Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This note explores clustering in cross country GDP per capita using recently developed model based clustering methods for panel data. Previous research characterizing the components of the overall distribution of output either use ad hoc methods, or methods which ignore/subvert the panel nature of the data. These new methods allow the characterization of the possible autoregressive relationship of output between time points. We show that traditional static clustering decade by decade gives mixed results regarding clustering over time, while the application of longitudinal mixtures presents three distinct clusters at all periods of time.