Post-Walrasian economics is not a doctrine, but a slogan announcing that something has to change. In this paper, I explore a conservative version of post-Walrasian economics that can be summarized as back to the methodology of Alfred Marshall's - particularly to his essay, "The Present Position of Economics" (1885). The Walrasian approach is a bottom-up, engineering vision: economics must be built on secure foundations of primitive theory, like a building. Marshall's approach is a top-down, archaelogical vision: the structure of economics must be uncovered starting with the observable facts and empirically determining what lies behind them. Marshall's methodology places the relationship between theory and empirical tools on center stage. The dominant tools of macroeconometrics are the vector autoregression (VAR) and calibration techniques, which were developed as reactions to two nearly simultaneous reactions to the Cowles-Commission program in econometrics: the Lucas critique and the Sims critique. The various strands of macroeconometrics are examined through the competing Walrasian and Marshallian visions of the role of theory in econometrics. A suggestion for a truly Marshallian (and post-Walrasian) econometrics is then offered.