ESTIMATION RESULTS OF HOUSEHOLD WELFARE
III. ESTIMATION RESULTS: The household welfare generating potential of land
Table 6 presents the result of estimation from the household welfare equation. The result shows that the marginal welfare value of land is highly significant. An increase of 1 hectare will increase annually income by 2,700,000 VND. The result of increase of marginal welfare value of land is consistent with results have been reached in Ramo´n Lo´pez and Alberto Valde´s (2002); Frederico Finan, Elisabeth Sadoulet, Alain de Janvry (2002) and Dominique van de Walle and Dileni Gunewardena (2001).
For a better sense of this value, I compare with the poverty line adapted by GSO/WB is 2100 kl/person/day, it is equivalent to 1,916 thousands VND/year (price in 2002) and see that the value of 2,700,000 VND/hectare/year would be a considerable. That is why we say that annual crop land would be a valuable instrument for household in the process of poverty reduction in rural.
Characteristics of household head are associated with large welfare effects. As presented in the table 6 A marginal increase in the household head’s education level raises welfare by 2,498 thousand VND/year (unfortunately my estimation result does give statistically insignificant result of household head without education). Marginal welfare in term of education attainment sharply increase at higher level from 5,996 thousands if education attainment of the head at lower secondary to 8,648 thousands at higher secondary. This welfare would reach up to 14,929 thousands if the heads have education attainment from vocational program, colleague and university or higher, If compare the welfare of subgroup that the head attains only primary level with the highest subgroup we see that it is almost 6 times lower than, the gap is too large and it shows that education attainment would be significant human capital in the process of poverty reduction. There are many researches worldwide and in Vietnam has demonstrated that higher education attainment always raise income for households such
as in papers reached by give higher Behrman, J.R. and Knowles, J.C. (1999); Brown, P. and Park, A.
(2001); Moock, P.R, H.A Patrinos, and M. Venkatarama (2003); Glewwe, P. and Jacoby, H.G. 2004.
The result presents household that lead by female raise by 1,774 thousands a year. This is consistent with the result presented in the table 4 (Ethnic and gender dimensions of poverty in Vietnam) where poverty incidence was always much lower compare to household lead by male. The value is meaningful, it is about 6.5 times higher than income raised by increase in household age.
Labor force: Labor force used in the regression is number of individuals excluding household head in the household that those whose age ranging from 15 to 60 year old for both male and female.
Male labor force is in 5 subgroups based on their education attainment, from no education to those who have education at level of vocational/ technical, and college and higher degree. Below primary, which are included no education. It is the same division to female labor force. In my estimation, I include 10 independent variables consists of 5 subgroups for male at five level of education attainments and the same 5 subgroups for female . The results release only 2 subgroups that they have statistically significant, so the rest will be dropped. The result shows that an additional of an adult male with no education attainment increase welfare by 1,889 thousands VND a year and increase to 1,922 thousands VND if they have primary level. This value is the same with the poverty line used by GSO/WB.
There is a high cost associated with ethnicity as being minority ethnic group by 1,871 thousands VND a year. Dominique van de Walle and Dileni Gunewardena (2000) pointed out that there were larger differences in educational attainments, land assets, household size, infrastructure and other attributes between majority and minority group, above disadvantages of minority group would affect the quality of education and other infrastructure, accessibility to services and information, and off-farm income-earning opportunities. Poverty incidence of minority ethnic group is still so high (table 4).
