Access to this full-text is provided by De Gruyter.
Content available from Economics E-Journal
This content is subject to copyright. Terms and conditions apply.
Research Article
Song Jian*
Multidimensional Poverty in Rural China: Human
Capital vs Social Capital
https://doi.org/10.1515/econ-2025-0140
received April 24, 2024; accepted February 13, 2025
Abstract: Occupational stratification is the comprehensive
division and classification of various occupations under-
taken by members of society according to specific stan-
dards and methods. Based on China Family Panel Studies
data, we use the Alkire–Foster method to calculate the rural
multidimensional poverty index and empirically examine
the impact of human capital, social capital, and occupational
stratification on rural multidimensional poverty reduction.
The results show that the improvement of human capital
and social capital can affect the occupational stratification of
rural household members, thereby promoting the growth of
household income and reducing multidimensional poverty
in the household; occupational stratification is an interme-
diator in the poverty reduction effect of human capital and
social capital; compared to social capital, human capital has
a more substantial impact on occupational stratification and
rural multidimensional poverty; human capital has a long-
term dynamic impact on household multidimensional pov-
erty. On the other hand, social capital has a short-term
impact on household multidimensional poverty. At the same
time, occupational stratification has a long-term dynamic
impact on household multidimensional poverty and is also a
long-term poverty reduction mechanism. We delve into the
long-term mechanisms for addressing multidimensional
poverty through the lens of occupational stratification.
Furthermore, we compare the contributions of social
and human capital to occupational stratification and the
reduction of multidimensional poverty in Chinese rural
areas. This analysis enriches the existing literature on
poverty studies.
Keywords: human capital, social capital, occupational stra-
tification, multidimensional poverty
1 Introduction
This article will explore how rural households can achieve
poverty reduction directly by accumulating human and
social capital, but equally how human and social capital
affect career options and how career options affect poverty
in a two-step mechanism. This will be represented by three
equations (one the direct effect on poverty and two repre-
senting the indirect mechanism transmitting social and
human capital through occupational choice to poverty
levels). Measurements will be taken at the household level.
The innovation of this article is we compare social and
human capital’s role in occupational stratification and mul-
tidimensional poverty reduction in Chinese rural areas.
Using Chinese data from 2010 to 2018, we estimate our equa-
tions correcting for endogeneity and find that social capital
contributes very little to either occupational choice or pov-
erty reduction. Rather, human capital composed of both
health quality and educational level drives both direct and
indirect effects. As human capital is highly correlated across
periods, we find a long-term effect as well, although we do
not find a long-term effect of social capital accumulation.
The main deviation of our results from the existing
literature is that we find that social capital plays a rela-
tively small role in rural multidimensional poverty reduc-
tion relative to human capital. Although there are few
articles comparing the importance of human capital and
social capital in poverty reduction, a large number of lit-
eratures show that social capital is an important influen-
cing factor in poverty reduction. We speculate that the
poverty reduction effect of social capital may be related
to marketization and institutional environment. We com-
pared China’s coastal areas with inland areas, large cities,
and small cities and found that the effect of social capital in
coastal areas is weaker than that in inland areas and the
effect of social capital in large cities is weaker than that in
small cities. In addition, we interact social capital with
market-oriented scores and find that the interaction coeffi-
cients are positive, that is, the poverty alleviation effect of
social capital is weaker in provinces with a good market-
oriented environment. We find that the results of the
* Corresponding author: Song Jian, China Center for Special Economic
Zone Research, Shenzhen University, Shenzhen, 518061, China; Shenzhen
Real Estate and Urban Development Research Center, Shenzhen, 518061,
China; Economic School, Nankai University, Weijin Road 94, Tianjin,
300071, China, e-mail: 675542045@qq.com
Economics 2025; 19: 20250140
Open Access. © 2025 the author(s), published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License.
current literature on the significant effect of social capital on
poverty reduction are mostly found in the disadvantaged
areas of developing and developed countries, and combined
with the weaknesses of social capital pointed out in some
literature, our results can actually be coupled with the con-
clusions in the literature, which is a useful supplement to
the research on human capital and social capital.
2 Background
Occupational stratification is the comprehensive division
and classification of various occupations undertaken by
members of society according to specific standards and
methods. It is one of the most critical components of social
stratification research (Treiman, 1977). In the context of
China’s urban labor market, there is a wealth of evidence
indicating significant occupational segregation, which sig-
nificantly contributes to wage disparities among urban
residents and migrant workers. This segregation is fostered
by various institutional arrangements, including the hukou
registration system, enterprise hiring practices, and social
welfare provisions, which collectively give rise to an urban
dual labor market structure. Contrary to Piore’s dual labor
market theory, which attributes market segmentation to
market dynamics (Doeringer & Piore, 1971), China’s labor
market segmentation is predominantly shaped by institu-
tional legacies arising from its economic transition. The
distinctive feature of China’s labor market lies in its insti-
tutional barriers to labor mobility, particularly the con-
straints on rural-to-urban migration imposed by the hukou
system. Migrant workers experience discrimination and occu-
pational segregation within the labor market, which exhibits a
high degree of institutional segmentation. Migrant workers
encounter substantial barriers to urban employment, coupled
with lower incomes, in stark contrast to the more favorable
opportunities and benefits enjoyed by urban residents (Tsui,
2005). State-owned enterprises are known for their provision
of stable employment and higher welfare benefits, whereas
the non-state sector, while offering greater flexibility, is
often less secure (Zhang & Song, 2003). Internal labor mar-
kets, comprising state and large non-state enterprises, have
the autonomy to set labor prices independently of external
market forces (Bai, 2009). This situation is detrimental to the
formation of a harmonious social-occupational hierarchy
and leads to a substantial occupational wage gap among
migrant workers and urban residents (Meng & Zhang,
2001). Moreover, the cost of transitioning between occupa-
tions is substantial (Artuc & McLaren, 2015), and occupa-
tional stratification is constrained by institutional factors
such as career trajectory, occupational identity, and job
availability. Consequently, occupational stratification exerts
asignificant influence on long-term career progression
(King, 2005). Even when individuals switch careers, low-
income workers often gravitate toward new occupations
with lower average wages (Fane et al., 2015). Therefore,
occupational stratification may exert a sustained and long-
term impact on the multidimensional poverty experienced
by families.
The data of the third national agriculture census in
Table 1 showed that in 2016 owners of large farms
1
with
education at junior high school level and above were
65.8%, 9.1 percentage points higher than all farmers,
2
and
the proportion of agriculture firm/organization
3
with a
junior high school education or above is 74.6%, 17.9 per-
centage points higher than that of farmers.
4
The 2016
migrant workers
5
monitoring survey report released by
the National Bureau of Statistics shows that the proportion
of migrant workers with junior high school education or
above is 85.8%, 29.1 percentage points higher than that
of farmers in the same period. Unlike farmers, they are
employed in manufacturing (30.5%), construction
(19.7%), wholesale and retail (12.3%), and resident service
industries (11.1%).
6
With the improvement of rural resi-
dents’human capital, their careers are more diversified
1The standard for large farm in crop cultivation is the land for
growing crops outdoors should reach 100 mu (approximately 6.67
ha) or more. There are 3.98 large farms in 2016.
2Farmers refer to individuals who have been engaged in agricultural
production activities in agricultural farms or agricultural firms. There
are 314.22 million farmers in 2016, and family farms generally have a
small scale, which is related to the historical land reform and the
implementation of the Household Responsibility System with
Contract for Production starting in 1979. In 2016, the average culti-
vated land area per Chinese household was only 0.48 ha.
3The agricultural firm/organizations refer to legal entities primarily
engaged in agricultural production and business activities, and there
are 2.04 million agricultural firm/organizations in 2016.
4The third national agricultural census data fifth brief report. https://
www.stats.gov.cn/sj/tjgb/nypcgb/qgnypcgb/202302/t20230206_
1902105.html.
5In China, due to the constraints of the household registration
(hukou) system, the majority of residents lived in rural areas prior
to 1978. In 1978, out of a population of 963 million, 770 million were
rural residents. With the relaxation of population mobility restric-
tions, a large number of farmers migrated to cities for work, and
the migrated population primarily consisted of rural residents moving
to cities for employment. However, due to limited changes in the
hukou system, these individuals remained registered as rural resi-
dents in terms of household registration, and they are referred to as
“farmer workers”in official Chinese circles and media.
6The 2016 migrant workers monitoring survey report. https://www.
stats.gov.cn/sj/zxfb/202302/t20230203_1899495.html.
2Song Jian
than farmers. In 2016, rural residents’per capita dispo-
sableincomereached12,363yuan,whiletheaverageincome
of migrant workers was 39,300 yuan in the same period,
indicating that with the improvement of rural residents’
human capital, they have more employment choices and
income. The practice has shown that rural laborers with
higher human capital have more careers to choose from;
that is, they have more diverse occupational stratification
(Figure 1).
Since 2013, China has pursued a targeted poverty alle-
viation strategy, initially conducting comprehensive sur-
veys to identify impoverished areas and households, and
establishing a national poverty database for accurate tar-
geting in 2013–2015. Subsequently, a range of measures
including financial aid, infrastructure development, and
industrial growth were implemented to enhance the income
and living conditions of the targeted households in
2016–2018. From 2019 to 2020, support for impoverished
areas was intensified, alongside strengthened monitoring
and evaluation to prevent poverty recurrence. These efforts
have yielded significant results: the poverty headcount ratio
decreased from 10.2% in 2012 to 0.6% in 2019,
7
and by the
end of 2020, all poverty-stricken counties and regions in
China had escaped poverty, achieving the scheduled goal
of poverty eradication. With the effective implementation
of China’s targeted poverty alleviation policy, rural poverty
alleviation has entered the “high-quality poverty allevia-
tion”stage, which requires the Chinese government not
only to win the battle against poverty resolutely but also
to establish a long-term mechanism to solve relative pov-
erty.
