Ownership and Non-Neutral Technological Change: Evidence from China's State-Owned Enterprise Privatization

URL SCAN: Ownership and Non-Neutral Technological Change: Evidence from China's State-Owned Enterprise Privatization
FIRST LINE: # Economics > General Economics

The Dissection

The paper isolates how Chinese SOE privatization altered capital-, labor-, and material-augmenting productivity between 1998 and 2008. Its central finding is narrower and uglier than a standard privatization success story: labor gains were immediate and front-loaded, capital gains did not improve, materials productivity declined modestly, and part of the labor improvement came from closing overstaffed plants. The apparent efficiency gain was therefore partly restructuring and selection—not a universal transformation of productive capacity.

The Core Fallacy

The paper is not a refutation of the Discontinuity Thesis. Its fatal limitation, if used as a guide to the AI transition, is historical extrapolation. Labor-augmenting productivity in a labor-intensive manufacturing era means the surviving workers became more productive. It does not mean human labor remains economically necessary.

The paper measures how ownership affected firms while labor was still embedded in production. DT concerns the later break: cognitive automation makes much of that labor economically redundant. A firm can become more labor-augmenting immediately before making most of its labor unnecessary. The paper tracks the improvement of the remaining workforce; it does not test whether a mass workforce remains indispensable.

Hidden Assumptions

• Historical productivity effects can be projected into an AI-dominated production regime.
• Labor-augmenting productivity benefits workers rather than merely increasing output per surviving worker after displacement.
• Privatization is the decisive causal lever, despite the simultaneous roles of restructuring, plant exit, credit access, market conditions, and state policy.
• Firm-level productivity gains aggregate into broad employment and wage gains.
• Displaced labor can be absorbed elsewhere through the wage-consumption circuit.
• Ownership reform can preserve productive participation once cognitive tasks become cheaper and better performed by machines.
• The observed persistence among 1998 incumbents represents durable transformation rather than cohort selection.

The abstract itself undermines the cleanest success narrative: exit of overstaffed plants contributed to the labor gain. Some of the improvement came from removing workers and firms from the measured system.

Social Function

Classification: partial truth and transition management; prestige signaling when deployed as a general defense of privatization.

The paper provides a legitimate historical result: ownership and restructuring can change the direction and distribution of productivity gains. Its anesthetic use is to turn a civilizational employment problem into a management variable. That move works while labor remains scarce or institutionally protected. Under P1–P3, it becomes irrelevant to the central question. Better ownership may determine who captures automated output, but it does not restore mass productive participation.

The Verdict

This is evidence about a specific Chinese manufacturing transition, not evidence that ownership reform can save the post-WWII labor system. It shows that productivity gains can coexist with plant closures, worker exit, and uneven technological effects. That is compatible with the Discontinuity Thesis, not a challenge to it.

When AI severs the labor-to-wage-to-consumption circuit, privatization merely changes the identity of the owner holding the machine. It does not turn displaced workers back into economically necessary participants.