Вы находитесь на странице: 1из 43

Executive MBA 2007-

2008
emba bridge- 2006/2007 1
Basic Econometrics
Basic Econometrics
Christopher
Christopher
Grigoriou
Grigoriou
Executive Executive MBA MBA- -HEC Lausanne HEC Lausanne
2007/2008 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 2
Overview
Overview
Objectives of the day
Objectives of the day

Interpreting Econometric Applications


Interpreting Econometric Applications

Understanding how it works


Understanding how it works
Program of the day
Program of the day

Introduction (why? how? basic concepts)


Introduction (why? how? basic concepts)

Interpretations
Interpretations

Case
Case
-
-
study
study
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 3
1.
1.
Introduction to Econometrics
Introduction to Econometrics
=> What for ?
=> What for ?

Impact Analysis
Impact Analysis

To evaluate the success/failure of a


To evaluate the success/failure of a
project, reform, law,
project, reform, law,

To test any economic theory


To test any economic theory
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 4
1.
1.
Introduction to Econometrics
Introduction to Econometrics
=> How ?
=> How ?

Apply statistical methods to economic


Apply statistical methods to economic
data
data

Econometric approach:
Econometric approach:
-
-
Develop working model from an
Develop working model from an
economic theory
economic theory
-
-
Estimate model with real world data.
Estimate model with real world data.
Real world data is not perfect
Real world data is not perfect
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 5
Some examples
Some examples

Keynesian Consumption Function


Keynesian Consumption Function

Price and Quantity


Price and Quantity

Philips Curve
Philips Curve

Production Function
Production Function
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 6
Keynesian Consumption Function
Keynesian Consumption Function

Theory: people increase consumption as


Theory: people increase consumption as
income increases, but not by as much as the
income increases, but not by as much as the
increase in their income.
increase in their income.

Marginal Propensity to Consume (MPC) is the
Marginal Propensity to Consume (MPC) is the
change in consumption divided by change in
change in consumption divided by change in
income.
income.
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 7

C =
C =

+
+
.
.
I
I

C = Consumption
C = Consumption

= Intercept
= Intercept

I = Income
I = Income

= slope (how much C changes for a given


= slope (how much C changes for a given
change in I)
change in I)

Not an econometric model
Not an econometric model

Assumes a deterministic relationship
Assumes a deterministic relationship
Keynesian Consumption Function
Keynesian Consumption Function
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 8

C =
C =
+
+

I
I
+
+



= error term
= error term

Error term captures several factors:
Error term captures several factors:
omitted variables omitted variables
measurement error in the dependent variable measurement error in the dependent variable
randomness of human behavior randomness of human behavior
Keynesian Consumption Function
Keynesian Consumption Function
=> Econometric Model
=> Econometric Model

Expected Results: Expected Results:

> 0 and 0 < > 0 and 0 <

< 1 < 1

represents the MPC represents the MPC


Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 9
Ordinary Least Squares
Ordinary Least Squares

Estimate from the least squares Estimate from the least squares
the line of best fit minimizes the sum of the the line of best fit minimizes the sum of the
squared deviations of the points on the graph squared deviations of the points on the graph
from the points on the straight line. from the points on the straight line.
Minimize Minimize ( (CA CA
i i
- - CP CP
i i
) )
2 2
CA CA
i i
= Actual Consumption for = Actual Consumption for obs obs i i
CP CP
i i
= Predicted Consumption for = Predicted Consumption for obs obs i i

How to estimate the model? How to estimate the model?


