Bảng 1. presents summary statistics for the gender, ethnicity and language of the household head as well as the poverty status of the household. Unsurprisingly given that we follow the same households over time, most of the general characteristics of households do not change much between 2008 and 2010, with around 78 percent of households headed by a male, 80 percent of households of Kinh ethnicity, 99 percent of households speaking Vietnamese and 84 percent of households where Vietnamese is the main language. Nevertheless, there are some small changes (with the exception of the gender of the household head) which are statistically significant.
The gender of the household head shows some variation across provinces, with only 68 percent of households headed by a male in Khanh Hoa and up to 92 percent in Lai Chau. As the ethnicity of households is often province specific, it is not surprising that the percentage of Kinh households varies strongly by province, with as few as 10 percent of households of Kinh ethnicity in Dien Bien, and 14 percent in Lai Chau (the North western provinces) to as many as 92 percent, 98 percent, 99 percent or even 100 percent in Khanh Hoa, Quang Nam (the South central coast provinces), Ha Tay and Long An respectively (the delta provinces). The majority of sampled households speak Vietnamese and this is highly correlated with ethnicity.
160
160
Table 1.1: General Household Characteristics by Province No. of
HH’s surveyed
Percent Gender of HH headc
Ethnicity of HH head
HH head speaks Vietnamese
Vietnamese main language of
HH
HH classified as poor by authorities (percent male) (percent Kinh) (percent) (percent) (percent) Provincea
Ha Tay (RRD) 480 21.8 76.4 99.0 100 100 10.8 Lao Cai (NE) 87 4.0 87.4 23.0 90.8 34.5 24.1 Phu Tho (NE) 305 13.9 75.4 80.0 100 93.8 8.9 Lai Chau (NW) 112 5.1 91.9 14.3 90.2 33.9 28.6 Dien Bien (NW) 105 4.8 90.5 9.5 99.0 10.5 32.4 Nghe An (NCC) 192 8.7 82.3 87.5 99.5 88.5 16.1 Quang Nam (SCC) 290 13.2 71.7 98.3 100 98.6 21.4 Khanh Hoa (SCC) 38 1.7 68.4 92.1 100 92.1 23.7 Dak Lak (CH) 135 6.1 83.7 68.1 97.8 77.0 14.8 Dak Nong (CH) 103 4.7 85.4 73.8 99.0 76.7 14.6 Lam Dong (CH) 67 3.1 77.6 62.7 100 65.7 14.9 Long An (MRD) 286 13.0 73.1 100 100 100 12.9
Total, 2010 2,200 100 78.4 79.5** 98.9*** 84.0** 15.9***
Total, 2008 2,200 100 78.7 79.0 97.4 83.2 20.0
Total, 2006b 2,193 100 80.3 80.6 97.0 84.7 22.9
Total, 2010w 1,314 100 78.4 81.4 99.0 84.8 15.2
a Region between brackets: RRD (Red River Delta), NE (North East), NW (North West), NCC (North Central Coast), SCC (South Central Coast), CH (Central Highlands), MRD (Mekong River Delta) – no households from the SE (South East) were included in the survey.
b Small differences between 2006 and 2008 totals are observed due to some methodological changes and measurement errors.
c Sample used is 2,198 due to two missing observations.
** Difference between 2008 and 2010 is significant at 5% level; *** significant at 1% level.
The most notable change over time shown in Table 1.1 is in the percentage of households classified as poor by MoLISA. This declined from 23 percent in 2006, to 20 percent in 2008 and 16 percent in 2010.71 In Figure 1.1 we explore this change in more detail by province, showing that the downward trend in poverty (as classified by Vietnamese authorities) has occurred in nearly all provinces, with the exception of Dien Bien, Khanh Hoa and Long An where poverty rates on this measure are higher than in 2006. The largest decrease in poverty is observed in Lai Chau with a 20 percentage point fall (from 49 percent to 29 percent), followed by Phu Tho with a decrease of around
71 It should be noted that the poverty line was raised by MoLISA in October 2010. This change is not reflected in our data given that our households are sampled in June/July/August of 2010.
161
161
12 percentage points (from 21 percent to nine percent). In our 2010 sample, Phu Tho has the lowest poverty prevalence, while Dien Bien has the highest (32 percent). The provinces of Lai Chau (29 percent), Lao Cai (24 percent) and Khanh Hoa (24 percent) also have significant proportions of their populations classified as poor by MoLISA.
Figure 1.1: Changes in Poverty Status between 2008 and 2010 by Province (percent)
N=2,200
Bảng 1. presents further statistics on household characteristics, including support from children living outside the household and place of birth of the household head or spouse. Statistics are disaggregated by household socioeconomic status, which we proxy here using food expenditure quintile. The percentage of male headed households is more or less the same over all expenditure quintiles, while there is a clear positive correlation between household wealth and Kinh ethnicity.
Furthermore, while 91 percent of the female headed households are of Kinh origin, only 76 percent of the male headed households are.
