With contributions from Vivien Lee, Aaron Ma and Eric Savoie

The latest

There is an awful lot happening in the global economy right now, much of it challenging. Tariffs are on the rise, as discussed in the last MacroMemo. So are bond yields – addressed in another section of our last report. Oil prices are continuing to go up – discussed next. And AI concerns have returned, though of the “will it take over the world?” variety, rather than the prior “will spending meet expectations?”

On the net, this all paints a somewhat more difficult macro picture, though we still see enough momentum and macro tailwinds to believe the economy can make it through this adversity.

-EL

Inflation, yields and the Federal Reserve

Given rallying oil prices, it is not a great surprise that the latest U.S. inflation data came in hot. The producer price index is now rising at 5.4% year-over-year (YoY). Even the ex-food and energy version is up by 4.6% YoY. The U.S. CPI (Consumer Price Index) landed at an elevated +3.4% YoY, with a fourth big monthly gain out of the last six months (see next chart).

Figure 1. U.S. inflation accelerates

As of August 2026. Sources: U.S. Bureau of Labor Statistics (BLS), Macrobond, RBC GAM

Core inflation is not as elevated, up by a less extreme 2.4% YoY. But there too, the monthly sequence is unfriendly. We saw a 0.3% month-over-month (MoM) rise in August and 10 of the last 12 monthly core prints rose by more than the rate needed to achieve a 2.0% annual growth rate (see next chart).

Figure 2. U.S. core inflation has been running above the monthly rate consistent with the Fed's 2% annual target

As of August 2026. Sources: U.S. Bureau of Labor Statistics (BLS), Macrobond, RBC GAM

And now oil prices have risen further in September, to US$101 for West Texas Intermediate (WTI) and US$106 for Brent crude. This is due to a deterioration in Middle East stability. The U.S. and Iran have continued to exchange sporadic fire with one another, hitting ships and tactical assets on both sides.

Meanwhile, the conflict has expanded somewhat as the Iran-allied Houthi’s captured a series of strategic islands in the Red Sea, and Saudi Arabia’s crucial East-West Crude Oil Pipeline has been temporarily sidelined by drone strikes.

This all points to further inflation heat in the September data.

That, alongside a multi-year history of inflation exceeding target within generally healthy economies, is catching the eye of the developed world’s central banks. The European Central Bank (ECB) just tightened by 25 basis points for a second time and the Danish central bank similarly raised rates.

The U.S. Federal Reserve is now squarely in focus and seems similarly inclined to initiate a tightening cycle on September 16. The market now prices this with a high (91%) level of conviction. Recall that three of twelve Fed voters dissented in favour of a rate hike at the last meeting, and that Fed Chair Warsh’s Jackson Hole speech was also hawkish.

If the Fed wants to get going on rate hikes in the next few months, the September meeting probably makes more sense than the October one - both because it provides an opportunity to set out new expectations more formally via the accompanying quarterly dot plot publication, and because it would avoid muddying the water in the days leading up to the midterm elections.

These expectations for tighter monetary policy, combined with more general concerns about U.S. debt, have pushed the U.S. 10-year yield up to as high as 5.00% for the first time since 2023. The 30-year yield has increased to 5.37% – on the cusp of a 20-year high. The rising yield story remains intact, though we posit that a Fed rate hike could go some distance toward stabilizing the long end, it would lower inflation expectations and increase U.S. policy credibility (potentially reflected in a stabilization or narrowing of the term premium).

We take the “under” on the current consensus for four 25bps rate hikes over the next year in the U.S. In our view, the oil shock should de-intensify within that time frame. A modest amount of monetary tightening should be enough to keep the economy from overheating.

-EL

Divergent U.S. and China borrowing costs

Until quite recently, the norm was for Chinese borrowing costs to exceed U.S. ones. After all, China is a less developed country, has a faster real economic growth rate and sports a worse debt rating – all traditional contributors to a higher interest rate.

But all of that flipped over in the past several years. At approximately 5%, the current U.S. 10-year yield is now, remarkably, triple China’s meagre 1.68% 10-year rate (see next chart). That constitutes the biggest negative spread in memory.

Figure 3. U.S.-China yield spreads at record high

Chinese bond yields as of September 14, 2026. U.S. yields as of September 11, 2026. Sources: Macrobond, RBC GAM

Why has this inversion occurred? It is a function of developments in both countries. U.S. yields have risen sharply over the past six years and Chinese yields have fallen gradually since 2018.

U.S. upward yield drivers

The U.S. side of the yield divergence story is well known.

U.S. inflation remains persistently too high. The country’s chronically large deficits are adding to its already elevated public debt load. Now, the Fed appears to be on the cusp of a new tightening cycle - albeit limited, in our view.

Investors may also be building in a larger term premium that reflects, at least in part, concern about the unpredictable and unconventional U.S. public policy environment.

On the demand side, yield-depressing quantitative easing operations are no more and the appetite of international reserve buyers has diminished. On the supply side, there is crowding from AI-related bond issuers and other sovereigns running large deficits.

Finally, and the one major positive in the mix, the U.S. economy is strong, justifying a relatively robust bond yield.

Chinese downward yield drivers

Meanwhile, China is experiencing several downward yield pressures.

The country has flirted with deflation and is experiencing chronically low inflation, holding down yields.

The Chinese economy has recently underperformed, also depressing yields.

In response to the two factors above, there is the potential for further Chinese monetary easing. China’s policymakers are also sensitive to the fact that the housing bust has been quite painful. High interest rates would make this even worse.

China’s fiscal situation is trickier to assess. On the surface, its deficit target of 4% of GDP seems tame compared to the U.S. equivalent of approximately 6% of GDP.  But that excludes Chinese government-adjacent borrowing such as special local government bond issuance and other government-sponsored activities. The most expansive definition of the Chinese deficit is as high as 12% of GDP – a startling figure.

