Freakonomics, Chapter 4

            In chapter 4, Levitt and Dubner seem to add more evidence to their underlying thesis that the conventional wisdom in society is not always true or accurate.  In this chapter they reveal that the reduction crime in New York city in the 1980’s was not due to economic growth or an improved police strategy, but instead it was the result of a rising abortion rate caused by the famous abortion legalization known as Roe Vs. Wade. 

            The authors succeed in making a convincing and interesting argument with their statistics in what they believe not to be a cause of crime reduction and what they believe to be a cause.  They are especially convincing with their statistic about the correlation between each states abortion rate and its crime rate. 

            My only question for the author is that although the increase in the abortion rate seems to be a reasonable and good argument, are there any other explanations or theories besides the main arguments that the authors refute in the chapter?  This correlation between abortion and crime seems to be more of theory than a causal relationship.  Stating that most of the children who were aborted would have grown up to commit homicides seems questionable.  

           Econometrics and the ability to find a correlation between different variables of interest is a useful tool to tell a story.  But one of the things I have learned in my Quantitative Methods class this semester is that Statistics and the construction of economic models is complex and far from a definite science.  Just because someone finds a correlation in a data set does not necessarily make their story the true one either.  

Elevator Pitch

             For my research paper I have sought to understand what exactly lengthens the duration of unemployment for recently laid off workers.  I believe this is an important question because unemployment in the U.S is an incredibly relevant and important topic in the economy today.  Any incite as to what increases unemployment could have large policy implications.  More specifically, I want to know whether unemployment insurance has a direct and significant impact on how long the recently unemployed stay unemployed. 

            I have created a linear regression model where the duration of unemployment is my dependent variable.  I used the amount of unemployment insurance claims, the amount of job losers as a percentage of the unemployed, and GDP as my independent variables. 

            I have found through my research that unemployment insurance has a strong positive relationship on the duration of unemployment.  Increasing the amount of unemployment insurance greatly increases the duration of unemployment.  This becomes even accurate when I control for the effect on job losers instead of the general unemployed population. Also, by controlling for GDP I am able to control for time trends and recessions that also affect the duration of unemployment.  My results show that incorporating the GDP variable increase the power of my model.