Viewpoints

• Rising real yields reflect structural shifts: Long-end yields are climbing not just due to inflation fears but rising real rates, suggesting expectations of stronger trend growth, higher risk premia, or tighter global capital supply.

• Surge in capex demand: A coming wave of AI-driven investment, especially in the U.S., could significantly boost corporate capex, mirroring the dot-com boom and pushing the corporate sector into a net borrowing position again.

• Large fiscal deficits projected in most major economies: It’s not just a US story: China running an 8% government deficit through to 2030, and worsening budget positions in Germany and Japan.

• Global savings-investment imbalance tightening: Post-GFC changes, like reduced capital mobility, mean countries are increasingly reliant on domestic savings to fund investment and deficits—requiring higher interest rates.

• Sustained higher equilibrium rates likely: With large fiscal deficits across major economies and higher private investment needs, the bond market may be signaling a long-term rise in equilibrium interest rates (r*), ending the era of cheap money.

For better or worse markets had (more Trump trade headlines as I write) substantially reduced estimates for a recession this year. Betting markets have reflected the same sentiment. Our own measure of cross market implied growth has recovered smartly pushing implied vols back down in many markets.

But across all major markets back end yields have continued to trend higher - a process that began post Covid, paused for a bit and then started again last year before accelerating over recent weeks.

April saw little change in state labor markets. Five states saw statistically significant gains in jobs in March, with none larger than .4 percent. Texas’s .3 percentage point gain amounted to 37,700 jobs. A number of states had insignificant declines.

Four states (DC, Iowa, Massachusetts, and Virginia) had statistically significant increases in their unemployment rates, with DC and Massachusetts up .2 percentage points and Iowa and Virginia up .1 percentage point. Nevada’s rate fell .1 percentage point and Indiana was down .2. The highest unemployment rates were in DC (5.8%), Nevada (5.6%), Michigan (5.5%) California (5.3%), and Kentucky (5.2%). Hawaii, Montana, North Dakota, South Dakota, and Vermont had unemployment rates under 3.0%, while South Dakota’s 1.8% was yet again the lowest in the nation.

Puerto Rico’s unemployment rate was again unchanged at 5.3% and the island’s job count moved up by 1,100.

Geopolitics, AI, and the New Digital Divide In the final part of this series, we turn from the physical and technological constraints on growth to the geopolitical ones. Part I highlighted structural headwinds in the world’s labor, capital, and energy markets. Part II explored the potential of AI to act as a macro workaround—while warning that it is not immune to physical limits, particularly energy. Part III brings these threads together: AI is not just a labor-saving innovation or a capital redeployment tool. It is increasingly becoming a geostrategic asset—one whose deployment depends heavily on energy availability, infrastructure, and political control.

1. AI as a Tool of Statecraft What oil was to 20th-century power, processing power is becoming in the 21st. Nations are now racing to dominate the AI value chain—from chip design and fabrication, to data access, to the energy systems needed to power high-performance computing. Export controls, industrial policy, and national security reviews are no longer confined to oil tankers and pipelines—they now apply to GPUs and data centers.

The result is a world increasingly defined not by technological openness, but by fragmentation. From Trump's protectionist agenda—including tightened semiconductor export controls—to the strategic ambitions of the US CHIPS Act under the previous administration, and China’s intensified homegrown AI efforts, digital sovereignty has become an explicit geopolitical objective. The once-celebrated ideal of borderless innovation is yielding rapidly to a harsher reality: national power now hinges increasingly on digital dominance and technological self-reliance.

2. The Energy Arms Race Beneath the Silicon If the second in this series taught us anything, it’s that AI is not ethereal—it’s physical. It runs on electricity, minerals, and metals. It demands stable grids, high-capacity cooling, and energy infrastructure on a scale few countries currently possess. AI isn’t just compute-intensive. It’s energy-intensive. And that shifts the locus of global competition. The new AI geopolitics is not just a chip race. It is quite literally a power race. Countries that control abundant, cheap, and low-carbon energy will gain a disproportionate advantage in AI scalability. Those that don’t, risk falling behind, regardless of their tech ambitions.

