Credit rating agencies are poorly understood institutions and thus far efforts to govern them through rule-making have been ineffective. In this paper, Timothy Sinclair and Ginevra Marandola argue attention to rules governing behaviour is actually mistaken when it comes to finance and rating agencies. The global financial crisis occurred not because of rule-breaking but because some relatively simple but crucial social relationships came apart and prevented market actors from transacting with each other, as they had prior to the crisis. This breakdown involves quite different sorts of rules to those normally considered by regulators. Regulation has not made rating more sound and transparent, and there seems little appreciation of the barriers to improvement. The authors highlight the weaknesses of the typical resort to regulative rules, and show why it has not worked in the case of the credit rating agencies. Rating cannot become a science, but the diachronic and constitutive approach advocated here should help prevent credit rating agencies becoming a catalyst to another global financial crisis.