DescriptionProblem Set II

Data Driven Economic Analysis (DSE)

Microeconometrics Module

Prof. Michele De Nadai

Trimester II 2023

1. The following table reports the value of potential outcomes Y0 and Y1 , along with the observed outcome

Y , the treatment value D and a control X (gender) for a population with 10 statistical units:

unit

1

2

3

4

5

6

7

8

9

10

Y0

10

4

10

1

11

20

10

3

2

5

Y1

12

8

10

3

9

21

12

5

15

12

Y

12

4

10

3

9

20

12

5

2

12

D

1

0

0

1

1

0

1

1

0

1

X

male

male

male

male

male

male

female

female

female

female

a) Compute the value of the individual treatment effect of D on the outcome Y for the first unit in this

population.

b) Compute the value of the individual treatment effect of D on the outcome Y for the fifth unit in this

population.

c) Compute the value of the Average Treatment Effect (ATE): E(Y1 − Y0 ) in this population.

d) Compute the value of the Average Treatment Effect on the Treated (ATT): E(Y1 − Y0 |D = 1).

e) Compute the value of the Average Treatment Effect conditional on X: τ (X) = E(Y1 − Y0 |X).

f) Compute the quantity ∆ = E(Y |D = 1) − E(Y |D = 0). Under which condition can you expect ∆ =

ATE?

1

2. Suppose we have availability of a binary instrument Z. Below you find an augmented version of the table

above where you also find the values of the instrument and potential treatments for each unit in the

population.

unit

1

2

3

4

5

6

7

8

9

10

Y0

10

4

10

1

11

20

10

3

2

5

Y1

12

8

10

3

9

21

12

5

15

12

Y

12

4

10

3

9

20

12

5

2

12

D0

0

0

0

1

0

0

1

1

0

1

D1

1

0

0

1

1

1

1

1

0

1

Z

1

0

1

0

1

0

1

0

1

0

D

1

0

0

1

1

0

1

1

0

1

X

male

male

male

male

male

male

female

female

female

female

a) Is the monotonicity assumption satisfied for this population?

b) Compute the Local Average Treatment Effect (LATE): E(Y1 − Y0 |D1 > D0 ) in this population.

c) What is the proportion of compliers in this population?

3. Using data from question 4 in the previous problem set we are interested in estimating the parameters of

the following linear model:

read = α0 + α1 female + α2 cathhs + u,

(1)

where read is the (standardized) score on reading; female

is a binary variable equal to one for females

and zero otherwise and cathhs is a dummy equal to one if the student attended a catholic High School.

a) Is there any reason to believe that the OLS estimator of α2 might be biased?

b) Some researchers are suggesting the use of parcath, a binary variable equal to one if parents are

catholic and zero otherwise, as an instrumental variable for cathhs in equation (1). Under which

conditions is this justified?

c) Describe the OLS regressions you would run to obtain the Two Stage Least Squares (2SLS) estimate

of α2 using parcath as an instrument for cathhs.

d) Suppose you have availability of two separate binary variables:

• mothcath: equal to one if the mother is catholic and zero otherwise

• fathcath: equal to one if the father is catholic and zero otherwise

Consider using both mothcath and fathcath as instruments for cathhs. Could you test for the

validity of the instruments in this scenario? How would you run the test and what are the assumptions

behind such a test?

e) Under which conditions is the 2SLS estimator for α2 using parcath as an instrument for cathhs

estimating the Average Treatment Effect of attending catholic high school on reading scores?

f) Describe in your own words who are the compliers when using parcath as an instrument for cathhs?

Also, how would you describe the no-defiers assumption (monotonicity) in this context?

2

Problem Sets

Data Driven Economic Analysis (DSE)

Microeconometrics Module

Prof. Michele De Nadai

Trimester II 2023

1. Consider a linear regression of y on X:

y = Xβ + u

The OLS estimator for β is obtained as (X′ X)−1 X′ y. It follows that ŷ = X(X′ X)−1 X′ y Obtain the OLS

estimator for a regression of ŷ on X.

2. The orthogonal matrix in a linear regression of y on X has the following expression:

M = I − X(X′ X)−1 X′

Show that M is idempotent: M = MM.

