The designation of “sub-prime” had a lot to do with the Byzantine structure of credit ratings agencies, which at times appeared arbitrary and at other times seemed to follow a purely punitive logic, combining statistical models based on general data with information about individuals’ past borrowing history as well as socio-demographic data and with the more subjective impressions of remote functionaries crunching numbers and assessing risk in offices half a continent away.

Max Haiven, Cultures of Financialization, loc. 194