Endogeneity and the IV estimator Flashcards

1
Q

Sources of Endogeneity(4)

A
  • unobservable omitted variables
  • reverse causation
  • simultaneous equations
  • measurement error bias
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Fundamental assumption of least-squares estimators(1)

A

The model error is unrelated to the regressors

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is the problem with endogeneity?

A

All estimates of βj become inconsistent

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is the endogenous model?

A

y=β_0+x_1 β_1+x_2 β_2+…+x_K β_K+u_i

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Property of instrument z

A

Changes in z affect x but not y (although indirectly through x)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What are the conditions to observe a variable, zt, to generate an exogenous form of xt?

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

When is the Instrumental Variables estimator consistent?

A

When z is correlated with x and uncorrelated with the error term

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Are IV estimators more efficient and consistent than OLS?

A

They are usually more consistent, but less efficient - especially if instruments are weakly correlated with the variable being instrumented

This will be reflected in increasing standard errors

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What is the order condition for IV estimator?

A

r >= K

where:

r is a number of instruments

K is the number of independent endogenous components

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Properties of the Pooled Model

A

y_it=β_0+x′_it β+u_it “ i”=”1,…,N, t”=”1,…,T.”

Coefficients are Constant

Pooled OLS is consistent when cov[u,x]=0

Panel-corrected standard errors and t-statistics must be used

Pooled OLS is inconsistent if the FE model is appropriate

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Fixed Effects Methods

A

Measures association between individual-specific deviations of regressors from their time-averaged values and individual-specific deviations of the dependent variable from its time-averaged value

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is the variance of the IV estimator equal to? (in terms of OLS estimator variance)

A

It is equal to the variance of the OLS estimator over the sample correlation coefficient between x and z

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What does the 2 stage least squares (2SLS) estimator do?

A

Takes all of the possible linear instruments for the variable xK and chooses the instruments that is most highly correlated with xK given by the linear projection:

How well did you know this?
1
Not at all
2
3
4
5
Perfectly