As such, energy security is fast becoming the ultimate gating factor in the AI revolution. Just as the post-WTO growth boom in the 2000s was fueled by a surge in real energy consumption and prices, the AI boom may repeat that pattern—only faster, and more unevenly distributed.

Among the many issues facing the public and financial markets, the impact of President Trump’s tariffs on inflation and how they affect the Federal Reserve’s monetary policy stand out. Estimates of the economic and inflation impacts of the tariffs are quickly evolving as the twists and turns of Trump’s policies change the magnitudes and character of the tariffs. This short note doesn’t speculate on the magnitude of the tariffs, rather it considers how CPI and PCE inflation will likely be affected differently. These differences may influence Fed behavior.

PCE inflation measures the percentage change in prices of all consumer goods and services. This includes all goods, including those paid for by consumers as well as those financed by third party payments like Medicare, Medicaid and medical services paid for by employer insurance. CPI inflation measures the percentage change in prices of goods and services directly paid for by consumers and excludes goods and services paid for by third parties.

As such, PCE and CPI inflation measure the price changes of different baskets of goods and services. PCE inflation is a broader measure of inflation, while the CPI is a better measure of consumer out-of-pocket expenses. The CPI comprises a much larger share 35.4% of shelter costs (including rental costs and owners’ equivalent rent, OER) than the PCE, and a smaller share of medical care (8.3%). Both measures of consumer prices include items like personal care and education expenses.

Of note, both the PCE and CPI inflation measures are adjusted for estimates of quality improvement. The quality adjustments are estimated by the Bureau of Economic Analysis within the U.S. Department of Commerce, based on hedonic regression analyses that try to capture the quality improvements of new products, other estimating techniques, and some degree of subjectivity.

While CPI and PCE inflation are the most widely used measures of inflation, the GDP deflator is a broader measure of inflation that captures the price changes in other components of GDP, including business fixed investment, imports and exports, and residential investment (new construction and home improvements). This is important in the current context, since a sizable portion of imported goods that are subject to tariffs are capital goods purchased by domestic businesses and used in production.

An historical note. The Fed’s semi-annual Monetary Policy Report to Congress and its central tendency projections of economic growth and inflation, which began in 1980 as required by the Full Employment Act of 1978, used the GDP deflator as its inflation measure until 1990, when it switched to the CPI; Fed Chair Greenspan subsequently switched it to the PCE. The Fed’s first-ever Strategic Plan of 2012 officially established the PCE inflation target of 2%.

All other government agencies--and all other central banks in the world--rely on CPI inflation, and the CPI is an important “policy variable” used in various functions, including to index the benefits of Social Security, other pensions and an array of other programs. The Fed began focusing on core inflation early on: Fed Chair Arthur Burns instructed his research staff in the mid-1970s to calculate the CPI excluding food and energy following the spike in agricultural prices in 1971-1972 and surging oil prices generated by the first Arab Oil Embargo in November 1973. His primary interest was to tell the public that the high inflation was caused by exogenous forces. He was also attempting to quell rising inflationary expectations and bond yields without having to tighten monetary policy too much (Remember, before May 1983, mortgage rates were calculated directly in the CPI).

Recent CPI and PCE inflation trends. CPI inflation has been materially higher than PCE inflation in the post-Covid economy, as shown in Chart 1. CPI inflation was 0.6 percentage points higher than PCE inflation in 2024 (3.4% vs 2.8%), 0.7 ppt higher in 2023 (4.8% vs 4.1%) and 0.8 ppt higher in 2022 (6.1% vs 5.3%). So far in 2025, they have tracked closer. The cumulative differences have added up: since December 2019, the CPI index has risen by 3.8 percentage points more than the PCE price index Chart 2). As we all know, the large rise in the general price level has resonated with American citizens and created a communications problem with the Fed.