Then again, no one includes U.S. municipal bond issuance in the U.S. federal deficit. And, ultimately, China runs a large current account surplus, which means that its private-sector saving is more than enough to cover its public-sector borrowing. Further, Chinese capital controls give domestic investors few alternatives. Thus, the Chinese fiscal situation is not great, but also probably isn’t exerting a large upward pressure on yields.

In an effort to summarize all of these U.S. and Chinese drivers, one might cheekily observe that it is primarily bad forces that are pushing U.S. yields higher, and also primarily bad forces that are pushing Chinese yields lower. So, contrary to what one might imagine, this interest rate divergence does not reflect virtue versus vice, but instead one type of problem versus another type of problem.

Implications

Even if the underlying drivers themselves are not worth celebrating, is it better to have low yields or high yields? Unsurprisingly, the advantage accrues to China and its low interest rates:

• China’s borrowing costs are much lower – this allows the government and risk-taking businesses to save on debt-servicing costs.

• Even more importantly, China’s effective hurdle rate for investment decisions is lower. It is easier to justify infrastructure and private-sector investments in China. To be sure, (largely) Chinese lenders are earning less on their loans, but the economic literature is clear that lower interest rates are ultimately a net positive for growth despite the existence of both winners and losers.

Under normal circumstances, one might add a third driver to the list: the potential for a softer Chinese exchange rate. But the yuan has bucked this trend, appreciating due to China’s large trade surplus and undervalued currency.

-EL

Data centre backlash

Growing opposition to data centre construction in the U.S. has caused billions of dollars of projects to be delayed or canceled. It has also prompted some states to pause new construction or impose additional regulatory hurdles. This adds to a list of factors that could increase costs and extend timelines for the AI buildout. These include electricity, memory chip and labour and equipment shortages. Separately, there are growing calls from tech leaders and policymakers to slow the pace of AI development amid increasing concern about model safety and risks to humanity.

Globally, opinions about AI are mixed (see the chart below). A recent Ipsos survey found respondents in Asia are generally more excited about the technology. The major English-speaking economies are more nervous and Europeans have mixed views (less excited and less nervous than those other regions).

Figure 4. Global views on AI are mixed

As of September 10, 2026. Sources: Ipsos AI Monitor, RBC GAM

But in most countries, respondents are less excited about AI products and services in 2026 compared to 2025. As AI becomes increasingly capable and adoption increases, concerns about labour disruption, cybersecurity, privacy, corporate power, disinformation and even risks to humanity are on the rise.

AI is physically manifest in big data centres that can raise local electricity costs, use significant amounts of land and water, generate emissions and noise, and don’t create many local jobs after construction is complete. The U.S. remains the epicentre of the data centre buildout with 45% of global operating capacity and 60% of planned capacity. Yet several surveys highlight Americans’ growing concern regarding AI and data centres:

• A 2026 Gallup poll showed 39% of Americans think AI will do more harm than good, up from 31% in 2025. Only 9% think it will do more good than harm. Nearly 80% of respondents expect AI will reduce the total number of jobs in the economy, and most don’t trust businesses to use AI responsibly.

• A 2025 Pew Research survey of 25 countries found Americans were the most concerned and least excited about the increased use of AI in daily life – a view that is now shared across party lines. Another Pew survey showed Americans are split on whether they trust the U.S. government to effectively regulate AI.

• A 2026 Ipsos survey found only 38% of Americans think products and services using AI have more benefits than drawbacks, compared with 55% globally. Relative to the global average, Americans were less likely to think AI will make the economy and job market, and their own jobs and health better in the next 3-5 years.

• A recent YouGov poll found half of respondents think data centre construction is bad for the country and don’t think local facilities are desirable for jobs and tax revenue. More than 60% think new data centres increase electricity costs and would oppose data centres being built in their communities.

• Similarly, a recent survey by Heatmap Pro and Embold Research found 61% of Americans would strongly oppose a data centre being built near their home. That’s up from 54% three months earlier and 24% a year ago (see chart below).

• A 2025 YouGov poll showed nearly 40% of respondents are very concerned that AI could eventually pose a threat to humanity, A similar share is somewhat concerned.

Figure 5. Americans’ opposition to data centres is growing

As of September 10, 2026. Sources: Heatmap Pro, Embold Research, RBC GAM

That last point has become more salient following several incidents in which developers have lost control of rogue AI systems AI researchers have also issued fresh warnings that the technology could pose an existential threat to humanity.

CEOs of the top model developers are now calling for a coordinated slowdown in AI development to allow regulation and risk prevention measures to catch up. But the Trump administration seems more concerned about the risk of ceding AI leadership to China, and some critics are worried about regulatory capture by big tech companies.

Beyond safety risks, Americans’ souring views toward AI are reflected in growing local resistance to data centre construction. Data Center Watch reports 75 projects worth US$130 billion were blocked or delayed in Q1 2026 alone – about as much as in all of 2025. But some of those projects were likely speculative, without firm customer commitments.

At a more local level, the Wall Street Journal reports dozens of cities and counties have temporarily paused new construction.

The Trump administration remains staunchly supportive of the AI buildout. President Trump even suggested communities that oppose data centres will end up being “backwards and poor.” But zoning, permitting and utility regulation takes place at the state and local level, and data centres are becoming a key issue in upcoming congressional and gubernatorial elections.

Texas, Pennsylvania and New York rank first, fourth and eleventh respectively among states in the number of data centres operating, announced or under construction. Governors in these states have now announced stricter regulatory requirements, pauses or even moratoriums on new data centre projects. Candidates in other states have proposed similar measures.

