Research Ideas and
econometric models
Nguyễn Hoàng Bảo
Outline
1. A Logistic Regression Model of Drop out of
High School
2. A Logistic Model of Rural–Urban Migration
3. Choice of domestic or foreign product
4. A Multinomial Logistic Regression Model of
Determinants of Divorce
5. Who receives remittances
6. How to construct demand for a single good
7. Who smokes cigarettes
1. A Logistic Regression Model of
Drop out of High School
•
Droup out of school is, of course, not occurred
randomly.
•
There are many factors which might affect the
problem of drop out prematurely.
•
Dependent variable: Drop out? (Y=1)
•
Independent variables: If the positive/negative
sign of the specific explanatory variable means
increase/decrease the probability of dropout and if
the question mark of the specific explanatory
variable means that there is no clear pattern.
There are four groups of explanatory variables.
Individual characteristics

divorced, being separated, being widowed (?)
School characteristics
•
School fees charged (+)(c)
•
Distance to school (+)
•
Teacher educational level (-)
•
Teacher experience (-)
•
Teacher motivation (-)
•
Pedagogical techniques used (?)
•
Condition of school building (-)
•
Availability of learning materials (textbooks) (-)
•
Quality of learning materials (-)
•
Characteristics of head of school (?)
•
Hours and days that school is opened (-)
•
Number of pupils in the class, and/or teachers/students ratio, space
for each pupils (proxy measurement of quality of schooling) (-)
Regional characteristics
•
Urban/rural (-)

−
=
−=
n
t
rt
R
R
n
t
rt
U
U
dt
dt
e
t
Y
Ce
t
Y
0
0
0
)(
)(
ω
ω
3. Choice of domestic or foreign
product

•
Religion
•
Number of children
•
Being single of relatives
•
Marital status of parents
•
Job characteristics (out of sight, out of mind)
•
Social ties
5. Who receives remittances
•
Self–adjustment mechanism
•
Community and family spirit

5. Who receives remittances
•
Household size (-)
•
I head male? (Yes=1) (-)
•
Is head employed? (-) (causality flows?)
•
Regional effects (Ho Chi Minh, Hue, and Da Nang)
•
Education spending per capita (+)
•

•
Price of complementary products (-)
•
Income per capita (?)
•
Advertising costs (+)
•
Tastes (?)
•
Desire (+)
•
Externality network (?)
•
Seasonal/Cyclical affects

20
Characteristics of good
•
Durable or undurable good
•
Expensive or unexpensive good
•
Storage or nonstorage good
•
Fashionable or unfashionable good
•
Necessary or unnecessary good
•
Inferior, common, and luxury good
•

22
Demand function (1)
lnQ
D
= a + b
j
(P
j
/P
i
) + … + b
k
(P
k
/P
i
) + c ln(Y/P
i
) +
other variables
Q
D
is the demand of rice (quantity)
P
i
is the price of rice
P
j
…P
k

is spending on rice
P
i
is the price of rice
y is total expenditure or income per capita
lnP
i
.lny captures the structural change of consumption
effects
P
index
is the consumer price index
24
Elasticity of demand with respect to
its own price
E = (%ΔQ
D)
/ (%ΔP)
If E=-0.2, when price increases 1%, demand
decreases 0.2%
25
Elasticity of demand with respect to
income
E = (%ΔQ
D)
/ (%ΔY)
If E=0.6, when income per capita increases 1%,
demand increases 0.6%
If E=-0.2, when income per capita increases 1%,
demand decreases 0.2% (the good is inferior