Research

Published Papers

🎗️(Social Science History, Special Issue, 2026) Quantifying Patenting by Women in the U.S., 1845-1924 with Mike Andrews and Enrico Berkes 

🎗️ (AEA Papers & Proceedings, 2025) Historical Differences in Female-Owned Manufacturing Establishments:  The United States, 1850--1880 with Richard Hornbeck, Martin Rotemberg, Anders Humlum (NBER Working Paper)

Related media coverage:

Podcast by Marketplace NPR

YouTube Video by Chicago Booth Review

Blogpost at Chicago Booth Review

🎗️The Interaction of Real and Financial Markets in the Global Economy: What Role Does China Play? (2019), Handbook of Global Financial Markets Transformations, Dependence, and Risk Spillovers, with Sumru Altug and Cem Cakmakli

Working Papers

🎗️(Revise & Re-submit, Research Policy)  Mapping International Technological Trajectories: Evidence from Multiple Patent Offices over Four Centuries with John Van Reenen, Antonin Bergeaud  

(CEPR) (CEP-LSE) (NBER) (VoxEU Piece)

We introduce a methodology to measure cross-country trends in innovation capability - "technological trajectories'' and implement this on a rich new patent dataset from 1836 to 2016. Intuitively, trajectories are revealed by a country's sustained increases in patenting across multiple jurisdictions. We first describe the data patterns, showing the relative decline of the UK, the rise of the US and Germany, and then Japan and China. We then econometrically estimate trajectories on (i) the post-1902 period for France, Germany, Japan, the UK and US, and (ii) the post-1960 period for a wider sample of 40 countries. Our trajectories are strongly positively correlated with TFP growth, and also (but less strongly) with the growth of labour productivity and capital intensity. We show that trajectories are predicted by several country attributes such as R&D, human capital, and defence spending.

🎗️(Reject and Re-submit, Research Policy) AI-Skilled Labor, Intangible Capital and Firm Productivity: Evidence from the TensorFlow Shock with Suleyman Faruk Gozen and Yusuf Ozkara 

This paper studies how the complementarity between intangible capital and AI-skilled workers shapes firm-level productivity. We exploit the 2015 release of Google TensorFlow as a plausibly exogenous shock to AI effectiveness. Using a difference-in-differences framework with value-added labor productivity and total factor productivity, we show that firms with a higher AI-worker share increase their productivity. The gains rise with firm size and intangible capital, and the effect is concentrated in the upper part of the productivity distribution. This evidence indicates that the complementarity between intangible capital and AI-skilled labor widens productivity dispersion. We also provide TensorFlow-specific evidence using job postings and worker profiles from Revelio Labs proprietary data.

🎗️(Revise and Re-submit, Journal of Corporate Finance) Artificial Intelligence and Workplace Misconduct with Suleyman Faruk Gozen, Mehmet Furkan Karaca, and Monika Tarsalewska

This paper studies how firms’ artificial intelligence (AI) capabilities, measured by the share of their workforce with machine-learning, natural-language-processing and computer-vision skills,  shape the effect of a major AI technology shock on workplace misconduct. Exploiting the open-source release of Google Brain’s TensorFlow machine-learning framework in a difference-in-differences design, we estimate intention-to-treat effects and find that firms with higher ex-ante AI intensity experience significant and persistent declines in workplace violations and penalty amounts. The  decline is accompanied by higher labor productivity and human capital investments, while we find no evidence of a labor-adjustment channel. Benefits are concentrated among larger, intangible- and organizational-capital intensive firms, which highlights uneven gains from AI capabilities and suggests that AI may widen compliance inequality among firms.