=> Fit a line through the data. => Fit a line through the data.
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 10
Ordinary Least Squares optimization process
Ordinary Least Squares optimization process
=>search for a
o
and a
1
that minimize the sumof the
squared residual (=the global error of the model)
y
i
= a
o
+ a
1
.x
i
+ u
i
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 11
Suppose we get C = 1000 + 0.8I
= 1000
= 0.8
Sample income levels
I = 0, Consumption = 1000
I = 1000, Consumption = 1800
If I increases by 1 dollar, then C increases on average by 0.8 dollars.
Interpretation (1)
Interpretation (1)
These estimates are consistent with theory since >0 and 0<<1
Suppose was 0.9?
< 1, but is it due to the sample? => tests + Confidence Interval
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 12
Interpretation (2)
Interpretation (2)
1 1- - Basic specification : Basic specification : Y Y
i i
= = + + . .X X
i i
+ + . .Z Z
i i
+ +
i i
= = marginal impact: marginal impact:
=> => an an increase increase of of 1 1 unity unity in X in X implies implies ceteris ceteris paribus paribus an an
increase increase of of unities unities in Y in Y
2 2- - Log Log- -log specification : log specification : ln lnY Y
i i
= = + + . .lnX lnX
i i
+ + . .lnZ lnZ
i i
+ +
i i
= = elasticity elasticity: :
=> => an an increase increase of of 1 1 per per cent cent in X in X implies implies ceteris ceteris paribus paribus an an
increase increase of of percent percent in Y in Y
3 3- - semi semi- -log : log : ln lnY Y
i i
= = + + . .X X
i i
+ + . .Z Z
i i
+ +
i i
= = semi semi- -elasticity elasticity: :
=> => an an increase increase of of 1 1 unity unity in X in X implies implies ceteris ceteris paribus paribus an an
increase increase of of per per cent cent in Y in Y
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 13
Other examples (1)
Other examples (1)
Price and Quantity Price and Quantity
Demand and Demand and elasticities elasticities of demand of demand
ln ln Q = Q = + + lnP lnP + +
Phillips Curve Phillips Curve
Relationship between change in money wages and Relationship between change in money wages and
unemployment unemployment
w w = f ( = f ( u) u)
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 14
Production Function Production Function
Relationship between inputs and outputs. Relationship between inputs and outputs.
Y = f (K,L) Y = f (K,L)
Cobb Douglas Y = Cobb Douglas Y = AK AK
a a
L L

Wage equation Wage equation
ln ln W = W =
0 0
+ +
1 1
EDUC + EDUC +
2 2
EXP + EXP +
3 3
GENDER + GENDER +

4 4
RACE + RACE + + +
i i
Other examples (2)
Other examples (2)
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 15
General Terminology (1)
General Terminology (1)
Pooled data: mixture of cross Pooled data: mixture of cross- -sectional and time series data sectional and time series data
Panel data: follow a microeconomic unit over time Panel data: follow a microeconomic unit over time
Quantitative data: continuous data Quantitative data: continuous data
Qualitative data: categorical data Qualitative data: categorical data
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 16
General Terminology (2)
General Terminology (2)
Y
Y
i it t
=
=


+
+


.
.
X
X
i it t
+
+


i it t
Y: dependent variable Y: dependent variable
X: independent or explanatory X: independent or explanatory
: : Error Error- -term term
subscript i: refers to subscript i: refers to ith ith observation observation
t: for time series data at time t t: for time series data at time t
Cross Cross- -sectional data: collected at 1 point in time sectional data: collected at 1 point in time
Time series data: collected over a period of time Time series data: collected over a period of time
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 17
What
What
do
do
we
we
know?
know?
From From a a theoretical theoretical economic economic hypothesis hypothesis to an to an
econometric econometric validation validation
Econometric Econometric methods methods = = evaluate evaluate an an average average
relationship relationship between between Y and X Y and X
Estimates Estimates are are done done with with error error
The The Ordinary Ordinary Least Squares: Least Squares: method method aiming aiming
at at assessing assessing the the parameters parameters that that minimize minimize this this
error error
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 18
Statistical
Statistical
Definition
Definition
Basic Concepts
Basic Concepts
Two Two basic basic ways ways to to characterize characterize a a statistical statistical serie serie : :
- - central central parameter parameter => => mean mean, , median median
mean mean : :
- - dispersion dispersion parameter parameter => variance, => variance,
standard standard- -deviation deviation
standard standard- -deviation deviation : :
1
1
i n
i
i
i
X X
n
=
=
=

