Mihai Codreanu: The Innovation Effects of Vertical Integration: Competition and Human Capital
Mihai Codreanu: The Innovation Effects of Vertical Integration: Competition and Human Capital
Our project examines how large companies, known as "superstar firms," replace their suppliers with in-house production—a practice called backward vertical integration. We first create a novel empirical measure of technological backward vertical integration (henceforth, TBVI) based on historical patent data, allowing us to trace back events to at least 1976. Then, using the universe of online work profiles, job postings, job reviews, and matched firm-to-firm inventor movements, we document that the firms are increasingly hiring knowledgeable employees with previous experience in the targeted technology fields from incumbents (also the firms’ suppliers in most cases) to complete TBVI. Using several case studies and a simple model, we pose a possible negative direct effect of TBVI through poaching human capital on innovation productivity for upstream incumbents, which may further translate into a decrease in the quality of final products for competitor downstream firms. On the other hand, increased competition for the supplier market may incentivize innovation. In the case studies analyzed to date, the negative human capital channel seems to trump the positive competition effects.
I am grateful to The Europe Center for supporting my research with a grant, which was instrumental in advancing my study on the innovation effects of vertical integration in high technology sectors. Our project has an important empirical component: we collect and match data on patents, employee work profiles, job postings and reviews, mergers & acquisitions, regulatory changes, and final consumer products. The grant helped me collect and assemble part of this data, analyze it using cutting-edge methods such as AI classifiers, and talk to European and US experts about data sources, institutional context, and implications of this research.
This Research Project
Our project (joint with Winston Xu) examines how large companies, known as "superstar firms," replace their suppliers with in-house production—a practice called backward vertical integration. We first create a novel empirical measure of technological backward vertical integration (henceforth, TBVI) based on historical patent data, allowing us to trace back to at least 1976. In simple terms, we define TBVI events as a company starting to patent in new technology classes that its existing technologies heavily rely upon.
Using the universe of online work profiles, job postings, job reviews, and matched firm-firm inventor movements, we document that TBVI firms increasingly hire significant numbers of knowledgeable employees with previous experience in the targeted technology fields. In a series of case studies, we document many of these employees are directly hired from much smaller firms and previous suppliers. We also find that hiring these experienced employees significantly pre-dates the TBVI events we infer using our algorithm. Therefore, preliminary data suggests that hiring experienced workers is an important source of knowledge for firms expanding to new fields. Surprisingly, our sample shows very similar trends between European and US companies.
In a simple model, we then show that a firm adopting TBVI through poaching human capital can lead to less innovation for affected suppliers, which may translate into a decrease in the quality of final products produced by the other customers of these suppliers (often, the TBVI firm’s competitors). This eventually could result in fewer choices for consumers and higher markup charged to them. Using the fact that many of the supplier relationships were at the national level, we exploit variation in overlapping local labor markets to show that, indeed, in a number of case studies, firms that had their employees poached are significantly negatively affected. These experienced employees may not be easily replaced with new labor. Using new scraped third-party product review data and cutting-edge AI classifier methods, we show descriptive negative effects on consumer goods quality produced by firms that still rely on previous suppliers’ technology (in the limited settings we analyzed). Nonetheless, increased competition (from new entry) in the supplier market may theoretically increase innovation in this intermediate stage. Therefore, we aim to analyze more data and settings in the near future. This is not to say that all parties are potentially negatively affected. Using imputed wage data from online sources, we find suggestive positive effects on moving workers’ wages, who may be able to extract higher rents for their specialized knowledge. However, while workers who move to these big firms might see short-term wage increases, wages may decrease in the long run for both movers and stayers, if the human capital channel trumps the competition channel and market concentration increases.
Steps forward
Looking ahead, we are trying to expand our research to more settings and conduct a deeper, more nuanced analysis of these events. Our immediate plans include obtaining exogenous predictors of TBVI to better establish causal relationships and enhance the empirical robustness of our findings. By integrating theoretical frameworks from the existing literature on vertical integration, we aim to strengthen the connection between our empirical results and economic theory. We also plan to utilize the extensive regulation and M&A data we have collected to examine how policy changes and merger activities influence the dynamics of vertical integration and its impact on innovation.
Ultimately, we hope that our work will contribute valuable insights into how vertical integration affects innovation, supplier dynamics, and consumer welfare. These insights could inform both industry practices and policy decisions, particularly in the context of high-technology sectors where such dynamics are most pronounced.
I am deeply grateful for the resources and support provided by The Europe Center, which have been instrumental in advancing this work. I am enthusiastic about the potential discoveries that lie ahead and look forward to sharing our findings with the broader community.