This paper jointly estimates firm labor market power and non-neutral productivity for Chinese manufacturing, and the effects of ownership transitions on each. I model the labor wedge as monopsony power net of an owner's political cost of unemployment, because public owners face employment goals from the government. At the same county employment share and size, state-owned enterprises (SOEs) carry a lower wedge than private firms. SOEs have lower labor-augmenting productivity than private firms. Privatization raises labor-augmenting productivity and the labor wedge, whose rise I read as a fall in the political cost of unemployment. SOEs' higher labor share mostly reflects lower labor-augmenting productivity. The wedge explains nearly half the labor share's fall after privatization.

Does Artificial Intelligence (AI) adoption bias technological change toward labor? We estimate the effect of AI adoption on labor-augmenting and Hicks-neutral productivity for Korean firms from 2017 to 2024. AI adoption raises labor-augmenting productivity by 6.1% and leaves Hicks-neutral productivity indistinguishable from zero, and the adoption effect on output is 1.6%. Manufacturing carries twice the labor-augmenting coefficient of information and communication (ICT). The labor efficiency comes through production processes in manufacturing and through product and service development in ICT. We recover firm markups and find no evidence that AI adoption raises them.

I estimate firm-level capital-, labor-, and material-augmenting productivity for Chinese manufacturing, 1998–2008, and the effect of state-owned enterprise (SOE) privatization on each. Results show technological change is strongly biased toward capital and labor. Private firms operate above SOEs on capital and labor, but below on materials. Privatization raises labor-augmenting productivity right after the transition, a front-loaded gain that fades on the full panel but persists among 1998 incumbents. It leaves capital-augmenting productivity no higher and modestly lowers materials. The SOE sector’s catch-up runs through restructuring of the plants the state kept, with exit of overstaffed plants adding to the labor gain.

Recent structural work identifies factor-augmenting productivity at the firm level, and industry-level aggregations and decompositions of these components have begun to appear. This paper asks what the industry aggregate of a factor-augmenting component should be. Two criteria—consistency with the industry’s effective input and consistency with growth accounting—both call for weighting each firm by its share of the augmented factor rather than by its share of output. The factor-weighted and output-weighted aggregates answer different allocation questions and differ by the revenue-weighted covariance between relative factor intensity and factor-augmenting productivity. When firms with higher labor-augmenting productivity are less labor-intensive, as when technology is labor-saving across firms, the covariance is negative and output weighting yields a larger aggregate. We extend cross-sectional, dynamic, and ownership decompositions to these aggregates and organize them in an allocation matrix that shows whether each factor is concentrated in firms that use it efficiently. Estimating a translog production function with capital-augmenting, labor-augmenting, and Hicks-neutral productivity for five Chinese manufacturing industries over 1999–2008, we find that revenue weights give faster growth than capital weights for capital-augmenting productivity in every industry, by 1.8 to 5.6 percentage points per year, and faster growth than labor weights for labor-augmenting productivity in four of the five. Revenue weights also attribute more of this growth to reallocation among surviving firms. That reallocation moved revenue, labor, and capital toward the same firms, those with high Hicks-neutral productivity, and was rarely factor-specific.