2
1
1
( )
1
i n
n i
i
X X
n

=
=
=


Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 19
Example
Example
: 2
: 2
different
different
classrooms
classrooms
Exam Exam of of Statistics Statistics
2 groups 2 groups with with on on average average exactly exactly the the same same mark mark => 11.5 => 11.5
What What information information does does it it provide provide on on your your own own result result? ?
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 20
Class A
Class A
5.9 11.5 Mean
34.3 Var. 149.5 Sum
72.25 8.5 11.5 20 13
49 7 11.5 18.5 12
30.25 5.5 11.5 17 11
20.25 4.5 11.5 16 10
12.25 3.5 11.5 15 9
6.25 2.5 11.5 14 8
0.25 0.5 11.5 12 7
2.25 -1.5 11.5 10 6
6.25 -2.5 11.5 9 5
20.25 -4.5 11.5 7 4
30.25 -5.5 11.5 6 3
72.25 -8.5 11.5 3 2
90.25 -9.5 11.5 2 1
(Xi-mean) Xi-Mean Mean Xi Rank
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 21
Class B
Class B
1.04 11.5 Mean
1.08 Var. 149.5 Sum
2.25 1.5 11.5 13 13
2.25 1.5 11.5 13 12
1 1 11.5 12.5 11
1 1 11.5 12.5 10
0.25 0.5 11.5 12 9
0 0 11.5 11.5 8
0 0 11.5 11.5 7
0.25 -0.5 11.5 11 6
0.25 -0.5 11.5 11 5
0.25 -0.5 11.5 11 4
1 -1 11.5 10.5 3
2.25 -1.5 11.5 10 2
2.25 -1.5 11.5 10 1
(Xi-mean) Xi-Mean Mean Xi Rank
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 22
=> To
=> To
characterize
characterize
a
a
serie
serie
you
you
need
need
The The mean mean of of the the serie serie (central (central parameter parameter) )
The The standard standard- -deviation deviation (dispersion) (dispersion)
=>
=>

of
of
course,
course,
the
the
answer
answer
also
also
depends
depends
on
on
the
the
dispersion (
dispersion (
standard
standard
-
-
deviation
deviation
)
)
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 23

Same
Same
thing
thing
with
with
a coefficient
a coefficient
estimate
estimate

=>
=>
the
the
coefficient
coefficient
is
is
an
an
averaged
averaged
impact
impact
=>
=>
its
its
significance
significance
depends
depends
on
on
its
its
dispersion
dispersion
,
,
i.e. i.e. the the accuracy accuracy associated associated to to the the estimate estimate
Don
Don

t
t
forget
forget
!
!
the
the
predictions
predictions
are
are
done
done
with
with
error
error
!
!
Y = Y = + + X X + +
Given the error in the estimate or the inaccuracy in the Given the error in the estimate or the inaccuracy in the
estimate (assessed by the dispersion) estimate (assessed by the dispersion)

is
is

significantly different from


significantly different from
zero?
zero?
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 24
Estimation
Estimation
of
of
the
the
education
education
return (1)
return (1)
One One ( (perhaps perhaps you you?) ?) wants wants to to know know the the impact impact of of an an
additionnal additionnal year year of of education education on on his his wage wage
Economic Economic Theory Theory : Mincer : Mincer s s Equation Equation
Econometric Econometric point: point: how how big big is is ? ?
=>
=> lwage lwage
i i
= = + + . .educ educ
i i
+ + . .exper exper
i i
+ + . . expersq expersq
i i
+ +
i i
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 25
=> 428 observations in => 428 observations in the the sample sample
=> =>R R- -squared squared = 15.68% = 15.68%
Our Our model model predicts predicts 15.68% 15.68% of of the the fluctuations fluctuations of of the the wages wages
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 26
lwage
educ