🌟(Under Revision) Property Rights and Women's Inclusion: Innovation and Business Ownership in the 19th Century  by Ruveyda Nur Gozen (SSRN)

Related media coverage: 

CEP Centrepiece Magazine 

LSE USAPP Blog

To what extent do inclusive institutions explain innovations and business dynamism? This paper examines the effects of Married Women’s Property Acts, which granted US women independent property rights during the nineteenth century, on women’s innovations and businesses. Using the universe of patents and a staggered-adoption design, I show that the reforms led to a substantial increase in women’s patenting, with effects emerging about six to eight years after enactment and growing afterwards. A back-of-the-envelope calculation suggests that these laws account for up to 59% of women’s patents after the reforms. Women’s inventions added to the volume of innovation without displacing men’s, and they were equally novel. The number of women who patented repeatedly rose even more than the number of one-time inventors, suggesting greater inventive effort by women. After the reforms, women were also more likely to assign their patents to others, which points to greater commercialization of their ideas and a wider market reach. Beyond patenting, the reforms also changed business ownership. Using the establishment-level Census of Manufactures, I show that the number of women-owned establishments rose by 30% to 81% after the reforms, with no change in the establishments owned by men, and that this increase came from new owners rather than inherited businesses. Finally, the effects reached beyond the owners themselves: after the reforms, the share of women among the workers of women-owned establishments rose, extending the gains to women in the wider labor force.

🌟Business Dynamism and Innovation: Evidence from a Novel 19th Century Microdata with Antonin Bergeaud, and Suleyman Gozen 

We construct a novel establishment-level dataset combining the nineteenth-century U.S. Census of Manufactures (1850–1880) with patent records from CUSP to examine how innovation shaped firm performance in early industrial America. Patenting establishments were substantially larger than comparable non-patenters across all dimensions—employment, capital, output, and output per worker—with premia further amplified among firms that produced breakthrough inventions. Estimating establishment-level production functions using a Levinsohn-Petrin approach, we find that patenting is associated with higher revenue total factor productivity (TFPR) while leaving markups unchanged on average. Breakthrough innovations, however, are associated with gains on both margins, with a size decomposition revealing the strongest amplification among the largest establishments — consistent with superstar dynamics — yet small breakthrough innovators also capture meaningful gains, suggesting the returns to invention were not exclusive to large firms.

🌟 Board Characteristics and AI Adoption with Suleyman Gozen and Sezer Yasar

We study the relationship between corporate board characteristics and firm-level artificial intelligence (AI) adoption. Linking BoardEx director data to Compustat for U.S. public firms, we first create a AI Board Index based on board members' breadth of experience in technology firms and technicality of their roles. Preliminary results suggest that firms whose directors score higher on this index adopt substantially more AI than their industry-year peers. We find that the more experience board members have at distinct technology firms and at leading technology firms, the more likely the firm is to adopt AI. We also find that firms with higher sales growth, greater R&D investment, and higher liquidity adopt more AI. Studying the roles of CEOs and boards separately, AI adoption is most likely when both board members and the CEO have stronger networks with technology firms. The results are robust to excluding the technology sector and to controlling for board and firm characteristics, industry-by-year fixed effects, and firm fixed effects. Finally, we use the deaths of board members as a quasi-random shock to board composition and find that the unexpected death of a more technical director while in office slows down the firm's AI adoption. 

🌟 Brexit and the Falling Innovation Dynamism with Ralf Martin, Esther Bøler, and Maxwell Read  

We study the effects of Brexit on innovation and international innovation collaborations in the UK. Our key finding is that overall innovation in the UK has declined and patent collaborations with the European Union fell significantly with Brexit. We find similar effects for research projects funded by the EU. Firms involved in collaborative innovation with scientists in the EU see a larger decline in patenting activity compared with other firms. We suggest that our results show a lower-bound estimate of the impact of Brexit on innovation in the UK, and that barriers to innovation could have more serious negative implications for UK economic outcomes in the long run.

Work-in-Progress

🌟 Immigration, Industrialization, and Innovation: Evidence from the Age of Mass Migration with Thor Berger (Lund University & CEPR) and Suleyman Faruk Gozen (University of Bristol)

🌟 Steam Power as General Purpose Technology with Thor Berger (Lund University & CEPR) and Suleyman Faruk Gozen (University of Bristol)