I always like to run along the Tagus river whenever I’m in Lisbon. There are many murals in Lisbon, and during one of these runs on that visit, I stumbled across a certain mural near the river, with the date “25 Abril 1974”, which I later found out is a national holiday in Portugal, celebrating the Carnation Revolution. I have to admit I didn’t know much about modern Portuguese history at the time, because most of the books I had read mostly covered earlier periods of Portuguese history or were simply about other topics, so it got me thinking that I should read more about it. Since then I’ve been reading numerous books about modern Portuguese history, including The Carnation Revolution: The Day Portugal’s Dictatorship Fell by Alex Fernandes, which I’m close to finishing at present, and is a thrilling hour by hour account of the events.

It explains how the revolution was literally kicked off, when one of the signals was sent through the playing of the 1974 Portuguese Eurovision entry “E Depois de Adeus” (And After the Farewell – on YouTube here if you’re interested in hearing this Frank Sinatra like track) by Paulo Carvalho, and the words introducing the song “It’s five to eleven”. Whilst the song never won the contest (a certain Swedish group called Abba won that year), it did literally start a revolution.

It is interesting how every book I’ve about what are the same series of events (the Carnation Revolution and the decades preceding), tell very different stories: despite the fact they are over fifty years ago, the biases still show. It is inescapable how the biases come out not so much in what they mention, but also in what they tend to gloss over or simply leave out. Whether it’s a narrative or a model, it’s the information or data that is missing that can have the most impact.

If we fast forward to the present day, the narrative about events happening in real-time is just as prone to bias as those from history. Traders in financial markets need to have unbiased information to make decisions. It’s the reason, traders pay for real-time news from sources like Bloomberg News or newspapers like the Financial Times: fact based reporting has a premium in the sea of opinion and bias, in particular on social media, which has a habit of drowning out everything. This is particularly relevant at difficult times like the present war in the Middle East: indeed “in war, truth is the first casualty” is a quotation we have all heard. As an aside, the ancient Greek dramatist Aeschylus was said to be the origin of this, although a quick Google search suggested other potential sources, including Samuel Johnson and more recently US Senator Hiram Johnson.

If we look more broadly at economic commentary, trying to strip away bias from a narrative difficult. We all know the permabear, who mysteriously flips their viewpoint when their party is in power, or the economist who is forever a hawk, regardless of whatever inflation is doing. Trying to make economic forecasts without bias is difficult. At Turnleaf Analytics, by employing a data driven approach and curating a thoroughly comprehensive dataset to describe the economy, we can strip away the bias and let the data speak. Of course there is an element in discretion is curating the dataset, and also in terms of putting together the forecast in a model, and any final tweaks. However, I’d argue that there is far more bias in having a view and then constructing a forecast around it. Furthermore, given how fast moving the situation at present, a data driven approach allows to quickly update a data driven forecast to be as fresh as possible. An opinion based view tends to be more static and more difficult to adjust.

So next time, you visit Lisbon, if you’re running along the Tagus, maybe take a step back and see if you can find the Carnation Revolution mural.

Figure 1

The inflation impact of an oil shock in Latin America can be understood by considering three layers of transmission:

1. Exposure – whether a country is a net importer and how heavily fuel weighs in the CPI basket
2. Extent of intervention – whether a country decides to leverage subsidies or administrative price controls or whether they already have automatic stabilizers
3. Cross-elasticities – second-round responses to food, transport, FX, and domestic supply disruptions

Most governments in the region are actively smoothing pass-through, so the immediate inflation effect is being shaped more by administrative choices than by the external shock itself. But that insulation is uneven. At one end, Mexico and Brazil remain relatively buffered. At the other, Peru is absorbing a largely unfiltered shock. In between, Colombia is actively offsetting global price pressure, Chile faces a policy-driven regime shift risk, and Argentina is allowing pass-through, but gradually. In the following sections, we go into cross-country pass-through in greater detail.

Mexico — Most Insulated

Mexico remains the most insulated economy in the region. President Sheinbaum has capped retail gasoline prices at MXN 24/litre, while the IEPS excise tax continues to operate as an automatic stabiliser, compressing as international prices rise. Together, these mechanisms sharply reduce the transmission of Brent into consumer fuel prices.

The constraint is fiscal. The longer oil prices stay elevated, the more IEPS revenue is sacrificed, raising the eventual probability of adjustment.

Brazil — Contained, but with a Diesel Tail Risk

Brazil has also contained near-term pass-through. Petrobras has signaled that it will not fully transmit the shock to domestic prices, and the government is moving to eliminate the PIS/Cofins levy on diesel imports. On the gasoline side, the increase in the ethanol blend to 30% provides an additional cushion.

