Section 8 zoom - midterm review PDF

Preview

Summary

midterm review...

Description

Page 1 of 11

C171/151 Economic Development Practice Midterm Exam Spring 2021 (1 hour and 20 minutes, 25 percent of final grade) ---BCourses Quiz---

This exam is open-book. Do not communicate with anyone during the exam. Answer all questions (total of 100 points in this exam) Question 1 Define in one or two sentences any 5 of the following 6 concepts. If you use an equation or a graph, make sure that you explain what it shows. Do not answer more than 5 (4 points each, total of 20 points).

A) Total factor productivity Lecture 1-B Measurement, Development Accounting and Impact Evaluation Total Factor Productivity (TFP) is the portion of output not explained by the amount of inputs used in production.

B) Demand-side determinants of education outcomes Lecture 5a Education & Development

Page 2 of 11

C) Omitted variable bias Lecture 2-B Review of Regression Analysis II

Page 3 of 11

D) Median voter theorem Lecture 4-A Political Economy & Development and Randomized Control Trials (RCTs)

E) Purchasing power parity adjustment Lecture 1-B Measurement, Development Accounting and Impact Evaluation Purchasing power parity (PPP) is a popular metric used by economics that compares different countries' currencies through a "basket of goods" approach, adjusting for differences in currencies and prices. It allows for economists to compare economic productivity and standards of living between countries.

Page 4 of 11

F) T-statistic T-statistic is the ratio of the departure of the estimated value of a parameter from its hypothesized value to its standard error.

Question 2 True or false, and briefly explain why. Select any 5 of the following 6 statements and do not answer more than 5 (6 points each, for a total of 30 points). Zero points for true/false labels without an explanation. A) In “The Colonial Origins of Development” Acemoglu et al (2001) find that the causal effect of institutions on economic development (estimated using their IV approach) is much weaker than the correlation between these two variables (estimated using OLS regression). False.

Page 5 of 11

IV:

OLS

Page 6 of 11

B) In “Schooling and Labor Market Consequences”, Duflo (2001) estimates the returns to education in Indonesia by comparing the wages of men in 1995 who grew up in districts with more schools relative to districts with less schools. False.

1st Difference:1995 outcomes of individuals born in districts with more vs less school construction in 1973-1979. 2nd Difference No 2:1995 outcomes of 2-6 year old in 1974 vs 12-17 year old in 1974.

Because children aged 13 and older in 1974 did not benefit from the program

Page 7 of 11

C) In “Worms: Identifying Impacts on Education and Health in the Presence of Treatment Externalities” Kremer and Miguel (2004) estimate the “spillover” effect of de-worming class mates by randomizing de-worming treatments across students within each school. False. (phase-in across schools)

The randomized phase-in across schools of the deworming intervention allows them to capture the overall effect of deworming even in the presence of externalities across individuals within schools. School-level randomization also naturally generates local variation in the density of treatment that they use to estimate spillovers across schools.

D) In “Teacher Performance Pay” Muralidharan and Sundararaman (2011) find that most of the effect on student test scores was driven by the fact that teachers in the incentive group showed up for work more frequently than teachers in the control group. False. Individual incentives appear more powerful than school incentives (effectsizes of 0.28 vs 0.15 standard deviations). Teacher behavior is affected in terms of better teaching effort, rather than attendance.

Page 8 of 11

E) In “Strengthening State Capabilities” Dal Bo et al (2013) randomize the wages posted across recruitment sites. They then let the government decide whom to make which job offer to. This allows them to analyze which positions are more or less attractive. False. Two stages of randomization (like 2 separate RCTs). The first stage: Randomly divide 106 sites into high wage and low wage posted. The second stage: Randomly assign job offers to pool of eligible candidates. (not let the government decide whom to make which job offer to).

Page 9 of 11

F) In “Strengthening State Capabilities” Dal Bo et al (2013) find that the effect of higher wages on the number of applicants is positive and statistically significant. False. (is positive and but statistically insignificant) Higher wages attract smarter and more motivated applicants. Higher wages led to larger pool of applicants and higher acceptance rates.

Page 10 of 11

Question 3 Answer all questions (50 points in total). Suppose you were interested in analyzing the effects on test scores of a scholarship program that was randomly assigned in equal proportions to a group of 5th and 6th graders. To test these effects, you collected data on test scores and created the following variables: treatment – 1 if in a treatment school and 0 if not raven – this is the score on a Raven’s test (a test used to measure intellectual aptitude). grade5 - 1 if the pupil began in grade 5 and 0 if the pupil began in grade 6 treat_grade5 - the interaction between treatment and grade5 Now suppose we ran the following command in Stata, which yielded the output below:

A) What does the constant term measure? (10 points)

Page 11 of 11

B) Based on these estimates, what is the effect of the intervention for 6th graders? What is the level of statistical significance of this estimate? What is the average test score among 6th graders in the treatment group? (15 points)

C) What is the treatment effect on 5th graders? Is it statistically significantly different from that of 6th graders? At what level? (10 points)

D) If the compliance for both 5th and 6th graders had been 100% in the control group and 50% in the treatment group, how large would the average treatment effect on the treated (ATT) be for each group? (15 points)...