At a more local level, the Wall Street Journal reports dozens of cities and counties have temporarily paused new construction. And several state legislators are now reviewing tax breaks for data centres. Ohio, which ranks sixth in data centres and has the lowest effective tax rate for data centres, is considering scrapping US$2.5 billion in annual incentives.

It should be noted that legislated and proposed pauses or moratoriums on new data centres don’t halt projects that are already under construction. And in most cases, lawmakers aren’t aiming to permanently stop new data centres from being built. Rather, they’re buying time until new rules and conditions can be put in place to address voters’ concerns.

Another poll from Global Strategy Group found Americans are 2/3 more likely to support a data centre being built in their community if it is powered by clean energy rather than fossil fuels.

Politicians’ tough talk on data centres might soften after the November elections. Texas’s data centre pause is viewed by some as campaign posturing and could end later this year. Even if the facilities are unpopular, lawmakers might be enticed by tax revenue, infrastructure investment, temporary construction jobs, contracts for local businesses and lobbying.

Building data centres somewhere in the U.S. is necessary from an AI leadership and data sovereignty perspective, which is central to the Trump administration’s supportive stance. A survey by Pew Research found most respondents think it is important for the U.S. to be the world leader in developing AI. But a plurality think China is more advanced than the U.S., and most Americans don’t trust the Chinese government to effectively regulate AI.

What could mitigate local resistance? An Embold Research poll found 37% of respondents are more likely to support a local data centre if a substantial portion of its power is generated on-site. Some states are imposing “bring your own power” rules to protect ratepayers or closely scrutinizing data centre energy consumption. Another poll from Global Strategy Group found Americans are 2/3 more likely to support a data centre being built in their community if it is powered by clean energy rather than fossil fuels.

Given prevailing sentiment, stricter AI regulation will likely play well with voters in the leadup to the 2028 elections.

Technological developments could also help. Ongoing improvements in GPU efficiency and packaging, more efficient AI models, and edge (on device) inference would reduce required data centre square footage. It may prove practical to engage in some amount of geographic arbitrage, not just between states but across nearby countries.

Further down the road, space-based data centres could get around the issue of local opposition. However, as we discussed here, that technology is still uncertain and perhaps decades from commercial viability.

Overall, the unpopularity of data centres is a growing challenge for hyperscalers and adds to a list of obstacles to the data centre buildout. But actual moratoriums aren’t as restrictive or widespread as headlines might suggest.

We think stricter regulation and short-term pauses are more likely than widespread bans on data centre construction. That could add to costs and timelines for new projects but might not substantially slow the pace of investment or growth in AI compute given the pipeline of projects that are already under construction or planned in states where opposition is less fierce.

At the federal level, AI regulation is more likely to focus on model safety, privacy and cybersecurity rather than data centre construction. Nonetheless, how the data centre backlash evolves post-midterms bears monitoring. Given prevailing sentiment, stricter AI regulation will likely play well with voters in the leadup to the 2028 elections.

-JN

S&P 500 seasonality and U.S. midterms

The U.S. midterm elections have been somewhat overshadowed this year by the Iran conflict, the AI CapEx boom, and tariff uncertainty. But November 3 is now less than two months away and prediction markets signal a significant shift in Washington. Kalshi gives Republicans just a 16% chance of holding the House, while the Senate sits at roughly a coin flip (53% Republican).

A divided government in which the Democrats control both the House and Senate is now deemed the most likely scenario. We penned a preview of the midterms nearly two months ago. What follows is additional perspective through a stock market lens.

History explains why investors may want to pay close attention to the election. In 31 midterm elections since 1902, the president's party has lost House seats in all but four instances, with a median loss of 28 seats for Democratic presidents and 27 seats for Republican presidents (see chart below). Only Theodore Roosevelt in 1902, Franklin Delano Roosevelt in 1934, Bill Clinton in 1998 and George W. Bush in 2002 bucked the trend.

This erosion of legislative control injects policy uncertainty, making it harder for the administration to advance its agenda on taxes, trade and regulation. Markets tend to price in that prospect and the associated uncertainty ahead of the vote.

Figure 6. President’s party has lost House seats in most U.S. midterm elections

As of September 2026. Sources: Ned Davis Research, RBC GAM

The seasonal data reinforces the view that equity investors often worry ahead of midterm elections. The chart below overlays the 2026 S&P 500 (blue line) against the average midterm-year path since 1930 (yellow line), both indexed to 100 at the start of the year. The average shows a soft patch in midterm years, beginning in late summer and extending into October.

Figure 7. S&P 500 often hits a soft patch as midterms approach

As of September 8, 2026. Sources: Bloomberg, RBC GAM

This year, the S&P 500 has thus far bucked that historical trend, sharply outperforming that average by rallying more than 20% from its March lows. Of course, the market's year-to-date strength does not make it immune to the midterm-related headwinds that could continue to blow through early November.

The table below breaks down the average S&P 500 path from September 1 through year-end, based on the outcome of the midterm election. Across the 24 midterms since 1930, the average path shows a 2.1% maximum drawdown between September 1 and Election Day, with the trough occurring around October 1. From that trough, the index rallies an average of 5.6% into year-end, suggesting that markets tend to recover once the election outcome is known and the policy picture becomes clearer.

The equity market’s reaction to the election has, in the past, differed marginally based on the outcome. A divided government, in which the opposing party wins full control of Congress, has historically produced the best post-election recovery, with a 7.6% rally after a modest 2.2% drawdown. Again, this is currently the most likely scenario. A unified outcome (the same party controls both the presidency and Congress) is also reasonably attractive, though unlikely to happen this time.

