The Strange and Awful Path: Why Construction Workers Make Less Than They Did in 1970

Source: Goolsbee, A. & Syverson, C. (2023). “The Strange and Awful Path of Productivity in the U.S. Construction Sector.” NBER Working Paper, University of Chicago Booth School of Business.

Every industry I’ve examined in this series shows intelligence suppression producing measurable damage. Manufacturing: NUMMI’s worst-to-best reversal. Healthcare: 98,000 preventable deaths a year. Hospitality: 70-80 percent annual turnover. Retail: $262 billion in lost sales from poor customer experience.

But no industry demonstrates the cumulative cost of suppression more dramatically than construction. Because in construction, the damage isn’t just visible in the current year’s numbers. It’s visible across half a century of declining productivity — a trend so anomalous that economists have struggled to explain it.

I think the explanation is sitting on every jobsite in America, holding a tool, waiting to be asked what they know.

The Strange and Awful Path

In 2023, economists Austan Goolsbee and Chad Syverson of the University of Chicago’s Booth School of Business published a paper with a title that tells you everything you need to know: “The Strange and Awful Path of Productivity in the U.S. Construction Sector.”

Their findings:

A construction worker in 2020 produced less than a construction worker in 1970. The value added per worker declined approximately 40 percent over 50 years.

This is not a measurement problem. The researchers tested for deflator bias and other statistical artifacts. They examined physical measures of productivity — housing units built per worker — and found similar stagnation or decline. The trend is real.

Every other major sector improved. Manufacturing productivity more than tripled. Agriculture soared. Even government services showed gains. Construction went backward. In the auto industry, output per worker went from roughly 5 cars per employee per year in 1939 to approximately 25 by 2020. Construction moved in the opposite direction.

Capital investment wasn’t the problem. The construction sector invested in capital at rates comparable to the overall economy. The money was being spent. It just wasn’t producing results.

The Federal Reserve Bank of Richmond, in a 2025 analysis, estimated that construction productivity could be 60 percent higher if construction firms operated at the scale of manufacturing firms. McKinsey estimated that the sector’s productivity could increase by up to 60 percent through widespread innovation adoption — and explicitly noted that “technology alone will not address poor productivity” and that “a fundamental culture change” is needed.

A $13 trillion global industry, going backward for half a century. The explanation is not capital, not technology, not measurement error. The explanation is suppression — an operating model that structurally prevents the people building the project from improving the design of the project. That’s the argument.

Why Construction Is Different

The suppression mechanism in construction operates through a distinctive architecture that makes it, in some ways, even more entrenched than Taylor’s factory model.

Adversarial contracting. The dominant procurement model — design-bid-build — separates the people who design the project from the people who build it, then pits the builders against each other on price. The lowest bidder wins. This creates an adversarial relationship from day one, where the contractor’s primary incentive is to minimize cost (including labor investment) and maximize change orders when things go wrong.

In this model, the trade worker — the carpenter, electrician, ironworker, or heavy equipment operator — has no formal connection to the design process. By the time the worker sees the drawings, the decisions have been made. If the design creates a constructability problem that the experienced tradesperson would have spotted in five minutes, the system has no mechanism for that intelligence to reach the designer. The worker adapts, works around the problem, and the cost shows up as a schedule delay or a change order — both of which are attributed to execution, not design.

Fragmentation. Construction is one of the most fragmented major industries. A single project may involve dozens of subcontractors, each managing their own workforce, each focused on their scope, with minimal coordination across trades. The institutional knowledge that accumulates within one project — the sequencing insights, the coordination solutions, the material behavior observations — walks off the site when the project ends and doesn’t transfer to the next one.

The knowledge drain. One in five construction workers is over 55. The industry has roughly half a million field vacancies in the U.S. alone. An entire generation of tacit knowledge — developed over careers of building things — is retiring. And because the system never created mechanisms to capture that knowledge, it’s disappearing.

The experienced ironworker who knows how a specific steel connection behaves in cold weather, the electrician who can read a panel layout and predict where the interference will occur, the superintendent who has seen a hundred foundations poured and knows what the soil conditions mean for the schedule — none of this knowledge exists in any database, any document, any system. It lives in the workers. And the workers are leaving.

The Constructability Gap

There is a specific, measurable failure mode that captures the suppression pattern in construction perfectly: rework.

Rework — doing work over because it was done wrong the first time, or because the design didn’t account for field conditions, or because trades interfered with each other due to poor coordination — consumes an enormous percentage of construction labor hours. Estimates vary, but studies cons