Competition and Privacy in Online Markets: Evidence from the Mobile App Industry

Abstract

Policy makers are increasingly concerned about the combination of market power and massive data collection in digital markets. This concern is fueled by the theoretical prediction that more market power causes firms to collect ever more data from their users. We investigate the relationship between market power and data collection empirically. We analyze data about more than 1.5 million mobile applications in several thousand submarkets of Google's Play Store. We observe these data for over two years and combine information on an app's data collection with information about its competitive environment. Our analysis highlights a robust positive relationship between market power and data collection. We find that more data are being collected in concentrated markets, and apps with higher market shares collect more data. This pattern robustly emerges across a series of cross-sectional and panel regressions as well as a series of specifications that exploit exogenous variation.

Competitors’ Reactions to Big Tech Acquisitions: Evidence from Mobile Apps

“…Google suggested up to 24 and later 50 of them according to the similarity. While Google does not disclose the exact algorithm of how it identifies the set of relevant apps, the selection, also according to app developers, seems to be based on similarity in app characteristics (see Kesler et al (2019) for a discussion). Although this can be considered quite narrow, we use this as our first definition of very close competitors following other studies (Kesler et al, 2019;Wen and Zhu, 2019).…”

“…While Google does not disclose the exact algorithm of how it identifies the set of relevant apps, the selection, also according to app developers, seems to be based on similarity in app characteristics (see Kesler et al (2019) for a discussion). Although this can be considered quite narrow, we use this as our first definition of very close competitors following other studies (Kesler et al, 2019;Wen and Zhu, 2019). 8 Taking only the similar apps at the time of acquisition complicates the analysis, as some apps are not observed regularly pre-and post-acquisition due to late entry, early exit, and missings in our dataset.…”

“…The outcome variables Y imt of competing app i, in market m, at time t are the different variables measuring innovation, quality, and, to a lesser extent, prices. In the main results, we present regressions on the different update dummy variables, the privacy-sensitive permissions dummy 16 There are some theoretical papers suggesting that more market power leads to less privacy and more data collection (Casadesus-Masanell and Hervas-Drane, 2015; Dimakopoulos and Sudaric, 2018), while there is only scarce empirical evidence confirming this positive correlation (Preibusch and Bonneau, 2013;Kesler et al, 2019).…”

“…For a detailed overview of all variables contained in the Play Store Data, seeKesler et al (2019). 7 See, as an example for Google, https://en.wikipedia.org/wiki/List_of_mergers_and_acquisitions_by_ Alphabet.…”

An integrated approach to competition regulation and data protection in India

“…The growing consensus around the fact that privacy protection could be a parameter of competition on product quality, particularly in zero-price markets, points to an alternative [21]. Kesler et al [29] provide empirical evidence for this, assessing two million apps in the Google Play Store to find a positive correlation between app market share and more data collection. The corollary is that the quality of data protection and privacy will be better in competitive markets.…”