~With contributions from Vivien Lee, Aaron Ma and Eric Savoie
Iran war update
An imminent deal?
The prospect of a peace deal between the U.S. and Iran spiked higher over the May 23-24 weekend. Accordingly, the price of West Texas Intermediate crude oil fell to just US$90 per barrel, down sharply from the US$109 that had prevailed a week ago. Stock markets have similarly surged on the prospect of a deal.
This is big news and hopefully resolves favourably. Of course, there have been innumerable false dawns over the past three months, providing a reminder that nothing is truly final until large numbers of ships are steaming safely through the Strait of Hormuz.
Still, there is greater reason for optimism about a resolution this time than with the prior false starts:
There are reports from both sides about advancing negotiations – not just from the White House as per prior claims. Iran is also apparently seeking guarantees from China that if the Strait of Hormuz is reopened, China will commit to a long-term energy purchase contract, make infrastructure investments into Iran and/or provide diplomatic support for Iran at the international level.
Iran reports that it permitted 33 ships to sail through the Strait of Hormuz over a 24-hour span on the weekend. If true, this would be the greatest volume since the start of the war. (Then again, the official tracker for ship crossings shows no such surge – see next chart). At a minimum, as per our AI-assisted analysis, Iran has not attacked a transiting ship in over a week (see subsequent chart).
Strait of Hormuz remains largely closed despite ceasefire
As of May 24, 2026. Sources: Bloomberg, RBC GAM
No Iranian attacks on Strait of Hormuz traffic in over a week
As of May 25, 2026. Data compiled by RBC AI tool web scan of news reports and third-party analyses. Source: RBC GAM
Probability markets now assign a non-trivial 30% chance that the U.S. blockade will be lifted by the end of May – essentially, within the next week – and a 69% likelihood that it is gone by the end of June (see next chart). The flow of ships through the Strait of Hormuz is now expected to normalize around the end of June.
Prediction markets expect progress on reopening the Strait by mid-year
As at May 25, 2026. Strait of Hormuz normalization = 7-day average of Strait of Hormuz transit calls (arrival of ships) is 60 or greater by specified date. Sources: Polymarket, RBC GAM
The specific nature of the deal and the points still being negotiated are subject to some debate. Key elements are thought to include:
A commitment to stop fighting. Of course, this has already essentially happened between the two main antagonists (see next chart).
Daily count of missiles and drones launched by Iran remains at or near zero
As of May 25, 2026. Specific counts of confirmed missile and drone attacks on United Arab Emirates (UAE) on May 5th and 6th not available. Data compiled by RBC AI tool web scan of news reports, third-party analyses and ministry of defense official statements from Gulf countries. Source: RBC GAM
Get Hormuz flowing again. Both countries would remove their blockade. It remains possible that Iran would continue to charge its recently instituted fee on ships making the voyage. This would constitute a flawed outcome but may prove politically tolerable given the imperative of unsticking the global energy market.
Sanctions relief for Iran. Certain sanctions would still apply to Iran, but the country would again be allowed to export energy and might even be permitted to export more than before the conflict began. One analysis figures an additional 1.0-1.5 million barrels per day of oil might eventually be exported. This would help to restrain global oil prices over the long run.
Halt Iran’s nuclear enrichment program. The reports here are contradictory, but several amount to Iran agreeing not just to halt its enrichment program but to hand over its existing materials to either Russia or China. Then again, another version says that nuclear program negotiations will be saved for a later stage of talks.
-EL
Resilient economic data
In the meantime, economic data continues to pleasantly surprise. The energy shock has not done too much damage at the economy-wide level. U.S. jobless claims remain low and relatively steady (see next chart).
U.S. jobless claims are low and steady
As of the week ending May 16, 2026. Sources: U.S. Department of Labor, Macrobond, RBC GAM
Our composite U.S. economic barometer has actually been improving for most of 2026. It remains decent even after a recent slip (see next chart). Most compellingly, the closely-watched Atlanta Fed GDPNow metric is pointing to an astonishing +4.3% annualized real GDP increase in the second quarter for the U.S. – roughly twice the “normal” rate of growth, despite the energy shock.
Our composite U.S. economic barometer remains surprisingly positive
As of May 2026. Shaded area represents recession. Source: RBC GAM
Internationally, conditions are also proving less dire than feared. Manufacturing purchasing managers’ indices (PMIs) across a range of developed markets trended higher through April and only gave up a fraction of their earlier gains in May (see next chart).
Manufacturing activity continued to expand in most developed countries but at a slower pace
Eurozone, UK and Japan as of May 2026. U.S. as of April 2026. PMI refers to Purchasing Managers’ Index for manufacturing sector, a measure for economic activity. Sources: Haver Analytics, RBC GAM
Similarly, the Citi Data Change Index argues that the run of economic data is stronger now than it was a year ago, both in the U.S. and across the G10 (see next chart).
Economic growth has improved globally
As of 05/08/2026. Sources: Citigroup, Bloomberg, RBC GAM
To be sure, there is still reason to imagine economic data should be slightly weaker than normal over the next few months as the aftereffects of the war play out. But so far the economic resilience on display has been pleasantly surprising.
-EL
Inflation rising
Whereas it is hard to detect the war’s impact on economic growth, its effect on inflation has been undeniable. Headline consumer prices have soared right around the world (see next chart).
Global inflation has risen sharply since the start of Iran War
As of March 2026. Shaded area represents U.S. recession. Sources: Organisation for Economic Co-operation & Development (OECD), Macrobond, RBC GAM
The impact of higher energy costs and Hormuz-induced fertilizer shortages is pushing agricultural commodity prices upward (see next chart). This confirms a broadening of the inflation impact beyond the direct consequences of higher energy costs. (That said, the magnitude of the rise so far is no greater than the sort of perturbation that regularly occurs every few years.)
Agricultural commodity prices soar on higher input costs – fertilizer and energy
As of 05/12/2026. Shaded area represents U.S. recession. Sources: S&P Global Macrobond, RBC GAM
Further to the theme of inflation spillovers, our metric of U.S. inflation breadth has increased greatly in recent months, pointing to the fact that this inflation shock is no longer purely about energy costs (see next chart).
