econ 140 11 - ECONOMICS 140 Professor Enrico Moretti...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
ECONOMICS 140 Professor Enrico Moretti 4/19/2010 Lecture 11 ASUC Lecture Notes Online is the only authorized note-taking service at UC Berkeley. Do not share, copy or illegally distribute (electronically or otherwise) these notes. Our student-run program depends on your individual subscription for its continued existence. These notes are copyrighted by the University of California and are for your personal use only. D O N O T C O P Y Sharing or copying these notes is illegal and could end note taking for this course. ANNOUNCEMENTS Today is the last official lecture where I will introduce new material. We are going to finish chapter 18. I basically just have one more example left from last week. Then we will go on to chapter 14 and 16. You should skip 14.2.6, 14.3.5, and 16.4. There is no new problem set for this week; instead there are practice problems that I will post either tonight or tomorrow. The week of April 26 th , we will have a review session in class where the GSIs will solve the practice problems. We will still have sections that week. We will have sections and office hours the week of May 3 rd . The final is on May 14 th . LECTURE Cross-Sectional Data Last week we were talking about ways to deal with the omitted variable problem. There are many ways to deal with one, one of them being panel data. You can use the longitudinal nature of panel data to deal with the omitted variable bias. You can also used pooled cross-sectional data to deal with omitted variable bias. Cross-section is when you observe a set of individuals over time. Suppose that we want to estimate the effect of gun control laws on crime. Crime st = α + βlaw st + Ε st We expect β≤ 0 if gun control laws have an effect on controlling crime and β=0 if gun control laws have no effect. This equation has a potential for reverse causality or omitted variable bias because states that pass these laws might be states that already have high levels of crime. Therefore, when you run the regression, beta might actually be greater than 0. Crime st = α + βlaw st + Ε st + s d s We can avoid this by including 49 state dummies to absorb all the differences between states. Because we have 50 states, we will have 49 dummy variables. Without the dummy term, the beta makes a broad comparison across different states. In the resulting model, beta doesn’t compare differences between one state and another but rather compares differences between one state at one period of time and that same state at another period of time. Note that the dummy variable only picks up permanent differences between states and not time varying differences. It might be a good idea to control explicitly for the time varying variables Another set of dummies that we want to include is year dummies. Say we are looking at a period of 10 years. We will have nine year dummies. Crime
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 08/03/2010 for the course ECONOMETRI 05 taught by Professor Wood during the Spring '10 term at University of California, Berkeley.

Page1 / 6

econ 140 11 - ECONOMICS 140 Professor Enrico Moretti...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online