Historians have recently recounted the ‘invention of the economy’ by experts, statisticians and economists in the early twentieth century as a measurable, techno-political object to be managed by national governments. Focusing on innovations in national accounting, these studies tend to treat the relationship between governmental enumeration and economic knowledge as unproblematic. However, these practices also have a history. Both the idea that statistics offered an objective numerical representation of reality, and that these could be interpreted as economic facts, also had to be invented. This paper explores these transformations in the context of nineteenth-century imperial governance in colonial NSW to broaden the chronological, geographical and intellectual parameters for understanding the pre-history of the economy. As nineteenth-century European governments produced an ‘avalanche in printed numbers’, colonies were important sites to experiment with the epistemological status of numerical information. Initially, the penal colony was monitored with musters and Blue Books as strategies in imperial surveillance. As these activities became publicly disputed, colonial bureaucrats began organising numbers with generalisations drawn from political economy to interpret government counting with a new validity, culminating in an annual Statistical Register (from 1857). This instituted new kinds of official memory from which ‘the economy’ was later constructed.