因果关系评估的准实验设计与实证经济学的可信性革命——2021年度诺贝尔经济学奖得主主要经济理论贡献述评

The Quasi-experimental Design Approach to Causality and the Credibility Revolution in Empirical Economics: A Review of Main Contributions by 2021 Nobel Economics Laureates

李宝良,郭其友

2021 第 43 卷 第 11 期 页码 140-152

DOI: 10.16538/j.cnki.fem.20211101.101

PDF全文下载 HTML全文

摘要

如何使用观察数据对变量之间的因果关系进行可信评估是实证经济学要面对的一个普遍问题。瑞典皇家科学院将2021年度诺贝尔经济学奖授予大卫•卡德(David Card),以表彰其对劳动经济学实证研究的贡献,以及约书亚•安格里斯特(Joshua D Angrist)和吉多•因本斯(Guido W Imbens),以表彰其对因果关系分析的方法性贡献。他们借助自然实验对因果关系评估进行准实验设计,显著提高了从观察数据中进行因果关系评估的可信性。对准实验设计方法和实证研究的回顾表明,各种准实验设计都有相同的基本逻辑,即在社会经济中找到一种场景也就是“自然实验”来模仿随机控制实验,评估反事实结果从而推断经济变量之间的因果关系;准实验设计的质量是决定实证研究质量和结论可信性的关键,高质量的准实验设计来自对自然实验背后相关制度运作过程的深刻理解和透彻分析。国内经济学者在借助准实验设计以科学规范的经济学研究方法讲好中国故事时,要注重对自然实验背后的相关制度进行深度的调查分析,规范准实验设计过程,避免P值操纵,并要加强影响机制的理论研究;政策制定者在利用准实验研究结论时,必须关注准实验设计研究结论的局限性和外部有效性。

How to evaluate the causal relationship between economic variables credibly using observational data is a common problem faced by empirical economists. The Royal Swedish Academy of Sciences awarded the 2021 Nobel Prize in economics to David Card for his empirical contributions to labor economics, and Joshua Angrist and Guido Imbens for their methodological contributions to the analysis of causal relationships. This paper reviews how they devise quasi-experimental designs for causality evaluation with the help of natural experiments and the credibility revolution in empirical labor economics with emphasis on the analysis of century-old questions: the employment effects of minimum wage, the labor market impact of immigration and educational investments. The main conclusions are as follows: Firstly, various seemingly different quasi-experimental designs of causality evaluation have the same basic logic. They all use the scene of natural experiments to imitate the way of randomized controlled trials (RCTs), and look for research objects with the same or similar characteristics as the control group that constitutes the counterfactual result of causality evaluation. Moreover, a survey of the classical literature of empirical labor economics shows that the credibility of causality evaluation does not depend on the data and econometric methods themselves, but on the in-depth understanding and thorough analysis of the operation mechanism of social and economic system. Therefore, the in-depth investigation and discussion of natural experiments is essential. Finally, China’s economic reform provides abundant natural experimental resources for quasi-experimental designs. Researchers, who want to tell the Chinese stories with scientific and standardized economic research methods, should pay attention to the in-depth investigation and analysis of the systems behind natural experiments, standardize the quasi-experimental design process to avoid p-hacking, and strengthen the theoretical research of influence mechanism. Policy-makers should pay attention to the limitations and external validity of the research conclusions when learning from quasi-experimental designs to evaluate the effective of policy.

参考文献

1 安格里斯特, 皮施克. 基本无害的计量经济学[M], 郎金焕、李井奎译, 格致出版社, 2012.
2 安格里斯特, 皮施克. 精通计量: 从原因到结果的探寻之旅[M], 郎金焕译, 格致出版社, 2019.
3 李宝良, 郭其友. 因果关系的实地实验与新实证发展经济学的贫困治理之道——2019年度诺贝尔经济学奖得主主要经济理论贡献述评[J], 外国经济与管理, 2019, 41(11): 136-152.
4 陆毅, 孙天阳. 以科学规范的经济学研究方法讲好中国故事[J], 经济学(季刊), 2021, 21(5): 1877-1882.
5 Acemoglu D, Simon J, and James A. Robinson. The colonial origins of comparative development: An empirical investigation[J]. American Economic Review,2001, 91(5): 1369-1401. DOI:10.1257/aer.91.5.1369
6 Angrist J D, Krueger A B. Does compulsory schooling attendance affect schooling and earnings?[J]. Quarterly Journal of Economics,1991, 106(4): 979-1014.
7 Angrist J D, Lavy V. Using Maimonides’ rule to estimate the effect of class size on scholastic achievement[J]. Quarterly Journal of Economics,1999, 114(2): 533-575.
8 Ashenfelter O. Estimating the effect of training programs on earnings[J]. Review of Economics and Statistics,1978, 60(1): 47-57. DOI:10.2307/1924332
9 Card D. The impact of the Mariel boatlift on the Miami labor market[J]. Industrial and Labor Relations Review,1990, 43(2): 245-257.
10 Card D. Do minimum wages reduce employment? A case study of California 1987–1989[J]. Industrial and Labor Relations Review,1992, 46(1): 38-54.
11 Card D, Krueger A B. Does school quality matter? Returns to education and the characteristics of public schools in the United States[J]. Journal of Political Economy,1992, a,100(1): 1-40.
12 Card D, Krueger A B. School quality and black-white relative earnings: A direct assessment[J]. Quarterly Journal of Economics,1992, b,107(1): 151-200.
13 Card D, Krueger A B. Minimum wages and employment: A case study of the fast-food industry in New Jersey and Pennsylvania[J]. American Economic Review, 1994, 84(4): 772-784.
14 Imbens G W, Angrist J D. Identification and estimation of local average treatment effects[J]. Econometrica, 1994, 62(2): 467-475.
15 Jackson K, Johnson R, Persico C. The effects of school spending on educational and economic outcomes: Evidence from school finance reforms[J]. Quarterly Journal of Economics,2016, 131(1): 157-218. DOI:10.1093/qje/qjv036
16 Peri G, Vasil Y. The labor market effects of a refugee wave: Synthetic control method meets the Mariel Boatlift[J]. Journal of Human Resources, 2018. doi: 10.3368/jhr.54.2.0217.8561R1.