8
Xi Jinping emphasized that rural intelligence support
is the core of a long-term poverty reduction mechanism and
that basic educational conditions for weak schools in com-
pulsory education should be continuously improved and
high-quality educational resources should be distributed to
schools in weak areas to promote educational equity and cut
offintergenerational transmission of poverty.
9
Then, under
the background of sustained high growth in national educa-
tion poverty alleviation investment, can the improvement of
human capital obtained by the rural population through
education provide endogenous motivation for solving the
problem of multidimensional poverty?
Table 1: Indicator system and weight setting for multidimensional poverty measurement
Dimension Index Definition of indicators and deprivation Equal weight Entropy weight
Income and consumption Income The per capita net income of the family is lower than the World Bank’s poverty line of $1.9/day,
representing income deprivation
0.0667 0.0702
Consumption Household per capita consumption below $1.25/day represents consumption deprivation 0.0667 0.1324
Engel’s coefficient Engel coefficient above or equal to 60% represents consumption structure deprivation 0.0667 0.0877
Educational opportunities School-age children enroll in
school
School-age children between 7 and 15 years old who are not attending school represent educational
deprivation
0.2 0.1504
Health risks Nutrition Adults with a BMI below 18.5 represent nutritional deprivation 0.1 0.0697
Health function The health function of family members is deprived if they do not seek medical treatment in time when
suffering from serious diseases
0.1 0.1306
Living standard Cooking fuel The use of firewood as fuel represents fuel deprivation 0.0667 0.0330
Clean drinking water Unable to use tap water, mineral water, filtered water, and purified water represents clean drinking
water deprivation
0.0667 0.0317
Housing A family’s average per capita living space is less than 15 m
2
, representing housing deprivation 0.0667 0.0858
Assets Net family assets Following Haveman and Wolff(), household net assets are insufficient to support basic consumption
of $1.25 per person per day for 3 months, representing asset deprivation
0.1 0.1161
Mobile phone The lack of mobile phones among family members represents a deprivation of communication 0.1 0.0924
7english.scio.gov.cn.
8The Decision of the Central Committee of the Communist Party of
China on Some Major Issues Concerning Upholding and Improving the
Socialist System with Chinese Characteristics and Advancing the
Modernization of the National Governance System and Governance
Capacity https://www.gov.cn/zhengce/2019-11/05/content_5449023.htm.
9Xi Jinping’s speech during his inspection visit to Shaanxi. http://cpc.
people.com.cn/big5/n1/2020/0511/c431601-31704374.html.
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 3
Much literature currently focuses on the impact of
human capital accumulation on poverty reduction. For
example, Schultz (1961) used human capital to explain
the sustained growth of national income in the United
States in the first half of the twentieth century and believed
that the lack of human capital and the neglect of human
capital investment in less developed countries were the
reasons for their poverty. Becker (2009) found by com-
paring the return of human capital with that of physical
capital that education can increase personal income and
promote the improvement of other people’s production
efficiency through spillover effects. Sen (1982) analyzed
the causes of poverty and found that people experiencing
poverty are poor because of their insufficient human
capital; that is, the low level of education and health leads
to the deprivation of the ability or opportunity to create
income. In the past 20 years, researchers have applied
random field experiments to anti-poverty research, and
the effects of health and education in developing coun-
tries have become the focus of attention (Banerjee et al.,
2007, 2010; Duflo, 2001) and believe that the impact of
human capital on the income of migrant workers mainly
includes the following aspects: first, the improvement of
human capital can make farmers easier to adopt new
technologies, carry out independent innovation, improve
labor productivity, and thus increase personal income
(López & Valdés, 2000; Parman, 2012). Second, human
capital, including education, vocational training, health,
and working experience, plays a vital role in narrowing
therelativewagegapbyexpandingnon-agricultural
employment opportunities for farmers (Démurger & Li,
2013; Kurosaki & Khan, 2001; Teulings, 2005).
Although human capital is essential for farmers to
overcome poverty, its full play depends on the social envir-
onment. Social capital is one of the sociological concepts
that refers to the social network, trust, and norms that can
improve economic efficiency through coordinated actions
(Putnam, 1993). For individuals, social capital refers to the
ability to acquire and use resources embedded in social
networks in actions (Lin et al., 2001). Social capital has
also been studied in poverty research (Abdul-Hakim
et al., 2010; Grootaert et al., 2002). Social capital is “the
norms and networks that enable people to act collectively,”
which can reduce rural poverty and improve the welfare
of poor households by lowering transaction costs (e.g., trust
can lower the negotiation time and cost) and promoting
cooperation (Woolcock & Narayan, 2000). Much of the lit-
erature discusses the role of social capital in the labor
market, such as facilitating information exchange (Grano-
vetter, 1974; Waldinger, 1996), reducing transaction costs
(Abraham & Medoff, 1983; Eccles & Crane, 1988; Waldinger
& Bozorgmehr, 1996), and improving efficiency (Burt, 1997).
China is a traditional relational society (Bian, 1997; Yang,
1994), and social capital, as an informal institution, has
largely shaped the reciprocal norms of Chinese society
(Hwang, 1987). Social capital helps promote employment
(Montgomery, 1991; Munshi, 2006) and increase income
(Granovetter, 1999; Narayan & Pritchett, 1999). Bian et al.
(2015) demonstrated the causal effect of social capital on
job opportunities and wage income in the labor market
6.40% 3.60% 3.50% 1.00% 1.30% 0.70%
37.00%
30.60% 21.80%
13.20% 16.20% 10.00%
48.40% 55.40%
47.00% 59.40% 58.60%
60.20%
7.10% 8.90%
19.60% 17.00% 16.80% 17.20%
1.20% 1.50% 8.00% 9.40% 7.10% 11.90%
All Large farm
owner
Agriculture
firm/organization
All Local rural
worker
migrant rural
worker
rekrowtnargimremraf
Never attended School Primary school Junior school
High School/vocational school College or more education
Figure 1: Comparison of education of farmer and migrant workers.
4Song Jian
from the perspective of information and preferences. Due
to information transmission (social ties offer inside infor-
mation about jobs and qualifications of job seekers, leading
to desirable job assignments) and resource mobilization
(ties to government officials are the key channels to
acquire economic resources), social capital is essential in
China’s economic and social activities and labor market.
Social capital is an important factor in poverty
research (Knight & Yueh, 2008). Asadi et al. (2008) suggest
that developing countries improve social capital to inte-
grate the environment and people to alleviate poverty
and receive sustainable development. Hassan and Birungi
(2011) access to social capital defined in terms of member-
ship in social organizations positively affects household
income and reduces poverty in Uganda. Saracostti (2007)
found social capital contributes to poverty reduction in
most Latin American countries. Andriani and Karyampas
(2015) social capital contributes to the poverty transition in
Italy. The effect of social capital is also verified in other
countries, such as Indonesia Rustiadi and Nasution (2017)
and Brazil (Marques, 2012). The poverty reduction effects of
social capital are also verified in China. Zhang et al. (2017)
found social capital, such as business ties, political ties, and
appropriable social organizations can contribute signifi-
cantly to poverty reduction in western China. Hong et al.
(2019) social capital shows a significant association with the
absence of poverty in border regions of China. However,
social capital is not necessarily the perfect solution. Neira
et al. (2009) found social capital does not guarantee growth
in itself, but rather serves as a lever on the fulcrum of
social capital to stimulate economic progress and there
may be negative forms of social capital at a given moment
in time in a particular society. Méreiné Berki (2017) finds
bonding ties as tools for everyday survival easily overwrite
system integration efforts for poverty alleviation. Callahan
(2005) relates corruption and vote buying to social capital
in the research on the reform in Thailand.
Most existing literature focuses on the poverty effect of
human capital and social capital accumulation and provides
valuable insights for this study, but it also has some limita-
tions. Wang et al. (2023) examined the role of social capital
on poverty, but without controlling for human capital fac-
tors, which may lead to an overestimation of the poverty
reduction effect of social capital. Fan et al. (2023) simulta-
neously investigated the impact of human capital and social
capital on the re-poverty risk of rural residents, but the
sample was limited to a single county in Henan Province
and did not directly compare the magnitude of the effects of
human and social capital. Emran et al. (2023) explored the
intergenerational impacts of occupational stratification and
human capital but overlooked the effect of social capital. It
needs to include the impact of human capital and social
capital accumulation on career choice and its poverty reduc-
tion effect, which cannot help us fully understand the
relationship between human capital and social capital accu-
mulation and poverty reduction.
After achieving the phased goals of poverty alleviation,
China will focus on solving the problems of relative pov-
erty alleviation, which requires building long-term poverty
reduction mechanisms to achieve short- and long-term
goals. However, the current dilemma confronting poor
rural households in China is that their family members
need higher vocational and technical skills, while simulta-
neously facing employment discrimination in the non-agri-
cultural labor market. This restricts their career options
significantly. Notably, existing research has insufficiently
emphasized the role of career choice.
3 Multidimensional Poverty State
Identification and Poverty
Measurement
3.1 Measurement of Multidimensional
Poverty
There have been many definitions of poverty. Oppenheim
and Harker (1996)defined poverty based on the perspec-
tive of “lack”; the World Bank (2001) defined poverty based
on the view of “social exclusion”; Sen (2006) defined pov-
erty based on the lack of feasible capabilities. This article’s
poverty measurement is mainly based on the perspective
of “minimum living standard.”Combined with this conno-
tation of poverty, Alkire and Foster (2011) proposed a
multidimensional poverty measurement method, the Alkir-
e–Foster (AF) method. This article will use this method to
measure the multidimensional poverty of Chinese rural
residents.