Figure 8. A divided Congress has often produced the best post-election recovery

Includes 24 midterm elections since 1930. Figures are calculated from the average S&P 500 path for the midterm election scenarios. Divided means presidency and Congress are controlled by different parties. Unified means presidency and Congress are controlled by the same party. Sources: Bloomberg, RBC GAM

The notable outlier is a split Congress, in which the House and Senate are controlled by different parties. In those instances, the average drawdown is a steeper 5.0%, and the subsequent rally of just 3.0% fails to fully recover the lost ground by year-end. Kalshi currently prices the split-Congress scenario at roughly 39%, making it the second-most-likely outcome.

Although quite a variety of macro and market forces are currently holding sway over the stock market and may remain the dominant considerations, this midterm-election seasonality is worth monitoring. To summarize, the S&P 500 often underperforms leading up to the election and then outperforms in subsequent months.

-AM

Diagnosing and upgrading Canadian productivity

Canadian productivity growth has been top of mind in recent years. This was initially due to how poorly Canada was faring on that front., Now it’s in the context of a flurry of policy changes intended to lift it out of the doldrums.

It is fitting to publish this report today, as it coincides with the much-anticipated Canadian Investment Summit that is happening in Toronto. Here, a diverse group of deep-pocketed international investors are being presented with a menu of Canadian infrastructure and resource investment plans.

Canada’s productivity backstory

The backstory is that Canada’s recent productivity performance has been abysmal. Output per capita is currently no better than it was four years ago (see next chart).

Figure 9. Canadian GDP per capita has stagnated

As of Q2 2026. Sources: Statistics Canada, Macrobond, RBC GAM

This recent swoon was intense, but really just the latest chapter after decades of anemic productivity results. As a result, Canada has tumbled from a level of prosperity that was nearly on par with the U.S. in the early 1980s to just ~73% of the U.S. today (see next chart).

Over this period, Canadian productivity rose by just 0.9% per year, versus a 1.6% annualized gain in the U.S. Reframed, U.S. productivity growth was 78% better than Canada over the period.

Figure 10. Canada has become less productive relative to the U.S.

As of Q1 2026. Total economy real output per hour worked converted into Purchasing Power Parity-based (PPP-based) U.S. dollars. Shaded area represents U.S. recession. Sources: National Bureau of Economic Research (NBER), Haver Analytics, Macrobond, RBC GAM

At the sub-sovereign level, using 2024 data, it is striking that of the 60 states and provinces spanning the U.S. and Canada, only two Canadian provinces place in the top half of the GDP per capita rankings (Alberta is twentieth and Saskatchewan is twenty-fifth). British Columbia ranks just forty-sixth, Ontario is merely forty-eighth, and six Canadian provinces are clustered at the very bottom of the rankings. Ontario ranks below Louisiana and Quebec lands below Alabama.

It should be noted that the weakness of the Canadian dollar contributes heavily to these results, whereas more conventional productivity comparisons – including the national-level comparison, earlier – set this aside since the cost of living does not fluctuate one-for-one with the currency.

Lest we overstate Canada’s productivity woes, it should be mentioned that the U.S. is truly the exceptional party here. Most developed countries have badly underperformed the U.S., and Canada is actually middle-of-the-pack from the standpoint of productivity growth over the past quarter century (see next chart).

Figure 11. U.S. outpaces other G7 countries in productivity gains

As of Q2 2026. Productivity measured as total economy/business sector non-agricultural industry real output (in local currency) per hour of all persons. Shaded area represents U.S. recession. Sources: Have Analytics, Macrobond, RBC GAM

Still, Canada leaves something to be desired. For one, Canada’s productivity level (as opposed to growth rate) is toward the bottom of the international standings (see next chart). Second, the U.S. is arguably Canada’s best comparator. The two countries are neighbours and harbour many similarities, notwithstanding recent tensions.

Figure 12. Canada’s productivity level is low among international peers

Canada, Japan, South Korean, and UK as of 2024. 2025 for all other countries. Source: OECD, Macrobond, RBC GAM

Productivity versus competitiveness

But wait – how is it that Canadian workers remain gainfully employed in a competitive international market if they are so unproductive?

As it happens, productivity isn’t the primary determinant of this. Competitiveness is. Emerging economies are far less productive than developed ones yet can enjoy sparkling growth rates and low unemployment rates because they are highly competitive. This is to say, they have low productivity but even lower wages (and often also a cheap exchange rate) that render them highly competitive.

This has also been the story for Canada. We calculate that Canada has a normal level of competitiveness right now because the country’s weak wage growth has matched its weak productivity growth. An undervalued exchange rate has also helped.

Thus, having bad productivity does not necessarily mean high unemployment or an economy in recession. But it does limit how fast workers’ wages can rise, how fast corporate profits can go up, and what Canadians can afford to import (with a depressed exchange rate). Collectively, that’s essentially the financial well-being of the country – a matter of utmost importance.

Identifying the problem

“For every complex problem there is an answer that is clear, simple, and wrong.” – H. L. Mencken.

While tempting to ascribe Canada’s productivity woes to a single thing – “high taxes” is probably the most popular answer – the problem is instead, in our view, the result of a cascade of interconnected factors spanning six categories (see next graphic). These are: acute temporary factors, public policy settings, Canada’s culture, Canada’s economic structure, business decision-making and global forces. But (as per the strike-through in item number seven) we reject the notion that human capital is centrally to blame. To the contrary, it is of an unusually high quality in Canada.

Figure 13. Many factors are causing Canada’s productivity shortfall

Source: RBC GAM as of February 12, 2025

Let us run through each of these drivers.

1) Acute temporary factors

There are several acute temporary factors that help to explain Canada’s woeful productivity performance over the last half decade.