The main vulnerability is diesel. Brazil still imports roughly a quarter of its diesel needs, leaving part of the economy exposed to higher global prices. At the same time, rising fertiliser costs, particularly urea, create a secondary channel into food inflation if the shock persists.

Colombia — Offsetting the Shock

To read the rest, visit our latest Substack post, here.

Figure 1

To quantify the rise in airfare prices, Turnleaf leverages its proprietary daily airfare index, which tracks thousands of ticket prices across 35 countries for travel within the next four weeks. As shown in Figure 2, fares increased by as much as 45% across markets between the last week of February 2026 and the week of March 18, 2026.

Figure 2

The dispersion across countries is substantial and reflects a combination of geographic exposure, network structure, and demand shocks. The largest increases are concentrated in Asia-Pacific markets such as Australia (+45%), India (+41%), and the Philippines (+41%), where long-haul dependency and limited substitution toward alternative transport amplify the pass-through of higher jet fuel costs. These markets are also more exposed to rerouting effects and capacity tightening as airlines adjust flight paths and schedules in response to regional instability.

In contrast, North America and parts of Europe show more moderate increases, with the United States (+17%) and core European markets largely in the 10–18% range. These regions benefit from more competitive airline markets, denser route networks, and a higher share of short-haul travel, which dampens price sensitivity to fuel shocks. Notably, Canada (-7%) and Peru (-3%) are outliers, suggesting localized demand softness or excess capacity that is offsetting upward cost pressures.

We will continue to monitor airfare dynamics as the conflict evolves. Upward pressure is likely to persist, with Brent crude approaching $110 per barrel and showing little indication of near-term relief (Figure 3).

Figure 3

In particular, the Strait of Hormuz is a key transit route not only for crude oil but also for industrial inputs such as urea, which is essential for fertilizer production. These inputs play a non-trivial role in agricultural cost structures and, by extension, food price dynamics.

While energy price shocks typically transmit rapidly into headline inflation via transport and fuel costs, the pass-through into food prices is more gradual and operates through indirect channels. These include input costs (fertilizers), production cycles in agriculture, and distribution costs, all of which introduce meaningful lags.

Consistent with this, there is limited evidence so far of an immediate response in proprietary food price indices (e.g., the Eurozone). This likely reflects short-term frictions such as existing inventories, pre-contracted pricing, and the seasonal nature of agricultural production, rather than an absence of underlying pressure.

Given the persistence and uncertainty of the current shock, it is therefore necessary to distinguish between immediate and lagged transmission channels. Specifically, the relevant pathways include:

• Brent crude → fuel and transport costs → headline CPI (short horizon)
• Brent crude → production and distribution costs → food prices → CPI (medium horizon)

The heatmaps below show lagged correlations between changes in Brent crude and urea prices and food CPI inflation over a 0–12 month horizon. These are intended to characterize timing and relative strength of pass-through rather than establish causality.

For Brent crude (Figure 1), correlations are weak at short lags but strengthen at longer horizons, with a broad clustering in the 9–12 month range. This is consistent with a delayed transmission into food prices via transport, processing, and distribution costs. While the magnitudes are moderate, the consistency across countries points to a common lag structure.

Figure 1

Urea (Figure 2) shows a more direct but less uniform pattern. Correlations are generally stronger and emerge earlier, typically in the 4–7 month range, consistent with fertilizer application and crop cycles. The dispersion across countries reflects differences in fertilizer dependence and agricultural structure.

Figure 2

Negative correlations at short lags in several economies likely reflect temporary offsetting factors—such as favorable harvests or inventory adjustments—rather than true inverse relationships. The shift toward positive correlations at longer lags underscores the importance of delayed cost pass-through.

Taken together, the evidence points to a layered transmission process: energy effects dominate in the near term, while fertilizer-related pressures emerge later and with greater heterogeneity. This suggests that the current oil shock is unlikely to drive immediate food inflation but may exert upward pressure over a 6–12 month horizon, with fertilizer dynamics representing an additional, independent risk channel.

Looking ahead, the key risk is not the immediate CPI response, but the potential for a second-round effect as higher input costs gradually feed through into food prices. If elevated energy and fertilizer prices persist, the combination could lead to a more sustained and broad-based inflation impulse than currently reflected in near-term data.

Given the uncertainty surrounding the duration of the conflict, several countries including France, South Korea, and China have discussed or already implemented measures to limit the inflationary impact of an oil shock.

The duration of the war remains uncertain, and its long-term implications for inflation are still unclear. However, when we examine inflation-adjusted Brent crude prices, they remain below the levels observed during previous historical crises (Figure 1 – Source: Josep Pocalles). Brent crude prices may be rising, but they are still well below past crisis peaks.