Firstly, following Alkire and Shen (2017) and com-
bining the availability of variables in the 5 years from
2010 to 2018 in the China Family Panel Studies (CFPS)
data, we set the indicators of rural multidimensional pov-
erty as five dimensions: income and consumption, educa-
tion, health, living standards, and assets. Each dimension
has 1–3 deprivation indicators, and the number of indica-
tors under dimension jis labeled d
j
. The indicator system
and weight setting are shown in Table 1.
In the identification of multidimensional poverty and
the calculation of the multidimensional poverty index, we
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 5
denote the dimension of multidimensional poverty as j=1,
2, …, 5, and the corresponding deprivation status of vari-
able kunder dimension jas α
jk
, where α
jk
=1 represents the
deprivation status of the household at this level, and α
jk
=0
represents the non-deprivation status of the household at
this level. Then, the deprivation status of different indica-
tors under different dimensions is summed up according to
their weights, and the comprehensive deprivation index of
household his obtained:
∑∑
===
s
αw
.
h
jk
d
jk
hjk
1
5
1
j
(1)
Since the deprivation index reflects the overall level of
deprivation in different dimensions of a household, we can
use this indicator to identify the multidimensional poverty
status of households. Specifically, we consider households
with a deprivation index above a certain threshold to be in
a state of multidimensional poverty. Following the World
Bank’s approach, we set the threshold at 0.3, which means
the definition of multidimensional poverty status for
households is as follows:
()=>Is
m
poverty 0.3 ,
hh
(2)
where I() is an indicator function, taking the value 1 if the
inequality in the parentheses holds and 0 otherwise.
During the multidimensional poverty status identification,
weight changes will have unpredictable effects on the pov-
erty status. Therefore, to ensure the robustness of the
empirical results, we have set two types of weights for
the indicator system, resulting in two types of multidimen-
sional poverty status and poverty index measurement.
Method 1 is the default equal weight method in the AF
method, in which the weight of each dimension is 0.2,
and the weight of each indicator under each dimension
is equal. Method 2 is the entropy weight method,
10
which
calculates the weight based on the change in data, with
larger weights for indicators with greater dispersion.
Compared to the equal weighting method, the entropy
weight method utilizes the information contained within
the data, representing a more effective approach to lever-
aging information for assigning weights. In Table 1, the
entropy weighting allocates more weight to income and
consumption, which aligns with our intuitive under-
standing of poverty. Therefore, we use the multidimen-
sional poverty status of the entropy weighting as the
main explanatory variable, while we use the multidimen-
sional poverty status of equal weighting in robustness
checks.
In addition, to compare the results of multidimen-
sional poverty with those of income poverty and conduct
a robustness test on multidimensional poverty, we defined
three income poverty statuses, represented by poverty1,
poverty2, and poverty3. Among them, poverty1 refers to
the income poverty status identified using the World
Bank’s low standard (1.9 US dollars/day) as the poverty
line, poverty2 refers to the income poverty status identified
using the World Bank’s high standard (3.1 US dollars/day)
11
as the poverty line, and poverty3 refers to the income pov-
erty status identified using China’sofficial poverty stan-
dard as the poverty line.
Based on the multidimensional poverty status of
families, a group’s multidimensional poverty headcount
ratio can be calculated. That is, for a group with nfamilies,
the headcount ratio of multidimensional poverty in a
group can be represented by the Hindex provided by
the AF algorithm:
∑
==
H
n
1mpoverty
.
h
nh
1
(3)
However, this indicator does not meet the strict mono-
tonicity requirement for poverty measurement indicators.
For example, for a family h that is already in poverty, if the
family’s situation deteriorates and its deprivation index
increases, the Hindex will not increase. Therefore, the H
index can only measure the proportion of households in a
group in multidimensional poverty and cannot reflect the
severity of multidimensional poverty. Thus, the AF algo-
rithm also provides a similar index to measure poverty
depth:
∑∑
===
Asmpoverty / mpoverty
.
h
nhh
h
nh
11
(4)
In this way, the Hindex can be adjusted to satisfy the
strict monotonicity of poverty measurement, namely the
MHA index:
∑
==
=
ns
M
HA HA 1mpoverty
.
h
nhh
1
(5)
10 The fundamental concept of the entropy weighting method is to
utilize the entropy of information to determine the weights of various
indicators. Entropy is a physical quantity that measures the degree of
disorder in a system. In the context of the entropy weighting method,
the greater the information entropy of an indicator, the less informa-
tion it provides, and thus its weight is smaller. Conversely, when the
information entropy is smaller, it indicates that the indicator provides
more information, and its weight is correspondingly greater.
11 https://www.worldbank.org/en/topic/poverty/brief/global-poverty-
line-faq.
6Song Jian
This index increases with the proportion of multidi-
mensional poor households and the degree of deprivation
of poor households.
3.2 Data Sources
We use data from the CFPS from 2010 to 2018 for empirical
analysis. CFPS data provides information on household
income and expenditure, as well as personal education
and work information, which can effectively identify the
education information of rural residents. In the individual
samples, CFPS data provides detailed job information,
which helps identify career choices. Among them, the
CFPS data has detailed occupation codes and can calculate
and match the occupational prestige scores (ISEI) for each
Occupation, which facilitates the construction of measure-
ment for occupational stratification. In addition, CFPS pro-
vides rich information on rural household income, assets,
consumption, and living conditions, and its tracking data
advantage can reflect the dynamic changes of multidimen-
sional poverty in rural China. The above characteristics
provide data support for the empirical analysis. The
sample size of each year in the original data of the CFPS
is as follows: 14,798 households in 2010, 13,315 households
in 2012, 13,946 households in 2014, 14,019 households in
2016, and 14,241 households in 2018. This article studies
the multidimensional poverty of rural families, so only
the sample of rural residents is used. We define multidi-
mensional poverty status and social capital on the house-
hold level. For a family’s per capita human capital and
occupational stratification, we selected family members
aged 16–60. We averaged their human capital and occupa-
tional stratification to obtain the family’s human capital
and occupational stratification. While processing the
CFPS data, we calculated the multidimensional poverty
status, human capital, social capital, occupational stratifi-
cation indicators, and various control variables required in
the regression. After deleting samples with missing values
in income and consumption, education, health, living stan-
dards, and assets variables, the final sample contains 27,594
households. However, when conducting regression analysis
on the binary variable of household multidimensional pov-
erty status, we use the conditional logit model (clogit) with
year-fixed effects to control for household fixed effects.
Because the clogit model cannot handle samples whose status
of the dependent variable does not change across periods, the
actual samples used in the regression are those whose pov-
erty status changes across periods. The total sample size used
in the regression is 11,306.
3.3 Poverty Measurement Results
Using the AF method, the results of multidimensional pov-
erty measurement with two types of weights are shown in
Tables 2–4. The figure in each table indicates that the level
of multidimensional poverty in rural areas varies signifi-
cantly under different weights. Still, the degree of multi-
dimensional poverty under different weights decreases
over time. Table 2 shows that from 2010 to 2018, the multi-
dimensional poverty index MHA under the two weightings
decreased by 0.95 (MHA decreased from 9.2 in 2010 to 1.6 in
2018) and 0.575% annually, indicating that China’s poverty
reduction policies have achieved significant success. Com-
paring Tables 3 and 4 shows that the measurement of dif-
ferent weights shows significant differences in the Head
Count ratio of multidimensional poverty H and the depth
of poverty A. The incidence of poverty has been reduced.
From 2010 to 2018, the average Head Count ratio of poverty
decreased by 2.413 and 1.375 percentage points yearly.
Table 2: Rural poverty measurement MHA (%)
Years Equal weight Entropy weight
2010 9.2 [8.7, 9.6] 5.3 [5.0, 5.7]
2012 6.0 [5.6, 6.4] 3.2 [2.9, 3.5]
2014 3.5 [3.3, 3.8] 1.8 [1.6, 2.0]
2016 1.6 [1.4, 1.8] 0.9 [0.7, 1.0]
2018 1.6 [1.4, 1.8] 0.7 [0.6, 0.9]
Table 4: Rural poverty measurement A(%)
Years Equal weight Entropy weight
2010 38.1 [37.6, 38.6] 41.1 [40.4, 41.8]
2012 36.2 [35.7, 36.8] 38.2 [37.5, 38.9]
2014 35.6 [35.0, 36.2] 38.1 [37.1, 39.0]
2016 35.0 [34.3, 35.8] 36.3 [35.2, 37.5]
2018 35.0 [34.2, 35.7] 36.1 [34.9, 37.3]
Table 3: Rural poverty measurement H(%)
Years Multidimensional poverty Income poverty
Equal weight Entropy weight 1 2 3
2010 24.0 [23.0, 25.1] 13.0 [12.1, 13.8] 31.6 47.8 30.5
2012 16.6 [15.6, 17.6] 8.5 [7.7, 9.3] 26.9 38.4 26.2
2014 10.0 [9.2, 10.7] 4.7 [4.1, 5.2] 19.8 28.7 18.8
2016 4.6 [4.1, 5.2] 2.4 [2.0, 2.8] 10.6 28.7 10.8
2018 4.7 [4.1, 5.2] 2.0 [1.6, 2.4] 7.5 15.6 9.4
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 7
Further, we used different poverty lines to compare the
Head Count ratio of multidimensional poverty with income
poverty. We found that the Head Count ratio of multidi-
mensional poverty was lower than the income poverty in
all three categories, and compared with the equal weight
method, the entropy weight method calculated multidi-
mensional poverty Head Count ratio is relatively low.
The above result may be because when using the entropy
weight method to calculate the deprivation index s, greater
weight is assigned to income, education, and health.
In this article, when empirically testing the relation-
ship between household human capital, social capital,
occupational stratification, and multidimensional poverty,
we mainly use the multidimensional poverty status mpov-
erty2 identified using the entropy weight while using the
poverty status mpoverty1 identified using the equal weight
and the income poverty poverty1, poverty2, poverty3 for
robustness testing, to verify the robustness of the conclu-
sions of this article for poverty identification methods.
