Where I Part Ways with the Consensus (It’s a Long List)

Introduction and Executive Summary

Most economic commentary today leans negative: growth is fragile, tariffs are dangerous, a recession is a real possibility, and AI’s promise still feels distant. I see the same data but a different picture. I’m more optimistic on growth, less alarmed about tariffs, and more focused on AI-driven productivity and demographic trends than most forecasters.

Still, I understand that many people aren’t feeling the momentum. Not federal workers. Not undocumented immigrants. Not new college graduates. Not those who feel acutely threatened by AI automation. And not those whose outlook is shaped by persistently pessimistic media narratives.

But even with these very real sources of anxiety, it’s hard to argue the labor market is in crisis when the national unemployment rate is just 4.3%.

Several durable tailwinds are being underweighted. The AI infrastructure buildout, wealth effects from rising asset values, and steady healthcare demand point to an economy with more underlying momentum than headlines suggest. Tariff impact will likely be modest and delayed, making them less macro-relevant than the political energy surrounding them. Recession calls remain out of sync with the scoreboard; this spring’s slowdown resembled a scare, not the start of a contraction.

Where I diverge most is productivity. AI is already lifting productivity through two engines:

• Tech-producing sectors and AI-buildup with extraordinary output per worker, and
• AI-induced labor cuts across the rest of the economy, where firms are restraining hiring and laying off workers, rather than adding headcount.

That’s why in the second half of 2025 and 2026 we’ll likely see GDP growth around 3% even as job growth hovers near zero. Soft employment data won’t signal soft economic activity—it will reflect rising productivity and stalled labor-force growth.

And the implications continue:

• The unemployment rate will roughly hold steady. Much stricter immigration enforcement is limiting labor supply, offsetting weak hiring and preventing a meaningful rise in joblessness.
• Wage growth will remain somewhat elevated. Annual raises—still influenced by past inflation—remain strong, especially in occupations still facing tight labor markets such as blue-collar and manual services.
• RIP the 2% inflation target. The Fed’s longstanding goal is increasingly out of step with reality. Inflation expectations and actual inflation remain above target, and while tariffs are boosting prices less than feared, they still add upward pressure. A higher equilibrium inflation rate may quietly become the new norm. Expect fewer Fed rate cuts than consensus.

Finally, I put more weight than most on education and demographics. The college-educated workforce keeps expanding while the non-BA labor pool shrinks, producing persistent shortages in blue-collar and care work—and tougher early careers for many new college graduates.

The Economy Is Stronger Than It Looks

Despite persistent headlines warning of a looming recession, the U.S. economy continues to show remarkable resilience.

Let’s start with the positives. We are in the midst of an AI-driven buildup boom—one of the largest technology infrastructure buildouts ever. It’s plausible that the current contribution to annual GDP growth is above two percentage points.

Add to that the wealth effect—as equity markets continue to rise, household net worth has increased significantly. This supports consumption, especially among the top 10% of households, who own roughly 90% of U.S. equities.

Demographics are also playing a stabilizing role. The aging of the large baby boomer cohort is quietly pushing the economy forward. As the population ages, healthcare spending continues to grow steadily—providing a built-in source of demand that’s largely immune to short-term economic cycles.

In this environment, it would take a lot to knock the economy into recession.

That’s not to say there aren’t headwinds. Immigration policy has become a growing drag. Enforcement activity and fear of detection have led some undocumented workers to exit the labor market, contributing to slower population and labor force growth. Tighter border policies and heightened workplace scrutiny are reducing the flow and participation of undocumented workers in sectors like construction, agriculture, and hospitality.

Interestingly, the economic effect of tighter immigration may be larger than the effect of tariffs, which have drawn more public attention but have had a more nuanced impact. Overall inflationary and economic effects have been smaller than many predicted. One major reason: limited foreign retaliation, which softened the expected blow. In fact, overall U.S. manufacturing production is now higher than it was in the first quarter of 2025, defying earlier forecasts of steep declines.

Meanwhile, federal government job cuts are creating localized pain in some areas. But they haven’t meaningfully shifted the national economic outlook.

When you weigh the positives against the negatives, the conclusion is clear: the positive forces are winning. The Atlanta Fed’s GDPNow model pegs Q3 GDP growth at 4.0%, and consumer spending at 3.4%, a robust pace that’s hard to square with the current wave of recession pessimism.

Some of the confusion stems from a genuine, if temporary, slowdown earlier this year. April and May marked the softest stretch, driven partly by panic over the new administration’s willingness to sacrifice short term economic growth. But that uncertainty has largely receded, and forward-looking indicators suggest that the worst is behind us.

In short, recession calls are out of touch with the actual strength of the economy. Many of them are based on weak job growth—but as we discuss later, weak job growth does not reflect overall weak demand for goods and services.

Where the Productivity Boom Is Coming From—and Where It’s Headed

There’s growing evidence that U.S. productivity growth is strengthening. Whether you look at the past year, 2 years, or since the pre-pandemic period. But to understand what’s happening—and what’s coming—we need to separate two distinct channels of productivity growth in the age of AI.

Channel 1: Productivity Gains from Tech-Producing Industries

The first and most visible source of recent productivity gains comes from the industries that produce technology itself—and those driving the AI infrastructure buildout.

Labor productivity in the tech sector is high and accelerating, and its contribution to overall productivity growth is increasing. These sectors are delivering more output per worker because of their business models and capital intensity.

• Software, cloud, and AI platforms generate enormous revenue with relatively few workers, thanks to the scalability of digital products. A single model or tool can serve millions of new users with almost no additional labor.
• Infrastructure and hardware producers—from semiconductors to advanced electronics—are highly automated and capital-intensive.