Figure 1

At Turnleaf, we recognize that the pass-through from Brent crude prices to CPI is immediate in most countries and particularly strong where retail gasoline prices are largely unregulated (see our previous post on Brent crude pass-through here). In the United States, for example, pass-through is both strong and rapid, whereas in the Eurozone—where a tax wedge dampens price transmission—it is considerably more limited. Since February 28, 2026, Turnleaf’s proprietary fuel indices have recorded significant increases in retail gasoline prices.

Below, we highlight the United States (Figure 2) and the Eurozone (Figure 3) as examples, both exhibiting a clear upward trend. Fuel effects are explicitly captured through our proprietary fuel indices across roughly 20 countries, which have steadily increased since the onset of the war. Figure 4 summarizes the widespread increases across EU countries, many of which have already seen upward revisions to our 12-month forecasts following the recent oil shock.

Figure 2

Figure 3

Figure 4

Other countries for which we also provide fuel indices like South Africa and Israel are expected to increase once regulatory prices are updated in the next month. As we collect more data and the situation in Iran becomes clearer, Turnleaf will update its inflation forecasts accordingly.

Figure 1

Figure 2

Oil hits the US fast. The Brent-CPI heatmap (Figure 3) shows the US peaking at lag 1, the strongest reading in the sample. US retail fuel prices are set by competitive markets with no administered pricing. Gasoline and motor fuel carry a combined CPI weight above 4%, and price changes at the pump flow into the index within one reporting cycle. The $30/bbl Brent move visible in the vertical spike in Figure 1 will register in March–April CPI prints almost mechanically. Canada, Thailand, and Malaysia show similar rapid pass-through profiles in Figure 3 for the same structural reasons.

Figure 3

Europe’s oil response is muted by the tax wedge. The Eurozone row in Figure 3 peaks at lag 1 at roughly a third of the US equivalent. Fuel taxes (excise duties plus VAT) account for 50–65% of the retail price across France, Germany, Italy, and the Netherlands. Because excise duties are levied per litre rather than ad valorem, a doubling of crude translates into a far smaller percentage increase at the pump. Governments also retain the option to temporarily cut duties to absorb the shock, as France and Germany both did during the 2022 spike. The weak, diffuse colouring across the European rows in Figure 3 reflects this structural dampening.

The UK sits in between. Figure 3 shows the UK Brent-CPI correlation peaking at lag 2, reflecting Ofgem’s quarterly energy price cap. Wholesale energy cost changes are not passed to households immediately and instead, the cap is recalculated each quarter based on a backward-looking observation window. This regulatory structure introduces a built-in 1–2 quarter delay between a wholesale price spike and its appearance in CPI.

Gas is Europe’s real (delayed) vulnerability. The TTF-CPI heatmap (Figure 4) tells a strikingly different story from the Brent panel. European economies dominate the warm-shaded zone, but the peak correlations are concentrated at lags 5–8 for Italy, Germany, and the UK. Two primary mechanisms drive this delay. First, regulated utility tariffs in most European markets reset on fixed schedules (quarterly or semi-annually), creating discrete pass-through windows. Second, household heating costs depend on billing cycles and contract renewal dates, which vary by country but typically impose a 3–6 month lag between wholesale price movements and the consumer bill.

Figure 4

Implications on Turnleaf Forecasts

The spike visible in Figure 1 will produce a small near-term fuel price uptick dampened by the tax wedge (the muted European rows in Figure 3), while the spike in Figure 2 will drive a larger and more persistent gas-driven inflation impulse arriving in Q3–Q4 2026 as the strong lag 5–8 correlations in Figure 4 imply utility tariff resets, industrial cost pass-through, and heating bill adjustments feeding into CPI simultaneously. The Hormuz disruption is already feeding into our real-time inflation nowcasts via the proprietary alternative and high-frequency data we track across 36 countries, including daily shipping flows, port activity, and commodity prices.

The key factor to watch will be emerging regulatory intervention which will determine whether the historical pass-through coefficients hold or compress as they did in 2022 during the onset of the war in Ukraine. So far, several countries have already responded to higher energy prices:

• In the UK, Ofgem’s April–June price cap has already been set at £1,641, meaning households on variable tariffs are shielded from the current spike until at least July. However, Cornwall Insight’s forecast for the July–September cap has surged to £1,801, an increase of £160 on April’s level, driven by elevated wholesale gas prices. Each of these interventions will compress the historical pass-through coefficients documented in Figures 3 and 4, but only temporarily and at mounting fiscal cost.
• Hungary reimposed fuel price caps on Monday, with Orbán capping petrol at 595 forints per litre and diesel at 615 forints, alongside a release of state reserves.
• South Korea announced its first fuel price cap in 30 years, with President Lee Jae Myung ordering authorities to “swiftly introduce and boldly implement” a maximum price system for petroleum products. Seoul has also unveiled a 100 trillion won ($68.3 billion) stabilisation fund to subsidise energy costs for households and industry.
• China’s NDRC raised retail gasoline and diesel price caps by 695 and 670 yuan per metric ton respectively from Tuesday, the largest adjustment since March 2022, under its 10-working-day mechanism that links domestic prices to international crude. That mechanism includes a built-in ceiling at $130 per barrel, above which retail prices are generally frozen, meaning pass-through will cap out if Brent keeps climbing. Beijing also halted refinery exports of diesel and petrol on 5 March to preserve domestic supply.

