The factory renaissance is not the one politicians sold

THE PROMISE of tariffs was dignity for workers and a revival of factories. The reality is a renaissance of capital, not labour. The July employment report, released on August 7th by the Bureau of Labor Statistics, found manufacturing employment essentially flat, adding only 5,000 jobs. The broader picture is worse: the sector has shed approximately 66,000 positions over the past year, as the ISM's manufacturing index itself noted in June 2026, with payrolls contracting for 32 consecutive months. The headline may read "treading water." The substance is that reshoring has arrived, but it does not employ the people politicians think it will.

The reason is not hard to see. New American factories are nothing like the old ones. In 2024, companies announced 244,000 reshoring and foreign-direct-investment manufacturing jobs back to the United States, according to the Reshoring Initiative, an advocacy group at the University of Rhode Island. Eighty-eight per cent of those were classified as high- or medium-high-tech roles — robotics, machine vision, automation controls — rather than assembly-line work. The reshoring wave that began as pandemic-era crisis management has evolved into a fundamental restructuring. But the jobs it creates are for engineers, not the workers Mr Trump's tariffs were designed to save.

The incentive structure is clear. Tariffs raise the cost of imported goods, narrowing the labour-cost arbitrage that once made overseas production so attractive. But they do not eliminate the economic logic of automation. A modern factory in Arizona or Ohio runs collaborative robots, programmable logic controllers and IoT-connected quality systems. It needs fewer workers per unit of output than a plant built a decade ago, let alone one from the 1990s. The result is what the labour economist Alí Bustamante calls a "jobless manufacturing rebound": output grows while headcount stagnates.

Construction spending tells part of the story. According to the IoT Analytics Industrial Macro Pulse report of May 2026, manufacturing construction spending peaked at $239 billion in June 2024 and has since declined by 21% through March 2026. The computer, electronics and electrical sector — which accounted for over half of manufacturing construction at its peak, driven by the 2022 CHIPS Act — saw spending fall by 44% from its July 2024 high. Outside electronics, nominal construction spending grew by only 5.6% between February 2025 and March 2026, or roughly 2.3% after adjusting for inflation. This is not a building boom. It is a cautious expansion, concentrated in capital-intensive facilities that will hire technicians, not crowds.

To be sure, some reshoring is real. Companies such as RTX have announced billions of dollars in new American capacity. Whirlpool invested $300 million in Ohio operations, projecting hundreds of new jobs. Semiconductor fabs in Arizona and Texas are rising from the ground, backed by more than $645 billion in announced investment across 140 projects since 2020. Yet even these projects hire smaller, more skilled workforces than the headlines suggest. A $10-billion semiconductor facility might create 3,000 jobs, not 30,000. The math has changed.

The deeper problem is one of political arithmetic. Tariffs promise concentrated benefits to a specific constituency — manufacturing workers — while imposing diffuse costs on consumers through higher prices. The political appeal is obvious: benefits are visible and costs are spread. The trouble is that the benefits never materialise in the form voters expect. Instead of assembly-line jobs returning to the Rust Belt, the reshoring wave produces high-wage engineering roles in Arizona, Texas and the Southeast, alongside capital returns for shareholders. The politics of protectionism have delivered the economics of automation.

The labour market data reinforce this reading. The Manufacturing Institute, in a projection with Deloitte, estimates that the sector will need 3.8 million workers by 2033, with up to 1.9 million potentially going unfilled. Average manufacturing compensation reached roughly $135,500 in 2025, up 13.1% from the prior year. Wages are rising not because of competition among workers for existing jobs, but because there is a genuine shortage of automation-fluent technicians. The bottleneck is skills, not capital. A 2025 reshoring survey of 500 manufacturers found that respondents would reshore 30% of offshore products if skilled labour existed, compared with 23% if tariffs were at 15%. Workers matter more than duties.

The July report's revisions add further disquiet. The BLS cut May 2026 payroll growth by 66,000 and June by 37,000, meaning employment in those two months is 103,000 lower than previously reported. Average hourly earnings across all private nonfarm employees changed little at $37.62, with annual wage growth of 3.2%. The unemployment rate held at 4.1%. The labour market is neither collapsing nor thriving. It is adjusting — slowly and unevenly — to a structural shift that tariffs have not reversed.

The broader lesson for policymakers is not that reshoring is worthless. Supply-chain resilience has genuine value, and the concentration of critical production in geopolitically vulnerable locations was always a risk. The lesson is that tariffs are a blunt and misdirected tool for the job they claim to do. They raise prices, invite retaliation, and create a constituency for permanent protection — all while failing to deliver the low-wage employment that is their primary political selling point.

A wiser policy would decouple supply-chain strategy from employment rhetoric. Targeted investment in semiconductor, defence and medical-device capacity may be justified on national-security grounds, without pretending it will rebuild the factory jobs of the 1970s. The real task is workforce development: community-college partnerships, apprenticeship programmes and immigration reform that allows skilled technicians to fill gaps that domestic training cannot. These measures are less photogenic than a tariff announcement at a reopening plant. They are also more likely to produce the result that reshoring was always supposed to deliver.

The manufacturing renaissance is happening. Just not the one politicians sold.