In recent years government agencies have been encouraged to depend on other agencies to perform certain functions on their behalf. This ‘shared service’ model is said to allow agencies to focus on their core missions and achieve economies of scale. We investigate the application of the shared service approach to official statistics, wherein government statistical bureaus increasingly perform data collection and analysis for other agencies. A potential problem with this model is that statistical bureaus are supposedly independent from elected governments. To what extent are these bureaus willing to carry out work for other government agencies, many of which are under the direct command of the political executive? Drawing on interviews with senior officials in the Australian, Canadian, Dutch and Swedish statistical bureaus, we explore how statistical officials interpret and approach the issue of shared services. We argue that support for shared service models depends on how actors in autonomous agencies balance two competing aspirations: professional independence versus inter-organizational influence.