The Age of Constraints, Part 2, AI and the Energy Reality Check

If Part I of this series mapped the fractures in the global economy’s supply-side foundations—labor, capital, and energy—this second installment turns to the potential workaround: artificial intelligence.

AI has been frequently pitched as the ultimate macroeconomic escape hatch. It promises to plug labor shortfalls, redeploy idle capital, and inject new life into flagging productivity trends. And in some areas, it already is. From medical imaging to generative models in finance, AI is no longer theoretical—it’s live, scaling, and impressive.

But it’s not weightless. Or frictionless. And it is certainly not energy-free.

A look at the long-run relationship between real energy prices and global per capita energy consumption—illustrated in the chart below—potentially suggests that AI may not ease the Age of Constraints so much as intensify it. Particularly when it comes to energy.

President Trump has stated that he "would not be disappointed" if a deal with China is not reached swiftly, as "not doing business is also a good deal for the United States." However, this perspective is flawed or his economic advisors are misinforming him. Trump's tariff approach failed to consider the US economy's heavy reliance on consumer imports and direct and indirect economic effects on importers, trucking and rail companies and retailers.

The adverse effect on sales (or company revenue), employee earnings, and company profits might amount to approximately seventy percent of the total import value.

Based on 2024 trade data, the US imported consumer-related goods valued at $1.5 trillion. The Federal Reserve's report on industrial production estimated domestic production of consumer goods at $2.18 trillion, for a combined total of roughly $3.7 trillion. However, as consumer goods progress through the economic chain from ports and shipping to retail, significant value is added to their production and importation value.

Based on GDP data, consumer spending on goods in 2024 surpassed $6.3 trillion, which is $2.6 trillion more than the total value of imports and domestically produced goods. This indicates that for every dollar spent on imports and domestic production, there is nearly an additional 70 cents in value-added, associated with markup costs (such as shipping, distribution, labor, etc.) and profits.

Distinguishing the profit margins between imported consumer goods and those produced domestically is not possible. However, even if these margins are the same, the potential impact on the economy's revenue stream (GDP) and income stream (GDI or wages and profits) remains significant. That's why a lengthy and expanding roster of companies, including Apple, GM, Ford, UPS, Walmart, and Procter & Gamble, among many others, have either reduced or withdrawn their profit estimates for 2025.

It's crucial to recognize that the US imports over $1.6 trillion in industrial supplies and capital goods. To fully understand the economic impact of global trade flows, it's necessary to consider the potential loss of both consumer and business imports, though the latter are harder to assess. Even if the impact on business goods is only half as significant as on consumer goods, the overall effect on the US economy from China alone is about 2% of GDP. If imports from other key trading partners, such as Canada and Europe, are also lost, the impact doubles.

In other words, contrary to what President Trump and his advisors has said, the "US Needs A Trade Deal," and soon.

In 1973, matriculating from Princeton, I submitted to William Branson my senior thesis entitled “The Nature of the Phillips Curve” in which I examined the trade-off between inflation and unemployment. I’ve been fascinated with the curve ever since.

Thought on the curve has advanced over the years. Milton Friedman’s “natural rate” hypothesis became widely accepted, implying no long-run trade-off exists. Empirical representations of “sticky” prices and inflation expectations, once combined in lagged inflation rates, are now separated into a backward-looking component for sticky prices and a forward-looking component for (usually survey-based) long-term expectations. Successful monetary policy anchored inflation expectations near the Fed’s now-explicit 2% target, and with that success the slope of the curve flattened. Supply shocks, which temporarily worsen the short-run trade-off, have been added to the curve for food and energy prices, the exchange rate and, most recently, COVID-related disruptions to supply chains. But decades later, the dilemma for monetary policy presented by the Phillips Curve remains unchanged: in the short run, with expectations anchored, the Fed chooses between higher inflation or lower unemployment. The risk of choosing lower unemployment is expectations becoming unmoored, pushing inflation persistently above 2%. The risk of choosing lower inflation is recession.