162
162
Table 1.2: Household Characteristics by Gender of Household Head and Food Expenditure Quintile (percent)
Gender of HH
head (percent
male)
Ethnicity of HH
head (percent
Kinh)
HH head speaks Vietnamese
Vietnamese main language of
the HH
Support from children outside the
household
Born in commune
(head, spouse or
both)
HH classified as poor by authorities Household Head
Female 90.7 99.4 92.8 33.5 77.9 22.7 Male 76.4 98.8 81.7 21.4 81.3 14.0 Food Expenditure Quintile
Poorest 79.3 51.5 96.4 59.6 20.2 83.1 33.5 2nd poorest 79.6 74.0 98.9 79.2 18.3 83.8 19.4
Middle 78.3 87.9 99.3 91.1 22.7 82.2 10.3 2nd richest 77.7 88.9 99.8 93.2 25.7 78.4 9.8
Richest 77.0 95.7 100.0 97.5 33.2 75.5 6.4
Total 2010 78.4 79.5** 98.9*** 84.0** 24.0*** 80.6 15.9***
Total 2008 78.7 79.0 97.4 83.2 14.0 81.1 20.0
Total 2010w 78.4 81.4 99.0 84.8 26.8 77.0 15.2 N=2,200 (2,198 for data separated by gender of the household head).
** Difference between 2008 and 2010 significant at 5% level; *** significant at 1% level.
Between 2008 and 2010 we observe a large increase in the number of households who receive support from children living outside the household from 14 percent in 2008 to 24 percent in 2010 (a difference which is statistically significant). This observation is consistent with rapidly increasing levels of internal and international migration in Vietnam.72 Female headed households are more likely to receive support: 34 percent of female headed households receive such support while only 21 percent of the male headed households do. In 2010, 74 percent of the female heads who receive support from their children are widows (80 percent in 2008).73 As such, it appears more likely that children support their mothers following the death of a male head of household. There also appears to be a positive correlation with expenditure quintile: the percentage of families that receive support is increasing with socioeconomic status with 20 percent of the poorest households receiving support from children living outside the household increasing to 33 percent of the richest households. The direction of causation here is however not immediately clear, in other words, whether children of wealthier parents are better able to afford to help, or whether the support received leads to higher levels of wealth of the recipient cannot be determined from these summary statistics.
A negative correlation appears to exist between food expenditure quintile and whether the household head or spouse (or both) were born in the commune where they live. The percentage of
72 Vietnam has experienced an exponential increase in movements of people both inside as outside its borders over the last 20 years (UNDP, 2010). The impacts of this migration are not only limited to the migrants themselves, but benefits also fall to the large numbers of migrant-sending households through remittances.
73 Male headed households who receive support are mostly married (93 percent of households in 2008 and 2010).
163
163
“immobile” households is higher in the poorer quintiles with 83 percent in the poorest and 76 percent in the richest quintile. This may suggest that mobility is (positively) correlated with economic status.
We also observe that while there does not appear to be a correlation between the gender of the household head and food expenditure quintile, female headed households do seem more likely to be classified as poor by the authorities (MoLISA): 23 percent of female headed households are classified as poor compared to 14 percent of male headed households. Moreover, the percentage of households classified as poor by MoLISA varies significantly by food expenditure quintile, with 34 percent of poor households in the lowest quintile and only six percent of poor households in the richest quintile.
Our measure based on food expenditure and the measure used by the authorities would thus appear to be well-matched.74
Figure 1.2 shows changes in poverty and support from children over time, split by gender of the household head and food expenditure quintile. In part A (households classified as poor) we find that the decrease from 2008 to 2010, in percentage point terms, of poverty was more or less the same for male and female headed households. Moreover, the decrease in the proportion of households classified as poor has clearly been largest in the poorest food expenditure quintile. In the poorest three quintiles, the decrease in the percentage of households classified as poor was between 37 percent and 40 percent while this was only 24 percent in the second richest quintile and the percentage of poor households even increased by eight percent in the richest quintile. These results are suggestive of declining levels of inequality.
Part B of Figure 1.2 shows that the positive correlation between poverty status and receiving support from children living outside the household was not existent in 2008. Even though the percentage of households receiving support from children has increased in all food expenditure groups, there is a much larger discrepancy across quintiles in 2010. In particular in the richest expenditure quintile, the proportion of households receiving support has more than doubled from 14 percent to 33 percent.75
74 The agency responsible for classifying households as poor/non-poor in Vietnam is the Ministry of Labour, Invalids and Social Affairs (MoLISA). The official criterion for being classified as poor is based on an income per capita threshold.
Between 2005 and October 2010, the threshold in rural areas was a monthly income per capita of 200,000 VND. The household’s income situation is determined using a survey carried out by local authorities. The income situation is based on income and living standards (e.g. housing), size of the household, vulnerability etc. Results from the survey are discussed in village meetings, which result in a list of poor households being finalised and submitted to commune and then district level authorities. In village meetings, certain exceptions from the income per capita criterion may be decided. For example, households with valuable assets or support from non-household members may not be classified as poor, even if they are below the income threshold. In practice, therefore, local authorities have some discretion in the classification process, and in this sense it is possible that criteria such as gender may play a role.
75 As noted, the direction of causality of this relationship is not, at this stage, clear. It would be interesting for further work to analyse the nature of this.
164
164
Figure 1.2: Changes between 2008 and 2010 for Selected Household Characteristics C. Households classified as poor by the authorities
0 10 20 30 40 50 60
Female Male Poorest 2nd poorest
Middle 2nd richest
Richest Total
2008 2010
D. Support received from children living outside the household
0 5 10 15 20 25 30 35 40
Female Male Poorest 2nd poorest
Middle 2nd richest
Richest Total
2008 2010 N=2,200