When constitutional reforms delegated new responsibilities to Germany's states in 2006, many predicted they would lead to damaging competition between them in many aspects of policy. However, in terms of its impacts on prisons policy, care home regulation and civil servant pay, the story so far has been much more mixed.

There has been a trend towards greater decentralisation from the level of the nation state across the world in recent decades, but such transfers of power are seldom uncontentious. Advocates of decentralisation claim that it allows policies to be better tailored to their contexts, give greater capacity for experimentation, innovation and lesson-drawing, and also (more controversially) acts as a check on central power. Its opponents are fearful either of policy incoherence or – more problematically still – downward pressure on levels of service provision or entitlement, regulation and taxation, as a result of states competing with each other in order to remain competitive.

This essay examines the impact of Germany's 2006 federalism reforms, which transferred certain competencies from the level of the nation state to the regions, or L??nder. It starts by briefly examining the literature on the possible impacts of a decision to decentralise power, and then sketches the German context and the nature of these 2006 reforms. It then examines the changes that have occurred since responsibilities were transferred in three areas of policy: prisons, the regulation of care homes, and pay and conditions for Beamte, an employment category which includes most public servants.

It finds only limited evidence of 'races to the bottom', and rather more evidence of upward pressure on standards. It also uncovers some support for cooperation between regions, even where this is no longer legally required. Together, these findings give cause for scepticism about grand claims that decentralisation will lead to a 'race to the bottom'. Among other factors, the likely policy response to decentralisation depends to a large extent on the policy area in question, and on what the budgetary impact of policy variation would be.