The microeconometrics of household behaviour: building the foundations, 1920-1960

This thesis explores the early history of microeconometrics of household behaviour from the interwar period to the 1960s. The analytical framework relies on a model of empirical knowledge production that captures the scientific progress in terms of its materialistic supplies and intellectual demands. Under this framework, the thesis traces how microdata at the household level first appeared in the econometrician’s research agenda and how the discipline was consolidated by communities of econometricians. This study contains four substantive chapters. The first three chapters are selective case-studies charting three important approaches in the development of microeconometric practices. The first chapter reviews the interwar literature. Among those decentralised practices, Arthur Bowley’s analysis on family expenditure stands out as one of the earliest exemplars. The second and third chapters explore the formation of two communities in the post-war period: Richard Stone’s Department of Applied Economics (DAE) at Cambridge, and Guy Orcutt’s Social Systems Research Institute (SSRI) at Wisconsin. With the benefit of the new microdata and the introduction of computer-based calculation, Stone and his crew created a cooperative group that produced the first series of microeconometric publications driven by intellectual problems and economic questions. By contrast, Orcutt came to the analysis of microdata driven by his dream of microsimulation, a bottom-up method of microeconometric modelling, more heavily dependent on computing power and designed for revising public policies. After frustration at the SSRI, he finally finished a household simulation model at the Urban Institute. Taking the DAE and SSRI as examples, the fourth chapter assesses both the internal academic relationships of these groups, and the consolidation of both literatures using bibliometric data and network analysis. The results demonstrate the ways in which the DAE was a more interconnected network than the SSRI. The citation analysis offers an alternative way in understanding the formation of econometric knowledge based on community relations rather than the supply of materials or intellectual demands