Part two of this book explores the various challenges that must be confronted in defining a ‘benchmark’ of accountability. It commences by identifying the underlying rationales for accountability, and then addresses the various issues thrown up when seeking to transform the mechanistic description of accountability (ie who is accountable, to whom, for what and how) into a normative account (ie who should be accountable to whom, for what and how). If we were able to do so, we would then have a yardstick against which we could measure the performance of accountability mechanisms and find them wanting.