Immigration: In the early 2020s, Canada induced an immigration surge that resulted in astonishing population growth of up to 3.5% per year. While this temporarily added to headline economic demand, it also diluted the capital stock. New workers – regardless of their initial skill level – also tend to initially be unproductive as they learn their new jobs. Further, the country’s immigration policy intentionally targeted lower-skilled immigrants who compositionally dragged Canada’s productivity level downward. With the resulting surge in the labour supply, companies were also disinclined to invest in productivity enhancements.

Lingering pandemic distortions: The pandemic may have lowered Canadian productivity in several ways. Experienced workers fled high-touch service sectors, compromising the productivity of those sectors. Some businesses expanded or refocused with the assumption that certain pandemic distortions would persist indefinitely, but they did not – causing capital misallocation. Many businesses pivoted from just-in-time inventories to just-in-case-inventories, trading productivity for resilience. Finally, Canada had the most persistent work-from-home trend after the pandemic and some economic research would argue this may be hurting Canadian productivity.

Interest rate shock: Borrowing costs are now a lot higher than they had been prior to the pandemic. This, in turn, discourages capital expenditures and thus dampens productivity growth.

High uncertainty: Uncertainty has been high for several years – first, due to the pandemic and its chaotic aftermath, then because of tariffs, and now given the oil shock plus even more Canada-directed tariffs. U.S. public policy uncertainty has spilled over into uncertainty for Canada. This likely discourages long-term investments right now.

2) Public policy failings

The second anti-productivity factor is Canada’s public policy environment. There are several aspects of this.

Taxation: The most common policy complaint about Canada is its high tax rates. There is some truth to this, but it is overstated.

Canada’s top personal income tax rates are indeed quite high, exceeding 50%. This presumably contributes to a brain drain that diminishes Canada’s supply of top entrepreneurs and professionals. But note that the world-leading American hubs for finance, culture and technology are all in similarly high-tax jurisdictions.

Venture capital firm Leaders Fund estimates that just 32% of Canadian-led “high-potential” startups were headquartered in Canada as of 2024. That’s down from 67% from 2015-2019.

Superficially, at least, Canada’s corporate income tax rate is slightly higher than in the U.S. and no better than the middle of the pack among developed nations. But this may exaggerate the problem. A better and more comprehensive corporate taxation metric that factors in a myriad of interconnected taxes and deductions – the effective tax rate on capital – finds Canada may well have the lowest overall rate in the G7 (refer to our earlier MacroMemo on this subject).

Despite this, it is fair to level some criticism:

• Canada’s government has in recent years selectively taxed sectors enjoying particular success – a policy no-no and a warning sign for other businesses.

• Small Canadian businesses are pampered with exceptionally low income-tax rates, and then effectively discouraged from expanding lest they face much higher rates above a certain threshold.

• Until quite recently, the country’s R&D tax credit was generous but only narrowly applicable.

Regulatory: In practice, this is where Canadian policy badly stumbled over the past decade. Canada introduced significant additional red tape that rendered it exceptionally difficult to undertake new resource or infrastructure projects.

Simultaneously, post-Global Financial Crisis banking regulations discouraged bank lending and risk-taking.

In real estate, the approval process and development fees had also become quite unfriendly.

Alongside this, stubbornly large internal trade barriers between Canadian provinces further restrained productivity gains.

Size of government: Canada’s public sector employment grew by a remarkable 21% between December 2019 and today. The initial leap was perhaps understandable in the context of the emergency programs created during the pandemic. But even as those expired, the government worker share of the population never returned to normal (see next chart). To the extent the government services provided today have not risen proportionately to this employment boom, there has been a profound productivity loss.

Figure 14. Share of public sector employment in Canada has risen markedly after COVID-19 shock

As of August 2026. Statistics Canada, Macrobond, RBC GAM

Counterpoints:

To be fair, Canada also has certain public policy strengths that are probably enhancing the country’s productivity versus the U.S. and possibly other peers.

• The Canadian deficit and public debt load are substantially smaller than in many other developed markets. This means less money is being siphoned off to service the public debt.

• Similarly, universal health care is surely a productivity booster for Canada when compared to the U.S. system that requires far more money to achieve worse health outcomes. Canadians should also theoretically be more freely able to move between employers – improving job matching and thus productivity, whereas certain frictions still exist in the U.S. even with Obamacare now implemented.

Still, fundamentally, a small open economy such as Canada’s situated beside a gargantuan economy like the U.S. arguably needs to be materially MORE tax- and regulation-friendly to have a hope of motivating businesses to operate and expand in the smaller market, especially in this era of high scalability and network effects. It is not enough just to match policies.

3) Culture

Canada’s culture could be holding Canada’s productivity back. This is admittedly hard to quantify, so any such discussion exists primarily in the realm of conjecture. Further, some skepticism about this purported productivity headwind is appropriate since the country’s culture has not changed drastically over the decades. This factor would then struggle to explain why the country was on par with the U.S. in the early 1980s but not today.

Still, we posit that there may be some cultural elements to Canada’s productivity shortcomings.

Risk-averse culture: Canada puts an emphasis on consensus-building over debate and disagreement. There is a distaste for the aggression necessary for fierce competition. Failure is stigmatized, in contrast to the “fail fast/learn fast” U.S. model. Financing sources – from banks to venture capital to angel investors – tend to be conservative. Canadian entrepreneurs frequently sell to foreign buyers rather than attempt to scale their businesses domestically.