3. Consider a random variable y and denote θ = E(y 3 ).

a) Construct a moment estimator (θ̂) for θ.

b) Show that θ̂ is unbiased.

c) Derive an expression for V ar(θ̂)

d) Propose an estimator for V ar(θ̂)

e) Derive the asymptotic distribution for θ̂

4. You are interested in the relationship between attending catholic schools and students’ standardized scores

on reading. We have information from a cross-section of 7, 430 US students and we consider the following

econometric model:

read = β0 + β1 female + β2 cathhs + β3 lfaminc + β4 cathhs × female + u,

(1)

where read is the (standardized) score on reading on a scale from 0 to 70; female

is a binary variable

equal to one for females and zero otherwise; cathhs is a dummy equal to one if the student attended a

catholic High School; lfaminc is the logarithm of annual family income.

Call:

lm(formula = read12 ~ cathhs + female + lfaminc + I(cathhs *

female), data = catholic)

Residuals:

Min

1Q

-27.697 -6.745

Median

1.192

3Q

7.190

Max

23.129

Coefficients:

1

Estimate Std. Error t value Pr(>|t|)

(Intercept)

19.0234

1.3846

cathhs

1.5247

0.6001

female

1.9117

0.2165

lfaminc

3.0570

0.1328

I(cathhs * female)

0.6010

0.8795

–Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 9.029 on 7425 degrees of freedom

Multiple R-squared: 0.07947,

Adjusted R-squared: 0.07898

F-statistic: 160.3 on 4 and 7425 DF, p-value: < 2.2e-16
a) How can we interpret the coefficient β3 in equation (1)?
b) Is there any evidence of an effect of attending catholic high schools on reading scores?
c) Somebody is suggesting that females are benefiting more from attending catholic high schools than
men. How would you respond to this based on the available estimates?
d) Under which conditions can you interpret these effects of attending catholic schools as causal? Is
this a credible assumption?
e) Derive the expression for the effect of being a female, as opposed to being a male, using the above
specification.
5. With reference to the previously considered dataset we estimate a Probit model to analyze the relationship
between attending a catholic high school and actually graduating from high school (hsgrad). In this sample
93% of students actually graduated. The model we consider is
hsgrad = Φ (α0 + α1 female + α2 cathhs + α3 lfaminc + α4 cathhs × lfaminc) + e
Call:
glm(formula = hsgrad ~ cathhs + female + lfaminc + I(cathhs *
lfaminc), family = binomial(link = "probit"), data = catholic)
Deviance Residuals:
Min
1Q
Median
-3.0434
0.2671
0.3148
3Q
0.4033
Max
1.2110
Coefficients:
Estimate Std. Error z value Pr(>|z|)

(Intercept)

-2.442099

0.293242

cathhs

1.895165

1.632514

female

0.001609

0.051604

lfaminc

0.385023

0.028878

I(cathhs * lfaminc) -0.123952

0.156729

–Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

Null deviance: 3018.6 on 5969 degrees of freedom

Residual deviance: 2808.4 on 5965 degrees of freedom

(1460 observations deleted due to missingness)

AIC: 2818.4

Number of Fisher Scoring iterations: 7

a) How can we interpret the coefficient α2 in the above equation?

b) Provide an estimate for the probability of graduating from high school for a male student attending

a catholic highschool whose parents earn 30000 dollars in annual income.

c) What about an estimate of the same probability for the same student, but with parents earning

10000 dollars in annual income?

2

d) Provide an expression for the marginal effect of attending catholic schools on the probabilty of

graduating from highschool. What is the value of such marginal effect for a male student whose

parents earn 30000 dollars per year?

3

Purchase answer to see full

attachment

**We offer the bestcustom writing paper services. We have done this question before, we can also do it for you.**

#### Why Choose Us

- 100% non-plagiarized Papers
- 24/7 /365 Service Available
- Affordable Prices
- Any Paper, Urgency, and Subject
- Will complete your papers in 6 hours
- On-time Delivery
- Money-back and Privacy guarantees
- Unlimited Amendments upon request
- Satisfaction guarantee

#### How it Works

- Click on the “Place Order” tab at the top menu or “Order Now” icon at the bottom and a new page will appear with an order form to be filled.
- Fill in your paper’s requirements in the "
**PAPER DETAILS**" section. - Fill in your paper’s academic level, deadline, and the required number of pages from the drop-down menus.
- Click “
**CREATE ACCOUNT & SIGN IN**” to enter your registration details and get an account with us for record-keeping and then, click on “PROCEED TO CHECKOUT” at the bottom of the page. - From there, the payment sections will show, follow the guided payment process and your order will be available for our writing team to work on it.