Following a rich intellectual history from Bresnahan (1982) to @steventberry-@PhilHaile (2014) it proposes a new test to discriminate between models of firm conduct.
You may recall that the same set of authors also recently published an excellent paper in AEJ: Micro which documents the stunning increase in common ownership profit weights in the S&P 500 over the last few decades. chrisconlon.github.io/site/common_ow…
I'll let others speak to the importance of the methodological contribution. The empirical results from breakfast cereals are interesting in their own right.
They add to recent structural work on markups by Seo & Park (2019) and entry by Ruiz-Perez (2019) in US airlines.
Backus et al reject an *exact* version of the common ownership hypothesis in which firms set RTE cereal prices directly in accordance with their investors' common ownership weights κ_fg.
However, the test does not reject versions in which <30% of κ_fg is transmitted to managers.
In my opinion their results speak to 2 important issues:
First, common ownership models should recognize that managers rather than investors make firm decisions. Agency problems between owners and managers likely attenuate the direct transmission of common ownership incentives.
Second, common ownership models should recognize that prices are just one of many firm decisions.
If managers can enhance firm productivity but pricing specialists set prices then common ownership would not result in higher markups but in higher costs. florianederer.github.io/common_ownersh…
If top managers can choose which markets to enter but pricing specialists set prices given these entry decisions then we would expect common ownership to mainly affect entry decisions rather than pricing choices.