Whether that shift proves transitory or persistent depends on duration. Futures markets have already repriced, but the physical supply layer has not yet confirmed or contradicted what is being priced. Turnleaf is monitoring that gap using high-frequency AIS-based routing, port call data, and commodity prices.

Figure 1

How the Twelve-Day War Priced Through to CPI

The June 2025 Twelve-Day War provides the most recent comparable template for how an energy shock of this kind transmits into consumer prices. As Brent crude rose approximately 10 dollars per barrel over the course of that conflict, Turnleaf’s daily US CPI year-on-year NSA forecast curve shifted upward by 30 to 40 basis points (Figure 2). After the ceasefire, the forecast curve reverted toward its pre-conflict trajectory and tracked the subsequent path of realized US CPI relatively well, confirming that markets priced the shock as transitory once supply pressure eased.

Higher crude prices feed through to transportation, manufacturing inputs, and food production, but conflict also disturbs components not conventionally thought of as energy-linked. Airline fares are a prominent example, responding both through kerosene costs and through the pent-up demand that releases when routes reopen after an airspace closure.

The current episode differs in two important respects. The geographic scope is substantially wider, with simultaneous pressure on Hormuz, the Red Sea, and Gulf LNG infrastructure rather than a more contained exchange. And the macro backdrop has shifted, with tariffs already adding to goods prices and structural gas demand running higher than it was twelve months ago. Even a similar crude price move could produce a larger and more persistent inflation pass-through.

Figure 2

The Strait of Hormuz and What It Means for Energy Costs

Roughly 20 million barrels of oil pass through the Strait of Hormuz each day, approximately a fifth of global petroleum liquids consumption. A sustained impairment cannot be quickly rerouted, and the freight and insurance costs that tanker owners are already building into rates will eventually reach the pump and, from there, the CPI basket.

Recent tanker call data show a meaningful decline relative to recent norms, though the figures do not yet include complete data through March 2, 2026. Kpler tracking data showed that traffic fell roughly 20 to 25 percent within hours of the initial strikes, with the majority of vessels either performing U-turns, idling, or diverting toward alternative routes. By late Saturday UTC, total reductions were approaching 75 percent relative to baseline. We expect tanker transit calls through the Strait of Hormuz to fall considerably in the next few weeks (Figure 3).

Figure 3

European Gas and the CPI Pass-Through

The EU imports approximately 58 percent of its total energy consumption. The post-2022 diversification shifted dependency toward Norwegian pipelines, US LNG, and Qatari LNG. It is that last source now under direct pressure, with QatarEnergy halting all LNG production at its Ras Laffan and Mesaieed facilities after Iranian drone strikes on Monday.

Italy generates roughly 38 to 44 percent of its electricity from natural gas depending on the period measured, making it the most exposed large economy in the bloc to a direct CPI pass-through from higher wholesale gas prices. Germany relies on gas for roughly 15 to 20 percent of its electricity generation with limited storage headroom. Rising structural demand from data centres and the ongoing coal-to-gas transition in Central European power generation make Europe more sensitive to an LNG supply shock in 2026 than it was in 2022, and Turnleaf has flagged both dynamics in previous reporting.

The TTF spike and partial retracement are consistent with repricing while the market waits for confirmation on duration. Even if the supply disruption is ultimately limited, the path through freight, insurance, and procurement competition can still raise costs, complicate inflation prints, and increase volatility in forward power and industrial input pricing. We will continue to monitor movements in natural gas futures and will account for them in our updated forecasts (Figure 4).

Figure 4

Who Gets Hit Hardest

The inflation channel operates through direct household energy costs; industrial input costs across manufacturing and chemicals; transportation and logistics via freight and air routes; and food production, where energy-intensive inputs and distribution costs amplify the pass-through well beyond headline energy prices.

Japan imports approximately 97 to 99 percent of both its oil and natural gas needs, placing it among the most structurally vulnerable large net energy importers to a sustained increase in global energy costs. South Korea faces similar risks. Different denominators: the 20 mb/d figure above is global consumption; the shares below use Persian Gulf departures and destinations. In 2024, 84 percent of crude oil and 83 percent of LNG leaving the Persian Gulf transited the Strait of Hormuz. China, India, Japan, and South Korea accounted for a combined 69 percent of all Hormuz crude flows.

