.png)
Finding the cause of an event without confusing it with a simple correlation is a task that not only economists and scientists have to solve. We often face it in our lives. For example, officials have to deal with it in order to understand whether government spending has helped to solve a particular problem. NES Professor Olga Kuzmina picked books that reveal this fundamental topic in plain and simple language and help find cause-and-effect relationships.
MIT Professor Joshua Angrist, who received the 2021 Nobel Prize for his work on "natural experiments", and Professor of the London School of Economics Jorn-Steffan Pischke introduce a reader to the world of econometrics. They use fascinating, sometimes almost detective and sometimes humorous examples from real life, showing how exciting the research process in economics can be.
Does health insurance make you healthier? Should the state rescue banks during crises? Was there a chance to save the life of the wife of the famous football player O. J. Simpson? These are some of the questions that Angrist and Pishke talk about to explain how econometrics work.
After all, if you think about it, even a seemingly obvious fact that "higher education brings extra earnings" cannot be confirmed just by comparing average salaries of people with and without a bachelor's, master's or doctor's degree. A more ambitious or well-off person is more likely to get a higher education, but even without it, he or she would most likely earn more as compared to less bright or initially poorer people. In an ideal dream world, we would like to make a clone of an experimental agent at the moment of making a decision about getting higher education in order to compare his or her life paths: with and without higher education. But the laws of nature and ethics do not allow us to conduct such an experiment. What can we do then?
Inspired by Kung Fu masters, Angrist and Pishke suggest getting acquainted with the ‘furious five’ - a weapon that can eliminate obstacles arising on the way to knowledge. In the first 5 chapters, the scholars wittily, and, most importantly, clearly and in great detail, using both mathematics and real-life examples, describe five types of "weapons":
- Randomized experiment
- Regression
- Instrumental variables
- Discontinuity design
- Differences-in-differences
Each of the tools is an important and relevant method of modern empirical research. In the final chapter, using the new toolkit and knowledge about its strengths and weaknesses and considering a range of pitfalls, the authors show how to answer the very question about the relationship between education and future earnings. Importantly, on the book’s official website, a curious reader can find data and code for building tables and graphs from the book. Therefore, it can also be considered as a fascinating textbook on basic microeconometrics.
The book by mathematician Judea Pearl and science writer Dana Mackenzie shows the importance of the classic mantra of economics and natural sciences: "Correlation is not causation" on a variety of examples - from wet sidewalks to the effects of drugs. Along the way, Pearl and MacKenzie tell the story of economic thought, of the development of statistics and philosophy, explain the consequences of the "reliability revolution" and ask questions about the role of big data and artificial intelligence.
Let's say we want to find out how much smoking one cigarette a day will shorten our life or how effective a vaccine is. To do this, we need to understand the impact of cigarettes on our lungs and of a vaccine on our immunity. A simple comparison of life expectancy of smokers and non-smokers does not prove that it is cigarettes that shorten life. People do not start smoking for no reason, perhaps they had psychological problems that forced them to start smoking and shortened their lives.
This book will allow you to better understand the difference between the ill-fated correlation of data and the actual causality. The authors introduce the reader to many fundamental concepts from statistics, giving examples from real life. And the final chapter is dedicated to big data and artificial intelligence. It gives you a clue to how, instead of looking for data correlation (that is, acting like an owl watching a mouse), the artificial intelligence learns to look for cause-and-effect relationships (that is, tries to understand why this mouse runs exactly in the particular direction that it runs).
Judea Pearl is a great data scientist and winner of the A.M. Turing Award for "for fundamental contributions to artificial intelligence through the development of a calculus for probabilistic and causal reasoning," and Dana MacKenzie is a mathematician who has written several popular science books. The Book of Why combines sufficient evidence verification , with a number of equations for calculating probabilities, as well as clarity for a wide audience. Reading it will allow you to critically perceive the results of published research and independently find cause-and-effect relationships in the data.