4 Empirical Model and Variable
Selection
4.1 Empirical Model
The key question of this study is whether human capital
and social capital affect rural poverty reduction through its
impact on occupational stratification. We use a classic
three-step regression (Baron & Kenny, 1986) to test the
mediating effect of occupational stratification in the rela-
tionship between human capital, social capital, and pov-
erty. The three-step method for discussing mediating
effects involves three models, represented by equations
(6)–(8). In equation (6), the coefficients β
11
and β
12
repre-
sent the total effect of human capital and social capital on
poverty. In equation (7), β
21
and β
22
represent the effect of
human capital and social capital on the mediating variable
occupational stratification.Inequation(8),β
33
represents the
effect of occupational stratification on poverty, controlling for
the influence of human capital and social capital. Under the
path of “human capital and social capital →occupational
stratification →poverty,”the mediating effect is estimated
as the product of the path coefficients in the first half (β
11
,
β
12
)andthesecondhalf(β
33
) of the mediating path.
∑=+ + + ++
+
ββhβ cX μν
ε
mpoverty sc
,
it it it k k it i
t
it
10 11 12 1, 11
1
(6)
∑=+ + + ++
+
ββhβ cX μν
ε
career sc
,
it it it k k it i
t
it
20 21 22 2, 22
2
(7)
∑
=+ + +
++++
ββhβ β
cX μ ν ε
mpoverty sc career
,
it it it i
t
kkit itit
30 31 32 32
3, 333
(8)
where mpoverty is the multidimensional poverty status of
the household; career is the occupational stratification of
the household, including the proportion of migrant
workers in the household (mig_rate), average occupational
prestige in the household (isei_mean); h is the per capita
human capital, including knowledge capital and health
capital, including per capita years of education and health
level; sc is the social capital of the household; Xrepresents
control variables such as household characteristics and
community characteristics; μ
i
and ν
t
represent household
fixed effects and year fixed effects, respectively, and ε
represents the error term. Since the multidimensional pov-
erty status of households is a 0–1 variable, we use the
conditional logit model (clogit) to control for households.
4.2 Variable Selection
4.2.1 Human Capital and Social Capital
Human capital is the knowledge, abilities, skills, and
experiences in economic activities. Knowledge and health
are the two most concerned aspects in the study of human
capital. For knowledge, we use the years of education of
the household labor force to represent the household labor
force’s human capital level. For health indicators, we use
the self-reported health of the respondents in the CFPS
personal questionnaire as a proxy variable for the health
capital of the household labor force. Correspondingly, at
the household level, we average the years of education of
the household labor force within the household and use
the per capita years of schooling (feduuyear) to represent
the level of knowledge and human capital of the house-
hold. At the same time, the health status of the household’s
labor force is averaged within the household, with the per
capita health (fhealth) representing the level of health
human capital in a household. We measure social capital
at the household level. Social networks are an essential com-
ponent of social capital, including the size and density of
social or relational networks (Burt, 1997; Wasserman & Faust,
1994). We use a total indicator that comprehensively reflects
the size and density of social or relational networks: the total
amount of cash gifts received and paid by a household to
measure the social capital of the household. Besides, we use
8Song Jian
a binary indicating related to party member of local manager
as another proxy for social capital.
Due to the two-way causality between human capital
and household poverty status and the possible simulta-
neous influence of certain omitted variables on household
human capital and household poverty status, endogenous
problems may cause estimation bias. To address this issue,
we construct instrumental variables for household human
capital to correct the estimation bias caused by endogeneity.
First, we use the implementation of the Compulsory
Education Law to construct an instrumental variable for
the number of years of education received. The logic of using
this event to construct the instrumental variable is that it
satisfies the conditions of being correlated with the endo-
genous variable and is uncorrelated with the error term.
First, implementing the Compulsory Education Law can affect
the years of education of family members. Second, as an
exogenous event, implementing the Compulsory Education
Law does not affect family occupational stratification and
income in other ways. The construction of the instrumental
variable is as follows: first, we use the year in which each
province implemented the Implementation Rules of
Compulsory Education Law as the year in which the province
implemented the Compulsory Education Law. When the pro-
vinces began to implement the Compulsory Education Law, if
the age of the family member was greater than 16, it was
considered that they had completed compulsory education
and their education level was not affected by the
Compulsory Education Law. If the age was less than 16 at the
time, we assume their education was affected by the imple-
mentation of the Compulsory Education Law. We regress the
years of schooling of family members on whether they were
affected by the Compulsory Education Law and average the
resulting fitted values across family labor to obtain the house-
hold’spercapitayearsofeducation.Second,forthehealth
capital of family members, we use the average health status
of family members evaluated by interviewers in the question-
naire as the instrumental variable of family members’health
capital. Finally, like human capital, there may be two-way
causality and omitted variable problems between social capital
and household poverty status. Following Bentolila et al. (2010),
we construct the average social capital of all households in the
same community except for the household as the instrumental
variable of the household’s social capital.
4.2.2 Occupational Stratification
Previous literature has discussed occupational stratifica-
tion from the perspective of occupational prestige or the
economic status of occupations. The essential criteria for
determining occupational stratification are multidimen-
sional and subjective to some extent. The ranking order
of different occupations not only stems from corre-
sponding educational requirements and income mobility
but also depends on the general evaluation of relative
occupational status (occupational prestige) and the social
attitudes towards the occupation. There are various views
on the classification of occupational prestige. For example,
Treiman (1977) used empirical data from 60 countries and
cultural groups, including agricultural and post-industrial
societies, to construct the Standard International Occupa-
tional Prestige Scale by averaging the prestige scores of
these individuals. Ganzeboom (2008) proposed a hierarch-
ical model of occupational prestige based on the social
structure of a country and constructed the International
Socioeconomic Index (ISEI). Since the CFPS data provide
household members’occupation codes and ISEI, we use
the ISEI to measure occupational stratification. Many stu-
dies have shown that migrant workers improve rural labor
skills and income (Du et al., 2005; Liu, 2008). Therefore, we
use whether or not the household labor force migrates as
an important aspect to reflect occupational stratification.
In addition to the explanatory variables, this article
adds other control variables, including household and
community characteristics. Household characteristics
mainly include family size (familysize) and dependency
ratio (dependrate). In estimating the impact of human
and social capital on occupational stratification and multi-
dimensional poverty, it is also necessary to control the
overall human capital status of the community to control
the spillover effect of human capital. We use the village per
capita education years (ceduyear) as a proxy for the com-
munity’s overall human capital. In addition, due to specific
trends and fluctuations in economic development in dif-
ferent years, we need to control the fixed effects of the
year. Variable names, definitions, and statistical descrip-
tions are shown in Table 5.
5 Estimation Results
5.1 Human Capital, Social Capital,
Occupational Stratification, and
Multidimensional Poverty Reduction:
Benchmark Model
This section uses econometric models (6)–(8) to examine
the impact of human capital, social capital, and occupa-
tional stratification on rural poverty reduction and the
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 9
Table 5: Variable definition and summary statistics
Variables Variable definition Obs Mean Std. error Min Max
Poverty poverty 1 Poverty status defined according to the World Bank’s low standard (US$1.9/day) poverty line 33,179 0.195 0.396 0 1
poverty 2 Poverty status defined according to the World Bank’s high-standard (US$3.1/day) poverty line 33,179 0.312 0.463 0 1
poverty 3 Poverty status based on the Chinese official poverty line 33,179 0.202 0.401 0 1
mpoverty1 Multidimensional poverty status identified using the AK method (equal weight) 28,836 0.266 0.442 0 1
mpoverty2 Multidimensional poverty status identified using the AK method (entropy weight) 28,836 0.077 0.266 0 1
Human capital fhealth Household average health level 29,310 3.325 1.040 1 5
feduyear Household average years of education 28,690 6.655 3.563 0 19
feduex Household average years of education based on the impact of the implementation of the Compulsory
Education Law
35,855 5.848 1.246 4.56 8.85
fhealth_ev The average health status of the household evaluated by the interviewer 32,737 5.175 1.162 1 7
Social capital gc The total amount of cash gifts received and paid by a household 36,238 3.487 3.907 0 13.710
pc Related to party member of local manager 36,238 0.121 0.543 0 1
gc_n Average gift change of all households in the same community except for the household 36,238 3.542 1.545 1.434 12.231
pc_n Average political relation of all households in the same community 36,238 0.123 0.276 0.004 0.643
Occupational stratification mig_rate Percentage of household labor force who work outside of their resident county 36,238 0.115 0.241 0 1
isei_mean Family.Occupation per labor reputation score 36,238 12.068 10.293 0 88
Family and community
characteristic
Familysize Number of family members 35,421 4.108 1.909 1 26
dependrate Dependency ratio 35,065 0.376 0.328 0 1
ceduyear The average years of education per capita in the village 36,105 5.298 1.567 0 17.5
10 Song Jian
mediating impact mechanism. The results are shown in
Table 6.
Columns (1) and (2) use occupational stratification as
the explained variable to verify human and social capital’s
impact on occupational stratification. The results show that
both human capital and social capital have a significant
promoting effect on the occupational stratification of
family members. Among them, improving human and
social capital has encouraged rural residents to migrate
for work and significantly enhanced the occupational pres-
tige of family members. This result means that improving
human and social capital will benefit the occupational stra-
tification of rural family members, consistent with mul-
tiple studies, such as Bian et al. (2015), which demonstrate
that social capital provides individuals with more job
opportunities in the labor market.
Column (3) in Table 6 studies the impact of human and
social capital on multidimensional poverty in rural house-
holds. The results show that the coefficients of human
capital, including health and knowledge, are significantly
positive at the 1% level, indicating that human capital
significantly reduces the probability of multidimensional
poverty in rural households, similar to the findings of Chen
and Wang (2001), which verify that human capital,
including health, has a positive effect on income, thus
making the accumulation of human capital significantly
reduce the multidimensional poverty of rural households.
