World Food Policy Volume/Issue 2-2/3-1 Fall 2015/Spring 2016 | Page 116

Income Diversity and Poverty Transitions: Evidence from Vietnam

Mehta 2005). In addition, the education
attainment of the household head does
not contribute to the differences in the
probability of being in one or another
poverty trajectory. In fact, the more
the head is educated, the better his/her
access to production resources, labour,
and output markets is, and the more
efficient he/she is in managing household
resources. However, this type of human
capital is more likely to have a long-term
effect on a household’s well-being rather
than on the change in shorter period of
time.

As discussed earlier, the Kinh are
usually more able to access to market
and, hence, take advantage of public
service and the development process,
which allows them to have a higher
probability of being nonpoor, and lower
probabilities of being poor in one or
more periods than their households of
ethnic minority groups (see Table 4).

Household wealth as measured
by the asset index shows quite strong
effects on poverty dynamics. It prevents
households from being poor and is
negatively correlated with falling into
poverty, churning around poverty line, or
being poor. It is also positively correlated
with staying nonpoor and rising out of
poverty (see Table 4). These findings are
in line with the discussion of the role of
assets in the poverty transitions (Carter
and Barrett 2006) as well as with empirical
findings from Bhide and Mehta (2005),
and Imai, Gaiha, and Kang (2011).

There was little evidence of the
difference among households in the
three provinces in the vulnerability
to poverty. TTH is more dynamic in
terms of economic activities owing to
the development of the tourism sector,

and the convenience of transportation.
Therefore, households in the province
have a higher probability of moving out of
poverty and a lower probability of falling
into poverty than their counterparts in
the other two provinces (see Table 4).
Robustness Check

In order to check the robustness
of the MNL model for poverty dynamics,
the study applies to the transitions
of poverty as referred to the poverty
line of $2.5 a day (see Table A.2). The
MNL regression, the results of which
are shown in Table 4, and Table A.2
pass the Hausman tests or suest tests of
independence of irrelevant alternatives
(IIA), which means that assumptions
of IIA could not be rejected; hence,
estimates from MNL models are efficient.
The reference model, in general, shows
similar effects to those in the basic one.
However, there are differences in the size
of the effects in these models compared
to the basic model because poverty
dynamics in the additional model refer
to a higher poverty line. Additionally, the
results from Table 4 are in line with those
from previous studies. The results from
the MNL regression in this study are,
therefore, realizable.

Conclusion

T

his study uses panel data on
households from regions in
Vietnam and a multinomial logit
model to estimate drivers of poverty
transitions. The results show a large share
of the population is vulnerable to poverty
where 38 percent of households have a risk
of being either transient or chronically

116