0% found this document useful (0 votes)
101 views3 pages

MOOC Econometrics: Dennis Fok

This document discusses the concept of endogeneity in econometrics models. It provides examples to illustrate how endogeneity can arise, such as when an explanatory variable is correlated with the error term due to omitted variables, measurement error, or strategic behavior. The key consequences of endogeneity mentioned are that the OLS estimator is inconsistent and the parameters no longer have a causal interpretation. Upcoming lectures will cover how to test for, recognize, and solve endogeneity issues to properly estimate economic models.
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
0% found this document useful (0 votes)
101 views3 pages

MOOC Econometrics: Dennis Fok

This document discusses the concept of endogeneity in econometrics models. It provides examples to illustrate how endogeneity can arise, such as when an explanatory variable is correlated with the error term due to omitted variables, measurement error, or strategic behavior. The key consequences of endogeneity mentioned are that the OLS estimator is inconsistent and the parameters no longer have a causal interpretation. Upcoming lectures will cover how to test for, recognize, and solve endogeneity issues to properly estimate economic models.
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/ 3

Motivating example

MOOC Econometrics We want to explain


Lecture 4.1 on Endogeneity: Motivation Number of flights at an airport per month (y ) using
Number of travel insurances made in previous month (x)
Dennis Fok
Suppose OLS yields
y = 10, 000 + .25x + e

Test
How should we interpret the obtained coefficients?
What does the estimate .25 really mean?

Lecture 4.1, Slide 2 of 12, Erasmus School of Economics

Interpretation of parameters Endogeneity

Given the estimates (y : flights, x: insurances)


OLS requires some assumptions:
y = 10, 000 + .25x + e
explanatory variables should be exogenous
Correct: 4, 000 insurances sold → expected number of flights violation of this: endogeneity.
= 10, 000 + .25 × 4, 000= 11, 000
High x tends to go together with high y . In this set of lectures, you will learn to:
The identified correlation yields adequate predictions. 1 Understand/recognize endogeneity.
2 Know the consequences of endogeneity.
Incorrect: Selling 4, 000 additional insurances causes 3 Estimate parameters under endogeneity.
.25 × 4, 000 = 1, 000 additional flights 4 Know the intuition of the new estimator.
The regression does not identify a causal impact! 5 Argue/test assumptions underlying this new estimator.
A third variable (travel demand) affects y (flights) and x
(insurances).

Lecture 4.1, Slide 3 of 12, Erasmus School of Economics Lecture 4.1, Slide 4 of 12, Erasmus School of Economics
Stochastic vs. non-stochastic regressors Economic models

In economics:
Controlled (or natural) experiments are rare
Standard assumptions for linear model (y = X β + ε) include New data with same X cannot be obtained
A2 Explanatory variables are non-stochastic Explanatory variables are stochastic!

Implications: If X stochastic:
Obtain new data: X stays constant (and y changes) new data set → new X values
Need “controlled experiment” X can be correlated with other variables
If X correlated with ε
OLS estimator b converges to true coefficient β for n → ∞
I X is endogenous
(OLS is consistent) I There is another variable that affects y and X
I OLS does not properly estimate β (inconsistent)
If X uncorrelated with ε
I X is exogenous
I OLS consistent
Lecture 4.1, Slide 5 of 12, Erasmus School of Economics Lecture 4.1, Slide 6 of 12, Erasmus School of Economics

Other examples of endogeneity – Omitted variables Omitted variable – Example

True model is
Model student’s grade using attendance at lectures.
y = X1 β1 + X2 β2 + η
but we ignore X2 and perform OLS on
Test
y = X1 β1 + ε Which omitted factor would lead to endogeneity of attendance?

We have: ε = X2 β2 + η
Three possible omitted factors:
X1 correlated with ε (X1 is endogenous) if
I X1 correlated with X2 and
1 Difficulty of exam
I β2 6= 0 NO: not correlated with attendance.
Derivation: 2 Motivation of the students?
YES: correlates with attendance and affects grade.
Cov(X1 , ε) = Cov(X1 , X2 β2 + η) 3 Compulsory attendance yes/no?
= Cov(X1 , X2 )β2 + Cov(X1 , η) NO: does not directly impact the grade
| {z }
=0

Lecture 4.1, Slide 7 of 12, Erasmus School of Economics Lecture 4.1, Slide 8 of 12, Erasmus School of Economics
Other examples – Strategic behavior Other examples – Measurement errors

Consider a model explaining demand using price.


y (eg. salary) depends on x ∗ (eg. intelligence)
Strategic price setting: x ∗ (intelligence) difficult to observe
1 Sets high price when high demand is expected x = x ∗ + measurement error: noisy measurement (eg. IQ score)
2 Price and sales positively correlated measurement error: x is endogenous in y = α + βx + ε
3 Price will be endogenous in regression of demand on price.

Lecture 4.1, Slide 9 of 12, Erasmus School of Economics Lecture 4.1, Slide 10 of 12, Erasmus School of Economics

Summary & what’s next? TRAINING EXERCISE 4.1

Endogeneity is a common problem


Train yourself by making the training exercise (see the website).
OLS is not useful under endogeneity

After making this exercise, check your answers by studying the


Upcoming topics:
How to solve for endogeneity? webcast solution (also available on the website).
How to test for endogeneity?

Lecture 4.1, Slide 11 of 12, Erasmus School of Economics Lecture 4.1, Slide 12 of 12, Erasmus School of Economics

You might also like