Channel 2: The Organizational Impact of AI

While the most visible productivity gains have come from the tech-producing sectors, there is growing evidence that AI is beginning to enhance productivity across the broader economy. These effects are still in their early stages but are already reshaping organizational behavior in ways that go beyond traditional automation narratives.

Most discussions of AI-driven productivity focus on direct substitution: automating tasks, eliminating roles, and redesigning workflows. That is certainly happening. But a fuller view suggests three less visible mechanisms already at work:

1. Anticipatory Hiring Drag: Firms anticipating automation within the next one to two years are proactively adjusting headcount. This includes slowing new hiring, pausing backfills, and delaying role approvals—especially for entry-level positions. Even when automation is partial, the result is more output per worker.
2. Narrative Pressure as a “Soft Recession”: The growing expectation—from CEOs, boards, and investors—that firms must demonstrate AI-led efficiency is functioning as a form of narrative-driven austerity. Much like in past recessions, companies are responding by cutting labor costs and leaning more heavily on existing technology—except now without the trigger of an actual downturn. AI acts as an organizational wake-up call, accelerating productivity-enhancing behavior.
3. Quality Improvements as Productivity: AI is also raising the quality of output. Even with unchanged staffing levels, teams are producing more valuable work, often improving customer satisfaction and enabling stronger pricing power. These qualitative gains are no less real in productivity terms.

This broader channel—AI-enabled efficiency gains across non-tech industries—is still nascent but becoming more apparent. The labor market is already reflecting some of these shifts. There is strong evidence that employment in AI-exposed occupations, particularly among early-career workers in fields like software development and customer service, has declined.

Although adoption remains uneven and many firms are still experimenting, this second channel of AI-driven productivity growth is poised to become a major contributor to aggregate economic performance over the coming years.

The Politics of Evidence

Some observers demand strict proof that AI is reshaping labor markets and productivity. But short of asking 1,000 CEOs to swear on a Bible before Congress that they’ve eliminated jobs because of AI, it’s unclear what “proof” would satisfy them.

Meanwhile, the patterns are emerging—and they’re hard to ignore.

The broader public’s eagerness to believe that AI isn’t changing the economy may itself be distorting interpretations. For better or worse, debates over AI’s economic impact have become polarized, shaped as much by ideology and tribal priors as by evidence.

Why Weak Job Growth Doesn’t Mean Recession This Time

One of the main reasons many economists and commentators have been calling for a recession is the recent slowdown in job growth. Historically, that intuition has held up: every time job growth turned negative, a recession followed. But this time is different.

Job growth has clearly cooled. Over the past four months, total nonfarm payrolls have grown at roughly a 0.2% annualized pace—a rate rarely seen outside of recessions or jobless recoveries. And yet, GDP growth is strong. It does not seem like the weak job growth is a result of weak demand for goods and services.

So what’s driving the slowdown in job creation, if not slowing demand?

1. It Started Before the New Administration: Upcoming annual benchmark revisions to payroll data suggest that monthly job growth slowed to double-digit thousands as early as 2024—well before the current administration’s trade and immigration policies took effect. The trend was already underway. That points to the next explanation:
2. Strong Productivity Growth and Automation: Higher output per worker means businesses can meet demand without expanding headcount.
3. Lagged Demand Effects: The dip in business and consumer sentiment between March and May 2025 likely softened orders and hiring plans. That drag is still working its way through payrolls. Encouragingly, sentiment has since rebounded, as shown in recent surveys.
4. Tighter Immigration Policy: Since late 2024, interior enforcement has intensified. Arrests and removals skyrocketed, and unauthorized entries have dropped sharply. Labor supply growth has all but vanished—especially in lower-wage sectors.

In short, while job growth has slowed, the underlying drivers—ranging from productivity gains to immigration constraints—point to structural shifts rather than a cyclical downturn, making a recession far less likely than the raw numbers suggest.

Blue-Collar Labor Shortage and Glut of College Grads

America’s labor force is at a demographic inflection point. With Baby Boomers retiring and immigration sharply constrained, the working-age population (ages 20–64) has flatlined—and may already be declining.

But not all groups are affected equally. The number of college-educated workers continues to grow at 1–1.5% annually. That’s because younger generations are far more likely to hold a bachelor’s degree than the baby boom cohorts they’re replacing.

In contrast, the population without a college degree is shrinking—and fast. This contraction is especially acute given restrictive immigration policy in 2025. And it’s a problem, because people with a bachelor’s degree generally avoid blue-collar and manual service occupations—roles in transportation, construction, maintenance, food service, personal care, and more.

That leaves labor supply in these essential roles dwindling just as demand remains strong. The result is persistent shortages in blue-collar and manual service jobs.

Meanwhile, as discussed above, generative AI is beginning to erode demand for office jobs. That’s where most new graduates are headed—raising the risk of a growing glut of degree holders.

This imbalance—too few workers for hands-on roles, too many for office ones—is reshaping the labor market. Over time, it could flatten starting salaries for college graduates, while employers in trades and care struggle to fill vacancies.

A Few Predictions

Based on the framework outlined above, here’s where I expect things to go:

• GDP: In late 2025 and 2026, GDP growth will be about 3%, but job growth close to zero. Despite solid demand, payroll growth will stay soft as firms emphasize efficiency and automation.
• Unemployment: The unemployment rate will roughly hold steady. Stricter immigration enforcement is constraining labor force growth, offsetting slow hiring.
• Wages: Wage growth will remain somewhat elevated. Annual raises—still influenced by past inflation—remain strong, especially in sectors with ongoing labor shortages.
• Inflation Targeting: RIP the 2% inflation target. The Fed’s longstanding goal is increasingly unrealistic. Both actual and expected inflation remain elevated, and while tariffs are boosting prices less than feared, they’re still adding upward pressure. A higher equilibrium inflation rate may quietly become the new norm.