为什么 “魔鬼经济学” 未能改变经济学

经济学人：

The approach was fun, but has fallen out of favour

这种方法很有趣，但已经失宠了

“Economics is a study of mankind in the ordinary business of life.” So starts Alfred Marshall’s “Principles of Economics”, a 19th-century textbook that helped create the common language economists still use today. Marshall’s contention that economics studies the “ordinary” was not a dig, but a statement of intent. The discipline was to take seriously some of the most urgent questions in human life. How do I pay my bills? What do I do for a living? What happens if I get sick? Will I ever be able to retire?

“经济学是一门研究人类日常生活的学科。”阿尔弗雷德 · 马歇尔（Alfred Marshall）的《经济学原理》（Principles of Economics）就是这样开始的，这是一本 19 世纪的教科书，帮助创造了经济学家今天仍在使用的通用语言。马歇尔关于经济学研究 “普通” 的论点不是挖掘，而是一种意图的陈述。这门学科是认真对待人类生活中一些最紧迫的问题。我如何支付账单？我靠什么谋生？如果我生病了怎么办？我能退休吗？

In 2003 the New York Times published a profile of Steven Levitt, an economist at the University of Chicago, in which he expressed a very different perspective: “In Levitt’s view,” the article read, “economics is a science with excellent tools for gaining answers but a serious shortage of interesting questions.” Mr Levitt and the article’s author, Stephen Dubner, would go on to write “Freakonomics” together. In their book there was little about the ordinary business of life. Through vignettes featuring cheating sumo wrestlers, minimum-wage-earning crack dealers and the Ku Klux Klan, a white-supremacist organisation, the authors explored how people respond to incentives and how the use of novel data can uncover what is really driving their behaviour.

2003 年，《纽约时报》刊登了芝加哥大学经济学家史蒂文 · 莱维特（Steven Levitt）的简介，他在文章中表达了截然不同的观点：“在莱维特看来，” 文章写道，“经济学是一门科学，拥有获得答案的绝佳工具，但严重缺乏有趣的问题。莱维特先生和这篇文章的作者斯蒂芬 · 杜布纳（Stephen Dubner）将继续一起撰写 “怪胎经济学”。在他们的书中，几乎没有关于生活中的普通事务。通过以作弊的相扑选手、赚取最低工资的毒贩和白人至上主义组织三 K 党为特色的小插曲，作者探讨了人们如何对激励措施做出反应，以及使用新数据如何揭示真正驱动他们行为的因素。

Freakonomics was a hit. It ranked just below Harry Potter in the bestseller lists. Much like Marvel comics, it spawned an expanded universe: New York Times columns, podcasts and sequels, as well as imitators and critics, determined to tear down its arguments. It was at the apex of a wave of books that promised a quirky—yet rigorous—analysis of things that the conventional wisdom had missed. On March 7th Mr Levitt, who for many people became the image of an economist, announced his retirement from academia. “It’s the wrong place for me to be,” he said.

Freakonomics 很受欢迎。它在畅销书排行榜上的排名仅次于哈利波特。就像漫威漫画一样，它催生了一个扩展的宇宙：《纽约时报》的专栏、播客和续集，以及模仿者和评论家，决心推翻它的论点。它正处于一波书籍浪潮的顶峰，这些书籍承诺对传统智慧所遗漏的事物进行古怪而严谨的分析。3 月 7 日，莱维特宣布退出学术界，在许多人眼中，他已成为经济学家的形象。“这对我来说是错误的地方，” 他说。

During his academic career, Mr Levitt wrote papers in applied microeconomics. He was, in his own self-effacing words, “a footnote to the ‘credibility revolution’”. This refers to the use of statistical tricks, such as instrumental-variable analysis, natural experiments and regression discontinuity, which are designed to tease out causal relationships from data. He popularised the techniques of economists including David Card, Guido Imbens and Joshua Angrist, who together won the economics Nobel prize in 2021. The idea was to exploit quirks in the data to simulate the randomness that actual scientists find in controlled experiments. Arbitrary start dates for school terms could, for instance, be employed to estimate the effect of an extra year of education on wages.

在他的学术生涯中，莱维特先生撰写了应用微观经济学方面的论文。用他自己的话来说，他是 “'信誉革命'的脚注”。这是指使用统计技巧，例如工具变量分析、自然实验和回归不连续性，旨在从数据中梳理出因果关系。他推广了包括大卫 · 卡德（David Card）、吉多 · 因本斯（Guido Imbens）和约书亚 · 安格里斯特（Joshua Angrist）在内的经济学家的技术，他们共同获得了 2021 年的诺贝尔经济学奖。这个想法是利用数据中的怪癖来模拟实际科学家在对照实验中发现的随机性。例如，可以采用任意的学期开学日期来估计额外一年的教育对工资的影响。

Where the Freakonomics approach differed was to apply these techniques to “the hidden side of everything”, as the book’s tagline put it. Mr Levitt’s work focused on crime, education and racial discrimination. The book’s most controversial chapter argued that America’s nationwide legalisation of abortion in 1973 had led to a fall in crime in the 1990s, because more unwanted babies were aborted before they could grow into delinquent teenagers. It was a classic of the clever-dick genre: an unflinching social scientist using data to come to a counterintuitive conclusion, and not shying away from offence. It was, however, wrong. Later researchers found a coding error and pointed out that Mr Levitt had used the total number of arrests, which depends on the size of a population, and not the arrest rate, which does not. Others pointed out that the fall in homicide started among women. No-fault divorce, rather than legalised abortion, may have played a bigger role.

