We study how commodity booms affect productivity using administrative microdata from Chile that combine firm exports by products and destinations, employer-employee records, and firm-to-firm production networks. Exploiting differential Chinese demand for Chilean commodities, we construct firm-specific commodity price shocks. We find three micro-level mechanisms. First, more exposed firms experience larger revenue increases but no differential productivity gains, channeling revenues into wages and materials. Second, among exposed firms, low-productivity firms expand employment while high-productivity firms do not, hiring workers from more productive employers. Third, domestic suppliers with greater indirect exposure show larger increases in sales and productivity. We develop a model with heterogeneous export wedges and labor market frictions where commodity booms can reduce sectoral aggregate productivity by exacerbating input misallocation—consistent with both the firm-level and aggregate evidence. Calibrated to Chile, this mechanism explains half of the mining TFP decline from 2005 to 2013.

US multinationals formed joint ventures in China for market access and lower labor costs. However, these ventures transfer technology to Chinese partners, fueling future competition. While individual multinationals weigh the risks to their own profits, they disregard the negative impact on other US firms and the broader economy, resulting in an over-investment that may reduce US welfare. In our empirical analysis, across industries, US firms in those with more joint ventures in China experienced worse outcomes, and within industries, firms with greater product overlap with the participating multinationals declined more. We develop a two-country model with oligopolistic competition, innovation, and joint ventures. For the US, the short-run gains from joint ventures are outweighed by long-run losses due to rising Chinese competition. Joint ventures benefit participating US firms at the expense of non-participants and the real wages of workers. A ban on joint ventures since 1999 would have boosted US welfare by 1.1 percent.

When should policymakers transition from adoption-led to innovation-led growth? Using South Korea's historical technology-transfer data, we document three facts on adoption: the advantage of backwardness, strategic pricing, and knowledge diffusion. We build a two-country growth model consistent with these facts. As the gap narrows, productivity gains from adoption diminish and adoption fees rise, eroding the effectiveness of adoption subsidies. Adoption's contribution to Korea's TFP growth falls from 31% to 5%, while innovation's rises from 2% to 24% (1973--2023). Korea's state-dependent shift from adoption to innovation subsidies improved welfare by 6.2%. Initially higher tariffs further raise welfare by lowering fees.

This paper develops an occupation-level measure of Capital-Embodied Innovation (CEI) by matching patents with capital goods based on their text similarity. The impact of CEI on labor demand is heterogeneous, depending on the similarity between capital and occupational tasks. Specifically, CEI associated with task-similar capital reduces the relative labor demand, whereas CEI related to task-dissimilar capital raises it. Between 1980 and 2015, abstract and non-routine occupations experienced more innovations in task-dissimilar capital and fewer in task-similar capital. CEI can explain a significant fraction of the task-biased labor market changes and the decline in labor share.