To read the rest, visit our latest Substack post, here.

Figure 1 – visit our Substack to see the latest forecast, here

Within hours of the ruling, Trump signed a proclamation invoking Section 122 of the Trade Act of 1974, a statute never previously used for tariffs, imposing a 10% global surcharge effective today, February 24, 2026. On February 21, Trump announced via Truth Social his intention to raise this to the statutory ceiling of 15%, though no formal proclamation had been issued as of the time of writing (White House Fact Sheet). Unlike IEEPA, Section 122 tariffs are capped at 15% and expire after 150 days (July 24, 2026) unless Congress votes to extend them. Markets are treating the 15% as the operative rate pending CBP enforcement guidance, and our table below reflects that assumption. Given the administration’s prior use of multiple statutory pathways, further action under Sections 232 or 301 remains a high-probability scenario once the 150-day window expires.

Refunds as Fiscal Offset

Other countries who have already made deals with Trump are likely to resist any global tariff that exceeds their established amounts. As the table below illustrates, several partners (EU, UK, trade-deal countries) now face a higher effective rate under the flat 15% Section 122 surcharge than under their prior IEEPA deals, while others (notably China) see a net reduction as the stacked IEEPA fentanyl tariffs are vacated. IEEPA refunds also pose a significant fiscal challenge (Figure 2).

Figure 2: Current Tariff Policy as of February 18, 2026 (Source: Yale Budget Lab; Global Trade Alert) – Broad Tariffs under IEEPA Authority (REPEALED)

U.S. Customs and Border Protection collected approximately $133–$142 billion in IEEPA duties through year-end 2025, with the Penn Wharton Budget Model projecting up to $175 billion in total refund liability. These refunds flow to importers rather than consumers, functioning as a form of backdoor fiscal stimulus that partially offsets the inflationary drag from remaining tariffs, though the timing is highly uncertain, with TD Securities estimating a 12–18 month disbursement window. On net, the Yale Budget Lab estimates a short-run consumer price increase of ~0.6% under full pass-through, roughly $600–$800 per household, approximately half the impact had IEEPA tariffs remained in force. This is consistent with our view that the inflation impulse from the tariff transition will be modest and gradual, concentrated in metals, electronics, and motor vehicles where Section 232 tariffs continue to bite.

Implications for Inflation and the Fed

The temporary and legally constrained nature of Section 122 tariffs reduces the risk of a sustained inflation shock, allowing monetary policy to remain focused on underlying disinflation trends. Chairman Powell’s last day is set to be May 15, 2026, when he is replaced by Kevin Warsh, a former Fed Governor (2006–2011) nominated by Trump on January 30, 2026, who is expected to cut interest rates at a faster pace than his predecessor. Though interest rates remain on hold for now, markets are pricing in 2–3 cuts in 2026 under the Warsh-led Fed, a trajectory that, if realized, would provide some offset to any tariff-driven cost pressures visible in H2 CPI prints.

The table above shows that the shift to a flat 15% Section 122 surcharge produces mixed outcomes across partners. The effective tariff rate rises to approximately 13.7% (Yale Budget Lab, Feb. 21, 2026, a moderate increase from the initial 10% rate but well below the pre-SCOTUS level of 16.0%. Firms largely absorbed import costs through 2025 pending legal clarity, delaying pass-through to consumer prices. Any residual price pressures are likely to emerge gradually over the next few months, with delays possible depending on how Congress acts at the 150-day mark. We will continue to monitor high-frequency port data, PMI import price subcomponents, and survey data to gauge upside risk on inflation.

What’s Next?

The tariff transition does not materially change the macro outlook. The SCOTUS ruling removes tail risk from an uncapped IEEPA regime, while the Section 122 replacement is time-limited, legally constrained, and partially offset by the fiscal stimulus of IEEPA refunds. The inflation impulse remains modest and manageable. The more consequential variable is whether Congress extends Section 122 past July 24 or the administration deploys Sections 232 and 301 at scale, either of which would warrant an upward revision to our CPI forecasts.

The breakdown of the real yield anchor

For much of the past decade, real yields offered a reliable framework for thinking about gold. When inflation-adjusted returns on safe assets fell, gold typically benefited. Since 2022, that relationship has weakened. US 10-year real yields rose sharply to post-GFC highs, yet gold remained resilient and went on to set successive record highs through 2025. Figure 1 captures the shift. The rolling correlation between gold and real yields trends lower after 2022, while the correlation between gold and broad-dollar weakness rises into 2025, which is consistent with gold rotating from a rate-dominated signal toward a more FX-sensitive regime.

Emerging-market central banks have helped drive this decoupling by diversifying reserves away from dollar assets. Their buying changes the marginal source of demand and shifts the transmission mechanism. Gold increasingly responds through exchange rates and import-price dynamics via the cost of dollars and local-currency purchasing power, rather than primarily through the domestic real-rate channel.

