Top of Mind - Optimism and worry in Bangkok

Introduction

I spent the last week at the Annual Meetings of the IMF and World Bank in Bangkok. On the sidelines of these meetings, there were also conferences organised by the International Institute of Finance and World Economic Forum, while I also attended meetings with clients. This array of events and meetings gave me the opportunity to participate in thought-provoking discussions with central bankers, other policymakers, institutional investors, CFOs as well as economists working in many different sectors. In this short note, I will set out my main takeaways from these meetings.

The word ‘resilience’ was said a million times

The ability of the global economy to withstand shocks was a key theme of the meetings. This was underlined by the IMF’s world economic forecasts that have underestimated global growth three years in a row in the face of negative shocks. In the previous years, shocks had generally led to growth being overestimated. The reasons mentioned for this resilience was (1) that the scale of the shocks were generally manageable in terms of size (2) the impressive ability of the various sectors to adjust (3) the offsetting impact of powerful structural capex cycles, not least AI. The overwhelming consensus at these meetings was that the global economy would remain robust over the next year, but there was a range of worries expressed about the downside risks to this view, from the sustainability of the AI capex cycle, to geopolitics, rising interest rates and sovereign debt.

AI momentum strong, but majority see capex bubble

There was a strong consensus that AI capex would strengthen once again significantly next year (from around USD 800bn this year to a trillion or above next year). Even though in terms of contribution to economic growth, this year likely represented the peak (essentially because the growth rate between 2025 and 2026 was exceptional – at around 100% and this would not be quite as strong between 2026 and 2027). The confidence in AI capex for the coming year was based on the momentum in demand compared to supply expressed in numerous AI company reports suggesting that this situation would likely extend even beyond next year.

Despite this optimism, almost every one (that I spoke to or listened to at least) seemed to agree that the scale of capex did not justify any reasonable expectation for the revenues that would be generated. This is a point we have also made before (see here), that even if you are an AI productivity optimist, productivity growth would really need to be spectacular to justify what is looking like the biggest innovation-related capex cycle in history. To put it in context, the AI capex cycle looks like being 3-times the size of the ICT capex cycle relative to GDP. It was ironic that as these discussions were taking place, markets started to wobble on investor fears that the sell-off in data centre bonds might be a canary in the coal mine for the AI capex boom more widely.

Geopolitics may now be an underrated risk

Although everyone talked a lot about geopolitical risk, this did not seem to be incorporated in any meaningful way in base line scenarios. The most concrete risk at the moment reflects developments in the Middle East. There was consensus that there would be no US-Iran deal: ‘I have no clue how this will end’ was a common statement. Most forecasters seemed to be incorporating a profile for energy prices in line with futures markets (which point to a gradual decline) and taken together with the declining energy-intensity of GDP, this meant the impact on baseline forecasts was relatively limited. To be fair, this is exactly in line with the way this is incorporated into our own views on the economy. Having said that, this consensus does make me wonder whether we might all be underrating geopolitical risks. For starters, energy inventories are low, so any further significant disruption to ME energy supplies might lead to a much bigger price shock. Second, the shock in terms of crude oil so far is modest, but it is already much more sizeable in terms of refined products and natural gas prices. Meanwhile, all kinds of other geopolitical risks could have more meaningful impacts: a potential Eurozone-China trade war or a Russian aggression towards NATO countries. Could geopolitical risk be like ‘the boy who cried wolf?’

A gradual rate hike cycle seems to be on the cards

I was fortunate enough to be in the room for the (separate) fireside chats of Banque de France President Emmanuel Moulin and Fed Chair Kevin Warsh (It was a long queue). In addition, I participated in a number of closed-door meetings, which included several central bank officials. Reading between the lines, when it comes to Mr Warsh, and just listening to the lines when it comes to Mr Moulin, the rate hike cycle looks likely to remain moderate. The consensus of economists also is that the rate hike cycle will not go quite as far as markets are pricing. This reflects the view that energy prices will likely come down and that second round effects will be more limited than in 2022-2023 as supply side constrains are more modest and labour markets less tight. This is also in line with our own view, though like my colleagues, I note that a larger energy price shock could change all that leading to more aggressive hikes than the macro consensus is expecting.