Complacency: Canada arguably prioritizes the pursuit of equality over the pursuit of excellence, with a fixed-pie mindset rather than a growth mindset. Many of Canada’s largest sectors are dominated by a small number of players, and those players have remained largely the same for generations – in contrast to the U.S., there are few serious disruptors. Canada is not devoid of entrepreneurs but lags the U.S. profoundly when it comes to high-potential startups: the U.S. produced a remarkable 45 times more such startups than Canada in 2024. The Canadian share of the world’s total such startups is also declining.

Antagonism toward success: While great wealth has garnered an increasingly negative reputation around the world, Canada may suffer a particularly strong form of tall poppy syndrome. Ambition, self-promotion and substantial wealth accumulation are all frequently frowned upon. Reflecting this, the country has few celebrity CEOs and it would be difficult to name the wealthiest Canadians as they pursue a low profile.

4) Economic structure

There are certain elements of Canada’s productivity shortfall that are arguably rooted in the very structure of the economy and the country.

Canada has a challenging geography and climate. Cities are far apart. Much of the country is separated from navigable oceans by mountains or enormous physical distances. Relative to the size of the country, there is limited arable/conventionally habitable land. The climate is frequently cold and unforgiving.

The country’s population density is quite low. This makes transportation costs significant between regions and reduces the network effects that dense megacities and their associated regions enjoy.

Canada ultimately has a much smaller population than the U.S., China or the integrated European Union. At a time when scale and network effects are so important for many businesses, this is a significant challenge for Canada.

Canada’s economic structure also includes several long-standing sector tilts that may be hurting productivity or productivity growth:

• Countries with expensive housing markets tend to allocate more capital toward the sector (people must spend a lot of money to buy a home). This starves other more productive sectors of capital. Canada’s housing market has been particularly expensive.

• High oil prices – and by this we merely refer to oil prices above perhaps US$50 per barrel – incentivize oil extraction from less productive sources. This in turn compromises productivity even though the companies involved are making money and behaving perfectly rationally. The OECD finds that resource-rich nations have tended to experience moderately slower productivity growth than their peers since the year 2000, when commodity prices began their supercycle.

• Canada has several concentrated, quasi-oligopolistic sectors including banking, telecom, legal services, air transportation and rail transportation. The supply management system similarly limits competition in the dairy and poultry industries. This may therefore represent a drag on productivity.

• Canada has a disproportionate share of seasonal industries, including construction, tourism, agriculture, fisheries and forestry. While other countries, including the U.S., also have large construction, tourism and agricultural sectors, Canada’s harsh climate renders them more seasonally limited than elsewhere. In turn, the capital and labour associated with these industries is less productive because it is unable to be steadily deployed across the year.

5)  Business decisions

Tempting as it is to place all of the blame for Canada’s poor productivity on the policy regime and basic structure of the economy, businesses also arguably deserve some of the blame.

Yes, there is a tangled web of interconnectedness. Is a business failing to invest in capital because the regulatory environment is too onerous? Is it because the country is too spread out? Is it due to temporary factors such as high interest rates and high uncertainty? Or is it just because Canadian businesses aren’t thinking big enough?

We assert that at least part of the story is the latter. It is hard to fathom that tax, regulatory and other differentials are sufficient to explain why Canadian businesses are – astonishingly – doing less than half as much non-residential investment per worker as American firms, nor why that fraction would be actively falling year after year for decades (see next chart). It is hard to fathom how any business could be at the productivity frontier with half the capital stock per worker of its international competitors.

Figure 15. Business investment in Canada has fallen behind the U.S.

As of 2025. U.S. real private investment converted to Canadian dollars using purchasing power parity-based exchange rates. Sources: C.D. Howe Institute, Haver Analytics, RBC GAM

It struck us that perhaps the U.S. CapEx numbers were being flattered by the enormous sums being invested by a handful of tech giants. But when we controlled for sector mix, Canada somehow fared even worse. And recall that the AI boom wasn’t a thing up until about three years ago – and yet Canadian firms have been underinvesting versus their American peers for far longer than that.

It is a very similar story for Canadian businesses and their research and development (R&D) spending. Canadian R&D expenditures as a share of GDP are barely more than half the U.S. level. Canada also trails the international developed-world norm (see next chart).

For most of the past quarter century, Canada’s R&D share of GDP has been declining.

Figure 16. Canadian businesses are underinvesting in R&D

As of 2025 for Canada, 2024 for France, Germany, Italy, Japan, Spain and U.S., 2023 for Switzerland and UK, 2017 for Mexico. Sources: OECD, Macrobond, RBC GAM

One supposes that Canadian businesses can succeed in the domestic market when their domestic peers are similarly underinvesting, but this behaviour creates a productivity hole that leaves all of Canada relatively poorer for it – via softer profits, lower wages and weaker tax revenues.

6) Global forces

Sixth and finally, a substantial portion of Canada’s slow productivity growth in recent decades – though not its weakness versus other countries – can presumably be attributed to certain global forces.

Large tailwinds drove global productivity growth across much of the 20^th century. One-time structural drivers proved enormously helpful, including rising education levels, rising female labour force participation and rising urbanization. Major general-purpose technologies were also key. The list includes the diffusion of electricity, combustion engines, aviation, mass production, the modern corporation, ratio/television, the air conditioner, plastics, vaccines/antibiotics, the microprocessor, the computer and the internet.

Compared to that halcyon era, it is perhaps not a surprise that global productivity growth itself has been comparatively slower over the past 15 years given relatively fewer new general-purpose technologies.

The Canadian productivity outlook

Happily, there is reason to think Canadian productivity growth should improve somewhat. Part of the reason is that some modest revival is already visible. Productivity in Canada is already starting to edge higher again after an earlier period of decline (see the first chart in this report).

While it is daunting to grapple with the long list of productivity challenges we have identified, the more useful perspective is that there are plenty of ways to make incremental improvements. As it happens, such incremental improvements are already significantly underway.

