Teams and Text: Collaborative Innovation in the Knowledge Space

[latest version]

Job Market Paper 2024

Econ Best Job Market Paper Award 2024 UniCredit (Runner-up)

IIn this paper, I study the impact of an expanding scientific and technological frontier on team innovations. To do so, I present a novel framework that integrates inventor teams and their patent texts. I model collaboration directly through a Bayesian model of Natural Language Processing. Applied to patent text data, this model builds a map of inventors, teams, and research fields, referred to as the knowledge space. Trained on over 400,000 U.S. patents from the USPTO PatentsView database, this framework allows me to tackle unanswered questions on how teams create new knowledge. Specifically, I investigate the effect of prior work on a team’s ability to produce a breakthrough–an innovation that sparks a new and successful research field. Leveraging high-dimensional patent text data, I back out two new measures: breakthrough patents and a team’s knowledge field, the set of research fields accessible to the team. I combine this with data on premature inventor deaths as a quasi-natural experiment. This identifies how team innovations change as they pivot to more or less advanced research fields. The framework unifies key elements of collaboration. Teams build on existing knowledge, and prior work both supports and obstructs innovation. I show that teams generate more breakthroughs when building on enough prior work to incorporate valuable knowledge, but not so much as to stifle novelty.

Keywords: Teams, Innovation, Patents, Topic Modelling

From Shares to Machines: How Common Ownership Drives Automation

With Dennis C. Hutschenreiter, Felix Noth, Stefano Manfredonia and Tommaso Santini

[IWH-Discussion Paper Series]

Submitted & under review

Does increasing common ownership influence firms’ automation strategies? We develop and empirically test a theory indicating that institutional investors’ common ownership drives firms employing workers in the same local labor markets to boost automation-related innovation. First, we present a model integrating task-based production and common ownership, demonstrating that greater ownership overlap drives firms to internalize the impact of their automation decisions on the wage bills of their local market competitors, thereby fostering more automation and reducing employment. Second, we empirically validate the model’s predictions. By analyzing patent texts, the geographic distribution of firms’ labor forces at the establishment level, and exogenous increases in common ownership due to institutional investor mergers, we isolate the effects of rising common ownership within and across labor markets. Our findings reveal that firms experiencing a positive shock to common ownership with labor market rivals exhibit increased automation, coupled with a decrease in employment. Conversely, similar ownership shocks do not lead to heightened automation innovation if firms do not share local labor markets.

Keywords: Common Ownership, Automation, Local Labor Markets, Market Power.

Acquiring Talent: Using M&As to redirect technical change

When a firm makes an acquisition, they not only purchase the stock of patents, but also their inventor human capital. This paper develops a model of firm dynamics which captures a potentially important mechanism in how firms redirect their technological direction. By incorporating new inventors into their research teams, firms can increase profits by competing in new markets. I develop and simulate a model of firm choice within the Knowledge Space: a mapping of inventors and patents. This space defines an innovation production process, and this paper extends this by imposing barriers to collaboration through firm boundaries. This paper shows that a carbon tax causes firms to acquire green talent and shift their production process towards lower carbon emitting goods.

Keywords: Innovation, Mergers & Acquisitions, Teams, Technical Change

Learning from Others: Optimal Panel Design for Forecasting

With Hannes Mueller, Christian Fons-Rosen and Renato Vassallo

This paper examines how combining optimal panel design with non-linear machine learning models can enhance forecast accuracy in time series analysis. While a country’s historical data may reliably predict future outcomes during stable periods, unprecedented crises—such as inflationary shocks, banking crises, or conflicts—require data from other countries to provide meaningful insights. The study explores the balance between incorporating valuable cross-country information and minimizing noise, ultimately proposing methods to select an optimal set of countries for forecasting in different crisis contexts.