Data brokers co-opetition

Data brokers share consumer data with rivals and, at the same time, compete with
them for selling. We propose a ‘co-opetition’ game of data brokers and characterize
their optimal strategies. When data are ‘sub-additive’ with the merged value net of
the merging cost being lower than the sum of the values of individual datasets, data
brokers are more likely to share their data and sell them jointly. When data are
‘super-additive’, with the merged value being greater than the sum of the individual
datasets, competition emerges more often. Finally, data sharing is more likely when
data brokers are more efficient at merging datasets than data buyers.