is a growing sense that dispersion levels will continue to rise from here.

Looking ahead

We remain content to keep risk at modest levels for the time being. At a high level, markets appear to be trading in a sideways pattern. Risk assets should also remain well supported as there is no evidence of a meaningful slowdown in any major economic zone and inflation remains subdued. Volatility in currency markets is restrained, and this continues to support the case for staying long carry. However, this masks some of the more extreme volatility under the surface. On this basis, it would not come as a big surprise to see a somewhat more extended risk-off move occur, before stocks establish a firmer base from which to rebuild upside momentum. This being the case, we would still rather wait to add on weakness, under no pressure to chase after ‘ghost’ returns.

^{* The information contained in this material is correct as of the publishing date of this article and is subject to change frequently.}