This article presents an overview and critique of the two leading macroeconomic approaches from the last 70 years: reasoning using high-level aggregates detached from individuals and their choices, and modeling using so-called microfoundations. We judge the validity of both methods, showing their inherent limits and deficiencies as explanatory and predictive tools of economics. We also underline several vital improvements, which are required if the models are supposed to guide policy decisions – even if this means a more modest and less (...) conceited approach. (shrink)