Over a period of twelve years, Barbara Bergmann developed several models of the labor market using microsimulation, eventually integrated in a 'Transactions Model' of the entire US economy, built with Robert Bennett and published in 1986. The paper reconstructs the history of this modelling enterprise in the context of the debates on the microfoundations of macroeconomics and the role of macroeconomic expertise from the 1970s stagflation to the late 1980s. It shows how a political element-her focus on distributional effects of policies-was central to her criticism of macroeconomic modelling and how both her epistemic and political positions were increasingly marginalized in the 1980s.