Carney (2018): The Future of Work (Whitaker Lecture)

Summary

"The Future of Work" is a speech delivered by Mark Carney, then Governor of the Bank of England, as the 2018 Whitaker Lecture, hosted by the Central Bank of Ireland in Dublin on 14 September 2018, in honour of T.K. Whitaker, the civil servant who steered Ireland's post-1958 shift to export-led growth. Source-access note: the task brief described the venue as the "Public Policy Forum, Toronto"; the primary source (BIS Review reprint title page, corroborated by the Bank of England's speech archive and the Central Bank of Ireland's Whitaker Lecture records) states Dublin — corrected here rather than reproduced. The Bank of England's own hosted-slides URL is confirmed dead; the full text was fetched and read in full (13 pages) from the BIS Review mirror, which matches the Bank of England's PDF in content.

The speech asks whether the "Fourth Industrial Revolution" — AI, automation, interconnectedness — will repeat the pattern of the first three industrial revolutions (steam, electrification, digitization), and closes with implications for UK monetary policy.

The Core Argument / Findings

A three-effect framework, credited to Acemoglu and Restrepo's 2018 NBER working paper, splits technology's labour-market impact into a destruction effect (technology replaces labour, reducing wages/employment), a productivity effect (technology raises productivity — eventually wages and demand — of those still working, analogous to Say's law), and a creation effect (new tasks emerge for labour). Over the medium term, destruction tends to outrun productivity, reducing labour's income share; historically the creation effect has eventually offset this.

Engels' pause as historical precedent. Carney invokes the early-19th-century UK episode in which productivity picked up sharply following industrialization, yet wage growth stalled and the labour share fell for decades before catching up — "a period dubbed 'Engels' pause.'" Even revolutions that eventually raise wages broadly can impose a long, painful transition first.

Could this time differ? Carney weighs scope, scale, and speed. On scope, prior waves substituted mainly for "hands" (manual, routine tasks), while AI may reach further into "heads" (cognitive tasks). On scale, he cites Frey and Osborne's (2017) estimate that roughly half of US employment is at "high risk" of automation, then qualifies it as ignoring task heterogeneity and economic (not just technical) feasibility; more conservative estimates (Nedelkoska and Quintini, OECD 2018) put high-risk jobs at roughly 10% in the UK and 15% in Ireland — comparable to past revolutions' 10–20% employment-share declines in the most-exposed industries. On speed, adoption cycles have shortened across revolutions.

Labour share and inequality. Carney states that globally "labour has seen its share of income fall over the past twenty years," citing the IMF (2017) finding technological progress the biggest contributor during the Third Industrial Revolution — with the UK and Ireland outliers where labour share stayed roughly flat. He states the mechanism plainly: "the more new technology substitutes for labour rather than complements it, the more the gains would accrue to capital rather than labour. An unequal distribution of capital means that increasing automation will push up inequality by construction" — reasoning traced to Kalecki (1954) and Marx on differing savings propensities by wealth. He also cites labour-market polarization (hollowing of mid-skill jobs) as established since the 1980s, generally attributed to technological displacement of routine tasks.

Policy responses. Carney's prescriptions are institutional and supply-side, not distributive: skills reform (citing Whitaker's own 1960s Irish reforms), retraining (UK Flexible Learning Fund, Singapore SkillsFuture), labour-market institutions (unions, minimum wages, cooperatives, Taylor Review protections), financial-sector modernization for credit access, and a monetary policy that lets structural change occur without accelerating or offsetting it, while factoring its effects into the equilibrium interest rate, equilibrium unemployment, and the Phillips curve. Despite naming unequal capital ownership as the mechanism turning automation into rising inequality, Carney proposes no redistribution of ownership, wealth tax, or basic-income transfer — a full-text search confirms no mention of land, economic rent, land value taxation, a citizen's dividend, universal basic income, or wealth redistribution.

Relation to the Georgist Case

The tie to Georgism here is thin: the speech never discusses land, economic rent, or any rent-capture instrument. Its value is twofold. First, it supplies a central bank governor's authoritative, widely cited framing of the historical base rate for technological transitions (Engels' pause), useful context for the wiki's AI-and-labour-share lane. Second, Carney's diagnostic sentence — that unequal capital ownership makes automation's gains flow disproportionately to capital "by construction" — is a plain-language statement of the same logic that motivates the wiki's Georgist and rent-capture case for AI (see Korinek & Stiglitz, who formalize the claim that whatever factor remains scarce absorbs a growing share of automation's gains). Carney states the diagnosis but stops well short of endorsing rent-capture or ownership redistribution; citing him as Georgist-adjacent evidence would overstate what the speech argues.

Nuances and Limits

• Genre. A lecture, not a research paper: it synthesizes and heavily footnotes others' work (Acemoglu & Restrepo, Frey & Osborne, Autor/Dorn/Hanson, IMF, OECD) rather than presenting new empirical findings of Carney's own.
• UK/Ireland-specific monetary framing. The equilibrium-interest-rate and Phillips-curve analysis is built for UK monetary policy; cited evidence mixes global (IMF), US (Frey & Osborne), and UK/Ireland-specific (OECD) sources.
• Self-qualified estimates. Carney flags the widely quoted "half of US jobs at risk" figure as likely an overestimate once economic feasibility is considered — a caution against treating any single automation-risk number, including ones cited elsewhere on this wiki, as settled.
• Not peer-reviewed, and as a live address, not later revised; some cited findings (e.g., Frey & Osborne 2017) have themselves been contested in literature the wiki carries elsewhere (see the IMF and Korinek & Stiglitz entries).

Bears On

A companion to the wiki's AI/automation research trio: Gen-AI: Artificial Intelligence and the Future of Work (IMF, 2024) supplies labour-market exposure estimates Carney's speech predates; Korinek & Stiglitz formalize the "scarce factor captures the gains" logic Carney states informally; Leduc & Liu model a bargaining-power mechanism complementing Carney's framework. Carney adds none of their formal modeling, but supplies the widest-circulation, most authoritative statement of the historical base rate (Engels' pause) and a clean articulation of why unequal capital ownership turns automation into inequality — useful for framing the wiki's AI-and-inequality lane, without itself being rent-capture evidence.

See Also

• Gen-AI: Artificial Intelligence and the Future of Work (IMF SDN 2024)
• Korinek & Stiglitz — AI, Innovator Rents and Non-Distortionary Redistribution
• Automation, Bargaining Power, and Labor Market Fluctuations (Leduc & Liu)
• Economic Rent

Sources

1. Mark Carney, "The Future of Work," 2018 Whitaker Lecture, Central Bank of Ireland, Dublin, 14 September 2018, reprinted as BIS Review r180914b. PDF — fetched and read in full — used for the venue/date correction, the destruction/productivity/creation framework, the Engels' pause discussion, the automation-risk and labour-share figures, all direct quotations, and confirmation (via full-text search) that land, economic rent, and UBI/redistribution instruments are not discussed.