Like human capital, the coefficient of social capital is also
significantly positive at the 1% level, indicating that
improving social capital level significantly reduces multi-
dimensional poverty in rural households, consistent with
the results of Bian et al. (2015), which suggest that social
capital provides individuals with more job opportunities in
the labor market, increases the wage income of family
members, and thus reduces the risk of multidimensional
poverty in rural households. But compared to Wang et al.
(2023), our coefficient of social capital is much lower, indi-
cating social capital and human capital are positively cor-
related. After controling human capital, the importance of
social capital has decayed.
Based on column (3), we control the occupational stra-
tification variables in column (4). The results show that the
proportion of migrant workers and the per capita occupa-
tion prestige in the household has a significant multidi-
mensional poverty reduction effect. At the same time, after
controlling for occupational stratification, the impact of
human capital and social capital on multidimensional pov-
erty in rural households decreases, among which the mag-
nitude of the marginal effect of health on multidimensional
poverty in rural households decreases from 0.010 to 0.002
and is no longer significant; the magnitude of knowledge
on multidimensional poverty in rural households
decreases from 0.005 to 0.002; the magnitude of the mar-
ginal effect of gift change on multidimensional poverty in
rural households decreases from 0.003 to 0.002. Following
Preacher and Hayes (2008), we do the Sobel test by calcu-
lating Sobel, Aroian, and Goodman statistics.
12
The results
of the test show that among the effects of human capital
and social capital on household multidimensional poverty,
various Sobel statistics from different channels are signifi-
cant at the 5% level, indicating that occupational stratifica-
tion plays a mediating role in the effects of human capital
Table 6: Human capital, occupational stratification, and rural multidi-
mensional poverty
(1) FE (2) FE (3) clogit (4) clogit
mig_rate isei_mean mpoverty2 mpoverty2
fhealth 0.009*** 0.340*** −0.010*** −0.002
(0.002) (0.070) (0.001) (0.001)
feduyear 0.009*** 0.450*** −0.005*** −0.002***
(0.001) (0.029) (0.000) (0.000)
gc 0.001*** 0.030** −0.003*** −0.002***
(0.000) (0.014) (0.000) (0.000)
pc 0.002*** 0.021** −0.002*** −0.001***
(0.000) (0.000) (0.000) (0.000)
mig_rate −0.032***
(0.007)
isei_mean −0.002***
(0.000)
familysize 0.005*** −0.141*** 0.009*** 0.008***
(0.001) (0.050) (0.001) (0.001)
dependrate 0.024** −3.887*** 0.065*** 0.061***
(0.010) (0.326) (0.006) (0.007)
ceduyear −0.003 0.455*** −0.017*** −0.016***
(0.002) (0.075) (0.001) (0.001)
constant 0.024 8.614***
(0.015) (0.506)
N27,594 27,594 11,306 11,306
R
2
0.433 0.571
Ll −1,799.454 −1,887.765
Note: The reported values in the table are marginal effects, and the
numbers in parentheses are standard errors. ***, **, and * indicate
significance at the 1, 5, and 10% levels, respectively.
12 The Sobel test is a method to verify whether the mediation
mechanism is statistically significant by constructing a Z-statistic
(which follows a standard normal distribution under large sample
sizes) to test the significance of the indirect effect
ab
(i.e., β
21
×β
32
in our case). The form of the test statistic proposed by Sobel is
() ()=+
Z
ab aseb bsea/2222
, while Aroian uses
() () () ()=++
Z
ab aseb bsea sea seb/2222 22
and Goodman uses
() () () ()=+−
Z
ab aseb bsea sea seb/2222 22
, Aroian, and Goodman
statistics.
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 11
and social capital on household multidimensional poverty.
In addition, as shown in Table 6, some household and
community characteristics also significantly impact the
probability of household multidimensional poverty.
Among them, the family dependency ratio significantly
positively impacts household multidimensional poverty.
Dominance analysis, as technically detailed by
Sfakianakis et al. (2021), involves sequentially adding inde-
pendent variables to assess their relative importance while
mitigating the effects of correlations among them, through
an extension of the Shapley decomposition method. We
have utilized this approach on the models presented in
Table 6 to determine the average contributions of each
variable to the explanation of the dependent variable.
This allows us to evaluate the relative importance of
human and social capital in influencing occupational stra-
tification and multidimensional poverty.
The results of the dominance analysis show that,
among the influences of human capital and social capital
on occupational stratification, the importance of human
capital and social capital for the proportion of migrant
workers in the family labor force (mig_rate) are 95.01
and 4.99%, respectively; Among the influences of human
capital and social capital on occupational prestige, the
importance of human capital and social capital are 98.21
and 1.79%, respectively. Therefore, for occupational strati-
fication, human capital is the most critical factor, indi-
cating that human capital is far more important than social
capital in determining occupational stratification, and the
impact of social capital on occupational stratification is
relatively weak. For multidimensional poverty, the impor-
tance of human capital and social capital is 98.00 and
2.00%, respectively, indicating that for the impact of multi-
dimensional poverty in rural areas, the importance of
human capital is far greater than social capital, and the
impact of social capital on multidimensional poverty in
rural areas is relatively weak. Using column (4) in Table
7for dominance analysis, the results show that for multi-
dimensional poverty, the importance of human capital,
social capital, and occupational stratification are 53.06,
2.13, and 44.81%, respectively, indicating that considering
occupational stratification, the importance of human and
social capital on multidimensional poverty decreases, indi-
cating that human capital and social capital reduce
Table 7: Robustness test using instrumental variables
(1) IV-FE (2) IV-FE (3) IVprobit (4) IVprobit
mig_rate isei_mean mpoverty2 mpoverty2
fhealth 0.009*** 0.321*** −0.457*** −0.488***
(0.001) (0.065) (0.076) (0.042)
feduyear 0.006*** 0.487*** −0.132*** −0.145***
(0.002) (0.021) (0.003) (0.005)
gc 0.001** 0.032** −0.020*** −0.012***
(0.000) (0.032) (0.004) (0.001)
pc 0.000** 0.030** −0.030*** −0.010***
(0.000) (0.031) (0.002) (0.000)
mig_rate −0.201***
(0.000)
isei_mea −0.009***
(0.001)
familysize 0.004*** −0.121** 0.110*** 0.095***
(0.000) (0.050) (0.010) (0.009)
dependrate 0.021*** −3.780*** 0.684*** 0.532***
(0.010) (0.300) (0.071) (0.045)
ceduyear −0.000 0.433*** −0.132*** −0.160***
(0.001) (0.002) (0.011) (0.011)
N26,576 26,576 11,306 11,306
R
2
0.011 0.032
Endogeneity test 0.986 1.564 5.432 7.543
Underidentification test 44.990 76.454
Weak IV test 18.543 19.743 343.462 358.321
Note: In (1) and (2), the endogeneity test reports a C-statistic, the underidentification test reports Anderson canon. corr. LM statistic, and the weak IV
test reports Cragg-Donald Wald Fstatistic. In (3) and (4), the endogeneity test reports the Wald test of exogeneity, and the weak IV test reports the
Anderson-Rubin (AR) test statistic. ***, **, and * indicate significance at the 1, 5, and 10% levels, respectively.
12 Song Jian
multidimensional poverty in rural households by affecting
the occupational stratification of rural household
members.
5.2 Robustness Test
Table 6 examines the relationship between human
capital, occupational stratification, and multidimensional
poverty. This section mainly conducts a robustness test on
the above results from two aspects. First, we address the
endogeneity between human capital and social capital,
occupational stratification, and household income using
instrumental variables of human capital and social
capital. Second, we replace the measure of poverty,
including using income-defined poverty status and multi-
dimensional poverty status constructed using the equal-
weighting method.
First, we used the instrumental variable of human
capital and social capital described in Section 4 to per-
form instrumental variable regression on each column in
Table 7. For columns (1) and (2), we used panel GMM
estimation, and for columns (3) and (4), we used an IV-
probit model with household fixed effects and calculated
the marginal effects of each coefficient. Due to the insuffi-
cient development of instrumental variable methods for
logit models, we use the instrumental variable method of
the probit model to address endogeneity issues. Since all
binary models report marginal effects, the results of the
probit model and the logit model are comparable. The
results in Table 8 are consistent with those in Table 7,
indicating that both human and social capital signifi-
cantly improve the occupational stratification of family
members, thereby reducing the probability of multidi-
mensional poverty in the family. It is worth noting that
after using instrumental variables to avoid endogenous
bias, the magnitude of human capital, social capital, and
occupational stratification’scoefficients in each regres-
sion significantly increased, possibly due to local treat-
ment effects. In addition, the underidentification and
weak identification test showed that the instrumental
variables strongly correlated with the endogenous vari-
ables. The endogenous test results showed that there were
endogenous problems in the regression. The above results
indicate that the selection of instrumental variables is
reasonable.
The three income poverty states poverty1, poverty2,
and poverty3, as well as the equal-weighted multidimen-
sional poverty state mpoverty1, were used to clogit estima-
tion and IV estimation on models (7) and (8) for columns
(1)–(12) in Table 8, and we did Sobel tests. The results were
consistent with the results of the benchmark model. Both
human capital and occupational stratification had signifi-
cant poverty reduction effects, and occupational stratifica-
tion was a mediator in the impact of human capital on
poverty.
5.3 Compare the Effect of Social and Human
Capital in Different Areas
The main deviation of our results from the existing litera-
ture is that we find that social capital plays a relatively
small role in rural multidimensional poverty reduction
relative to human capital. We compared China’s coastal
areas with inland areas, large cities, and small cities, and
found that the effect of social capital in coastal areas is
weaker than that in inland areas, and the effect of social
capital in large cities is weaker than that in small cities. In
addition, we interact social capital with market-oriented
scores and find that the interaction coefficients are posi-
tive, that is, the poverty alleviation effect of social capital is
weaker in provinces with a good market-oriented environ-
ment (Table 9).