Figure 1

The dollar channel and EM currency clustering

In the current regime, gold strength often lines up with broad-dollar weakness because exchange-rate moves pass through quickly into traded goods and import prices. Figure 2 shows this by plotting gold returns against an inverted broad-dollar index return series. The relationship is episodic, but the co-movement becomes more visible in stress windows, including the large dollar move around the 2025 marker.

Figure 2

For emerging markets, the key issue is synchronicity within the basket rather than idiosyncratic single-country moves. Figure 3 pairs standardized broad-dollar index returns with an EM depreciation basket and a breadth measure that tracks how many currencies in this basket depreciate at the same time. Breadth stays low in most months but rises in stress episodes, which captures the clustering mechanism in practice: the distribution shifts toward shared drawdowns when the broad-dollar strengthens and funding conditions tighten.

Figure 3

Figure 4 checks whether the shift in transmission shows up in realized outcomes by plotting the cross-sectional distribution of headline CPI inflation across the sample and a trimmed dispersion metric defined as the P80–P20 spread. The sample combines advanced economies and major emerging markets (United States, Eurozone, United Kingdom, Japan, Canada, China, India, Brazil, Mexico, Turkey, South Africa). Dispersion rises into the 2021–22 inflation shock as inflation outcomes fan out across countries, then compresses sharply through 2023 and 2024 as inflation re-converges. The modest widening around 2025 suggests renewed fragmentation, but at a far smaller scale than 2022.

Figure 4

Country-level transmission

The shift from yield-driven to reserve-driven gold increases the scope for inflation dispersion based on structural archetypes.

China is a structural contributor. The PBoC has been an important contributor to the current regime through price-insensitive gold accumulation to reduce exposure to US Treasuries. While domestic inflation in China is muted by capital controls and managed exchange rates, reserve diversification affects the marginal demand for dollars and reshapes global pricing conditions for other economies.

Turkey illustrates regime-dependent pass-through. Turkey’s exchange rate is not a clean market price, because policy has often leaned on reserves, regulation, and administrative measures to smooth or delay depreciation. That does not remove pass-through, it changes its timing and its form.

To read the rest, consider subscribing to Turnleaf’s Substack, here.

In this article, I’ll try to discuss my takeaways from the event. Whilst, it is impossible to be totally comprehensive given the sheer number of sessions, I’ll try to give a flavour of the discussions from the event. Some discussions were under Chatham House rules, so in some cases, I’ll focus on more on the general takeaways.

Geopolitics: From the White House
Marc Gustafson (Eurasia Group) gave a fascinating presentation on the Trump presidency and what to expect for the rest of the term. He began with a few anecdotes from his time in the White House in the Situation Room, working for Obama, Biden and Trump, and how their very differing styles on camera and in private. He noted how in his first term, at the start there had been a measured approach from Trump. In his second term, he had moved at breakneck speed.

He felt that we were at “peak Trump”, with the various checks and balances he can encounter (indeed the recent striking down of tariffs by the Supreme Court is an example of that), from Congress and public opinion and, inflation was also another factor. Foreign policy could have some upsides, but also risk, noting that other presidents had made blunders, and it will happen to Trump. He noted that the schedule of a president is hard, and old age could count against Trump. On Iran, his base case was that there would likely be an attack.

Marc Gustafson (Eurasia Group)

Macro views
One of the keynote sessions on macro saw Michael Koester (Market Alpha Advisors) interview Mo Grimeh (Point72). Grimeh saw some stickiness in inflation because of tariffs, expecting it to be north of 3%. Just as an aside, the US inflation market is currently pricing in around 2.5% for a year out. Our models at Turnleaf Analytics see US CPI around 2.8% that far out, somewhat closer to Grimeh’s view.

Grimeh also saw some softening of the unemployment rate pushing up to around 4.6-4.7%. He thought that the Fed would cut 1-2 times this year. On the Fed, Grimeh noted that it is more than purely the Chair, hence, it would be unlikely that Walsh would be able to cut rates drastically. If there was a substantial softening in policy, whilst stock might react positive initially, it would have the potential to push up long term yields.

On a broader perspective, we are largely done with convergence in policy of the various global central banks he suggested. This would likely be good for the FX market and FX vol. Recently, vol has been drifting down, albeit punctuated with spikes (and not purely just in FX). Grimeh noted that several factors had been helping to push down vol. Retail investors had been buying structured notes which supressed volatility. There had been inflows into systematic strategies such as trend and vol selling. There had also been inflows in multistrat firms (as well as newcomers). Hence, hundreds of teams had been created to do similar things (eg. mean-reverting carry) resulting in a convergence. We had seen the lows in volatility in his opinion.