Higher inflation is gaining ground in U.S.
As of April 2026. Share of Consumer Price Index (CPI) with year-over-year % change falling within the ranges specified. Sources: Haver Analytics, RBC GAM
Similarly, the global supply chain pressure index shows a significant recent deterioration, presumably due to the energy shock. This should prove temporary and is not as big as the post-pandemic supply chain nightmare. But it shouldn’t be ignored (see next chart).
Global supply chain pressure rising sharply
As of April 2026. Shaded area represents U.S. recession. Sources: Gianluca Benigno, Julian di Giovanni, Jan J.J. Groen and Adam I. Noble, “A New Barometer of Global Supply Chain Pressures,” Federal Reserve Bank of New York, Macrobond, RBC GAM
China’s history of exporting deflation to the world is even under temporary threat due to more costly energy inputs: the Chinese Producer Price Index is now rising for the first time in a few years (see next chart).
Producer price inflation in China has accelerated on energy price shock and improving domestic industrial conditions
As of 04 2026. Shaded area represents U.S. recession. Sources: China National Bureau of Statistics (NBS), Macrobond, RBC GAM
The point is not that inflation is suddenly a big problem, but instead that price pressures are showing signs of spreading beyond energy products. These other sources of inflation may prove harder to unwind after the energy shock has ended. This supports the view that most developed-world central banks are now probably finished with monetary easing. Many may have to consider rate hiking in the year ahead.
-EL & VL
Oil and natural gas intensity by country
It remains fair to argue that the regions most adversely affected by this energy shock are Asia and Europe. North America is much less exposed. This is because the first two continents are the primary consumers of Middle Eastern energy, and energy markets are regionalized in a way that means not all countries are affected equally when the global supply-demand balance shifts.
But this is just a first-order approximation. Additional clarity can be achieved by considering further variables.
In a recent MacroMemo, we examined which governments are now subsidizing the cost of energy in their country. This has the effect of reducing the economic and inflation pain, albeit by creating an additional fiscal burden to be dealt with later. Asia and Europe have been doing this more than North America, reducing the North American economic advantage.
We now present a third layer to consider: the extent to which different countries actually use oil and natural gas relative to the heft of their economies. As it happens, this further erodes the North American advantage.
The U.S. is geographically large and famously car-oriented. It is thus not surprising that Japan and Europe use significantly less oil per unit of economic production than does the U.S. (see next chart). The UK only uses half as much as the U.S. Conversely, Canada is an outlier in consuming considerably more than the U.S.
Canada and U.S. lead G7 in oil used per unit of economic production
As at 05/11/2026. Data is for 2024. Oil intensity is measured as tonnes of oil consumption per $1,000 of constant 2015 US$ GDP. Sources: BP Statistical Review of World Energy, UN Trade and Development, RBC GAM
Meanwhile, emerging economies are broadly more oil-intensive than the U.S. – Russia, India, Mexico, Brazil, Indonesia and even China are all hit worse (see next chart).
Saudi Arabia leads world’s 20 largest economies in oil intensity
As at 05/11/2026. Data is for 2024. Sources: BP Statistical Review of World Energy, UN Trade and Development, RBC GAM
What about for natural gas? Interestingly, most developed countries are even less exposed to natural gas (relative to the U.S. baseline) than they are to oil (see next chart). This is presumably because other developed countries tend to rely more on renewable energy, and some are quite heavy users of nuclear power. As with oil, emerging-market economies are generally more exposed to natural gas than the U.S.
Canada leads G7 oil and gas intensity, followed by the U.S.
As at 05/24/2026. Data is for 2024. Energy intensity measured as megajoules (MJ) of oil and natural gas consumption per dollar of constant 2015 US$ GDP. Sources: BP Statistical Review of World Energy, UN Trade and Development, RBC GAM
Of course, an important distinction on the subject of natural gas is that the U.S. (and Canada) may be more exposed than most to an equivalent price movement. But there has been minimal price action in practice in North America versus a relatively large increase in Europe and Asia.
The conclusion is that there are a number of factors that somewhat narrow the gap between the most adversely affected regions (Europe and Asia) and the least affected (North America). But, in our view, an important gap still exists.
-EL & JN
Precarious energy markets
The flow of energy through the Strait of Hormuz has been heavily restricted for nearly three months. The International Energy Agency (IEA) says global oil supply was down 12.8 million barrels per day (mb/d) in April relative to February, a 12% reduction. Global oil inventories fell at a record 4 mb/d pace in March and April. Goldman Sachs estimates daily stock draws have more than doubled so far in May.
A supply shock of this scale and duration was supposed to lead to escalating price increases. Yet oil prices are no higher than they were two months ago and refined product prices in Asia and Europe have retreated from their earlier highs (see chart). But if a deal fails to be consummated and the Strait remains closed, this precarious balance might not be sustainable.
Crude oil and refined product prices are no higher than they were two months ago
As of 05/22/2026. Sources: Bloomberg, Macrobond, RBC GAM
The last pre-conflict barrels of oil exiting the Strait of Hormuz reached their destinations in Asia, Europe and the Eastern U.S. in the first half of April. That’s when we started to see a disconnect between the physical/spot market for oil and the paper/futures market (see chart below). The difference between the first-month futures contract for Brent crude (the “oil price” that is typically quoted) and the dated or spot price is typically only a few dollars in either direction. But it blew out to $35 per barrel of oil (spot exceeding futures) in the first half of April.
This was an indication that, while the paper market was becoming more hopeful that a U.S.-Iran ceasefire would eventually see the Strait reopened, the physical market was still struggling to clear and so higher prices were needed to balance supply and demand.
Spread between spot oil price and futures contract has narrowed after widening in April
As of May 25, 2026. Source: Bloomberg, RBC GAM
That spread has shrunk considerably over the past month. But rather than futures prices rising toward spot, it has mostly been falling spot prices narrowing the gap. The release of strategic petroleum reserves has helped relieve pressure on commercial inventories, and trade flows have adjusted to supply disruptions with refiners sourcing new supplies. The IEA notes Atlantic Basin crude exports have increased by 3.5 mb/d since February with notable gains from the U.S., Brazil and Venezuela – most of which is heading to Asia.
