This book attempts to meet that very standard by focusing on what works. It attempts to move forward the empirical efforts of Tinbergen, Goldberger, Klein, Eckstein, and Fair the past 80 years to determine what works. That is, the effort to convert economics from just theory to hard (by which I mean reliable) science. Doing so requires three things. First, it requires that the postulates we test have some economic meaning, and not be just some collection of variables we are “running up the flagpole,” to see what happens. Second, it requires that the theory-based postulates we test are structured loosely enough so that thedatadetermine what is real, i.e., the exact shape and content of the theory being tested. It is not for us to say a priori by how we structure the model we test, whether Keynes’ consumption function, whose principal determinant is current income, is correct, or whether Freidman’s, whose principal determinant is average income (permanent income) is correct. Third, it is not for us to claim some empirical result proves some theory iscorrect,simplybecause itexplainssome variationintheeconomy, insome timeperiod,insome economicmodel.Tobecorrect,itshould explainmost variance, in most or all time periods, in most or all models. This book tries to adhere to these three rules, we think successfully. To meet the first condition, its model is built around the theory that we found most consistent with the data. To meet the second, the shape (and inclusion)ofeachequationinthemodel isdata-determined, e.g.,thereare no predetermined assumptions about what drives consumer or investment spending. Third, a large-scale econometric model is needed to capture all the sources of economic variation, and that’s what is used. Extensive robustness testing was used to prove that any initial statistical finding was real and not just some spurious artifact of the time period or particular model tested.

本书2017年出版，480页，部分目录如下：