Among the acute temporary challenges cited earlier, Canada has already recalibrated its immigration regime. As evidence of the initial tentative success of this effort, the country’s capital stock per capita is rising again (see next chart). Pandemic distortions are also fading as time passes, including the gradual trickle of workers returning to their offices for additional days per week. We are hopeful that the high uncertainty presently constricting Canadian expansion plans will fade in the years ahead, though less convinced that interest rates will fall significantly.

Figure 17. Canadian capital stock per capita has rebounded

As of Q2 2026. Sources: Statistics Canada, Macrobond, RBC GAM

Turning to the aforementioned global forces that ceased to blow quite so favourably in recent decades: artificial intelligence now appears to be the long-awaited next major general-purpose technology. We believe it will increase productivity growth around the world in the years ahead.

Canada should also benefit significantly. Canada and its great resource wealth should also benefit as the world increasingly prioritizes resource security – another potential productivity driver.

Conversely, making Canada’s culture more amenable to risk-taking will be hard to achieve and at a bare minimum take a long time, so it is not central to our forecasts.

What about Canada’s economic structure? Much is plainly unalterable: the geography and climate are as they are. The low population density doesn’t have a short-term solution, nor does the present lack of scale. A few of the sector-based drags could be more malleable: the size of the government now seems to be shrinking back to more normal proportions, and the real estate sector’s outsized influence on the economy is diminishing as home prices ebb.

Promisingly, it appears that Canadian R&D spending as a share of GDP is starting to rise again after a lengthy decline.

Public policy mix is an important area of improvement. Taxes were marginally lowered over the past 18 months, including a broadening of the R&D tax credit. Even more importantly, the process for securing infrastructure and resource project approval has been significantly streamlined. Bank capital requirements are also declining modestly, and builders should benefit from reduced development charges and easier zoning rules.

Additional government funding is being directed toward major high-value infrastructure projects. As per the ongoing investment summit, serious efforts are being made to attract more private capital as well. The Canadian banks are lining up to commit their capital and lending might.

It’s tempting to proclaim that Canada’s recent trade-diversification spree should boost productivity – and all else equal, it should. But the reality is that the decline in access to the U.S. market materially outweighs those incremental gains and constitutes a new productivity headwind for Canada. The barriers between Canadian provinces are modestly declining but we do not expect miracles ahead.

Finally, one can hope that the most critical step occurs – that Canadian businesses will be willing to invest more in themselves as they see the policy environment improve, temporary headwinds fade and the prospect of a powerful global AI tailwind. Promisingly, it appears that Canadian R&D spending as a share of GDP is starting to rise again after a lengthy decline.

Many of these developments will take years to play out. That needs to be the timeframe over which any potential improvement in Canadian productivity growth is considered.

All else equal and at the margin, a period of faster productivity growth would translate into a stronger Canadian dollar, higher earnings, faster stock market gains and slightly higher interest rates.

Stylistically, we believe Canadian productivity growth is capable of accelerating from roughly 1.0% per year to perhaps 1.5% per year - and potentially beyond that if Canadian businesses opt to invest in themselves with the enthusiasm that their American competitors do, and if AI also proves to be as revolutionary as it seems destined to be.

For context, such an acceleration would mean that Canadian financial prosperity rises 50% faster than before, and Canadians would become 16% more prosperous within a generation.

It may be asking too much to predict an outright narrowing of the U.S.-Canada productivity gap, but halting the ongoing deterioration would be a welcome start.

All else equal and at the margin, a period of faster productivity growth would translate into a stronger Canadian dollar, higher earnings, faster stock market gains and slightly higher interest rates.

-EL

European resilience

Eurozone economic data took a sharp turn for the better over the summer. The currency bloc’s data surprise index – which measures whether key economic indicators are coming in stronger or weaker than expected – has risen to its highest level in more than 3 years (see chart below).

Figure 18. Eurozone data have been stronger than expected over the summer

As of 09/11/2026. Sources: Bloomberg, Macrobond, RBC GAM

The economy rebounded strongly in Q2 with a 0.6% non-annualized gain following a flat reading in Q1. Excluding a big swing in volatile Irish GDP, the H1 expansion was fairly even at around 0.3% per quarter.

Still, a solid increase in Q2 GDP is impressive in the face of a significant energy shock. The currency bloc is relatively more exposed to this type of shock as a net energy importer. Indeed, Eurozone survey data deteriorated sharply in March and April (see table below).

Figure 19. Eurozone economic survey data have deteriorated rapidly

Shading based on 3-year history as of 09/07/2026. Sources: European Commission (DG ECFIN), Sentix , S&P Global, ZEW (Leibniz Centre for European Economic Research)

Sentiment improved as energy prices came off their spring peaks. That trend continued in July and August even as the U.S.-Iran Memorandum of Understanding fell apart, and prices came under renewed upward pressure. In fact, most confidence measures are back to pre-conflict levels.

The rebound in the Eurozone manufacturing Purchasing Managers’ Index (PMI) has been particularly impressive. Output, new orders and export orders are all at their highest levels since 2022 (see chart below).

Figure 20. Eurozone manufacturing outlook is improving

As of August 2026. Sources: S&P Global, Macrobond, RBC GAM

Germany’s manufacturing PMI also points to the strongest production growth since 2022. This is a bright spot given the sector’s multi-year slowdown amid increasing competitiveness challenges from China. The latter is now by far the biggest exporter of passenger cars with shipments up 60% over the past year, while Germany’s auto exports are down by more than 10% (see chart below).