Falling demand has also helped balance the market, with imports declining as refiners reduce runs. Chinese seaborne crude imports fell by 3.6 mb/d in April relative to February. Imports into Japan, Korea and India cumulatively declined by a similar amount.
The IEA says the slowdown in global refinery activity has eased tensions in the market for crude oil, but tightness is spreading to product markets. The issue is compounded by lost refining capacity in the Gulf – about 20% of seaborne trade in refined products normally passes through the Strait of Hormuz. Unlike past oil shocks, the current crisis is a simultaneous loss of oil and refined product supply.
Demand destruction has prevented refined product prices from rising even further, with many countries – particularly in developing Asia – adopting measures to limit energy consumption (see chart). Airlines have cut back on flights by 18% according to one analyst. The IEA says demand from the price-sensitive petrochemicals sector is seeing the steepest decline.
Developing Asia is taking the most measures to limit energy consumption
As at 05/22/2026. Sources: International Energy Agency, RBC GAM
Still, the current degree of demand destruction might not be sufficient to contain prices if the Strait doesn’t reopen soon. The Energy Information Administration’s (EIA’s) latest short-term energy outlook assumed flows would slowly resume in the coming weeks (late May or early June) and that Brent crude will average $106/bbl in May and June. But if reopening is delayed by a month, it said near-term prices could be $20/bbl higher. Prices could rise even further over the summer as the IEA’s record 400 mb stockpile release runs its course. Forty percent of the stockpile was drawn by May 8.
Relative price increases for refined products threaten to be even greater – particularly middle distillates (diesel and jet fuel) that complex refineries like those in the Gulf are able to produce more of. While crude oil inventories were elevated heading into the current shock, European and U.S. distillate inventories were below average. Oil inventories are close to 100 days of demand – an 8-year low – and refined product stockpiles are less than half that.
Despite higher margins for products facing the most severe shortages, refiners can only rebalance production so much based on their configuration. And in some cases, restrictions on refined exports are preventing trade flows from rebalancing, exacerbating regional shortages. Europe is particularly reliant on Middle Eastern jet fuel imports and inventories are dwindling rapidly (see chart). While the IEA was overly pessimistic in its earlier assessment that Europe would run out of jet fuel by early June, severe shortages are still possible by late June or early July.
European jet fuel inventories are well below seasonal norms
Amsterdam-Rotterdam-Antwerp (ARA) jet inventory levels. As at 05/24/2026. Sources: Bloomberg, RBC GAM
Even if the Strait reopens soon, restarting and repairing Gulf production, refineries and export facilities, getting crude oil and products to market and rebuilding inventories will be a months-long process. Both the EIA and IEA think the oil market will remain in deficit (i.e., inventories being depleted) until Q4 with trade and production not returning to pre-conflict patterns until late this year or early next year. Refined product prices are likely to remain elevated just as the Northern Hemisphere’s summer travel season is getting underway.
Some further demand destruction seems likely. But with developing economies accounting for a disproportionate share of the adjustment, we think advanced economies – even net energy importers like Europe and Japan – can avoid recession. At this point, price increases for oil and refined products haven’t been as significant as feared, but that shouldn’t be taken for granted if the Strait remains closed and buffers continue to erode.
Disruption will intensify well before inventories reach zero. At the current pace of drawdowns, JP Morgan analysts suggest the operational floor for pipelines to maintain pressure and refineries to continue running could be reached in September.
As noted above, the current shock is extraordinary in that it encompasses crude oil and refined products. A third element is natural gas, with the Strait’s closure disrupting one-fifth of global Liquified Natural Gas (LNG) trade (slightly more than 10 billion cubic feet per day or bcf/d). Prices in Asia – the destination for more than 80% of Gulf LNG exports pre-conflict – are up 75% since late February while European natural gas prices are about 45% higher (both are down significantly from March’s highs). Yet there is less concern about natural gas shortages, for a few reasons:
The global economy is about one-quarter less reliant on natural gas relative to oil.
LNG only accounts for a fraction of natural gas supply – flows through the Strait represent less than 3% of global production.
New export capacity, particularly in North America, supported LNG supply through the 2025/26 heating season. The IEA estimates global LNG trade increased by 12% between October and February while benchmark prices in Europe and Asia fell by 25% over that period. U.S. LNG exports were up 3 bcf/d in February relative to a year ago and the country added another 0.9 bcf/d of export capacity in April.
A price premium in Asia helped divert LNG cargoes from Asia, helping to rebalance the market. Despite growing LNG trade, natural gas markets remain relatively regionalized – North America is well supplied, with prices little changed since the start of the conflict.
Several Asian countries have implemented fuel switching and policy measures to limit natural gas use. Milder weather has also helped reduce consumption. European natural gas consumption has declined as the power sector pivots to alternative sources of generation.
Recent steep price increases barely register on a 5-year chart of European natural gas prices (see below). The current disruption pales in comparison with the energy shock following Russia’s invasion of Ukraine. Europe is significantly less reliant on Gulf LNG than it was on piped Russian gas several years ago. German base power prices are little changed since the start of the conflict.
Natural gas prices rise sharply amid the Iran war
As of 05/20/2026. Sources: Intercontinental Exchange (ICE), Macrobond, RBC GAM
While the relative supply shock for natural gas is much less significant than for oil, the Iran conflict could have a prolonged impact on the market. As we discussed previously, natural gas infrastructure has sustained the most significant damage during the conflict. It could take several years to repair damage at Qatar’s Ras Laffan LNG facility. That leaves 3-4% of global LNG production offline in the meantime. The IEA thinks this will prolong tight market conditions through 2026 and 2027.
