Scribd
Upload
161 views4 pages

Econometrics Mock Exam

Mock_Exam_2 (2)

Uploaded by

BENSTOKE
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
161 views4 pages

Econometrics Mock Exam

Mock_Exam_2 (2)

Uploaded by

BENSTOKE
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Winter Term Exam

EC2C3
Econometrics 1

Suitable for all candidates

Instructions to candidates

This paper contains one question with 10 subquestions. Answer all subquestions. Each is worth 10 points.

Time Allowed Reading Time: 15 minutes


Writing Time: 2 hours

You are supplied with: No additional material

You may also use: No additional material

Calculators: Calculators are allowed in this exam

If at any point in this exam you feel that anything is unclear, please make additional assumptions that
you feel are necessary and state them clearly.

Mathematical derivations are not required.

Write concise answers. Precise arguments will be rewarded; however, accuracy is more important than
extensive details. Including irrelevant information in your answers will not improve your marks and will reduce
the time you have to write relevant answers. Marks will be deducted for incorrect irrelevant information.

© LSE January Exam EC2C3 Page 1 of 4


A large city is considering building a subway to improve transportation and has hired economists to evaluate the potential
benefits of a subway. The primary question is if building the subway will reduce road congestion. We gather data from
a set of 𝑁 cities and define a binary variable, 𝐷𝑖 , which takes the value 1 if the city has a subway. We also use data from
Google Maps to track the average speed of travel on roads in the city. This is our outcome, denoted 𝑦𝑖 , measured in
𝑘𝑚/ℎ𝑟. We have data on other city variables, 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑖 , 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑖𝑛𝑐𝑜𝑚𝑒𝑖 , 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑎𝑔𝑒𝑖 , and % 𝑐𝑜𝑙𝑙𝑒𝑔𝑒 𝑑𝑒𝑔𝑟𝑒𝑒.

1. Suppose we want to believe that subways were randomly assigned to the cities we have in our data. How could
we check if random assignment was successful? Why is this important?

Rather than using the binary variable 𝐷𝑖 , we gather data on 𝑙𝑒𝑛𝑔𝑡ℎ𝑖 , representing the length (in km) of subway in the
city as measured by satellite using Google Maps (𝑙𝑒𝑛𝑔𝑡ℎ𝑖 = 0 if there is no subway). We specify a regression equation,
(1) 𝑦𝑖 = 𝛽0 + 𝛽1 𝑙𝑒𝑛𝑔𝑡ℎ𝑖 + 𝛽2 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑖 + 𝛽3 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑖𝑛𝑐𝑜𝑚𝑒𝑖 + 𝛽4 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑎𝑔𝑒𝑖 + 𝛽5 % 𝑐𝑜𝑙𝑙𝑒𝑔𝑒 𝑑𝑒𝑔𝑟𝑒𝑒𝑖 + 𝑢𝑖
Results of estimating this equation are below.

2.
a. Interpret the estimate for 𝛽1 .
b. There are 3 Stata outputs below. Explain which one tests the null hypothesis, 𝐻0 : 𝛽3 = 𝛽4 . Evaluate the
results.

© LSE January Exam EC2C3 Page 2 of 4


3. Consider the previous regression. What assumption do we need for the estimate of 𝛽1 to represent the causal
effect of interest? Describe three distinct sources of potential bias. Explain how the source would cause bias and
describe the direction of the bias in context of your explanation.

4. We have a variable in our data, 𝑐𝑜𝑚𝑝𝑎𝑐𝑡𝑖 , which represents that the city is geographically small (has high
population density). It is a binary variable. Consider that the effect of subway length on congestion may differ
in a compact compared to a non-compact city. Specify a regression equation to estimate such an effect. Explain
the interpretation of coefficients of interest.

City councils decide the length of a subway built and may choose this taking into account road congestion. Suppose city
councils choose 𝑙𝑒𝑛𝑔𝑡ℎ𝑖 by,
(2) 𝑙𝑒𝑛𝑔𝑡ℎ𝑖 = 𝛿0 + 𝛿1 𝑦𝑖 + 𝛿2 𝑇𝑎𝑥 𝑅𝑒𝑣𝑒𝑛𝑢𝑒𝑖 + 𝛿3 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑖 + 𝑣𝑖
5. Discuss the properties of OLS estimation of (1) and (2) and any concern that is present.

6. Suggest an estimator of equation (1) that will overcome the concern you raised in question 5 and describe how
it would be implemented.

7. What assumptions are necessary for your estimator from question 6? Describe them in context and evaluate their
validity.

© LSE January Exam EC2C3 Page 3 of 4


Regardless of your suggested estimator in question 6, let us consider an alternative setting. Suppose we observe two
cities, 𝑖 = 0 and 𝑖 = 1 in two time periods, 𝑡 = 0 and 𝑡 = 1. Suppose neither city has a subway at time 0 and that city
𝑖 = 1 has a subway built and operating at time 1. Suppose you have a dataset with an observation for each road in each
city. Data have variables: 𝑖, taking the value 0 or 1 for the city, 𝑡, taking the value 0 or 1 for the time period, and 𝑦,
denoting the average speed of travel on that road in that time period.

8. What is the most basic manner in which the Difference-in-Differences estimator is implemented? Describe in
context the assumption it requires. Does the assumption seem plausible? Why might the assumption fail?

9. Stata Question: What Stata code would you run to calculate the Difference-in-Differences estimate?

10. Suppose instead of using 𝑦𝑖 as the outcome we used ln(𝑦𝑖 ).


a. How would the interpretation of our estimated effect change?

b. Regardless of your discussion in question 8, suppose the assumption you described in question 8 is true.
Re-evaluate the assumption in the context of using ln(𝑦𝑖 ) as the outcome.

© LSE January Exam EC2C3 Page 4 of 4

You might also like