Table 6 - Estimation of the welfare household index
Arg value Coefficient Standard
Error t Statistic P value [95% Conf. Interval]
Land assets 3944.787 2699.725 399.910 6.750 0.000 1915.815 3483.635
Social subsidy 2010.723 1.235 0.046 27.100 0.000 1.145 1.324
Head of household characteristics
gender 0.846 1774.459 559.165 3.170 0.002 678.377 2870.541
age 39.638 267.554 14.045 19.050 0.000 240.023 295.085
Ethnicity 0.133 -1871.464 301.564 -6.210 0.000 -2462.593 -1280.334
0< education <6 0.229 2498.406 664.090 3.760 0.000 1196.647 3800.164
6<= education <=9 0.284 5996.175 735.970 8.150 0.000 4553.517 7438.833
9< education <=12 0.093 8648.607 792.435 10.910 0.000 7095.265 10201.950
education >12 0.184 14929.090 743.760 20.070 0.000 13471.170 16387.020
Labor force (number of individual)
Male adults with education = 0 0.174 1888.898 729.769 2.590 0.010 458.394 3319.402
Male adults with 0< education <6 0.203 1921.769 770.308 2.490 0.013 411.801 3431.736 Locality characteristics
Cooperative 0.276 -2500.103 249.538 -10.020 0.000 -2989.251 -2010.955
Male self-employment wage (daily) 5.630 -1.674 1.745 -0.960 0.338 -5.094 1.747
Male no self-employment wage
(daily) 26.753 3.983 0.885 4.500 0.000 2.249 5.716
Female self-employment wage
(daily) 0.292 -14.192 5.784 -2.450 0.014 -25.529 -2.854
Female no self-employment wage
(daily) 6.403 5.088 1.471 3.460 0.001 2.205 7.971
Region
Red River Delta 0.188 1947.667 553.263 3.520 0.000 863.154 3032.180
Northeast 0.181 198.0105 463.113 0.430 0.669 -709.789 1105.810
Northwest 0.029 -920.951 610.888 -1.510 0.132 -2118.423 276.521
North Central Coast 0.096 -2076.030 502.002 -4.140 0.000 -3060.060 -1091.999
South Central Coast 0.096 1877.706 565.241 3.320 0.001 769.712 2985.699
Southeast 0.155 5981.269 582.402 10.270 0.000 4839.637 7122.902
Mekong Delta 0.202 2478.22 517.121 4.790 0.000 1464.553 3491.887
Constant -4627.079 836.239 -5.530 0.000 -6266.286 -2987.872
Number of observations = 9523 F( 23, 9499) = 137.54
Prob > F = 0.0000 R-squared = 0.4781 Root MSE = 10629
Regression with robust standard errors (White's test for heteroskedasticity) Source: Author's calculations based on VLSS 2002
In my result, accessing to cooperatives decrease a high value for household, by 2,500 VND a year. This result partly explains the successful process of de-collectivization in Doi Moi, this process has brought significant performance in agricultural production and poverty reduction. However Vietnam is still maintaining the existence of cooperative, but these are facing many difficulties and low efficiency (Vietnam Cooperative Association Report 2004).
Turning to returns to land, they are positive and significant for all regions except for NW and NCC. An increase in hectare annual cropland, for example, increased by 1,948 VND a year in RRD and the same in SCC. This value is quite higher in MKD by 2,478 thousands but the marginal value reach highest in SE by 5,981 thousands a year. While there is a cost in 920 and 2,076 thousands VND a year in NW and NCC respectively.
There is a wide range of difference in return to land among regions since the returns to land will be sensitive to the levels of input use (e.g., labor and capital), technology and environment (e.g., irrigation) in each region, the negative returns in the NU and NCC would suggest that at least one of these components was critically unfavorable there. In fact, the average irrigation ratio of the NU has been much lower than in the RRD (Dollar and Glewwe, 1998; MARD, 2000). According to Kazushi TAKAHASHI (2004) pointed out that the different returns to land caused by the difference in the returns to human capital variables of the RRD are higher than the NU, but lower than the MKD.
SUMMARY OF FINDINGS
In this section, I estimate a household welfare equation to explore the relationship between land and welfare. There are two main findings. First, Agricultural land contribute positive income to household in rural. The marginal value of land would change in all 8 regions that I estimate. Second, the marginal value of land differs according to a household’s complementary assets and contextual setting.
CHAPTER V