-JN
AI is driving unprecedented demand for memory
We previously discussed electricity generation as a key constraint to the AI data centre buildout. A shortage of memory chips has emerged as another major bottleneck. Rapidly rising prices are adding to hyperscalers’ CapEx bills, rippling through other electronic products’ supply chains and driving massive profits and equity price gains for a few key producers. The shortage looks set to persist for a few years before new supply comes online.
High bandwidth memory (HBM) is an essential component of AI chips. It allows GPUs to process large amounts of data at high speeds, enabling more efficient model training and faster inference. As AI chips become more powerful, their memory requirements are increasing (see chart below). NVIDIA’s latest generation Rubin chips require 3.6x as much memory as its H100 chips from 4 years ago.
Each new generation of NVIDIA chip requires more memory
As at 05/11/2026. Sources: Bloomberg, RBC GAM
Consulting firm TrendForce estimates HBM demand will rise by 70% this year alone, an increase that is running up against short-term inelastic supply. HBM is a specialized, 3-D stacked form of dynamic random access memory (DRAM) and 90% of DRAM supply comes from just three firms: South Korea’s Samsung and SK Hynix and U.S.-based Micron. All three have already sold out of their entire 2026 supply of HBM and have warned that significant memory shortages are likely to persist through at least 2027.
Given strong demand, these producers are expected to double their annual CapEx in the coming years (see chart below). The clamour for new capacity is so great that SK Hynix’s customers have reportedly offered to fund new production lines and finance expensive equipment purchases.
The big three memory producers are ramping up capital expenditure
As at 05/11/2026. Samsung and SK Hynix estimates converted from South Korean Won (KRW) to USD at market exchange rates. Consensus estimates shown for 2026-28. Sources: Bloomberg, RBC GAM
But some memory companies are taking a cautious approach to new investment given the boom-bust nature of the memory market and worries about eventual overcapacity. While not explicitly acknowledged by the firms, they also benefit (at least in the short run) when supply is constrained because it translates into higher profit margins.
It will take several years to bring new memory fabs online: industry analysts see new factories reaching full production in late 2027 or 2028. In the meantime, a supply shortage is resulting in sharp increases in memory prices (see chart below). The price of NAND flash memory, which provides longer-term storage and is also used in data centre servers, has increased alongside DRAM prices. The big three memory producers account for more than half of NAND production, but the market isn’t as concentrated as DRAM.
Memory prices are rising sharply
As at 05/11/2026. DRAM = DDR4 1Gx8 contract price. NAND = TLC 512Gb contract price. Sources: InSpectrum Tech Inc., Bloomberg, RBC GAM
Between rising prices and more memory-intensive chip design, memory now represents more than 60% of the component cost of AI chips (see chart below). NVIDIA and other chipmakers are passing along these higher costs, driving hyperscalers’ CapEx bills higher. Microsoft said $25 billion of its $190 billion CapEx guidance for 2026 reflects higher component pricing.
RBC Capital Markets estimates that rising memory prices will account for one-third of the year-over-year increase in hyperscaler CapEx in 2026. Critically, this means that the growth in hyperscaler CapEx is somewhat less extraordinary in inflation-adjusted terms. This is what matters for GDP growth and when gauging the expected return on investment.
Union Bank of Switzerland analysts say memory’s share of AI server rack costs could triple to 18% from 6% previously. While servers account for roughly half of an AI data centre’s upfront cost, memory price inflation is only likely to reduce the expected rate of return on a new data centre by a few percentage points. Still, it leaves less room for error if inference pricing or demand disappoints, or if GPU lifespan assumptions prove optimistic.
Memory represents a growing share of AI chip costs
As at 05/11/2026. Estimated quarterly spend by four leading U.S. AI chip designers (NVIDIA, AMD, Google and Amazon). Sources: Epoch AI, RBC GAM
Memory producers are increasingly focusing on HBM given rising margins. TrendForce estimates HBM will take up 23% of DRAM output this year, up from 19% last year. That is limiting DRAM supply for other tech products like smartphones, computers and servers – even auto producers are feeling the pinch.
Each of NVIDIA's latest generation Rubin chips requires as much memory as 24 high-end smartphones or 9 PCs. Several companies including Tesla and Apple have suggested the DRAM shortage will constrain production and compress margins. Mentions of memory prices on earnings calls more than doubled in the latest quarter according to Bloomberg.
Goldman Sachs expects higher memory prices will contribute to a 10% increase in phone and computer prices this year, although the overall effect on inflation is relatively small. Higher electronics and software prices, alongside rising electricity costs associated with the data centre buildout, likely added 0.1% to core Consumer Price Index (CPI) inflation over the past year. A similar contribution is expected over the next 12 months. The estimated effect on core Personal Consumption Expenditures price index inflation is somewhat greater.
If investors expect the memory producers to maintain current pricing power for years to come, there is a long history of boom-bust cycles in the industry.
While consumers, hyperscalers and other memory purchasers are suffering, memory companies themselves are seeing exponential earnings growth. Samsung and SK Hynix now account for nearly half of South Korea’s KOSPI, which is now the world’s 7th largest stock market and the best performing index year-to-date. Meanwhile, Micron is among the top performers in the S&P 500. Other memory producers like SanDisk and Western Digital have seen even bigger gains.
We recently discussed the outsized returns of semiconductor stocks within the S&P 500, and memory producers have been a key contributor to that trend.
Despite strong demand, there are clear risks for memory chipmakers. Hyperscalers might bristle at paying memory producer’s big margins and could begin producing their own memory chips, as several have done with AI chips to reduce their reliance on NVIDIA’s GPUs. They could also look for workarounds that reduce the memory intensity of their models and data centres.
If investors expect the memory producers to maintain current pricing power for years to come, there is a long history of boom-bust cycles in the industry. At a minimum, current extreme margins should shrink somewhat as new supply comes to the market.
Still, the memory shortage is worth monitoring as it is a critical input into AI CapEx, which itself has become a key driver of U.S. economic growth.
When it comes to data centre CapEx ambitions, the memory supply crunch doesn’t appear to be as significant or stubborn a constraint as electricity. The timeframe for adding memory chip fabs is shorter than bringing new power supply online. AI chipmakers and hyperscalers have been willing to pay premium pricing for memory chips, elbowing out other DRAM buyers.
