Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
ECON-UA 323: ECONOMIC DEVELOPMENT
Assignment
Please submit a PDF of this assignment at the start of class (12:30pm EST) on Brightspace. There
aren’t any graphs or unusual symbols on this problem set, so if you find yourself disliking overleaf
this one is totally doable in Google Docs or Word. Please make sure to number your answers.
Relative to other problem sets, this one is the most straightforward – you shouldn’t spend more
than three hours on it. I would like you to read two papers
Muralidharan, Karthik and Nishith Prakash (2017) “Cycling to School: Increasing Secondary
School Enrollment for Girls in India” American Economic Journal: Applied Economics 9(3):
321-350
Barteska, Philipp, and Jay Euijung Lee (2023). Bureaucrats and the Korean Export Miracle. Job
Market Paper
Parts of the papers are somewhat technical and you can skip those – for this problem set, I’m more
interested in the empirical analysis. In the next pages, I’ve included a template1 which will help
you think about how to read empirical analyses, and which you should fill out.
1 I’ve adapted the template from one designed by Dan Levy and Sue Dynarski.
Muralidharan, Karthik and Nishith Prakash (2017) “Cycling to School: Increasing Secondary
School Enrollment for Girls in India” American Economic Journal: Applied Economics 9(3):
321-350
After reading the paper, please watching the following (7ish minute) video:
https://www.youtube.com/watch?v=6nG63ISt_Ek
Population and Treatment
1) What is the main goal of the paper? (e.g., assess the effect of class size reduction on
students’ performance, assess the effect of a job training program on earnings, etc.)
2) What outcome measures are used?
3) What population is being studied?
Methodology
4) What method(s) is(are) being used for the evaluation (Randomized Controlled Trial, simple
OLS, Differences-in-Differences, Fixed Effects, etc.)? How is the control group formed?
5) What are the key threats to the internal validity of the study? That is, why would the
numbers they generate be biased estimates of the true casual effects of the population they are
studying? Explain why the difference-in-difference-in-difference strategy solves some of the
issues with the difference-in-differences (they are somewhat vague about this in the video, your
answer should be more detailed than what they say)
6) They discuss one potential source of difference-in-differences bias at timestamp 3:16 of
the video. Explain how this would bias the results direction of bias – do you think the difference
in differences would be “too big” or “too small” relative to the true causal effect? (Obviously their
strategy has an effect on the estimate, but they do not explain the intuition for what’s going on.)
Results
7) For the main outcome, what is the range of the effects estimated by the study? Their
statistical significance? Is there a preferred estimate designated by the author(s)?
8) How “big” are these effects, from a policy or practical perspective?
9) Is there a secondary outcome that you think is important? If so, what is the range of the
effects estimated by the study? Their statistical significance? Is there a preferred estimate
designated by the author(s)?
10) How “big” are these effects, from a policy or practical perspective?
11) If you had been writing the study, what other outcomes would you have wanted to look at?
Why?
12) In the paper, they discuss the role of distance and how it matters for the effectiveness of
the program. There was some confusion about this distinction on the exam: this is not bias. If you
wanted to know the average effect of the program, the fact that it works differently for different
types of people is not something you need to account for to measure causal effects. Explain why
this is, and why understanding heterogeneity is important for thinking about expanding the
program to different locations
Barteska, Philipp, and Jay Euijung Lee (2023). Bureaucrats and the Korean Export Miracle. Job
Market Paper
Population and Treatment
1) What is the main goal of the paper? (e.g., assess the effect of class size reduction on
students’ performance, assess the effect of a job training program on earnings, etc.)
2) What outcome measures are used?
3) What population is being studied?
Methodology
4) What method(s) is(are) being used for the evaluation (Randomized Controlled Trial, simple
OLS, Differences-in-Differences, Fixed Effects, etc.)? What is the kind of bias this method avoids?
That is to say, if you just used observational data of some kind, why might not you believe that the
relationship between x and y is causal?
5) What are the key threats to the internal validity of approach? That is, why would the
numbers they generate be biased estimates of the true casual effects of the population they are
studying? More practically for this paper, what is a reason why might migration not be random
in the way they hope?
6) How would these threats affect the estimate (direction of bias)?
Results
7) For the main outcome, what is the range of the effects estimated by the study? Their
statistical significance? Is there a preferred estimate designated by the author(s)?
8) How “big” are these effects, from a policy or practical perspective?
9) Is there a secondary outcome that you think is important? If so, what is the range of the
effects estimated by the study? Their statistical significance? Is there a preferred estimate
designated by the author(s)?
10) How “big” are these effects, from a policy or practical perspective?
11) If you had been writing the study, what other outcomes would you have wanted to look at?
Why?
12) I have seen Barteska present the paper, and he faced a lot of skepticism (which is normal! The
world is always trying to trick us). What do you think is perhaps not super compelling about his paper?