Freakonomics 方法的不同之处在于，正如该书的标语所说，将这些技术应用于 “一切隐藏的一面”。莱维特先生的工作重点是犯罪、教育和种族歧视。该书最具争议的一章认为，美国在 1973 年在全国范围内将堕胎合法化导致了 1990 年代犯罪率的下降，因为更多不受欢迎的婴儿在成长为犯罪青少年之前就被堕胎了。这是聪明屌丝类型的经典之作：一位坚定不移的社会科学家使用数据得出违反直觉的结论，并且不回避冒犯。然而，这是错误的。后来，研究人员发现了一个编码错误，并指出莱维特先生使用了逮捕总数，这取决于人口规模，而不是逮捕率，而逮捕率则不然。其他人则指出，凶杀案的下降始于女性。无过错离婚，而不是合法堕胎，可能发挥了更大的作用。

Other economists, including James Heckman, Mr Levitt’s colleague in Chicago and another Nobel prizewinner, worried about trivialisation. “Cute”, was how he described the approach in one interview. Take a paper on discrimination in the “The Weakest Link”, a game show in which contestants vote to remove other contestants depending on whether they think they are costing them money by getting questions wrong (in the early portion of the game) or are competition for the prize pool by getting them right (later on). That provided a setting in which Mr Levitt could look at how observations of others’ competence interacted with racism and sexism. A cunning design—but perhaps of limited relevance in understanding broader economic outcomes.

其他经济学家，包括莱维特在芝加哥的同事、另一位诺贝尔奖得主詹姆斯 · 赫克曼（James Heckman），都担心这种微不足道。“可爱”，这是他在一次采访中描述这种方法的方式。以 “最薄弱的环节” 中的一篇关于歧视的论文为例，这是一个游戏节目，参赛者投票决定是否删除其他参赛者，这取决于他们是否认为自己因为答错问题而花钱（在游戏的早期），或者通过答对问题来争夺奖池（后来）。这为莱维特提供了一个环境，让他可以研究对他人能力的观察如何与种族主义和性别歧视相互作用。这是一个狡猾的设计，但在理解更广泛的经济结果方面可能意义有限。

At the heart of Mr Heckman’s critique was the idea that practitioners of such studies were focusing on “internal validity” (ensuring estimates of the effect of some change were correctly estimated) over “external validity” (whether the estimates would apply more generally). Mr Heckman instead thought that economists should create structural models of decision-making and use data to estimate the parameters that explained behaviour within them. The debate turned toxic. According to Mr Levitt, Mr Heckman went so far as to assign graduate students the task of tearing apart the Freakonomics author’s work for their final exam.

赫克曼先生批评的核心是这样一种观点，即此类研究的从业者关注的是 “内部有效性”（确保对某些变化的影响的估计得到正确估计）而不是 “外部有效性”（估计是否更普遍地适用）。相反，赫克曼认为，经济学家应该创建决策的结构模型，并使用数据来估计解释其中行为的参数。辩论变得有毒。根据莱维特先生的说法，赫克曼先生甚至给研究生分配了一项任务，即在期末考试中拆解《怪胎经济学》作者的作品。

Did you know... 你知道吗...

Neither man won. The credibility revolution ate its own children: subsequent papers often overturned results, even if, as in the case of those popularised by Freakonomics, they had an afterlife as cocktail-party anecdotes. The problem has spread to the rest of the profession, too. A recent study by economists at the Federal Reserve found that less than half of the published papers they examined could be replicated, even when given help from the original authors. Mr Levitt’s counterintuitive results have fallen out of fashion and economists in general have become more sceptical.
两人都没有赢。可信度革命吞噬了自己的孩子：随后的论文经常推翻结果，即使像那些由Freakonomics推广的论文一样，它们作为鸡尾酒会的轶事有来世。这个问题也蔓延到了该行业的其他领域。美联储经济学家最近的一项研究发现，他们研究的已发表论文中只有不到一半可以复制，即使得到了原作者的帮助。莱维特的反直觉结果已经过时，经济学家普遍变得更加怀疑。

Yet Mr Heckman’s favoured approaches have problems of their own. Structural models require assumptions that can be as implausible as any quirky quasi-experiment. Sadly, much contemporary research uses vast amounts of data and the techniques of the “credibility revolution” to come to obvious conclusions. The centuries-old questions of economics are as interesting as they always were. The tools to investigate them remain a work in progress. ■
然而，赫克曼所青睐的方法也有其自身的问题。结构模型需要的假设可能与任何古怪的准实验一样难以置信。可悲的是，许多当代研究使用大量数据和“可信度革命”的技术来得出明显的结论。几个世纪以来的经济学问题一如既往地有趣。调查它们的工具仍在进行中。