Potential social value from data : an application of discrete choice analysis

There is great interest in understanding the value of data, and different empirical approaches have begun to be explored, in the academic literature and in commercial contexts. However, the existence of informational externalities and the non-rival character of data immediately imply that private markets and uses alone will not deliver an economically efficient social availability of data, nor will any market prices (if they exist) reflect the social value. What’s more, the value of any given data set is also fundamentally determined by the value of the uses to which it can be put, which are likely unknown until after the fact. In this paper we apply a form of discrete choice experiment, applying it to the question of how public sector controllers of data could estimate the potential social value other users would gain from the data and also help identify which attributes of the data have particular value to users. This could inform decisions such as whether public bodies should sell data access to the commercial sector to cover their costs, or how much to invest in maintaining open access data bases by estimating a social benefit for a cost-benefit analysis. We use a commercially available discrete choice analysis software package, conjoint.ly, widely used in marketing and show how it can elicit preferences and marginal values capturing, at least to some degree, the data externalities about which there are, at present, few empirical insights. The method is straightforward to use and could be readily applied in practical contexts