On a bigger picture, whilst there had been some reversal of globalisation, with less multilateral actions, and a move towards bilateral activity, on the theme of selling US assets, he didn’t believe that was happening. However, on the margin there was less allocation to the US.

Michael Koegler (Market Alpha Advisors) and Mo Grimeh (Point72)

Dori Levanoni (Capstone) was interviewed by Lee Ferridge (State Street) on the outlook for FX. One of the first questions was on the USD, and the oft discussed notion of “dollar debasement” and are we in downtrend? Levanoni suggested we’re not. However, we are at a pivot point. The USD has rallied since 2011, and we’re at pretty high valuations. Then there is the question of exceptionalism. Hedge ratios for USD was also lower than normal. On other currencies, he felt that GBP was at risk, with an uncertainty premia.

Dori Levanoni (Capstone)

There was a panel moderated by Michael Melvin (UCSD Rady School Of Management) on supply chains, tariffs and capital flows. Ankit Sahni (Element) gave a recap of what had happened over the past year, noting that tariffs had broadly been expected last January, albeit not the broad range of countries they would be applied to. He noted how the market view had shifted from overweight US based on growth and tariffs to a very short position at present.

Bill Campbell (Doubleline) said that Trump had been an accelerant of deglobalisation. Globalisation had provided cheap labour and a global savings glut. EM had a better fiscal situation, had higher real yields, and there was still room for compression (of yields) in EM.

So was this a sell US story or a buy EM story, Melvin asked? Colin Crownover (Fidelity) suggested it was more a sell US story than buy EM story. After all, tariffs did not alter the external balance. What it did change was the price at which investors wanted to buy US debt. Flows were always positive, it was how they are priced that had changed. People were overweight US assets, and supply chains are very complex, it was a story that might play out over the coming decade.

Bill Campbell (Doubleline), Ankit Sahni (Element Capital), Saket Selot (Dow Chemical), Colin Crownover (Fidelity Investment) and Michael Melvin (Rady School of Management UCSD)

Keeping to a macro theme, Michael Koegler (Market Alpha Advisors) moderated a panel focused on EM. Natalia Gurushina (VanEck) noted that it would be difficult to repeat the same extent of EM outperformance this year. That being said, EM markets tended to pay more (ie. yields), EM countries were less leveraged, and as a result EM central banks were less constrained by the fiscal factors. Geraldo Garcia (Banxico) echoed the bullish setup for EM, noting how growth had been strong in EM versus DM. Some factors which had been traditionally been problematic for EM, like inflation were less of an issue now. In G10, inflation was around 2-4%, whilst in EM it has been 3-5%. DM had lower rates, and EM had higher real rates. In terms of specific currency views in EM, long BRL seemed popular in the panel, with IDR as a short.

Natalia Gurushina (VanEck), Gerardo Garcia (Banxico), Michael Koegler (Market Alpha Advisors) and Chris Milonopoulos (Lazard Asset Management)

Allan Guild (Hilltop Walk Consulting) hosted an Oxford Style Debate on the USD and exceptionalism. Marc Chandler (Bannockburn Global Forex) took the bearish viewpoint of the debate. He suggested that a rubicon had been crossed with Greenland. He noted that the USD’s role was changing, with Trump’s actions with respect to the Fed. There didn’t necessarily need to be a replacement for the USD, it could be fragmented. Walter O’Leary (University of Miami) took the bullish viewpoint. The USD still was 60% of reserves. Indeed a reserve currency is not replaced by blips. There was a rule of law and transparency compared to other alternatives. The US had deep capital markets, and retained military strength, all factors to help preserve US exceptionalism.

Allan Guild (Hilltop Walk Consulting), Marc Chandler (Bannockburn Global Forex) and Walter O’Leary (University of Miami – Miami Herbert Business School)

FX market structure and execution
Michael Melvin (Rady School of Management UCSD) moderated a panel on FX market structure, which featured speakers from the sell side, buy side, vendors and official organisation. Anna Nordstrom (NY Fed)  discussed the role of her team. They operated in the FX market, frequently to do customer transactions for the US government. In recent weeks there has been news that the NY Fed had “checked rates” on USDJPY on behalf of the US Treasury. Whilst, it was not possible to comment on particular situations, Nordstrom noted that the US Treasury owned USD policy. If they wanted to do something in the market (eg. intervention), the NY Fed would be clear with market participants who they were dealing for.

There was also a lot of discussion of the FX Global Code, which was a collaboration between official organisations, and participants in the FX market. It was noted that it had cleared up a lot of bad practices. At the same time some trading groups haven’t signed the global code yet.