Rising electricity prices (and additional costs associated with behind-the-meter generation) and higher memory prices both eat into data centre margins and spill over into higher costs for other businesses and consumers. But memory-related price hikes aren’t as visible as higher electricity prices in the broader economy and aren’t creating the same backlash against data centre construction.
Still, the memory shortage is worth monitoring as it is a critical input into AI CapEx, which itself has become a key driver of U.S. economic growth. And memory stocks are likely to remain a source of volatility for investors.
-JN
Out of left field department
El Niño ahead
The term “El Niño” refers to time periods when the temperature of the Pacific Ocean rises significantly above normal for 9 to 12 months, distorting the world’s climate in various ways. A variety of weather models now predict a 98% probability of an El Niño event this year.
El Niño tends to develop during the April to June period, before peaking in strength from October to February and leaving lingering effects for the subsequent summer.
Forecasting models now suggest there is an approximately 50% chance that this year could be a particularly powerful edition – a “super” El Niño, as it were. The last such event was in 2015-2016, and these come along on average less than once a decade.
The last super El Niño is estimated to have cost the global economy about 0.7ppt of output per year over a five-year period – a loss of US$3.9 trillion.
Overall, a key implication is that this coming winter may be milder than normal, and the subsequent summer could be hotter. But the effects vary substantially by region. Western Canada and the northern U.S. tend to be distinctly warmer, while the southern U.S. tends to be wetter. Problematically for farming, India and Southeast Asia tend to be drier. So does Australia and part of Sub-Saharan Africa.
On the aggregate, El Niño is bad for the global economy. The last super El Niño is estimated to have cost the global economy about 0.7ppt of output per year over a five-year period – a loss of US$3.9 trillion. International Monetary Fund (IMF) analysis finds particularly negative implications for India, Australia, Japan, Indonesia, Chile, New Zealand and South Africa.
El Niño also tends to be slightly inflationary at the global level. This is mainly due to its effects on agriculture, metals and mining, power, transportation and insurance.
While El Niño holds the potential to materially affect various pockets of the global economy, it is unlikely to be the central theme for the year ahead.
For agriculture, India suffered a substantial 14% reduction in monsoon rainfall during 2015-2016, hurting its farm sector. Rice prices accordingly tend to rise. Dry weather in Southeast Asia may increase palm oil prices. Wheat prices often also rise due to drier weather in Australia and Western Canada. Sugar and cocoa prices are often also higher, to the horror of chocolate lovers.
Copper and aluminum prices could also rise as mines grapple with weather-related adversity, while electricity demand could be lower than normal in the winter and higher than normal the subsequent summer.
Interestingly, while insurers may encounter greater costs with regard to wildfires in drier regions and floods in wetter regions, El Niño has historically substantially reduced the amount of Atlantic hurricane activity, reducing global insured losses significantly.
A final word: while El Niño holds the potential to materially affect various pockets of the global economy, it is unlikely to be the central theme for the year ahead.
Pandemic fears
Two diseases with high fatality rates—hantavirus and Ebola – are presently experiencing outbreaks and capturing the public’s attention. Several years removed from the worst pandemic of the last century, there is naturally an elevated level of concern about such matters.
Even a relatively minor pandemic on par with the 2003 SARS outbreak would be consequential at the regional level. It would discourage travel, the consumption of restaurant meals, attendance at large public events and the use of public transit, while encouraging more working from home.
So how threatening are these new viral outbreaks? Neither is likely to have a major effect on the world. Let’s run through why.
Hantavirus
Hantavirus is primarily found in rodents and concentrated in South America. But a recent human outbreak on a cruise ship visiting the region has caused concern.
The bad news is that hantavirus has a reported human mortality rate of 20-50% and no vaccine has been developed. Further, human-to-human transmission is occurring in this instance, which is relatively rare for the virus (though it is known to be possible for the Andes variant responsible for this outbreak). Unfortunately, the outbreak was not identified until the cruise’s passengers had largely dispersed to their homes around the world, increasing spread.
But, all things considered, the good news outweighs the bad. This is not a new virus or a new mutation – it has been around for centuries. It would already have created a pandemic at some point in its history if it was capable of doing so. Five earlier outbreaks over the past 31 years were all quelled.
For those who remember their pandemic math, its basic reproduction number (R0) is somewhere in the range of 0-2. This is quite low, with the implication that only minor behavioural modifications are necessary to arrest its spread. The hantavirus is usually not contagious until just before symptoms appear and because of its long incubation period (2-8 weeks), it cannot spread very quickly.
In contrast, and at risk of surfacing bad memories, COVID-19 had many problematic attributes:
It was highly contagious, replicating in the upper airways and transmitting through the air. Its R0 was as high as 10 – nearly impossible to stop even with extreme precautions.
The coronavirus family (of which COVID-19 is a member) is far more adept at mutating than hantaviruses due to the former’s large RNA genomes. COVID-19 became significantly more optimized for humans over the span of the pandemic.
Infected people were contagious well before their symptoms appeared. Some were asymptomatic or had only mild symptoms and yet could still readily pass it along to others.
COVID-19’s short incubation period allowed for exponential growth in the number of cases.
COVID-19 was not as deadly as the hantavirus (or ebola, discussed next). But when combined with its incredible ease of transmission it was nevertheless extremely problematic because it reached so many people. It was also quite deadly for older people and those with compromised immune systems.
Ebola
Ebola sporadically flares up in Sub-Saharan Africa, seemingly emerging from fruit bats. The latest outbreak is ongoing and suspected of killing at least 130 people in the Democratic Republic of Congo and spilling over into Uganda.
The bad news is that Ebola has a high human fatality rate of 25-90%, and the current strain is less well known than others. Some variants have ready-made vaccines, but this one does not. The host countries were slow in identifying this outbreak, so it was relatively advanced before containment efforts began.
