Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
21 Cards in this Set
- Front
- Back
Sampling error
|
Arises by chance, esp. in small samples
|
|
Omitted variable bias
|
Bias is systematically correlated with error
|
|
Latent variables
|
Affect the model but are not directly observable
|
|
Endogenous variable
|
How aquired value provides info that tells us more about dependent variable
|
|
Exogenous variable
|
Process would not provide any additional information
|
|
Two conditions for instrumental variables
|
1) Correlation between z and x
2) z is uncorrelated with u |
|
Problem with 2SLS
|
Produces larger error terms
|
|
Structural equation
|
The original equation, including the endogenous variable(s)
|
|
Reduced form equation
|
Endogenous variable in written in terms of exogenous variables
|
|
First stage of 2SLS
|
Regress endogenous variable(s) on exogenous variables, including z
|
|
Exclusion restrictions
|
Need as many excluded instruments as there are endogenous variables
|
|
Finite sample bias
|
As n approaches infinity, endogeneity approaches 0
|
|
What happens when you do the first stage incorrectly?
|
The influence of the variables that were lefft out goes into error term.
|
|
Why 2SLS errors are bigger than OLS errors
|
1) Only use some info
2) Variation in the error term 3) The more corr there is in the orig. model, the harder it is to detect indiv. influences |
|
Test for endogeneity
|
Hausman Test
|
|
Hausman Test
|
2 estimators:
-consistent & efficient under Ho -consistent under Ho and Ha |
|
Consistenct estimator
|
As n --> infinity, coefficients converge toward true values
|
|
Inconsistent estimator
|
As n --> infinity, coefficients converge on wrong value
|
|
Efficient estimator
|
Results in smaller standard errors than any other estimator of its type (ie, linear)
|
|
Hausman Ho
|
Independent variables are exogenous
|
|
Hausman Ha
|
At least one independent variable is endogenous
|