Econ-T4 eng.pdf - Econometrics Introducing qualitative...

Econometrics
Introducing qualitative factors in RM
Dr. Toni Mora

Including dummy variables
¡
Nominal variables (levels of education). Also
called categorical variables (
¹
number of
categories)
l
Dichotomous or binary variables (two values)
l
Polytomous independent variables (more than two
categories)
Useful for changes in time, specific observations, outliers or
differentiating RM based on groups of individuals

Dichotomous variables
¡
Those variables only having two values, that is in
the form of binary information. Usually, these
values are substituted by 0 & 1 to obtain an easier
interpretation of OLS results
¡
In Catalan also called variables fictícies
¡
We can introduce these dummy variables additively
or by means of interaction terms

Additively
i i i i
D X y e b b b + + + =
2 , 1 1 0
i i
additive
i i i
X X y e b b b e b b b + + + = + + + =
, 1 1 2 0 2 , 1 1 0
) (
!" !# $
Introduced dummy
Example: D
i
=0 if the firm does not
export & D
i
=1 if the firm exports
i i i
X y e b b + + =
, 1 1 0

Relationship with the rest of
the exogenous variables
¡
The use of the dichotomous variables is not
equivalent to disentangle from the sample in two
parts and estimating both models
¡
The rest of coefficients differ based on the
estimated RM (Analysis of covariance-ANCOVA)