Today there is a new supply shock to consider: tariffs. Approximately 10% of “core” (excluding food and energy) personal consumption expenditures (PCE) are imports: 6% directly as final consumer goods, 4% indirectly as inputs to the production of final consumer goods and services. Hence, the 10% universal tariff threatened for July by the Trump Administration, if “passed through” entirely and immediately to consumers, would add approximately 4 percentage points to the annualized rate of core PCE inflation in the third quarter of this year. My own work suggests pass-through is delayed and incomplete, with about 75% of the tariff appearing in consumer prices within one year, and 88% within two years. Still, this would be a significant inflation shock. In could be squashed by tight monetary policy, but at what cost?

Using a “modern” Phillips Curve for core PCE inflation described in 2016 by (then) Fed Chair Janet Yellen at the annual meeting of National Association for Business Economics, I explored the horns of the Fed’s current dilemma. First, I generated a baseline forecast assuming no tariffs, unemployment at 4%, and expectations at 2% (Chart 1). After 2025, baseline inflation (4-quarter percent change) fades fairly quickly to the Fed’s target. Then, I introduced a 10% tariff shock while assuming the Fed maintains unemployment at 4%, again with expectations at 2%. When the tariff is thusly “accommodated” by monetary policy, inflation remains above 3% through 2026, and above 2.5% through 2027 – high enough and for long enough to make any central banker uncomfortable. Would expectations remain anchored at 2%? Perhaps, even probably, so: they did during the much bigger COVID-era price shock.

Labour, Capital, and Energy in a Fractured World

The latest surge in US tariffs hasn’t yet shattered the global economy—but it has more openly revealed the fractures that were already there.

Long before the new wave of US protectionism, the world economy was drifting into a more fragmented, frictional phase. The free flow of goods, capital, labour, and energy—pillars of the late 20th-century global order—had been quietly eroding for years. What the tariffs have arguably done is to make the drift official. They have marked a turning point, not because they started something new, but because they confirmed that the previous global economic model was no longer sustainable.

Still, if US trade policies have clarified the direction of travel, they have also accelerated the journey toward fragmentation—often in ways that undermine the very resilience they claim to restore. By targeting countries and sectors, the US has reignited a zero-sum logic in global trade: one where national security concerns override economic efficiency, and where long-term cooperation is sacrificed for short-term leverage. This approach may score political points domestically, but it risks entrenching the same vulnerabilities it seeks to eliminate—raising input costs, disrupting investment, and pushing allies and adversaries alike toward parallel, disconnected systems.

Yet even without this protectionist turn, economic models were already under strain. The foundations of growth had begun to shift well before tariffs re-emerged as a policy tool. Labour markets were misaligned, with ageing workforces in some economies and idle potential in others. Capital was abundant but increasingly abstract—flowing into intangible assets and financial engineering rather than productive investment. Energy was no longer cheap, predictable, or apolitical. And trend productivity, which once rose smoothly on the back of scale and specialisation, had become choppy and contested.

This is the Age of Constraints—not a crisis, but a condition. A world where the fundamental factors of production no longer reinforce each other, but strain against one another. A world where efficiency is harder to come by, and where growth increasingly hinges not on accumulation, but on adaptation.

This commentary launches a three-part series exploring how these structural shifts—in labour, capital, and energy—are reshaping the world economy. In the next instalment, we’ll explore why artificial intelligence, is in fact a deeply pragmatic response to these constraints. But first, we must understand the systemic pressures that have brought us to this turning point.

But first, we need to understand just how deep the constraints run.

I. Labour: A Global Workforce Out of Balance

In the 20th century, labour was an abundant and cheap input. Factories expanded. Cities grew. Consumption soared. But in much of the world, that labour force is now shrinking—and ageing fast.

• Japan’s working-age population has fallen by over 10% since 2000. • China’s population began declining in 2023, with projections pointing to a loss of 400 million people by the end of the century. • Europe and South Korea face similar demographic cliffs.

At the other end of the spectrum, Africa and South Asia are entering a demographic dividend phase. Nigeria is projected to surpass the US in population by 2050. India now has more young people than any nation on earth.