The great detour and the need for macroeconomic theory

This paper makes three points. First, big data, advanced econometrics and sophisticated computational techniques do not obviate the need for macroeconomic theory and small, analytical models. The design and interpretation of empirical studies and large-scale models are guided by economic theory. Second, the current macroeconomic theory employs simplistic microeconomic assumptions with little empirical support, imposes heroic restrictions on interaction structures and agent heterogeneity, and applies the same models to economies with widely different structures and institutions. Third, a more fruitful structuralist and behavioral approach builds on empirical evidence of systematic, macroeconomically relevant behavioral patterns, recognizes the effects of structural and institutional factors, and places the interaction across sectors and between different types of decision makers at the center of the analysis.