ETHNIC MINORITIES IN VIETNAM: A SOCIOECONOMIC PERSPECTIVE

1
ETHNIC MINORITIES IN VIETNAM A SOCIO-ECONOMIC
PERSPECTIVE

• Bob Baulch (Univ. of Sussex)
• Truong Thi Kim Chuyen
• (National University of Ho Chi Minh City)
• Dominique Haughton (Bentley College, Waltham)
• Jonathan Haughton (Suffolk University, Boston)
• Suffolk University, February 2002

2
Who?

• 54 ethnic groups
• Smallest has lt1,000 members

3
Where?

• Urbanization
• Kinh 27
• Others 2
• Minorities in
• Northern Uplands
• Central Highlands
• Scattered, south

4
Poverty

• For ethnic minorities
• High headcount poverty rate, falling slowly

5
Living Standards

• Education
• Enrolment rates low for some minority groups
• But rising faster
• Health
• Sought prenatal care 47 of minority, 70 of
  Kinh mothers
• Sought care for sick child 75 of minority, 88
  of Kinh households

• Nutrition
• BMI, mean consumption of calories only slightly
  lower for minorities
• Expenditure/capita
• Gap is large and rising

6
Minority vs. Majority HHs, 1993 and 1998
7
The Key Policy Question

• How close the gap in living standards between
  majority and minority households?

8
A finer disaggregation 5 groups
9
Kernel densities Expenditure per capita for 5
groups
10
Finest disaggregation Primary Sch. Enrol. Rates
by 12 Ethnic Groups
11
Intermarriage of HH heads

12
Government programs

• Implemented by
• CEMMA (Committee for Ethnic Minorities in
  Mountainous Areas)
• MOLISA, especially Hunger Eradication and Poverty
  Reduction Program
• But spread thin corruption.

• Subsidies
• Salt, radios
• Agricultural policy
• Poppies, RD, reforestation
• Land
• Titling, traditional rights often not recognized

13
Government programs (cont.)

• Migration
• New economic zones
• Education and Training
• Expanded in Vietnamese
• Health
• Expanded
• Cultural identity
• Dances, folklore, fashion

14
Why are minorities so poor?

• Endowments
• Land, water emphasized by minority hhs
  themselves
• Reproducible physical capita
• Human capital
• Remittances
• Objective factors.

15
Evidence on endowments
16

• Remoteness
• Inputs dearer
• Price received for output lower
• Schooling more expensive
• Language
• Politics
• Institutional

17
Evidence on Remoteness
18

• Knowledge, customs, culture
• Discrimination
• Picked up as lower returns on characteristics
• Subjective

19
Evidence on culture
20
Traditional Solution Assimilate

• e.g.
• Suppression of long houses
• Use of Vietnamese language in instruction

21
Regression results 1

• One more HH member, expenditure per cap falls 7
• Land Extra hectare of annual crop land raises
  expenditure per capita by 37 for minorities,
  by 16 for kinh.
• Consistent with same r of r on land
• Explains why minorities emphasize land

22
Regression results 2

• More education
• Associated with higher consumption per capita
• Higher relative (not absolute) return for
  minority hhs in full sample, up to 7 years
• Higher returns to KinhHoa in mixed-commune
  subsample

23
Blinder-Oaxaca decomposition

• Difference in log Difference due to Difference in
• of per capita differences in returns to
• expenditure characteristics characteristics
• where a,b and two groups (e.g. a for KinhHoa and
  b for minorities)

24
Decomposition results
25
Decomposition shows

• Even if minority characteristics were raised to
  KinhHoa level, this would still only eliminate
  about half of the gap in consumption per capita.
• Mirrors findings of van de Walle and Gunewardena
  using VLSS93.

26
Conclusions

• Ethnic minorities are poor, and falling behind
  relatively but gaining somewhat from economic
  growth
• The gap is smaller for the more assimilated
  groups, large for Hmong and Central Highland
  minorities

• Half the difference is attributable to
  differential returns to characteristics
• Challenge is to decrease the gap.
• Suggests need for anti-poverty interventions
  tailored to needs of different groups

27
Its surprisingly relevant

• Central Highland protests, Feb-Mar 2001
• The Son La debate
• Public rhetoric (backward)
• How CEMMA runs who leads it
• Long houses house churches
• Language policy in schools
• Surprise Nong Duc Manh

28
MARS (MULTIPLE ADAPTIVE REGRESSION SPLINES)

• For each continuous independent variable, MARS
  creates a piecewise linear function, first with
  too many change points (knots) and then prunes
  unnecessary knots
• For each categorical variable, MARS arranges
  categories to obtain the best possible fit
• MARS looks for suitable interactions between
  independent variables
• MARS ends up with a collection of Basis
  Functions, which are transformations of
  independent variables taking into account
  non-linearities and interactions
• MARS then estimates a least-squares model
  with its Basis Functions as independent variables

29
MARS insights

• Kinh Education substitutes for land
• Minorities Irrigated land an important
  determinant of living standards, but requires
  household labor as complement.

30
MARS model for KinhHoa
31
(No Transcript)
32
MARS model for ethnic minorities
33
(No Transcript)