Jeremy Smart (XTX), Anna Nordstrom (NY Fed), Chris Matsko (State Street), Yan Pu (Vanguard) and Michael Melvin (Rady School of Management UCSD)

In recent years, a major theme has been workflow automation in FX execution. Allan Guild (Hilltop Walk Consulting) chaired a panel on this subject. One big trend has been towards the consolidation of execution desks on the buy side from single asset only to multi asset. Marc Natale (Murex) noted that it was challenging, because whilst the objective was to have a cross asset desk, execution, risk etc. was all in different systems.

Matt O’Hara (360T) explained that the technology is here. Automating G20 was easy, and indeed 90% of spot is automated there. However, in forwards and swaps place, it was much less automated. When it came to NDFs deep and accessible liquidity was needed for automation, Jay Moore (LMAX) suggested.

Bart Joris (LSEG) said that discussions had to have end to end workflows had been had for years. However it was not that simple. How much automation would you like to do? TCA could be complex, interacting with them required programming skills, although AI was helping here. Indeed, I’ve seen some nice approached to TCA, such as TradeFeedr’s system, which has a Python based API, so you can do a lot of complicated querying via a Jupyter notebook. I would also expect MCPs to become important here for querying TCA.

Michael O’Malley (State Street), Jay Moore (LMAX), Bart Joris (LSEG), Matt O’Hara (360T), Allan Guild (Hilltop Walk Consulting), Marc Natale (Murex) and Gabino Roche Jr (Saphyne)

AI and its impact on markets
Iro Tasitsiomi (T Rowe Price) gave a presentation on AI and investing. She noted that the idea of AI is not new, tracing back its origins to the Dartmouth meeting in the mid-1950s. Since then, it has gone through various cycles, including the “winter of AI. Hence, the overnight sensation of the launch of ChatGPT was more the culmination of decades of work. Whilst Gen AI might be new, old style AI (ie. supervised learning) has been around for many years.

There was no free lunch when it came to using generative AI, such as hallucinations, non-determinism, and data leakage. There was also the temptation of using it for the sake of it, which was a big risk (hammer looking for a nail). Despite this AI could be an enormous accelerator.

Iro Tasitsiomi (T Rowe Price)

I also gave presentation about coding with Python, whilst addressing the key question we have over software engineering: why should we learn coding if we have AI to do it for us? LLMs can code very well these days. Indeed, I use Claude Code a lot, and a gave a quick demo of some analysis that I had done using it, to decompose US CPI.

Whilst these tools can make us much more efficient, it is still incredibly important to understand the code they generate, I suggested. I explained that the whole point of a human language is that it is incredibly expressive. The flipside, is that it is imprecise, unlike a programming language. Hence, you may well ask an LLM to code something, but what it produces is not precisely what you wanted, just because human language is easy to misinterpret. In my example, Claude Code did indeed right masses of code, but there was some mistakes which I had to fix, which resulted in the wrong or incomplete decomposition of US CPI. The code structure also needed to be refactored to make the code easier to read and maintain.

Furthermore, one of the major costs of software is not writing it initially, but the maintenance, particularly if it has been designed poorly. Software engineering as a discipline is not purely about writing the code. Indeed over the years, software engineering has been more about writing at higher and higher levels of abstraction (we don’t write assembler code for example). In a sense, LLMs are just another level of this abstraction for developers to adjust to.

There was an AI and data focused panel moderated by Lee Ferridge (State Street). Akshay Padmanabha (Balyasny) noted how AI was being used extensively. It could be used to automate many different things, such as data cleaning, screening, information extraction etc. Ed Tostanoksi (PGIM) noted how it was challenging from a multi asset perspective, to use some AI techniques. With many tools, the sample size for multi asset was too small. When looking at regimes, you had to go a long way back. However, classical statistics was designed for situations like this. But with large datasets, these newer techniques could add value. Padmanabha added that it was not about decision making. You had to figure out what AI was good at. The black box nature of these tools made it difficult, although prompting to explain can help. There was also the question of overfitting.

Akshay Padmanabha (Balyasny), Edward Tostanoski (PGIM) and Lee Ferridge (State Street)

Conclusion
There were many different discussions at TradeTech FX USA. On macro, the consensus view was short USD, to some extent based on the sell US theme. Although perhaps the bearish USD view was not as prevalent as I would have thought. The other strong view from discussions, was broadly long EM, on the back of stronger growth, high real yields and better fiscal picture. There was also a lot of focus on inflation, with many participants expecting stickier inflation than perhaps is priced into the market. Indeed, Turnleaf Analytics is expecting inflation towards the end of the year to be above what is priced by the inflation fixing markets in the US.

Outside of macro, AI was a big theme at the conference, and it is making inroads both from an idea generation perspective, to help screen and structure data. On the execution side, a lot more houses are taking advantage of electronic execution, but we are still very far from the stage of fully automated end to end workflow, especially when it comes to areas outside of G10 FX spot.

I will be interested to see what discussions will be like next year. Let’s see!