# UCB Economics Regression Analysis Question

Question Description

I’m working on a economics multi-part question and need a sample draft to help me learn.

1. What were the most important new things you learned from the lecture that you did not know before class?
2. What were points discussed in the lecture that you are still confused/unclear about and would like some further clarification on?
the content covered in the lecture?

1 attachmentsSlide 1 of 1attachment_1attachment_1.slider-slide > img { width: 100%; display: block; }
.slider-slide > img:focus { margin: auto; }

Unformatted Attachment Preview

DATA
PANEL
Panel data
key
la I
concept
Sw
Pawel
data consist of observations
units
at two
Data
or
more
Xit Yit
it
and
in
i
refers to unit
subscript
t
refers to time
435
same n
time periods
subscript
Ex
for the
Y for unit i 3
Balanced panel
all
Unbalanced panel
missing data
t l
it
where
in period 7 5
variables are observed
for all i and t
for at least
one
i ort
Motivating example
Goal
estimate effect of alcohol taxes
Data
state level panel data
Xit
Regression in
beer
4
2 01
traffic deaths
Yit traffic fatality rate
tax
82
on
of’s
0.13
x
Y
Regression in 88
Q
Does
an
increase
1.86
0.44 X
O 13
more
traffic deaths
A
No because of OVB Ex of omitted vars
rural us urban driving social attitude towards
drinking and driving
OUB using panel data
Eliminating
Suppose
Zi
Ex
that
Zi
is
does not change
a
factor determining
over
time
fatality rates
Zit Ziti Zi
attitude towards drinking and driving
changes slowly
approx constant btw
82 and 88
model
Regression
Yit
pot B XittpaZituit
February 18
Because 2
does not change
be eliminated
by
looking at
over
time its
changes over
effect
time
can
I
Yise
t
pot f Xi82 p Zi tuisz
Po t p Xi88 PzZituigg
Yi88
i
Subtracting
Yigg
D
g
from
2
the effect of
Xi82
Ui88 Ui82
Specifying the regression in changes
OVB from variables that
The
regression
Yg Yaz
line
2
eliminates
P Xi88
Yi82
l
over
Zi
time eliminates
constant over time
are
becomes
0.072
1.04
Xs
1182
0.36
Including
the
an
mean
intercept allows
of fatality rate
February
for
possibility
changes over
time
23
FE assumptions
where
Yit p Xittai
hit
ist
n
that
t t it
E uit
1
2
Xit ail
Xiii
Xiii y Xit
uit
O
iid
are
suit
3 Large outliers are
unlikely
4 No perfect multicollinearity
For multiple
Xk
Xi it
regressors
Xit
should
be replaced by
it
Asst
concerns
Ass 2 requires
past
variables to
If
not within units
correlated
Xit
present
future
be iid
Xit and
is said to
across
Xis for stt
units but
be autocorrelated
are
or
serially correlated
Inference
Under
and
Ass 1 4
the FE
estimators
asymptotically normal
are
consistent
If
regression
are
errors
heteroscedasticity robust
use
are
not valid
clustered SE
Here
unit
both
heteroscedasticity
units
I recommend
too
SE
the standard
autocorrelated
small
cluster
Clustered SE
are
robust to
and autocorrelation within
clustered
SE
if
n
is not

attachment