( (average average) coefficient ) coefficient


=> 0.1075 => 0.1075
=> => Any Any additional additional year year of of education education implies implies on on
average average an an increase increase of of 0.1075% in 0.1075% in the the wage wage
Not Not only only a a mean mean (coefficient), but (coefficient), but also also a a standard standard
deviation deviation (0.0141465) (0.0141465) as as any any other other statistical statistical serie serie
The The standard standard deviation deviation provides provides information on information on the the
accuracy accuracy of of the the estimate estimate. .
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 27
Coefficient estimated with error
standard deviation
Confidence Interval :
95% chances for the coefficient to be in the interval
(

2 = 0.07968 ; + 2 = 0.1352956)
2 = 0.07968 ; + 2 = 0.1352956)
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 28
Is Is significantly significantly different different from from zero zero? ?
A test A test classically classically used used to compare to compare averages averages: t : t- -test. test.
=>Compare the actual coeff.( ) with the restricted coeff.( )
weighted by the dispersion (= a measure of the
accuracy of the estimate)

= 0.1075; = 0 !! = 0.01414 =>


0.1075
7.60
0.01414
t

= =

So So what what? ?

Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 29
H H
0 0
: : = 0 = 0
Compute Compute a a t t- -statistic statistic
| | t t- -statistic statistic | | > 2 => > 2 => reject reject H H
0 0
=> => significantly significantly different different from from zero zero ( (what what we we expected expected!!) !!)
| | t t- -statistic statistic | | < 2 => < 2 => cannot cannot reject reject H H
0 0
=> => significantly significantly different different from from zero zero ( (no no impact impact of of
studying studying one one more more year year on on my my wages wages !!) !!)
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 30
1
1
-
-
Coefficient =
Coefficient =
the
the
impact
impact
studied
studied
2
2
-
-
Either
Either
the
the
standard
standard
deviation
deviation
or
or
the
the
t
t
-
-
statistic
statistic
or
or
the
the
p
p
-
-
value (
value (
critical
critical
probability
probability
i.e.
i.e.
the
the
type 1
type 1
error
error
)
)
3
3
-
-
T
T
-
-
test =>
test =>
= 0
= 0
for
for
each
each
coefficient
coefficient
4
4
-
-
P
P
-
-
value
value
associated
associated
= Type 1
= Type 1
error
error
.
.
Whatever
Whatever
the
the
econometric
econometric
results
results
&
&
the
the
purpose
purpose
of
of
the
the
study
study
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 31
Comparing Comparing the the t t- -statistic statistic to 2 = a 5% type to 2 = a 5% type- -1 1 error error
A type I A type I error error = = the the probability probability to to reject reject H0 H0 while while it it s s true true
=>i.e. 5% chances to =>i.e. 5% chances to be be wrong wrong when when rejecting rejecting H H
0 0
A more A more accurate accurate way way: : the the p p- -value = value = the the exact type 1 exact type 1 error error: :
Less Less than than 1% chance to 1% chance to be be wrong wrong when when rejecting rejecting H H
0 0
( (H H
0 0
= = the the coefficient coefficient is is not not significantly significantly different different from from zero zero ) )
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 32
Case
Case
-
-
Study
Study
:
:
Feldstein
Feldstein
and
and
Horioka
Horioka
(1980)
(1980)
From From the the liberalisation liberalisation of of the the capital capital flights flights
=> => how how did did the the capital capital really really move move ? ?
Feldstein Feldstein and and Horioka Horioka (1980) : (1980) : correlation correlation between between
savings savings and and investments investments
The The impact impact of of economic economic policy policy depends depends on on the the degree degree
of of mobility mobility of of the the capital capital
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 33
Correlation between Savings and Investment
and the openness of the economy
1- Correlation between Savings and Investments close to 1:
Closed economy
any increase in national savings induces an identical increase in investments
=> low degree of capital mobility
2- Correlation between Savings and Investments close to 0:
Openned (integrated) economy:
National savings respond to investment opportunities on the world market/
national investment is financed by savings fromthe rest of the world => high
degree of capital mobility
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 34
Econometric
Econometric
model
model
(
(
Feildstein
Feildstein
and
and
Horioka
Horioka
)
)
.
i i
i
i i
I S
Y Y
= + +
Testing Testing the the correlation correlation between between Investments Investments and and Savings Savings
=> =>
T T- -test on test on the the
(i) (i) H H
0 0
: : = 0 = 0
( (ii ii) ) H H
0 0
: : = 1 = 1
The The sample sample : : 19 countries 19 countries of of the the OECD OECD
(a) (a) long long- -term term effect effect (1970 (1970- -1998) 1998)
(b) (b) short short- -term term ( (three three 10 10 year year periods periods) )
1970 1970- -1979; 1980 1979; 1980- -1989; 1990 1989; 1990- -1998 1998
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 35
Saving
Saving
and
and
Investment
Investment
: long
: long
period
period
(1970
(1970
-
-
98)
98)
0.58 0.58 R R- - Squared Squared
19 19 Observations Observations
0.03 0.03 ( (standard standard- -error error) )
0.08 0.08 Constant Constant
0.13 0.13 ( (standard standard- -error error) )
0.62 0.62 s s
i i
Saving Saving and and Investment Investment: :
1970 1970- -1998 1998
1 1- - Is Is b b significantly significantly different different from from
zero zero? ? H H
0 0
: : = 0 = 0
t t- -test (to compare test (to compare means means) )
| | 4.85 4.85 | | > 2 > 2
At At the the 5% 5% level level H H
0 0
is is rejected rejected
Note Note the the p p- -value ( value (computed computed by by the the
software) software) is is inferior inferior to 1% to 1%
In In the the long long- -term term we we cannot cannot
conclude conclude to a to a perfect perfect degree degree
of of capital capital mobility mobility