The good news is that this is believed to be an existing variant of Ebola, not a new mutation. This variant may actually have a slightly lower fatality rate than some of the others. There have been 40-50 outbreaks since Ebola was first discovered in 1976, all ultimately contained. The biggest one was in 2014-2016, reaching U.S. shores and spreading briefly within a U.S. hospital.
Fortunately, as with the hantavirus, Ebola is not especially contagious. It has a basic reproduction number of 1.5-2.5. It does not transmit via the air, instead requiring close contact with bodily fluids. It often infects medical care personnel who work closely with the sick. Traditional African burial practices, which involve mourners touching the deceased, are thought to be a contributor to its spread. Countries are now banning such practices during this outbreak.
Bird flu
While not in the news, those fretting about a pandemic should arguably pay closest attention to H5N1 bird flu. As its name suggests, it spreads primarily in birds. While this used to occur seasonally, it is now constant and widespread. It has also spilled over into a variety of mammals including cows, with the occasional human infection. The human mortality rate is around 50%.
It isn’t presently capable of transmitting efficiently between humans. But it is a form of influenza and so has the theoretical capability of mutating in a way that does permit easy spread. Fortunately, there is already an H5 vaccine, but it might need to be modified and would certainly need to be mass-produced if it became problematic among humans.
Broad conclusions
Any pandemic concerns are probably overblown at present. Financial markets appear to agree – the stocks of vaccine-makers briefly rose on the news but have since retreated back to prior levels. Probability markets assign just a 6% chance of a hantavirus epidemic in 2026, an 8% chance of an Ebola pandemic, and a 12% chance of any pandemic in 2026. Note that the term “pandemic” applies at a threshold well below what the world experienced in 2020.
Further, mRNA technology has now advanced to the extent that developing a vaccine against the next scourge could be accomplished quite quickly. Suffice it to say that a pandemic in the year ahead is not central to our economic forecasts and indeed not even on the periphery.
-EL
Tariff potpourri
Global trade growth accelerating
U.S. tariffs that have been in place for more than a year are reshaping rather than restraining global trade flows. In fact, global trade volumes were up 7% year-over-year in January and February. Outside of post-recession rebounds, that is the fastest pace of trade growth since 2006 (see chart).
This growth has been driven by trade flows outside of the U.S., which were up 9% year-over-year. Tariffs have weighed on U.S. trade volumes, though a 5% year-over-year decline in the first two months of 2026 is exaggerated by pre-tariff front running a year ago. But even compared with two years earlier, ex-U.S. trade volumes have been growing 1.6x as fast as U.S. trade.
This challenges the narrative that globalization is slowing or going into reverse. Countries have been able to find alternative destinations for products that were previously exported to the U.S. and a number of high-profile trade deals were struck last year and early this year in an effort to further diversify trade flows.
Global trade growth is rising, U.S. is falling
As of February 2026. Shaded area represents U.S. recession. Sources: CPB Netherlands Bureau for Economic Policy Analysis, Macrobond, RBC GAM
Section 122 ruling
The U.S. Court of International Trade (CIT) once again ruled against the Trump administration, this time saying its Section 122 tariffs are unlawful. Section 122 of the Trade Act of 1974 is the legal foundation for the 10% baseline tariff introduced in late February after the Supreme Court struck down the administration’s IEEPA tariffs. It accounted for about 2.7 ppts of the 6.8% effective tariff rate levied on U.S. imports in March.
Section 122 permits tariffs of up to 15% for up to 150 days to address balance of payments issues. But in a 2-1 decision, the CIT ruled that the large trade deficit, current account deficit, and negative net international investment position cited by the administration don’t constitute a balance of payments deficit within the context of the original legislation (we discussed this issue back in March).
The CIT’s decision only applied to the plaintiffs in the case, unlike its IEEPA ruling last year which imposed a universal injunction. As with the IEEPA decision, the Department of Justice appealed to the Federal Circuit court which temporarily paused the lower court’s decision while it evaluates the case. The 10% baseline tariff remains in place and we’re unlikely to see a final ruling before the tariffs are set to expire on July 24.
The court case might not impact effective tariff rates – after July, that will largely be dictated by the outcome of Section 301 investigations (more on that below). But the final ruling will set a precedent regarding the president’s tariff powers. These were curtailed by the Supreme Court’s IEEPA ruling and could be further constrained if the lower court’s ruling is upheld.
It could also determine whether tariff revenue already collected under Section 122 (an estimated $25 billion as of the CIT’s ruling) will ultimately have to be refunded to importers. This process is now underway for IEEPA tariffs.
IEEPA tariff refunds starting to flow
Following the Supreme Court’s IEEPA ruling in February, the CIT ordered U.S. Customs and Border Protection (CBP) to repay $166 billion in tariffs that were unlawfully collected under that legislation. In April, CBP launched the first phase of its refund process, approving $35.5 billion in claims covering 8.3 million shipments as of May 11. More than one-third of those refunds have already been issued. Another 6.8 million shipments have been deemed eligible, so total refunds under the initial phase could exceed $60 billion.
With the low-hanging fruit being processed first, the pace of refunds is likely to slow. It’s doubtful the full $166 billion will be claimed and returned. Still, more than $60 billion in refunds would add to the government’s budget deficit. The CBO has already projected this at $1.85 trillion this year prior to the Iran conflict (which will further inflate the deficit via increased defense spending).
Refunds will be a nice cash injection for importers who paid the tariffs, but we expect the economic multiplier will be relatively small. Most refunds will go to businesses rather than consumers.
Section 301 investigations
Seeking a more durable and legally sound replacement for IEEPA and Section 122 tariffs, the Trump administration announced Section 301 investigations into global manufacturing overcapacity and failure to prohibit imports produced with forced labour – “unfair” trade practices that hurt U.S. businesses – in 60 trading partners covering 99% of U.S. imports. Those investigations were launched in March with public hearings taking place in late April and early May. Section 301 investigations must be completed within 12 months, but the Office of the United States Trade Representative is seeking to expedite the process, allowing tariffs to be put in place by July 24 when Section 122 tariffs expire.