5.4 Analysis of the Long-Term Impact of
Human and Social Capital and
Occupational Stratification on
Multidimensional Poverty in Rural Areas
Establishing a long-term mechanism for sustainable pov-
erty alleviation is a crucial task of poverty alleviation in the
post-2020 period.
13
Due to the multidimensional poverty
reduction effects of human capital and social capital
through occupational stratification in rural households,
we need to examine the long-term dynamics of human
capital, social capital, and occupational stratification on
multidimensional poverty to explore the long-term
mechanism of multidimensional poverty reduction in
rural areas. We examine the correlation between human
capital, social capital, and occupational stratification in
different periods. Table 10 shows that the correlation
between different periods in variables of human and social
capital, and occupational stratification was positive, while
13 http://www.xinhuanet.com/politics/2019-11/15/c_1125236445.htm.
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 13
education exhibits the highest intertemporal correlation.
The above results indicate that human and social capital
and occupational stratification have a certain persistence
in the long run. As human capital has significant poverty
reduction effects, sustained human capital investment and
long-term human capital accumulation have important
implications for household poverty. Current household
human capital may affect future household human capital
accumulation through education investment, thereby
affecting poverty’s long-term persistence. Social capital
and occupational stratification may also have similar
long-term mechanisms.
In Table 11, we added the lag of human capital, social
capital, and occupational stratification to the regression of
household multidimensional poverty status to examine the
long-term impact of human capital, social capital, and
occupational stratification on household multidimensional
poverty. In column (1), we added the lag of human capital
and social capital (2 years ago), and the results showed that
human capital had a significant negative impact on the
current household multidimensional poverty status, indi-
cating that human capital has a long-term dynamic impact
on household multidimensional poverty. On the other
hand, the lagged value of social capital had no significant
impact on the household multidimensional poverty status,
indicating that social capital has a short-term impact on
household multidimensional poverty. In column (2), we
added the lagged value of human capital and social capital
(2 and 4 years ago), which once again confirmed the results
of column (1). The coefficients of the lagged value of human
capital were significantly negative, while the coefficients of
the lag of social capital were not significant. Social capital
Table 8: Robust test using substitution variable and instrumental variable
(1) clogit (2) clogit (3) clogit (4) clogit (5) clogit (6) clogit (7) clogit (8) clogit
poverty1 poverty1 poverty2 poverty2 poverty3 poverty3 mpoverty1 mpoverty1
fhealth −0.018*** −0.003 −0.032*** −0.005* −0.018*** −0.004* −0.017*** −0.002
(0.002) (0.002) (0.003) (0.002) (0.002) (0.002) (0.002) (0.002)
feduyear −0.012*** −0.005*** −0.018*** −0.005*** −0.013*** −0.005*** −0.008*** −0.003***
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
gc −0.009*** −0.007*** −0.009*** −0.006*** −0.009*** −0.007*** −0.005*** −0.003***
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
pc −0.003*** −0.002*** −0.003*** −0.002*** −0.003*** −0.002*** −0.003*** −0.002***
(0.000) (0.000) (0.000) (0.000) (0.001) (0.000) (0.000) (0.000)
mig_rate −0.281*** −0.315*** −0.273*** −0.053***
(0.013) (0.013) (0.013) (0.009)
isei_mean −0.006*** −0.008*** −0.006*** −0.002***
(0.000) (0.000) (0.000) (0.000)
N11,306 11,306 11,306 11,306 11,306 11,306 11,306 11,306
(9) IVprobit (10) IVprobit (11) IVprobit (12) IVprobit (13) IVprobit (14) IVprobit (15) IVprobit (16) IVprobit
poverty1 poverty1 poverty2 poverty2 poverty3 poverty3 mpoverty1 mpoverty1
fhealth −0.170*** −0.103*** −0.156*** −0.117*** −0.193*** −0.100*** −0.372*** −0.348***
(0.041) (0.046) (0.039) (0.041) (0.041) (0.045) (0.051) (0.045)
feduyear −0.190*** −0.106*** −0.191*** −0.100*** −0.179*** −0.009*** −0.164*** −0.108***
(0.010) (0.012) (0.010) (0.010) (0.011) (0.013) (0.016) (0.013)
gc −0.033*** −0.029*** −0.028*** −0.019*** −0.034*** −0.030*** −0.026*** −0.014***
(0.003) (0.003) (0.002) (0.002) (0.003) (0.003) (0.004) (0.003)
pc −0.022*** −0.019*** −0.024*** −0.018*** −0.023*** −0.019*** −0.019*** −0.017***
(0.002) (0.002) (0.003) (0.003) (0.002) (0.002) (0.003) (0.002)
mig_rate −1.057*** −0.951*** −1.039*** −0.204***
(0.059) (0.049) (0.058) (0.050)
isei_mean −0.024*** −0.025*** −0.025*** −0.007***
(0.002) (0.001) (0.002) (0.002)
N11,306 11,306 11,306 11,306 11,306 11,306 11,306 11,306
Endogeneity test 5.11 6.43 5.63 9.76 4.56 4.98 9.07 8.55
Weak IV test 145.32 165.98 145.11 134.54 143.86 176.87 197.32 189.06
Note: All columns control the fixed effects of family size, depend rate, ceduyear, and family and year, and the rest are the same as above. ***, **, and
* indicate significance at the 1, 5, and 10% levels, respectively.
14 Song Jian
is less sustainable in reducing household multidimensional
poverty than human capital. In columns (3) and (4), we gradu-
ally added the lag of occupational stratification, and the results
showed that the lag of isi_mean in occupational stratification
was significantly negative in both lags of 2 and 4 years ago,
indicating that occupational stratification has a long-term
dynamic impact on household multidimensional poverty.
6 Conclusion
Rural intellectual assistance is vital for China to implement
targeted poverty alleviation strategies and strategies to
address long-term relative poverty. Based on the CFPS
data, this article uses the AF method to calculate the rural
multidimensional poverty index, empirically examines the
impact of human capital and social capital and occupa-
tional stratification on rural multidimensional poverty
reduction, and discusses the long-term characteristics of
the above impacts. The main conclusions are as follows:
the improvement of human capital level and social capital
can affect the occupational stratification of rural house-
hold members, thereby promoting the growth of house-
hold income and reducing the probability of multidimen-
sional poverty in the household; occupational stratification
is the intermediator in the poverty alleviation effect of
human and social capital; compared to social capital,
human capital has a more substantial impact on occupa-
tional stratification and rural multidimensional poverty;
human capital has a long-term dynamic impact on house-
hold multidimensional poverty, which is a long-term pov-
erty alleviation mechanism. On the other hand, social
capital has a short-term impact on household multidimen-
sional poverty. At the same time, occupational stratification
has a long-term dynamic impact on household multidimen-
sional poverty, which is also a long-term poverty alleviation
mechanism.
Table 9: Compare the effect of social and human capital in different areas
Small city Big city Inland areas Coastal areas Marketization
(1) clogit (2) clogit (3) clogit (4) clogit (5) clogit
mpoverty2 mpoverty2 mpoverty2 mpoverty2 mpoverty2
fhealth −0.005*** −0.012*** −0.007*** −0.013*** −0.010***
(0.001) (0.001) (0.001) (0.001) (0.001)
feduyear −0.003*** −0.009*** −0.004*** −0.008*** −0.005***
(0.000) (0.000) (0.000) (0.000) (0.000)
gc −0.003*** −0.001*** −0.004*** −0.002*** −0.003***
(0.000) (0.000) (0.000) (0.000) (0.000)
pc −0.003*** −0.002*** −0.004*** −0.001*** −0.003***
(0.000) (0.000) (0.000) (0.000) (0.000)
gc_mark 0.000*** 0.001*** 0.001*** 0.000*** 0.001***
(0.000) (0.000) (0.000) (0.000) (0.000)
pc_matk 0.001*** 0.000*** 0.001*** 0.000*** 0.001***
(0.000) (0.000) (0.000) (0.000) (0.000)
N3,391 10,967 3,957 7,349 11,306
ll −598.874 −1,283.656 −731.685 −1,097.545 −1,798.454
Note: All columns control the fixed effects of family size, depend rate, ceduyear, and family and year, and the rest are the same as above. ***, **, and
* indicate significance at the 1, 5, and 10% levels, respectively.
Table 10: The correlation between values in 2012–2018 and in 2010 of human capital, social capital, and occupational stratification
Values in 2012–2018 Human capital sc Occupational stratification
Education Health mig_rate isi_mean
2012 0.78 0.31 0.1 0.18 0.42
2014 0.72 0.30 0.1 0.10 0.29
2016 0.62 0.24 0.09 0.11 0.20
2018 0.56 0.22 0.07 0.10 0.20
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 15
We propose the following suggestions based on the
findings and considering the current situation in China
and the need to address relative poverty in the future.
Our research confirms that human capital accumulation
significantly contributes to multidimensional poverty reduc-
tion in rural areas. Despite the narrowing educational gap
between urban and rural China, disparities in educational
development persist, with rural education quality remaining
a concern. For instance, in 2021, the national average teacher-
to-class ratio in primary schools was 2.02:1, but rural areas
only had 1.88:1, indicating a shortage of full-time teachers
(Guo & Li, 2024). Additionally, the aging of rural teachers is
a serious issue, with the proportion of teachers over 55 years
old being significantly higher in rural areas than in urban
areas (Shi & Sercombe, 2020). Additionally, the growth rate of
investment in rural compulsory education is below the
national average. Revitalizing rural education requires
addressing these challenges. Specificmeasuresinclude:
strengthening the rural teacher workforce through profes-
sional development and improved compensation (Li et al.