Drivers of bond sell-off a matter of contention

The other big theme at the various meetings were the drivers of the bond sell off and risks around sovereign debt more generally, with France in the spotlight. The majority judged that the sell-off – at least when it came to US and German government bonds – was simply a reflection of markets pricing in a higher real rate over the long run. This was interpreted as being a rise in the equilibrium rate, or r-star, which made sense because of the capex super-cycle (higher planned investment relative to savings). Term premia or inflation compensation had not moved so much. Others did put more weight on fiscal and inflation concerns. Others suggested that the rise in real rates was not an r-star story, because of the correlation between central bank expectations and energy prices. In that view, markets were simply extrapolating the likelihood of higher central bank rates in the short-term into the long-term. For the record we are in the r-star camp, but judge that although r-star has probably risen, markets might be overstating the extent of the rise. On the other hand, we do think that term premia are currently not fully reflecting fiscal risks and bond supply, so do have room to rise further. Long yields could remain under upward pressure even if the eventual central-bank hiking cycle proves shallower than forwards currently imply.

Spectrum of worries about France

Meanwhile, there were various degrees of concerns about France and sovereign risk more generally. Some took the view that Moulin expressed in his fireside chat, that the degree of fiscal consolidation that France had to achieve was manageable and that it was encouraging that a consensus about the need for consolidation was emerging in the political establishment even though the situation was concerning. However, many others took a much less benign view of the situation. First, the fiscal consolidation that France had to achieve just to stabilise debt was enormous (in the region of 4-5% GDP). Second, this would be difficult given the fragmented parliament. Third, political uncertainty would likely persist for much of next year at least, given the presidential elections would likely be followed by parliamentary elections. Fourth, consolidation would need to be mostly on the expenditure side, make it more difficult politically, while rising debt service costs meant other spending had to be squeezed even more. I fall more in the latter camp than the former camp and to those concerns I would add the possibility of a President Le Pen clashing with Europe over France’s contributions to the budget. Government clashes with Europe tend to make markets nervous.

There was general agreement that French government bond spreads were nowhere near levels that would justify ECB intervention. Any eventual ECB intervention was generally seen coming in the shape of the TPI (the ECB’s bond buying programme to guard against fragmentation) or a pausing of QT. Still it was also acknowledged that France did not currently meet the eligibility criteria for the TPI, as it was not in compliance with the fiscal rules. Interestingly, some European officials highlighted that theoretical intervention options were not limited to the ECB, as the ESM also had policy options. These were in the form of a precautionary credit line, primary and secondary market purchases and loans. It is worth noting that all of these would come with various forms of conditionality.

Knot a supporter of ‘whatever it takes’ policy

Since former ECB President Draghi’s ‘whatever it takes intervention’ to calm fragmentation and redenomination risks in eurozone government bond markets, there has been a general acceptance in the leadership of the ECB that the central bank had to maintain a lender of last resort function. ECB President Lagarde admitted after originally misspeaking that the ECB was there to ‘close the spread’ in the case of rising spreads not justified by fundamentals. One of the current top contenders to be her successor Klaas Knot (former President of the Dutch central bank) seemed to support this line at an IIF panel in Bangkok. Paraphrasing him, he noted that Draghi’s intervention was a paradigm shift for the eurozone, which was now in a much better place because of it. Of course that does not mean he (or the current ECB Governing Council) would support intervening to cap bond spreads in any circumstances. My interpretation remains that they would intervene if (a) the spread moves are large and/or quick enough and (b) if they are unjustified by fundamentals and/or France agreed to take concrete actions.

At least we got a trick

Bullishness on the US economy versus the eurozone economy was another common theme. When a European economist mentioned that the US was a ‘one trick pony’ (referring to its extreme reliance on AI capex), a US economist leaned over to me and said ‘at least we have got a trick’. Short-term, it is difficult to make the case that the US-eurozone growth gap will start to close. However, I would make a few points to balance the argument. European capex on clean energy (USD 550 bn this year according to the IEA) is running ahead of US capex on clean energy. Indeed, many of the round tables I attended on the energy transition and on the industrial sector referred to Europe in a favourable light in these sectors and an example to follow. Second, Europe will not win the AI model or capex race, but it can greatly benefit from AI adoption. Finally, the US-eurozone growth differential may depend on the risks that materialise over the coming months. Europe would almost certainly do (even) worse if the interest rate or energy shock becomes bigger. On the other hand, Europe would be more resilient than the US if we were to see an AI shock.

To sum up, the global economy has been remarkably resilient, but the mechanisms producing that resilience are simultaneously creating the next vulnerabilities. I hope you enjoyed this ‘on the ground briefing’.