Section 301 tariffs have generally been upheld by the courts and seem to stand on more solid legal ground than IEEPA and Section 122 tariffs. The U.S. has levied Section 301 tariffs on some Chinese imports since 2018, collecting $35.5 billion in tariff revenue last fiscal year. But given the potential breadth and scale of new Section 301 tariffs – assuming the administration intends to largely replace lost IEEPA revenue – they could still be subject to legal challenges, particularly given the courts’ willingness to push back against the White House’s tariff agenda thus far.
Our working assumption is that Section 301 tariffs will be imposed this summer. This will lift the U.S.’s effective import tariff rate (the yellow line below, based on actual tariffs paid rather than announced tariff rates) back up toward the lower end of the 9-11% range that prevailed prior to the IEEPA ruling.
IEEPA ruling and Section 122 replacement nets out to a lower effective tariff rate
As of 05/21/2026. Sources: Bloomberg, RBC GAM
Tariff threats and U.S.-China talks
With the IEEPA ruling hampering President Trump’s ability to quickly impose new tariffs, U.S. trade policy uncertainty has receded to levels last seen prior to the November 2024 presidential election. Still, a few tariff threats have been issued in recent weeks.
Trump threatened to increase tariffs on auto imports from the European Union (EU) to 25% in May in response to the trade bloc not complying with a deal struck last year. He subsequently delayed the threatened tariff hike, giving the EU until July 4 to ratify the trade deal and threatening “much higher” tariffs if it didn’t do so. It was unclear whether those tariffs would apply to autos or more broadly.
Ratification of the deal was delayed by Trump’s threats to annex Greenland earlier this year and again following the Supreme Court’s IEEPA ruling. In response to Trump’s threats, the European Council recently agreed on legislation that would eliminate tariffs on U.S. industrial goods – living up to a key commitment in last year’s deal. But the European Parliament still needs to give its stamp of approval.
In April, Trump also threatened 50% tariffs on countries supplying Iran with military weapons. While no countries were named, it was seen as a warning against China and Russia, though the two countries have denied sending weapons to Iran recently. Trump said the tariffs would take effect immediately, but no new levies have been imposed and it’s unclear how the administration would do so without IEEPA. Even adding to existing Section 301 tariffs on Chinese imports would require a consultation period.
Trump’s two-day summit with President Xi failed to extend the two countries’ one-year trade truce that is scheduled to expire in November. But the talks did result in some easing in trade tensions:
China agreed to purchase 200 Boeing jets – its first major order in nearly a decade – and resume some imports of U.S. beef.
The U.S. will lift automatic detention measures on Chinese dairy products.
The two countries will also work toward reciprocal tariff reductions on $30 billion worth of imports.
There was no further progress on easing China’s restrictions on rare earth exports or U.S. export controls on cutting-edge GPUs.
-JN
Chinese housing revival?
China’s housing market has been exceptionally weak in recent years. Several of its builders tumbled into financial distress. Home sales are less than half what they were five years ago (see next chart). And the country’s new and existing home prices continue to fall (see subsequent chart).
Housing supply in China remains elevated while demand sits at multi-decade low
As of April 2026. Sources: China National Bureau of Statistics (CNBS), Macrobond, RBC GAM
China’s home prices falling at a slower rate
As of April 2026. Sources: China National Bureau of Statistics (CNBS), Macrobond, RBC GAM
To the extent real estate was long the most popular savings vehicle for Chinese households, this had adversely impacted household wealth and discouraged consumer spending. The lack of housing and consumption growth has in turn significantly dimmed China’s overall rate of economic growth.
But what if China’s housing market were starting to stabilize? This is hardly certain, but there is undeniably a fresh green shoot: new and existing home prices in China’s Tier 1 cities – its biggest metropolises – are again rising (see next two charts).
New home prices rise in major cities across China
As of April 2026. Sources: China National Bureau of Statistics (CNBS), Macrobond, RBC GAM
Resale home prices also rise in major cities across China
As of April 2026. Sources: China National Bureau of Statistics (CNBS), Macrobond, RBC GAM
It is not uncommon for Tier 1 cities to act as leading indicators, with the home prices in smaller cities moving higher with a lag. Consistent with this, home prices in lower-tier cities are starting to fall less quickly.
But there are tempering factors:
There have already been a few false starts over the past several years where Tier 1 city home prices briefly rose before slipping back beneath the waves.
While Chinese housing affordability has improved quite significantly), this is still quite expensive relative to other countries (see next chart). It now takes 25 years of median household income to purchase the median home, down from 38 years.
Housing affordability has improved in China but still high compared to other countries
Price-to-income ratio measured as median apartment prices vs. median household disposable income. Sources: Numbeo, Macrobond, RBC GAM
The story is similar at the city level – housing affordability for China’s Tier 1 cities is much worse than for other world cities, even if it has improved (see next chart).
Housing affordability has improved in Chinese cities but still among the world’s most unaffordable
Price-to-income ratio measured as median apartment prices vs. median household disposable income. Tier 1 cities include Beijing, Guangdong, Shanghai and Shenzhen. Tier 2 cities include Chengdu, Chongqing, Dalian, Hangzhou, Nanjing, Qingdao, Shenyang, Suzhou, Tianjin, Wuhan, Xi’an and Xiamen. Sources: Numbeo, Macrobond, RBC GAM
Chinese buyers may have been willing to tolerate ultra-high prices when they felt reasonably assured that the price would continue to rise during their stewardship. It may be a different story today, especially with a shrinking population that should theoretically result in deflationary home price pressures over the long run.
On the other hand, the country’s rural-to-urban migration continues. China could opt to raze its lower quality housing stock, and Chinese households no doubt wouldn’t mind more square footage per apartment. Affordability is also capable of continuing to improve even if home prices keep inching higher, thanks to robust household income growth.
Let us watch to see if China’s housing market is truly turning the corner. If it is, the country might be more capable of hanging onto real GDP growth approaching 5% per annum for longer than otherwise imagined.