2023); promoting curriculum and teaching reforms in rural
schools, and facilitating partnerships with urban schools to
share resources; enhancing rural school infrastructure and
investing in educational informatization; developing teaching
materials and resources tailored to rural students’needs in
collaboration with educational publishers and higher educa-
tion institutions; and establishing a diverse curriculum
system that includes courses related to rural career develop-
ment. The government should encourage social participation
in rural education and form a collaborative investment
mechanism. These measures will help cultivate rural talents,
promote rural development, and support long-term poverty
reduction efforts.
The 50–50 division in high school admissions has exa-
cerbated the educational divide between urban and rural
regions, a consequence of the initial disparities in educa-
tional quality. Since 2016, the Ministry of Education has
mandated that enrollment in general and vocational edu-
cation at the high school level should be approximately
equal.
14
This policy, in effect, results in a 50–50 split in
high school admissions, which exacerbates the educational
quality gap between urban and rural areas, leading to a
higher proportion of rural students being directed into
vocational education. However, this policy, coupled with
the urban-rural educational quality gap, directs more rural
students into vocational education, depriving them of edu-
cational choices and causing widespread anxiety. This
aligns with Johannes Giesinger’s assertion that “it is unrea-
sonable to prematurely close the doors of future choices
for children without necessity. what we need is to promote
the development of students, not to simply judge their talents
and abilities (Giesinger, 2017).”Therefore, it is recommended
that the restrictions on high school admissions be relaxed,
and enrollment quotas be determined based on actual
teaching resources and demand. Additionally, vocational edu-
cation grapples with challenges such as low social recogni-
tion, parental and student resistance, misalignment with
industry needs, talent–employer mismatch, and graduate dis-
satisfaction (Chen, 2020). Insufficient infrastructure, a weak
teaching staff, and funding disparities further impede its pro-
gress. To bolster vocational education, it is crucial to legisla-
tively elevate its status within the national education system,
align professional settings and training with industry
Table 11: Analysis of the long-term impact of human and social capital
and occupational stratification on multidimensional poverty in rural
areas
(1) clogit (2) clogit (3) clogit (4) clogit
mpoverty2 mpoverty2 mpoverty2 mpoverty2
fhealth −0.002 −0.001 −0.001 −0.001
(0.002) (0.002) (0.002) (0.002)
feduyear −0.001* −0.000 −0.001 −0.000
(0.001) (0.001) (0.001) (0.001)
sc −0.002*** −0.002** −0.001*** −0.001*
(0.000) (0.000) (0.000) (0.000)
fhealth
t–2
−0.007*** −0.003* −0.003*** −0.001*
(0.002) (0.002) (0.002) (0.002)
feduyear
t–2
−0.002*** −0.001* −0.001*** −0.001
(0.001) (0.001) (0.001) (0.001)
sc
t–2
−0.001 −0.000 −0.001 −0.000
(0.000) (0.000) (0.000) (0.000)
fhealth
t–4
−0.005*** −0.002***
(0.002) (0.002)
feduyear
t–4
−0.001 −0.001
(0.001) (0.001)
sc
t–4
−0.001 −0.001
(0.000) (0.000)
mig_rate −0.019*** −0.023***
(0.007) (0.008)
isei_mean −0.001** −0.000
(0.000) (0.000)
mig_rate
t–2
−0.003 0.008
(0.007) (0.007)
isei_mean
t–2
−0.001*** −0.001***
(0.000) (0.000)
mig_rate
t–4
−0.005
(0.008)
isei_mean
t–4
−0.000*
(0.000)
N9,674 8,014 9,674 8,014
***, **, and * indicate significance at the 1, 5, and 10% levels,
respectively.
14 https://www.gov.cn/xinwen/2016-09/20/content_5110023.htm.
16 Song Jian
demands (Zhe, 2023), strengthen school–enterprise coopera-
tion, enhance teaching qualityandcareerprospects,and
increase funding and optimize resource allocation, particu-
larly in western regions.
Our research confirms that occupational stratification
significantly contributes to multidimensional poverty
reduction in rural areas. To enhance rural vocational edu-
cation and non-agricultural training, it is crucial to address
issues such as the disconnect between training and practice,
monotonous content, insufficient qualified teachers, and low
education levels among farmers. Strategies include refining
training content and methods to align with practical needs,
improving the training system to increase diversity and effec-
tiveness, strengthening teacher team development with a
focus on practical experience, emphasizing agricultural and
information technology education to uplift farmers’educa-
tion levels, and promoting rural vocational and adult educa-
tion to enhance employability. Implementing a lifelong
vocational skill training system and enhancing employment
guidance and entrepreneurship support are also vital. These
measures will effectively advance rural vocational education,
improve farmers’professional skills and employability, ele-
vate the occupational stratification of rural residents, and
provide a robust talent foundation for sustained rural pov-
erty reduction.
Although social capital reduces the multidimensional
poverty of households through occupational stratification,
some literature (Bentolila et al., 2010) also points out that
social capital may cause negative externalities by dis-
torting the labor market and reducing overall productivity
besides its benefit to individuals’career choices. The dis-
tribution and role of social capital widen the income gap
among rural households and need to be more conducive to
equal opportunity. Therefore, the government should not
encourage residents to improve their occupational stratifi-
cation fully through social capital. In fact, according to
empirical results, the role of social capital in occupational
stratification and poverty reduction is far less than that of
human capital. We should build formal institutions to pro-
mote information exchange, reduce transaction costs, and
improve the employment efficiency of rural residents,
leveraging the development of state and social institutions
to partially replace the role of family social capital in occu-
pational stratification to inclusively improve the occupa-
tional stratification of rural households and reduce the
multidimensional poverty of rural households.
Funding information: The author states no funding is
involved.
Author contributions: The author confirms the sole
responsibility for the conception of the study, presented
results, and manuscript preparation.
Conflict of interest: The author states no conflict of interest.
Data availability statement: The data supporting this
study’sfindings are available from the Institute of Social
Science Survey of Peking University. Restrictions apply to
the availability of these data, which were used under
license for this study. Data are available from the author
or applied from the Institute of Social Science Survey of
Peking University at http://isss.pku.edu.cn/cfps/index.htm
with the permission of the Institute of Social Science
Survey of Peking University.
Article note: As part of the open assessment, reviews and
the original submission are available as supplementary
files on our website.
References
Abdul-Hakim R, Abdul-Razak N A, Ismail R. Does social capital reduce
poverty? A case study of rural households in Terengganu, Malaysia.
European Journal of Social Sciences, 2010, 14(4), 556–566.
Abraham, K. G., & Medoff, J. L. (1983). Length of Service and the Operation
of Internal Labor Markets (No. 1085). National Bureau of Economic
Research, Inc.
Alkire, S., & Foster, J. (2011). Counting and multidimensional poverty
measurement. Journal of Public Economics,95(7–8), 476–487.
Alkire, S., & Shen, Y. (2017). Exploring multidimensional poverty in China:
2010 to 2014. In Research on Economic Inequality: Poverty, Inequality
and Welfare (pp. 161–228). Emerald Publishing Limited.
Andriani, L., & Karyampas, D. (2015). Capital social, pobreza e exclusão
social na Itália. Revista Debates,9(2), 77–113.
Artuc, E., & McLaren, J. (2015). Trade policy and wage inequality: A structural
analysis with occupational and sectoral mobility. Journal of International
Economics,97(2), 278–294. doi: 10.1016/j.jinteco.2015.06.001.
Asadi, A., Akbari, M., Fami, H. S., Iravani, H., Rostami, F., & Sadati, A.
(2008). Poverty alleviation and sustainable development: the role of
social capital. Journal of Social Sciences,4(3), 202–215.
Bai, N. (2009). The dual labor market in urban China: Migrants,
townspeople, and allocation of jobs. The China Review,9(2), 69–93.
Banerjee, A. V., Cole, S., Duflo, E., & Linden, L. (2007). Remedying
education: Evidence from two randomized experiments in India. The
Quarterly Journal of Economics,122(3), 1235–1264.
Banerjee, A. V., Duflo, E., Glennerster, R., & Kothari, D. (2010). Improving
immunisation coverage in rural India: Clustered randomised
controlled evaluation of immunisation campaigns with and without
incentives. British Medical Journal,340, c2220.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable
distinction in social psychological research: Conceptual, strategic,
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 17
and statistical considerations. Journal of Personality and Social
Psychology,51(6), 1173.
Becker, G. S. (2009). Human capital: A theoretical and empirical analysis with
special reference to education. University of Chicago Press.
Bentolila, S., Michelacci, C., & Suarez, J. (2010). Social contacts and
occupational choice. Economica,77(305), 20–45. doi: 10.1111/j.1468-
0335.2008.00717.x.
Bian, Y. (1997). Bringing strong ties back in: Indirect ties, network
bridges, and job searches in China. American Sociological Review,
366–385.
Bian, Y., Zhang, L., Wang, W., & Cheng, C. (2015). Institution-spanning
social capital and its income returns in China. In Handbook of
research methods and applications in social capital (pp. 344–357).
Edward Elgar Publishing.
Burt, RS (1997). The contingent value of social capital. Administrative
Science Quarterly,42, 339–65.
Callahan, W. A. (2005). Social capital and corruption: Vote buying and the
politics of reform in Thailand. Perspectives on Politics,3(3), 495–508.
Chen, Y. (2020). Problems with rural vocational education in China and
countermeasures: Learning from the experience of german dual-
system vocational education. 2020 3rd International Conference on
Humanities Education and Social Sciences (ICHESS 2020) (pp. 192–196).
Atlantis Press.
Chen, S., & Wang, Y. (2001). China’s growth and poverty reduction –
trends between 1990 and 1999. Policy Research Working Series 2651,
The World Bank.
Ching-Hwang, Y. (1987). Class structure and social mobility in the Chinese
community in Singapore and Malaya 1800–1911. Modern Asian
Studies,21(3), 417–445.
