Why do we get technology policy so wrong, so often? As governments rush to develop new identity policies they fail too often in answering essential questions: are identity policies capable of addressing a diverse range of policy goals? Are the techniques we imagine to be necessary in fact helpful? Instead, policy makers remain fixated on expensive and sexy 'biometrics' and vast new centralised databases to solve problems they do not understand. This survey seeks to resolve why policy makers repeatedly commission identity schemes based on obsolete knowledge of modern technological capabilities. We argue that policy making requires an understanding of technological issues as well as more traditional political and organisational concerns, and a little less bravado. As a result, policy makers can set about developing effective solutions that are citizen friendly and actually address pressing policy goals.