-EL & VL
Earnings acceleration propels equity bull market
Global stocks have been especially resilient so far this year. As many indices climb to fresh records, investors may be wondering whether the rally can be sustained. Recent gains have been supported by unusually rapid profit growth and profits continue to accelerate. Massive and ever-increasing artificial intelligence (AI) related investments have meaningfully brightened the earnings outlook in the near term.
While the durability of AI-related capital expenditures remains uncertain, stocks can continue climbing as long as earnings expectations keep rising.
Strong earnings outlook
Profit growth was surprisingly strong in the first quarter. Analysts expect earnings increases to accelerate in the quarters ahead. In the Q1 2026 earnings reporting period, the incidence of S&P 500 company results exceeding analysts’ estimates rose to 84%, which is the highest reading since mid-2021 (see next chart).
Profits grew by an impressive 19.1% in the first quarter compared to the same quarter last year. Analysts expect that rate to keep improving to as much as 24% by the end of the year (see subsequent chart). Earnings growth is expected to moderate only slightly in 2027 and into 2028 to rates that are still expected to remain in double digits.
S&P 500 companies reporting results above consensus forecasts
As of 05/24/2026. Sources: Refinitiv, RBC GAM
S&P 500 Index earnings per share expected to improve to the end of 2026
As of 05/22/2026. Sources: Thomson Reuters, RBC GAM
Positive revisions are unusual
The positive revisions to earnings estimates so far this year have been extreme relative to history. The next chart plots the typical path of S&P 500 consensus earnings-per-share (EPS) estimate revisions over the span of a year. The shaded area encompasses a full standard deviation from the norm. Estimates usually fall throughout the year as analysts tend to start out overly optimistic.
This year, however, estimates are going up rather than down, and at a pace that sits significantly above the one-standard-deviation threshold. This sizeable upgrade to earnings forecasts argues the outlook is much rosier than previously imagined.
This year’s positive revisions to S&P 500 consensus earnings estimates have not followed the typical path
As of 05/24/2026. Based on monthly data since 2026. Sources: Bloomberg, RBC GAM
Sector drivers
The main driver of these profit upgrades is coming from the Information Technology sector, which is benefitting from the enthusiasm for AI. The following table lists the earnings growth expectations for each sector in the S&P 500, as well as a comparison to expectations at the beginning of the year.
Growth of S&P 500 Index earnings per share is expected to remain in double digits
As of 05/22/2026. Sources: Refinitiv, RBC GAM
While the Energy sector saw the biggest upgrade in percentage point terms given the surge in oil prices due to the war, that sector represents a relatively small 3.4% share of the S&P 500. Information Technology, which is a far larger 37.4% share of the index, experienced the second-largest earnings upgrade of all the sectors, bringing its 2026 earnings growth estimate to a gigantic 50.6% versus the already-strong 30.8% that was expected at the start of the year. Other sectors with noteworthy earnings upgrades were Communication Services and Materials.
Despite the relatively narrow base of profit upgrades, the intensity of the increases that did occur have propelled index level earnings growth expectations to 24.5%. That’s up from the 15.6% that prevailed less than five months ago.
Economic cycle says…
This earnings acceleration is especially fascinating because it is uncommon to see such rapid earnings growth at the current stage of the economic cycle. One way to assess the economic cycle is using the year-over-year change in the Conference Board’s U.S. Leading Economic Index (see next chart). This index is comprised of various economic indicators. It currently suggests the U.S. economy is in an early acceleration phase (year-over-year change above -4% and rising).
Historically, earnings have grown an average of 2% during this phase, with future expectations rising an average of 7% per year (see subsequent chart). Since 1991, earnings have risen more than 20% just 5% of the time when the U.S. economic cycle was in early acceleration.
Moreover, across all phases of the economic cycle, earnings growth has averaged 8% per year and only risen at a rate of more than 20% around a tenth of the time. In the context of these statistics, this year’s expectation of 24.5% earnings growth is a major positive outlier.
Economic Index suggests U.S. economy is in an early acceleration phase
As of 03/31/2026. Sources: Conference Board, Bloomberg, RBC GAM
This year’s expectation of 24.5% earnings growth is a major positive outlier across all phases of the economic cycle
As of 05/22/2026. Sources: Conference Board, Bloomberg, RBC GAM
That said, despite the “early acceleration” designation as per the Conference Board’s leading indicator, the current economic expansion is not exactly young. It is extending into its sixth year. Another way of framing the economic cycle – via our own business cycle scorecard – makes the case that this could actually be a late-cycle or at least mid-cycle moment (see next chart).
It is notable that more advanced phases of the economic cycle as per the Conference Board’s leading indicator – “late acceleration” and “early deceleration,” see above – are often associated with stronger earnings growth (refer back to table above). Perhaps the market is already in one of these phases.
Our business cycle scorecard suggests we’re in late-cycle or at least mid-cycle
As of 05/08/2026. Calculated via scorecard technique by RBC GAM. Source: RBC GAM
Strong earnings outlook = strong stock market
Regardless of the exact point in the cycle, history implies that a strong earnings outlook bodes well for stock market returns. The next chart plots the 1-year returns of the S&P 500 alongside the annual change in earnings estimates. The two lines on the chart move closely together and suggest a strong relationship between the series.
Moreover, the table in the chart lists return statistics based on the trajectory of earnings estimates in different environments. When growth in earnings was expected over the year ahead, stocks gained an average of 13.6% per year and rose 88.9% of the time. If earnings were forecast to rise rapidly (i.e. more than 10% annually), then the stats were even better. In those situations, stocks averaged 16.7% yearly returns and gained 94.2% of the time.
Conversely, when earnings were expected to fall, stocks declined an average of 4.5% per year and rose only 45% of the time.
Strong S&P 500 index earning outlook suggests strong stock market returns
As of 05/22/2026. Sources: Bloomberg, RBC GAM
The persistence of AI-related capital expenditures remains an open question that could determine whether this profit acceleration proves durable or transitory. Alongside this, it will be important to monitor the breadth of earnings contribution from various sectors and industries for any sign that earnings growth is broadening. At this time, though, as long as earnings expectations continue their upward trajectory, the stock market rally is fundamentally well supported.
-ES