You need to estimate the version of the gravity model mentioned in
Krugman textbook in order to give it
more flexibility:
c
ij
b
j
a
i
ij
D
Y
Y
K
T
- Where Y
i
and Y
j
are the GDP of your country and
jth trade partner, respectively. Tij and Dij
represent the trade
and distance. You have already collected a
ll this information. On the hand a, b,
and c represent
Elasticities
of trade with respect
to the related variables. Coefficients b
and c is the
ones that you need to estimate. B
y the way, K is a fixed numb
er.
- In order to make the relati
on ready for estimation, let’s talk
e natural logarithm of both sides of the
gravity equation:
c
j
b
j
a
j
D
Y
Y
K
T
)
D
(
Ln
c
)
Y
(
Ln
b
)
Y
(
Ln
a
)
K
(
Ln
)
T
(
Ln
j
j
j
The term
)
Y
(
Ln
a
)
K
(
Ln
is a fixed number, so let’s ca
ll it A. Also let’s B = b, C = c
to be
consistent with estimation part.
)
D
(
Ln
C
)
Y
(
Ln
B
A
)
T
(
Ln
j
j
j
Where T
j
is the trade (sum of import
and export) between your country a
nd partner j, D
j
is the
distance between your country and partner j, Y
j
is the nominal GDP of the partner country j
- Organize your data in three columns of Excel spreadsheets. In
new three columns of Excel
spreadsheet calculate the natural logarithm of the three variables.
(5 points)
- You should use the new three columns of Excel spreadsheet in
LINEST function to estimate A, B,
and C coefficients, a
nd other statistics.
- If you have a problem working with
LINEST function, search YOUT
UBE, for tutorials. You can
learn it within a few minutes, us
ing the tutorials and my examp
le file.
- To see one example of using LINES
T function for estimations,
check the posted example on the
portal.
- As it was mentioned before, B (
or b) and C (or c) are elasticities, so do a little research in order to see
how you should interpret the result of your regression (elasticities) in your writing.
- All used data, estimated coefficients, and statistics should b
e fully presented. The
computer output of
the estimation should be include
d. Discuss the size and sign o
f coefficients, and the statistics of the
model (like R
2
).