This is really big. A machine learning study has found that the political leanings of economists not only influence the language they use to present their results in academic research, but also the very estimates they produce for several “objective” variables, such that the implied policy prescription is consistent with their partisan leaning. This might shake economics to its core. Economics has built much of its credibility on the presumption that its analyses are largely objective and that the role of politics is to choose trade-offs between objective benefits and costs on qualitatively different dimensions. This article demonstrates that these “objective” benefits and costs may not actually be objective but instead a biased function of the economist’s political orientation. Paper: