Abstract
We model the relationship between privacy concerns and consumer welfare in data provision under three types of data ownership. Data firms and intermediaries collect and process the data provided by consumers. Consequently, consumer welfare is influenced by the manner in which data are collected and processed. Our findings reveal that a compromised ‘partial data ownership’ regime, wherein consumers retain fundamental personal data rights and keep moderate data privacy concerns, is more favourable for optimising consumer welfare. This suggests that both full data ownership (i.e. consumer data utilisation is hindered by technical barriers and consumers have the control over data) and minimal data ownership (i.e. data are controlled by monopolistic data intermediaries) can result in a decline in consumer welfare, thus restricting the long-term growth of the data economy.
Notes
1 ISO/IEC, ‘Information Technology – Vocabulary’, see https://www.iso.org/obp/ui/#iso:std:iso-iec:2382:ed-1: v1:en:en.
2 However, Farboodi and Veldkamp (Citation2021) argue that data exhibit diminishing returns when they are used in predictions, primarily because improving prediction quality depends on reducing randomness, rather than on enlarging the dataset limitlessly.
3 Another modelling approach in economic growth is to take data as a kind of knowledge in the R&D sector, see e.g. Cong, Xie, and Zhang (Citation2021), and Jones and Tonetti (Citation2020).
4 Another reason for not adopting the assumption of a functional relationship between individual data volume and consumption is that once this form of assumption is adopted, the range of positive consumer utility may be narrowed. In other words, there would be fewer function forms that would satisfy the existence of a positive interval of consumer utility. In Cong et al. (Citation2022), although a functional relationship between data volume and consumption is set out, this relationship is not discussed much. Furthermore, this situation can affect the model solving and the portrayal of the data economy to some extent.
5 We consider the data intermediaries as a sort of processing departments, which process data products as an intermediary input in the final production process. Thus, we exclude the risk attitude problem of the intermediaries. In other words, the data intermediaries in our model are risk neutral. In some other studies, the data intermediaries are regarded as financial intermediaries, and so can be modelled as risk-averse entities. See Haddad and Muir (Citation2021).
6 Specifically, is required if
holds constantly.
7 Figures 1c and 1d have a similar situation. In those two cases, the crossing point appears when equals 0.56341 and 0.55648 respectively.