0.62 0
4.85
0.13
t


= = =
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 36
Saving
Saving
and
and
Investment
Investment
: long
: long
period
period
(1970
(1970
-
-
98)
98)
0.58 0.58 R R- - Squared Squared
19 19 Observations Observations
0.03 0.03 ( (standard standard- -error error) )
0.08 0.08 Constant Constant
0.13 0.13 ( (standard standard- -error error) )
0.62 0.62 s s
i i
Saving Saving and and Investment Investment: :
1970 1970- -1998 1998
1 1- - Is Is significantly significantly different different from from
one one? ? H H
0 0
: : = 1 = 1
t t- -test (to compare test (to compare means means) )
| | 2.94 2.94 | | > 2 > 2
At At the the 5% 5% level level H H
0 0
is is rejected rejected
Note Note the the p p- -value ( value (computed computed by by the the
software) software) is is inferior inferior to 1% to 1%
In In the the long long- -term term we we cannot cannot
conclude conclude to to closed closed economies economies

0.62 1
2.94
0.13
t


= = =
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 37
Saving
Saving
and
and
Investment
Investment
:
:
short
short
-
-
term
term
Three Three 10 10 year year periods periods 1970 1970- -79; 1980 79; 1980- -89; 1990 89; 1990- -98 98
0.40 0.65 0.82 0.82 s s
i i
0.15 0.14 0.14 0.14 ( (standard standard- -error error) )
0.11 0.08 0.08 0.08 Constant Constant
0.03 0.03 0.03 0.03 ( (standard standard- -error error) )
0.30 0.55 0.61 0.61 R R- - Squared Squared
19 19 19 19 Observations Observations
(3)
1990-1998
(2)
1980-1989
(1)
1970-1979
Periods
Saving and Investment
=> => What What can can you you say say about about openness openness of of OECD countries OECD countries
over over each each 10 10 year year periods periods? ?
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 38
Exercise
Exercise
:
:
Saving
Saving
and
and
Investment
Investment
:
:
short
short
-
-
term
term
Three Three 10 10 year year periods periods 1970 1970- -79; 1980 79; 1980- -89; 1990 89; 1990- -98 98
What What can can you you say say about about openness openness of of OECD countries OECD countries
for for each each of of the the three three periods periods? ?
(for (for each each period period: : perfect perfect capital capital mobility mobility? ? Closed Closed economies economies? ? Etc Etc ) )
What What would would you you say say regarding regarding the the evolution evolution of of the the
capital capital mobility mobility over over the the whole whole period period? ?
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 39
Exercise
Exercise
:
:
Saving
Saving
and
and
Investment
Investment
:
:
short
short
-
-
term
term
Three Three 10 10 year year periods periods 1970 1970- -79; 1980 79; 1980- -89; 1990 89; 1990- -98 98
Results Results: :
Conclusion: Conclusion:
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 40
Exercise
Exercise
:
:
Saving
Saving
and
and
Investment
Investment
:
:
short
short
-
-
term
term
Three Three 10 10 year year periods periods 1970 1970- -79; 1980 79; 1980- -89; 1990 89; 1990- -98 98
Results Results: :
Period Period 1 : 1 : (i) (i) H H
0 0
: : =0 => t = 5.93 => |t|>2 => rejection =0 => t = 5.93 => |t|>2 => rejection of of H H
0 0
( (ii ii) ) H H
0 0
: : =1 => t = =1 => t = - -1.32 => |t|<2 => 1.32 => |t|<2 => non non- -rejection rejection of of H H
0 0
Conclusion: Conclusion:
- - None None of of the the three three periods periods with with a a perfect perfect capital capital mobility mobility
- - Even Even a a behaviour behaviour of of closed closed economies economies over over the the first first period period
=> => A change A change toward toward openness openness over over the the 2 2 last last periods periods? ?

