The tension between the goals of integrated, seamless public services, requiring more extensive data sharing, and of privacy protection, now represents a major challenge for UK policy-makers, regulators and service managers. In Part I of this article ( see Public Administration volume 83, number 1, pp. 111-33), we showed that attempts to manage this tension are being made at two levels. First, a settlement is being attempted at the level of general data protection law and the rules that govern data-sharing practices across the public sector. We refer to this as the horizontal dimension of the governance of data sharing and privacy. Secondly, settlements are also being attempted within particular fields of public policy and service delivery; this we refer to as the vertical dimension. In this second part, we enquire whether risks to privacy are greater in some policy sectors than others. We do this, first by showing how the Labour Government's policy agenda is producing stronger imperatives towards data sharing than was the case under previous administrations in three fields of public policy and services, and by examining the safeguards introduced in these fields. We then compare the settlements emerging from differing practices within each of these policy sectors, before briefly assessing which, if any, principles of data protection seem to be most at risk and in which policy contexts. Four strategies for the governance of data sharing and privacy are recapitulated - namely, seeking to make the two commitments consistent or even mutually reinforcing; mitigating the tensions with safeguards such as detailed guidelines; allowing privacy to take precedence over integration; and allowing data sharing to take precedence over privacy. We argue that the UK government has increasingly sought to pursue the second strategy and that the vertical dimension is, in practice, much more important in defining the settlement between data sharing and privacy than is the horizontal dimension. This strategy is, however, potentially unstable and may not be sustainable. The conclusion proposes a radical recasting of the way in which the idea of a 'balance' between privacy and datasharing imperatives is conceived.
