The Capital-Labor Split

In the AI Story

The arithmetic of AI-augmented production is uniquely lopsided. Segal describes the cost of Claude Code with the Max plan: one hundred dollars per person, per month. The productivity gain: a twenty-fold multiplier for a significant class of work. The surplus — the difference between the trivial cost and the enormous value of the additional output — is captured by whichever party has the bargaining power to claim it. In the absence of countervailing institutional mechanisms, the bargaining power belongs to capital: shareholders receive the productivity gain as margin expansion; workers receive their contracted compensation regardless of how much more they produced.

The historical mechanisms for moderating this dynamic — strong unions that bargained for wage increases reflecting productivity improvements, profit-sharing requirements that mandated distribution of augmented gains, progressive taxation that captured capital appreciation for public investment — are at their weakest point in the post-war era. Union density in the United States private sector is below seven percent. Profit-sharing is voluntary and concentrated among firms that were already generous. Capital gains taxation applies preferential rates lower than ordinary income. The countervailing mechanisms have been eroded precisely as the technological substrate that most rewards them has arrived.

The decision that Edo Segal describes in his quarterly board conversations — if five people can do the work of a hundred, why not have five? — is the capital-labor split expressed at the firm level. The decision to keep the team or to cut it is not operational but distributional: it determines how AI-augmented productivity is allocated between capital and labor within that specific organization. Segal's choice to keep the team is admirable. The structural incentive rewards the opposite choice: a firm that cuts reports higher margins, attracts investment on better terms, and gains competitive advantage. Individual ethics do not override structural incentives at the level of the economy.

The capital-labor split is the single most consequential distributional variable of the AI transition because it determines the default allocation of the surplus. Every other distributional mechanism — taxation, social insurance, international coordination — operates on whatever the split leaves behind. If capital captures ninety percent of the surplus, no realistic tax policy can redistribute enough to match the scale of the underlying concentration. The split must be addressed at the point of production, through bargaining structures and profit-sharing institutions that claim a share for labor before the surplus reaches the taxation stage.

Origin

The concept has classical roots in political economy — Ricardo, Marx, and later the post-Keynesian tradition all theorized the factor shares as a central variable in economic dynamics. The modern empirical study of labor's share was revived by work in the 2010s, particularly Loukas Karabarbounis and Brent Neiman's 2014 paper documenting the global decline in labor's share, and Piketty's Capital in the Twenty-First Century situating the decline in a longer historical frame.

Milanovic's contribution is connecting the capital-labor split to the homoploutia framework: the decline in labor's share matters differently in an economy where the same individuals hold both capital and labor income, because the political coalitions that might resist the shift are structurally different from the coalitions that resisted earlier capital-favoring shifts.

Key Ideas

Labor's share has declined for forty years. The trend predates AI but is structural and sustained, indicating that capital-friendly institutional arrangements have been accumulating for a generation.

AI accelerates the shift. The lopsided arithmetic — trivial tool cost, enormous productivity gain — generates a surplus whose default allocation flows to capital absent institutional redirection.

Countervailing mechanisms are weakened. Unions, profit-sharing, progressive capital taxation — the historical tools for redirecting productivity gains to labor — are at post-war lows precisely when they are most needed.

Structural incentives override individual ethics. The firm that captures the surplus as margin outcompetes the firm that shares it with labor. Individual leadership choices do not change the equilibrium.

The split is the foundational variable. Taxation and social insurance operate on what the split leaves. Addressing inequality at the tax stage alone is structurally inadequate when the split itself is the dominant mechanism of concentration.

Debates & Critiques

Economists disagree on how much of the declining labor share reflects measurement issues (the treatment of housing, software capitalization, healthcare benefits) versus genuine distributional shifts. The scholarly consensus has settled on 'a real decline of significant magnitude,' while acknowledging that precise magnitudes depend on definitional choices. For the AI context, the magnitude debate matters less than the direction: whichever adjustment is applied, the trajectory has been toward capital for four decades.