management of the ‘digital exhaust’ that we all emit by using the services provided by Facebook, Google and other online platforms (Gelb & Clark, 2013; Neef, 2015; USAID, 2017; Zuboff, 2019). For this article, however, I restrict myself to digital forms of governmentbacked identity documents or other credentials: this is the way that the term is used by the World Bank and UN agencies in their policy discussions (World Bank, 2018c). Nonetheless, the boundaries between ‘digital’ and ‘legal’ identity are blurry. The World Bank’s 2016 World Development Report (WDR), which focused on the development benefits from digital technologies, recommended that the best way to achieve the SDG legal identity target was ‘through digital identity systems, central registries storing personal data in digital form and credentials that rely on digital, rather than physical, mechanisms to authenticate the identity of their holder’ (Atick et al., 2016). The Bank argued that digital forms of official identity can increase access to both public and private services where civil registration is weak, or help to reduce some forms of corruption, such as double-dipping for entitlements or ghost workers in public employment. The increased availability of affordable technology to capture biometric details provides new ways to authenticate identity and helps to ensure uniqueness, creating much stronger levels of certainty that the person holding a document is the person to whom it was issued, or removing the need for a document altogether. Although the WDR also emphasised the importance of strengthening the ‘analog foundations of the digital revolution’ (World Bank, 2016, p. 2), and the Bank has provided important support to the strengthening of existing civil registration systems, it was suggested that low-income countries could also ‘leapfrog’ the paper-based stage, and move straight to digital identification (Gelb & Clark, 2013, pp. 3, 46; World Bank, 2016, p. 154). Of course, no modern civil registration or identification system could remain wholly paper-based; but the way in which digitalisation is conceptualised is important. A distinction is increasingly made between foundational and functional identity systems: a foundational identity is a system ‘primarily created to manage identity information for the general population and provide credentials that serve as proof of identity for a wide variety of public and private sector transactions and services’ (World Bank, n.d.; first proposed in Gelb & Clark, 2013). This foundational system is then drawn upon for various public ‘functional’ identities, such as registers for drivers’ licences, civil service employees, pensions, or public health insurance; as well as private sector applications such as SIM card registration and bank accounts. A model the Bank has supported for new digital identity systems in countries with low existing rates of official registration is the creation of a ‘foundational’ identity layer for all adults resident in the country without distinction, comprising a national identity number issued on the basis of biometric identification, leaving questions of legal status for later decision – as proposed for example in the project ‘West Africa Unique Identification for Regional Integration and Inclusion’ (WURI) (World Bank, 2018e, 2020). This approach draws heavily on the ‘Aadhaar’ biometric registration and identity number developed in India, where the ‘foundational’ layer (‘aadhaar’ means ‘foundation’) is not linked to legal status in the country: the only criteria for enrolment are residence in India and provision of biometric identifiers (Raghavan et al., 2019). The view is that the simple ability to prove who you are, and that you are the same person on the next interaction with the authority concerned, has benefits both for the person concerned who was previously unable to participate in the formal economy; and for the state, which can reduce fraud and 6