0.82 =

0.65 =

0.40 = Period Period 3: 3: (i) (i) H H


0 0
: : =0 => t =2.67 => |t|>2 => rejection =0 => t =2.67 => |t|>2 => rejection of of H H
0 0
( (ii ii) ) H H
0 0
: : =1 => t = =1 => t = - -4.04 => |t|>2 => rejection 4.04 => |t|>2 => rejection of of H H
0 0
Period Period 2: 2: (i) (i) H H
0 0
: : =0 => t = 4.62 => |t|>2 => rejection =0 => t = 4.62 => |t|>2 => rejection of of H H
0 0
( (ii ii) ) H H
0 0
: : =1 => t = =1 => t = - -2.46 => |t|>2 => rejection 2.46 => |t|>2 => rejection of of H H
0 0
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 41
=> => A change A change toward toward openness openness over over the the 2 2 last last periods periods? ?
In other words
=> are the two coefficients significantly different?
=> H
0
:
8089
=
9098
?
=> not exactly the same t-statistic as usual
because the both terms are estimated (with error)

8089 9098
8089 9098
2 2


0.40 0.65
* 1.2419 2
0.14 0.15
t




= = = <
+
+
=> No significant decrease in over the two last periods, we cannot
conclude to an increased liberalisation of the capital market for this
sample of countries and these periods.
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 42
Conclusion
Conclusion
-
-
What
What
have
have
we
we
learnt
learnt
?
?
-
-
-
-
-
-
-
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008
Executive MBA 2007-
2008
emba bridge- 2006/2007 43
Conclusion
Conclusion
-
-
What
What
have
have
we
we
learnt
learnt
?
?
1 1- - Basic Basic methodology methodology regarding regarding econometrics econometrics
- - Economic Economic problem problem => data => => data => econometric econometric validation validation
2 2- - Characterizing Characterizing a a statistical statistical serie serie
- - Central Central parameter parameter, dispersion , dispersion character character
3 3- - The The most most common common econometric econometric estimator estimator
- - Ordinary Ordinary Least Least Squares, concept Squares, concept of of error error- -term term
4 4- - Reading/ Reading/interpreting interpreting econometric econometric results results
- - R R- -squared squared, Marginal impact, , Marginal impact, elasticity elasticity, , semi semi- -elasticity elasticity, ,
confidence confidence interval interval, p , p- -value, value,
5 5- - Statistical Statistical test test of of the the coefficients coefficients
- - t t- -test ( test (student student test): test): against against a constant, a constant, against against another another estimate estimate
Executive Executive MBA MBA HEC Lausanne 2007/2008 HEC Lausanne 2007/2008

Вам также может понравиться