Figure 21. China has come out as the top passenger car exporter

Based on latest data available as of August 2026. Sources: China General Administration of Customs (GAC), Korea Automobile Manufacturers Association, U.S. Bureau of Economic Analysis (BEA), Japan Automobile Manufacturers Association (JAMA), German Association of the Automotive Industry (VDA), Macrobond, RBC GAM

But offsetting that headwind, the latest PMI report noted German manufacturers are seeing higher demand related to defence spending and the data centre buildout. On the former, analysts at Morgan Stanley found Germany’s defence industrial base is best positioned (among 7 countries including the U.S., UK, France, Germany, Italy, Sweden and Poland) to take advantage of higher defence spending.

On the latter, Germany’s expertise in precision engineering and electrical equipment manufacturing makes it a beneficiary of the AI buildout, even if the country itself isn’t a leader in data centre construction. Over the past 18 months, German exports of tech and electronic products have increased by nearly as much as auto and parts exports have declined (see chart below).

Figure 22. German tech/electronic exports are making up for a pullback in auto exports

As of July 2026. Sources: German Federal Statistical Office (Statistisches Bundesamt), Macrobond, RBC GAM

Despite the resilience in survey data, higher energy prices still present a downside risk to the Eurozone outlook. European natural gas and electricity prices continue to climb to multi-year highs (see chart below). While prices are still well below levels seen in 2022 when European natural gas supply was more significantly and directly disrupted, rising input costs will act as a headwind to energy intensive manufacturing industries. We’ll see if the pickup in manufacturing sentiment can be maintained in the coming months.

Figure 23. European natural gas and electricity prices at fresh highs but well below 2022 levels

As of 09/11/2026. Sources: Bloomberg, Macrobond, RBC GAM

For European consumers, delayed pass through to regulated utility prices means most of the pain from higher prices will be felt later this year and early next year. Eurozone inflation has already climbed to 3.3% in August from 2.0% at the end of last year and could rise to 4% in early 2027. That would push real wage growth into negative territory and dampen consumer spending. However, the household sector’s high savings rate could act as a shock absorber.

The ECB has been on the front foot in managing inflation risks. It has raised interest rates by 50bps so far this year and signaled that another hike is likely by year end. The central bank’s proactive approach has helped contain upward pressure on long-term bond yields. German 30-year yields are up less than G7 peers – although greater focus on fiscal sustainability has driven up borrowing costs in France and to a lesser extent Italy. Overall, Eurozone financial conditions are only slightly tighter relative to the start of the year. Bank lending continues to pick up strongly (see chart below).

Figure 24. Eurozone bank lending is accelerating, financial conditions are stable

As of July 2026. Sources: European Central Bank (ECB), Macrobond, RBC GAM

Higher energy prices and competitiveness challenges present cyclical and structural headwinds for the Eurozone, but the economy has proved surprisingly resilient of late, supported by fiscal/defense spending and the AI buildout. There is a risk that further increases in utility costs sap economic momentum. But we think recent trends support our modestly above-consensus forecast for Eurozone growth.

UK also shows resilience but faces other challenges

UK data have also generally been coming in stronger than expected, particularly the latest GDP report. It showed the economy expanding by 0.4% month-over-month in July (consensus was flat). One-third of the increase came from computer programming, consultancy and related activities. According to the Office for National Statistics, many of the businesses seeing the strongest sales growth are involved in activities related to AI and cloud computing. Programming and consultancy output has increased by 14% over the past year, nearly 9x as fast as overall economy (see chart below).

Figure 25. UK computer programming and consulting is benefiting from AI adoption

As of September 14, 2026. Sources: UK Office for National Statistics, RBC GAM

The pickup in economic activity over the past six months is somewhat surprising in the face of higher energy prices and given the UK’s reliance on LNG imports (albeit not directly from the Middle East). As in the Eurozone, PMI data have largely rebounded following a spring swoon. However, the manufacturing sector has lost some momentum recently as prices have come under renewed upward pressure.

Some of the UK’s resilience can be attributed to delayed pass through of higher energy prices to households. Domestic utility bills only reflected an earlier increase in wholesale gas and electricity prices in July when regulated utility prices rose by 13%. The regulator has announced a further 4% increase effective in October, which would have been greater were it not for the government’s decision to remove Value Added Tax on domestic electricity bills.

Interestingly, despite higher energy inflation, UK CPI is nearly 1 percentage point lower than it was a year ago, although about half of that can be attributed to the government’s affordability measures that took effect in the spring. However, with nominal wage growth slowing by even more, real wage growth is now close to flat. That presents a headwind for the consumer sector. It could worsen if energy prices remain elevated through the fall (setting up for another jump in regulated utility prices in January).

The silver lining has been that with wage growth slowing, underlying inflationary pressure has eased enough to keep the Bank of England from raising interest rates. Another steady rate decision is expected in September.

However, the futures market still points to rate hikes later this year and the bond market has exerted its own tightening pressure. Long-term Gilt yields have seen the largest increase in the G7 year-to-date. 30-year Gilt yields are closing in on 6% (see chart below).

Figure 26. The UK has the highest long-term yields in the G7

As of September 14, 2026. Sources: Bloomberg, RBC GAM

Investors remain hyper-focused on UK fiscal sustainability. New PM Burnham and Chancellor Healey are set to unveil their first budget in late October. So far, they have been tightly lipped on any new fiscal measures (aside from the aforementioned electricity VAT cut) but haven’t ruled out potential tax hikes. The government has the tough task of addressing affordability challenges while putting the country’s finances on a more sustainable track.

In brief, the UK economy has demonstrated similar resilience to the Eurozone in the face of higher energy prices. But it isn’t benefiting from the same fiscal tailwind and the risk from tighter financial conditions seems greater. We are penciling in slightly slower GDP growth in the UK compared to the Eurozone in the second half of the year.

-JN