Démurger, S., & Li, S. (2013). Migration, remittances and rural
employment patterns: Evidence from China. Research in Labor
Economics,37(13), 31–63.
Doeringer, P., & Piore, M. (1971). Internal labor markets and workforce
analysis. M. E. Sharpe.
Du, Y., Park, A., & Wang, S. (2005). Migration and rural poverty in China.
Journal of Comparative Economics,33(4), 688–709.
Duflo, E. (2001). Schooling and labor market concolumns of school
construction in Indonesia: Evidence from an unusual policy
experiment. American Economic Review,91(4), 795–813.
Eccles, R. G., & Dwight B. C. (1988). Doing deals: investment banks at work.
Harvard Business School Press.
Emran, M. S., Ferreira, F. H., Jiang, Y., & Sun, Y. (2023). Occupational
dualism and intergenerational educational mobility in the rural
economy: Evidence from China and India. The Journal of Economic
Inequality,21(3), 743–773.
Fan, Y., Ning, J., & Qin, H. (2023). Investigating the effectiveness of
livelihood capital in reducing re-poverty risk: An empirical analysis
of policy withdrawal and income structures in rural China. Frontiers
in Environmental Science,11, 1175315.
Fane, G., Philipp, K., & Iourii, M. (2015). The U-shapes of occupational
mobility. Review of Economic Studies,82(2), 659–692. doi: 10.1093/
restud/rdu037.
Ganzeboom, H. (2008). ISCO-08: Perspectives, prospects and
applications. In ponencia presentada en la Spring meeting of ISA-RC28
(Vol. 16).
Giesinger, J. (2017). Educational justice, segregated schooling and
vocational education. Theory and Research in Education,15(1), 88–102.
Granovetter, M. S. (1974). Members of two worlds: A development study
of three villages in western Sicily. American Journal of Sociology,
79(4), 1063–1066.
Granovetter, M. (1999). Coase encounters and formal models: Taking
Gibbons seriously. Administrative Science Quarterly,44(1), 158–162.
Grootaert, C., Oh, G. T., & Swamy, A. (2002). Social capital, household
welfare and poverty in Burkina Faso. Journal of African Economies,
11(1), 4–38.
Guo, Y., & Li, X. (2024). Regional inequality in China’s educational
development: An urban-rural comparison. Heliyon,10(4), e26249.
Hassan, R., & Birungi, P. (2011). Social capital and poverty in Uganda.
Development Southern Africa,28(1), 19–37.
Haveman, R., & Wolfe, B. (2000). Welfare to work in the US: a model for
other developed nations?. International Tax and Public Finance,7(1),
95–114.
Hong, L., Tisdell, C., & Fei, W. (2019). Poverty and its reduction in a
Chinese border region: is social capital important?. Journal of the
Asia Pacific Economy,24(1), 1–23.
King, Z. (2005). The ‘bounded’career: An empirical study of human
capital, career mobility, and employment outcomes in a mediated
labour market. Human Relations,58(8), 981–1007.
Knight, J. B., & Yueh, L. Y. (2008). The role of social capital in the labour
market in China. Economics of Transition,16(3), 389–414.
Kurosaki, T., & Khan, H. (2001). Human capital and elimination of rural
poverty: A case study of the North-West Frontier Province, Pakistan.
Hitotsubashi University: Institute of Economic Research.
Li, J., Xue, E., Cao, J., He, Y., Wu, Y., & Hou, H. (2023). Knowledge mapping
of the rural teacher development policy in China: A bibliometric
analysis on web of science. Sustainability,15(9), 7057.
Lin, N., Fu, Y. C., & Hsung, R. M. (2001). Measurement techniques for
investigations of social capital. Social Capital: Theory and Research,
4,57–81.
Liu, Z. (2008). Human capital externalities and rural–urban migration:
Evidence from rural China. China Economic Review,19(3), 521–535.
López, R., & Valdés, A. (2000). Fighting rural poverty in Latin America:
New evidence of the effects of education, demographics, and
access to land. Economic Development and Cultural Change,49(1),
197–211.
Marques, E. (2012). Social networks, segregation and poverty in São
Paulo. International Journal of Urban and Regional Research,36(5),
958–979.
Meng, X., & Zhang, J. (2001). The two-tier labor market in urban China:
Occupational segregation and wage differentials between urban
residents and rural migrants in Shanghai. Journal of Comparative
Economics,29(3), 485–504.
Méreiné Berki, B. (2017). The role of social capital and interpersonal
relations in the alleviation of extreme poverty and spatial
segregation of romani people in Szeged. Journal of Urban and
Regional Analysis,9(1), 33–50.
Montgomery, J. D. (1991). Equilibrium wage dispersion and interindustry
wage differentials. The Quarterly Journal of Economics,106(1),
163–179.
Munshi, K. (2006). Nonmarket institutions. Understanding Poverty., ed.
Abhijit Vinayak Banerjee, Roland Bénabou and Dilip Mookherjee,
389–399.
Narayan, D., & Pritchett, L. (1999). Cents and sociability: Household
income and social capital in rural Tanzania. Economic development
and cultural change,47(4), 871–897.
Neira, I., Vázquez, E., & Portela, M. (2009). An empirical analysis of social
capital and economic growth in Europe (1980–2000). Social
Indicators Research,92, 111–129.
Oppenheim, C., & Harker, L. (1996). Poverty: The facts (3rd ed). Child
Poverty Action Group (CPAG).
18 Song Jian
Parman, J. (2012). Good schools make good neighbors: Human capital
spillovers in early 20th century agriculture. Explorations in Economic
History,49, 316–334.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling
strategies for assessing and comparing indirect effects in multiple
mediator models. Behavior Research Methods,40, 879–891.
Putnam, R. D. (1993). The prosperous community. The American Prospect,
4(13), 35–42.
Rustiadi, E., & Nasution, A. (2017). Can social capital investment reduce
poverty in rural Indonesia?. International Journal of Economics and
Financial Issues,7(2), 109–117.
Saracostti, M. (2007). Social capital as a strategy to overcome poverty in
Latin America: An overview. International Social Work,50(4), 515–527.
Schultz, T. W. (1961). Investment in human capital. The American Economic
Review,51(1), 1–17.
Sen, A. (1982). Poverty and famines: An essay on entitlement and deprivation.
Oxford University Press.
Sen, A. (2006). Conceptualizing and measuring poverty. In D. Grusky &
R. Kanbur, (Eds.), Poverty and inequality. Stanford University Press.
Sfakianakis, G., Agiomirgianakis, G. M., & Manolas, G. (2021).
Macroeconomic determinants of NPLs using an extended sample
and dominance analysis. In Advances in Longitudinal Data Methods in
Applied Economic Research: 2020 International Conference on Applied
Economics (ICOAE) (pp. 285–296). Springer.
Shi, J., & Sercombe, P. (2020). Poverty and inequality in rural education:
Evidence from China. Education as Change,24(1), 1–28.
Teulings, C. N. (2005). Comparative advantage, relative wages, and the
accumulation of human capital. Journal of Political Economy,113(2),
425–461.
Treiman, D. J. (1977). Occupational prestige in comparative perspective.
Academic Press.
Tsui, A. Y. (2005). Must labor move? Journal of Comparative Economics,
33(4), 688–705.
Waldinger, R., & Bozorgmehr, M. (Eds.). (1996). Ethnic Los Angeles. Russell
Sage Foundation.
Wang, J., Xiao, H., & Liu X. (2023). The impact of social capital on
multidimensional poverty of rural households in China.
International Journal of Environmental Research and Public Health,
20, 217.
Wasserman, S., & Faust, K. (1994). Structures of capital: The social
organization of the economy.
Woolcock, M., & Narayan, D. (2000). Social capital: Implications for
development theory, research, and policy. The World Bank Research
Observer,15(2), 225–249.
World Bank. (2001). World Development Report 2000/2001: Attacking
Poverty. Oxford University Press.
Yang, M. M. H. (1994). Gifts, favors, and banquets: The art of social
relationships in China. Cornell University Press.
Zhang, Y., Zhou, X., & Lei, W. (2017). Social capital and its
contingent value in poverty reduction: Evidence
from Western China. World Development,
93, 350–361.
Zhang, J., & Song, X. (2003). The dynamics of urbanization in China. Urban
Studies,40(12), 2487–2500.
Zhe, Z. (2023). The dilemma and countermeasures of rural vocational
education in China under the background of rural
revitalization strategy. The Frontiers of Society, Science and
Technology,5(8), 86–90.
Long-Term Poverty Reduction Based on the Perspective of Occupational Stratification 19
Appendix
Table A1: Use the logit model and remove education and health-related
variables in poverty measure
Use logit model Remove education and
health-related variables in
poverty measure
(1) logit (2) logit (3) clogit (4) clogit
mpoverty2 mpoverty2 mpoverty3 mpoverty3
fhealth −0.020*** −0.001 −0.011*** −0.001
(0.000) (0.001) (0.000) (0.002)
feduyear −0.004*** −0.003*** −0.006*** −0.003***
(0.000) (0.000) (0.000) (0.000)
gc −0.004*** −0.001*** −0.004*** −0.002***
(0.000) (0.000) (0.000) (0.000)
pc −0.001*** −0.000*** −0.002*** −0.002***
(0.000) (0.000) (0.000) (0.000)
mig_rate −0.043*** −0.045***
(0.009) (0.008)
isei_mean −0.004*** −0.001***
(0.000) (0.000)
N29,395 29,395 11,306 11,306
ll −4,494.087 −4,756.065 −1,518.001 −1,678.032
Note: All columns control the fixed effects of family size, depend rate,
ceduyear, and family and year, and the rest are the same as above. ***,
**, and * indicate significance at the 1, 5, and 10% levels, respectively.
20 Song Jian
