econometrics final Flashcards
type I error
false positive
type II eror
false negative
What is Gauss Markov theorem “blue”
best= lowest variance
unbiased= expected value Betahat = true Beta
if you have a heteroskedastic disturbance term, what happens to to Beta estimate if you use robust standard errors?
Betahat will stay the same BUT standard errors will be less efficient.
if you have a homoskedastic disturbance term, what happens to to Beta estimate if you use robust standard errors?
Betahat will stay the same AND standard errors will also remain unchanged.
Three issues with Goldfeldt Quant test
- variance of error term must be fan shaped
- cannot perform on dummy variable
- tedious, only tests for heteroskedasticity one varaible at a time.
what are the two IV assumptions
relevance- Z affects / is correlated with y
exclusion- cov(X,disturbance term)= 0 in other words variable and error cannot be correlated.
what is an ADL model
time series model with p # of Y lags and q # of X lags.
what does low power for a hypothesis test mean?
type II error is more likley
when does low power in hypothesis testing occur?
when variance is high
what is exogeniety?
X is not correlated to the error of disturbance term cov(X, u)=0
what is heteroskedasticity?
variance of error term is different across all individuals.
why is heteroskedasticity important?
can cause standard errors of Betahat to be wrong.
what is attenuation bias?
when the Bhat estimate is combines to a value that is smaller than true Bhat value.
static model
studies contemporaneous data
