All rulers – no matter if autocratic or democratic – depend on administrators to implement their policies. Since perfect monitoring of bureaucratic agents involved in policy implementation by elected officials is impossible, the relationship between politics and administration is at the heart of government, and should thus also be at the heart of comparative politics. Yet, most cross-country research tends to overlook it, implicitly assuming policy implementation is a mechanical activity. A century old theoretical discussion – involving authors such as Max Weber (1978 [1921]), Woodrow Wilson (1887) and Frank Goodnow (1900) – has shown the fundamental importance of the relationship between politicians and administrations. Yet our knowledge of the more specific mechanisms of control and its empirical effects has so far been hampered by a lack of comparative studies. This panel therefore addresses the issue from a contemporary and historical perspective and offers papers that empirically investigate the effects on different outcomes (e.g. level of corruption, pro-growth policies) of different bureaucratic structures. In particular, the papers will explore the societal effects of the wide variety of ways used to control the administration, from extremely direct accountability (e.g. a patronage relationship between the ruler and the administrators) to a more rule based relationship (e.g. autonomous self-recruited civil servants). The papers include large-N cross-country (Parrado and Wegrich; Dahlström and Lapuente; Brösamle) and within-country (Lewis, Selin and Wood) comparisons as well as more in-depth case studies (González de Lara) aimed at uncovering the micro-foundations of the relationships between rulers and administrators.