The empirical record of labor displacement during past waves of technological innovation — primarily the British Industrial Revolutions — and what it teaches about how workers, firms, and economies adapt when machines replace human labor.

The Core Finding: Entry Collapse, Not Displacement

Hillary Vipond’s study of English bootmaking (1851–1911), using 170 million full-count census records, provides the most granular quantitative account of mechanization-driven job loss to date.¹ The headline numbers: approximately 152,000 artisanal bootmaking jobs disappeared as skills became obsolete, while 144,000 new jobs demanding different skills emerged. But the mechanism of adjustment was not what standard displacement narratives assume:

• Incumbent artisans were rarely displaced. They were not pushed out of the trade into unemployment or lower-paying work.
• Entry collapsed. Young men stopped entering the artisanal trade. The decline unfolded through erosion of entry, not expulsion of incumbents.
• New jobs went to young workers, but not in the same locations. The industry shifted geographically toward Leicestershire and Northamptonshire.

This pattern — slow adoption, persistent demand for old skills, incumbents “grandfathered out,” and the burden of adjustment falling on new entrants — has direct implications for how we think about AI-driven labor displacement today.²

Historical Waves of Displacement

Schneider and Vipond place the bootmaking case in a broader sequence:

• First Industrial Revolution (1750–1850): Textile machines replaced weavers and spinners. Hand spinning — which employed up to 20% of women and children in Britain — was eliminated with scant labor reinstatement; the resulting job loss persisted into the 1830s.
• Second Industrial Revolution (c. 1850–1940): Improved inanimate power and implements. Sewing and riveting machines replaced hand-sewing work of bootmakers. Steam engines replaced human and animal strength.
• Third Industrial Revolution (c. 1970–present): Digital technologies that primarily augmented rather than replaced labor.
• Fourth Industrial Revolution (emerging): AI aims to displace human cognition — pattern recognition, judgment, and decision-making.

The early decades of each revolution may have seen the most widespread displacement. If so, the early years of the fourth revolution could be the best analogue for the scale of disruption to come.³

Diminishing Returns to Displacement

A key historical pattern: each successive wave has had diminishing returns in terms of labor displacement. The First Industrial Revolution replaced muscle and dexterity; the Second improved on those methods; the Third introduced digital technologies that augmented more than they replaced. The Fourth’s displacement of cognition is qualitatively different — it targets the human brain’s core capacities — but the historical record suggests that early decades of each wave are the most disruptive.⁴

What History Teaches

1. Displacement is real, but its mechanism matters. The bootmaking case shows that “jobs destroyed” and “workers harmed” are not the same thing. Incumbents were shielded; young people bore the cost through lost opportunities.
2. Geography matters. Job creation and destruction were unevenly distributed. The industry relocated; workers who could not migrate lost access.
3. Retraining is not always the answer. When decline unfolds through entry collapse, the policy focus shifts from retraining displaced workers to creating new opportunities for young people.
4. Scarring effects are long-lasting. The destruction of hand spinning drove hundreds of thousands of women from paid work and contributed to the male breadwinner family model — a social restructuring that persisted for generations.
5. History is the only laboratory for long-run analysis. We cannot run controlled experiments on technological unemployment; historical episodes are the only way to study long-term impacts.⁵

Open Questions

• Does the “entry collapse” pattern generalize to AI-driven displacement of cognitive work, or does the speed of AI adoption overwhelm the slow-adoption shielding that protected incumbents in the 19th century?
• Will AI create enough new tasks and occupations to offset the ones it destroys, as happened in bootmaking (144K new jobs for 152K lost)?
• What happens when the “new jobs” require skills that are themselves quickly automated?

Connections

• ai-mathematical-practice — AI’s impact on mathematical research: a contemporary case of technological change reshaping a skilled profession
• ai-economic-growth-complementarity — the growth-theory framing: whether AI is bounded by tasks, tacit complements, or neither; bootmaking’s entry-collapse is complementarity-constrained adjustment in historical form
• universal-basic-income — one policy response to technological unemployment
• post-scarcity-economics — the theoretical endpoint of labor-replacing automation
• dynastic-small-business — another historical model of labor organization under economic pressure

Sources

• Benjamin Schneider and Hillary Vipond 2023 — The Past and Future of Work: How history can inform the age of automation

Footnotes

1. raw/papers/schneider-vipond-past-future-work-2023.md ↩

2. raw/papers/schneider-vipond-past-future-work-2023.md ↩

3. raw/papers/schneider-vipond-past-future-work-2023.md ↩

4. raw/papers/schneider-vipond-past-future-work-2023.md ↩

5. raw/papers/schneider-vipond-past-future-work-2023.md ↩