HOW FREAKONOMICS WORKS.

HOW FREAKONOMICS WORKS. There were several coded discussions in our house about who was going to give SUPERFREAKONOMICS to whom for Christmas. I have finally gotten a copy from Amazon and placed it on the coffee table for everybody. I posted here about how Professor John Meyer had predicted in the early sixties that computer models would run into difficulties because “Nature doesn’t run very good experiments”, and I commented that Steven Levitt (of Freakonomics) “is gifted at finding variables that move independently of other variables so that he is able to disentangle cause and effect.” In other words, nature sometimes runs good experiments and Levitt is good at finding them. Noah Scheiber wrote a good exposition of how instrumental variables—when nature has created them—can be used to show causation. He began with a paper by Josh Angrist and Alan Krueger. Does another year of schooling increase income” Or do people with more ability choose to stay in school longer and make more money because of their natural ability? Angrist and Krueger found an experiment that nature had run. Many states required a student to stay in school until they turned 16, but all students started school the year they turned six. Consider students who left school as soon as they turned 16; if they were born in December, they had almost one year more of schooling than those born in January, but they hadn’t made a choice to take the extra year. (The paper showed that the extra year of schooling did lead to greater earnings). Scheiber described the impact of Levitt on graduate students in economics at this time—when everybody was looking for natural experiments. He said that when a student discovered an instrumental variable (an “instrument”), “word of the discovery would rocket through the department. The discoverer would become instantly, if momentarily, famous, like the holder of a winning card at a Bingo hall….”

This entry was posted in Economics. Bookmark the permalink.

Leave a Reply

Your email address will not be published.