status. When selecting the households, a cluster sampling procedure was applied in
which, as a first step, a village-level sampling frame was constructed encompassing
all villages in the district,
2 including information on the number of resident
households. Twenty villages were randomly selected using the Probability Proportionate to Size (PPS) method (Carletto 1999), and based on the number of
households in each village. Within each selected village, 15 households were
then randomly selected using updated, village-level household lists as the sampling
frames. This sampling procedure results in a self-weighting sample, since the PPS
method accounts for differences in the number of resident households across
villages (Carletto 1999). After introducing the measures of poverty used, we will
describe the link between poverty and access to financial and natural capital
resources, and investigate farmers’ resilience to shocks, their attitudes towards
risk and their time preferences.
5.3.2 Poverty Measures Used
Two poverty measures are described in this chapter – one absolute and one relative
(cf. Sect. 5.2). The absolute poverty measure builds on an index of household daily
per capita expenditure. For the study, detailed data were collected on farmers’ food
and non-food expenditures in 2007 and in 2010, following the methodology of the
World Bank’s Living Standard Measurement Survey (LSMS), which is described in
detail by Grosh and Glewwe (1998, 2000). As there is a considerable amount of
seasonality in relation to agricultural production and incomes in Yen Chau district,
two expenditure survey rounds were implemented – one between March and April
during the lean season (period before the rice harvest) and the second between
December and January after harvesting of the farmers’ main crop, and using a recall
period of 2 weeks. The final estimate of per-capita daily expenditure was calculated
as an average of the expenditure elicited from the two survey rounds. The level of
expenditure obtained was used as a proxy of farmers’ incomes, and poor households
were identified using the official poverty line set by Vietnam’s Ministry of Labor,
Invalids and Social Affairs (MOLISA) for rural areas
3 (see also Chap. 12). In 2007,
and according to our estimations, 16.9 % of households were living under this line.
In addition to classifying households as poor and non-poor using the official
rural poverty line, we also used a relative poverty measure for some of our analyses,
classifying households into wealth groups based on a linear composite index
constructed using principal component analysis (cf. Dunteman 1994) from a
range of indicator variables capture multiple dimensions of poverty.
2 Except for the villages in four sub-districts bordering Laos, for which research permits are very
difficult to obtain.
3 The poverty line in 2007 was estimated to be 9,105 VND per-capita per day, then in 2010 was
raised to 11,030 VND per capita per day (Van Dinh 2012) (see Chap. 12).
180
C. Saint-Macary et al.
which, as a first step, a village-level sampling frame was constructed encompassing
all villages in the district,
2 including information on the number of resident
households. Twenty villages were randomly selected using the Probability Proportionate to Size (PPS) method (Carletto 1999), and based on the number of
households in each village. Within each selected village, 15 households were
then randomly selected using updated, village-level household lists as the sampling
frames. This sampling procedure results in a self-weighting sample, since the PPS
method accounts for differences in the number of resident households across
villages (Carletto 1999). After introducing the measures of poverty used, we will
describe the link between poverty and access to financial and natural capital
resources, and investigate farmers’ resilience to shocks, their attitudes towards
risk and their time preferences.
5.3.2 Poverty Measures Used
Two poverty measures are described in this chapter – one absolute and one relative
(cf. Sect. 5.2). The absolute poverty measure builds on an index of household daily
per capita expenditure. For the study, detailed data were collected on farmers’ food
and non-food expenditures in 2007 and in 2010, following the methodology of the
World Bank’s Living Standard Measurement Survey (LSMS), which is described in
detail by Grosh and Glewwe (1998, 2000). As there is a considerable amount of
seasonality in relation to agricultural production and incomes in Yen Chau district,
two expenditure survey rounds were implemented – one between March and April
during the lean season (period before the rice harvest) and the second between
December and January after harvesting of the farmers’ main crop, and using a recall
period of 2 weeks. The final estimate of per-capita daily expenditure was calculated
as an average of the expenditure elicited from the two survey rounds. The level of
expenditure obtained was used as a proxy of farmers’ incomes, and poor households
were identified using the official poverty line set by Vietnam’s Ministry of Labor,
Invalids and Social Affairs (MOLISA) for rural areas
3 (see also Chap. 12). In 2007,
and according to our estimations, 16.9 % of households were living under this line.
In addition to classifying households as poor and non-poor using the official
rural poverty line, we also used a relative poverty measure for some of our analyses,
classifying households into wealth groups based on a linear composite index
constructed using principal component analysis (cf. Dunteman 1994) from a
range of indicator variables capture multiple dimensions of poverty.
2 Except for the villages in four sub-districts bordering Laos, for which research permits are very
difficult to obtain.
3 The poverty line in 2007 was estimated to be 9,105 VND per-capita per day, then in 2010 was
raised to 11,030 VND per capita per day (Van Dinh 2012) (see Chap. 12).
180
C. Saint-Macary et al.