There is by now a huge literature on the increase in the college premium and other dimensions of inequality in the United States and many other Western nations (see Acemoglu andAutor [2011] for an overview of this literature). As I discuss in the following text, the focal explanation in this literature is that technological changes of the last 4 decades have increased the demand for skills and have pushed up premia to different kinds of skills, college education among them (though other factors including globalization and changes in labormarket institutions have also contributed to these trends). The paper by Jaimovich, Rebelo, Wong, and Zhang tackles an important topic anddevelops a relatively underresearched line of inquirywithin this broad literature. The main idea is that a major contributor to the increase in the demand for skills has been “trading up” (the authors’ term) by households to higher-quality products as they have become richer. Higher-quality products are argued to bemore intensive in skilled labor. As a result, this process has naturally brought a higher demand for skills as a by-product of economic growth. This is an important idea, and one I sympathize with a lot. The paper also has a noteworthy original contribution in providing compelling motivating evidence. It estimates product quality froma variety of sources, links these to establishment-level demand for skills from the microdata of the Occupational Employment Statistics (OES) data set of the Bureau of Labor Statistics, and verifies that higher-quality products are more skill intensive than products of lower quality. This empirical work alone is worth more than the price of admission. But the paper does not fully deliver on this very promising research agenda. The reason why it fails to do that is interesting and instructive. It is because it follows a methodology I call quantitative Friedmanite modeling. This approach combines Friedman’s (1953/2008) famous

中文翻译：

---

评论

现在有大量关于美国和许多其他西方国家大学保费增加和不平等的其他方面的文献（见 Acemoglu 和 Autor [2011] 对这些文献的概述）。正如我在下文中所讨论的，这篇文献的主要解释是，过去 4 年的技术变革增加了对技能的需求，并推高了不同类型技能的溢价，其中包括大学教育（尽管其他因素包括全球化劳动力市场制度的变化也促成了这些趋势）。Jaimovich、Rebelo、Wong 和 Zhang 的论文解决了一个重要的话题，并在这一广泛的文献中发展了一条研究相对较少的调查路线。主要观点是，技能需求增加的一个主要因素是随着家庭变得更加富裕，家庭“交易”（作者的术语）到更高质量的产品。更高质量的产品被认为需要更多的熟练劳动力。因此，这一过程自然带来了对作为经济增长副产品的技能的更高需求。这是一个重要的想法，我非常同情。该论文在提供令人信服的激励证据方面也做出了值得注意的原创贡献。它从各种来源估计产品质量，将这些与来自劳工统计局职业就业统计 (OES) 数据集的微观数据的企业级技能需求联系起来，并验证更高质量的产品比低质量的产品。仅凭这项实证工作就比入场费更有价值。但是这篇论文并没有完全实现这个非常有前途的研究议程。它未能做到这一点的原因很有趣，也很有启发性。这是因为它遵循一种我称之为定量弗里德曼建模的方法。这种方法结合了 Friedman (1953/2008) 著名的