Lectures & Workshops
The lectures and workshops are the heart of CISS 2026. Over the week, an international faculty shares current research and hands-on methods at the intersection of competition and innovation economics — from the macroeconomics of market power to merger policy, copyright, AI, and quasi-experimental research design. Expand any entry below to read the lecture title and abstract.
Since 1980, various aggregate measures of market power, such as markups, profit rates and market capitalization, have increased, in the US and around the globe. We discuss the challenges of measuring economy-wide market power, as well as a few leading likely determinants of this change. We find that the rise of aggregate markups is due in large part to the reallocation of market shares, as well as the rise of the inverse expenditure share on variable inputs. Instead, output elasticities have not risen. We then analyze the macroeconomic causes and consequences of this process, and use a model to evaluate the potential underlying mechanisms. Key in the transmission mechanism is the underlying heterogeneity in the distribution of parameters. This passes through to heterogeneity in the distribution of outcomes such as output, the labor share, profits and welfare. We illustrate that focusing on aggregate measures of market power is misleading as the same aggregate markups are associated with different output and welfare, depending on how the markups and underlying parameters are distributed. We conclude with a discussion of the role of policy, in particular, how competition policy can be a tool not only for restoring economic efficiency, but also for redistribution.
How do mergers reviewed by the European Commission affect innovation and markups? This lecture examines their effects by linking merging firms and their rivals identified in DG Competition’s merger case database to patent data from PATSTAT and financial accounts from Orbis. Using a firm-by-year panel, the analysis applies staggered event-study and difference-in-differences models with matched control groups.
The results show that citation-weighted patenting declines by about 18% for both merging parties and their non-merging rivals, on average over the first five post-merger years. Over the same period, markups increase by about 4% for merging firms and 3% for rivals.
Declines in innovation are concentrated among large firms and acquisition targets, appear across most broad technology fields, including several considered strategically important, and are not limited to the most concentrated markets. While efficiencies due to the combination of research activities could explain the decline among merging firms, the parallel decline among rivals is more consistent with weaker innovation competition. Likewise, simultaneous markup increases among merging firms and rivals are more consistent with greater market power than with merger-specific efficiency gains. Overall, the evidence provides little support for the view that EU merger enforcement has, on average, been too stringent.
This lecture considers what the effects of digitization can teach us about generative AI and copyright. It examines how new technologies affect demand for existing works, the creation and consumption of new content, and outcomes for established creators. Using evidence from the book market, the talk shows that digitization can both displace and increase demand for existing works. It shows evidence that AI-generated books have so far had little effect on human authors, although results for other industries may differ. It concludes by discussing whether and how copyrighted content could be licensed for AI training.
Mergers can weaken innovation incentives by reducing competition between existing firms or eliminating potential competitors. They can also bring together knowledge, expertise and complementary assets that make research more productive and help firms develop and commercialize new products. Technological closeness and product-market overlap can contribute to either outcome: they may create opportunities for cooperation and knowledge exchange, but also redundancy and incentives to protect existing profits. Assessing these opposing forces is a central challenge for merger control, particularly when innovation is uncertain and its consequences emerge over long periods. This talk examines that challenge in light of the European Commission’s proposed merger guidelines and recent theoretical and empirical research. It discusses how authorities can assess effects on ongoing research, future products and longer-term innovation, and what evidence can help distinguish potential harms from innovation benefits.
Empirical research in competition and innovation often relies on policy changes, regulatory interventions, eligibility rules, or other quasi-experimental variation to identify causal effects. This research workshop provides a short introduction to a set of causal inference tools particularly useful in these settings: matching and weighting methods, Difference-in-Differences, Synthetic Control methods, and Regression Discontinuity Designs.
The emphasis is on research design rather than implementation: when a method can provide a credible counterfactual, which identifying assumptions it requires, how these assumptions can be assessed, and which threats to identification are particularly relevant when studying firms, markets, innovation, and competition. The methodological discussion will be illustrated throughout with three recent applications in competition and innovation economics.
Being able to match large databases from different sources becomes more and more a required skill for empirical researchers as the low hanging fruits, meaning topics based on data from a single source, are often already harvested. Luckily, the additional effort not only opens opportunities for new but also deeper research. The main issue is the fact, that most data collectors have other things in mind than providing common keys for researchers. For example, a patent office does not collect VAT numbers for the applicants as these are not required for the legal status of the documents. Local administrations have to publish all EU funded projects on their webpages but do not provide links to firm level databases. Without a shared key, the match has to be based on mutual content, which usually is not harmonized. Whoever has tried to match by harmonization of firm address fields knows that this is a frustrating and time consuming procedure, especially for international data. This seminar empowers the participants to use a universal and free matching tool called SearchEngine. As the specific properties of every match are different and the quality of the result fundamentally depends on the customization of the SearchEngine, it also teaches the basics of the implemented heuristics and provides insights into the SearchEngine Machine Learning approach SEML to master matching projects of any scale.
