EN中文

Journal of Geographical Sciences >

2022 , Vol. 32 >Issue 7: 1383 - 1404

DOI: https://doi.org/10.1007/s11442-022-2002-9

Special Issue: Urban and Rural Governance Toward Sustainable Development Goals

Differences and dynamics of multidimensional poverty in rural China from multiple perspectives analysis

  • WANG Bingbing , 1, 2 ,
  • LUO Qing 3 ,
  • CHEN Guangping 1, 2 ,
  • ZHANG Zhe 1, 4 ,
  • JIN Pingbin , 1, 2, *
Expand
  • 1. School of Earth Sciences, Zhejiang University, Hangzhou 310013, China
  • 2. Institute for Geography & Spatial Information, Zhejiang University, Hangzhou 310013, China
  • 3. College of Resources and Environment, Henan University of Economics and Law, Zhengzhou 450046, China
  • 4. Zhejiang Provincial Key Laboratory of Geographic Information Science, Hangzhou 310013, China
*Jin Pingbin (1968-), PhD and Professor, specialized in human geography and rural vitalization. E-mail:

Wang Bingbing (1995-), PhD, specialized in rural vitalization and poverty alleviation. E-mail:

Received date: 2022-01-07

  Accepted date: 2022-05-17

  Online published: 2022-09-25

Supported by

National Natural Science Foundation of China(41771141)

Program of Philosophy and Social Science of Henan Province(2021BJJ002)

Fold

Abstract

Absolute poverty was completely eliminated in China in 2020. However, poverty measured by income does not fully reflect the actual situation. This paper analyses multidimensional poverty and its dynamics in rural China from perspectives of region, terrain, and geographical location during 2010-2018. We use the Chinese Family Panel Survey data, adopt the Alkire-Foster method and improve the multidimensional poverty index (MPI), calculating and comparing multidimensional poverty and its dynamics among 3009 rural households. In addition, the contribution of the indicator to multidimensional poverty is decomposed. The results indicate that multidimensional poverty has obvious regional differences, topographical differences, and geographical differences. Moreover, the targeted poverty alleviation policy has a significant impact on multidimensional poverty eradication, and the rate of decline of the MPI during 2016-2018 is significantly greater than that of 2010-2014. Education contributes more than 50% to the MPI. In general, the proportion of households with persistent multidimensional poverty is higher than temporary multidimensional poverty, and temporary multidimensional poverty is higher than no multidimensional poverty. These results obtained from a large scale, long time and multiple perspectives could offer new insights for the government to further consolidate the results of poverty alleviation while offering China’s experience to other developing countries.

Key words： multidimensional poverty; dynamics; regional differences; geographical location differences; topographical differences; rural China

Cite this article

WANG Bingbing , LUO Qing , CHEN Guangping , ZHANG Zhe , JIN Pingbin . Differences and dynamics of multidimensional poverty in rural China from multiple perspectives analysis[J]. Journal of Geographical Sciences, 2022 , 32(7) : 1383 -1404 . DOI: 10.1007/s11442-022-2002-9

1 Introduction

Poverty is a major social and practical problem in the world, and eradicating poverty is the common mission of the human race (Haushofer and Fehr, 2014; Padda and Hameed, 2018; Gava et al., 2021). By the end of 2020, all the rural poor had been lifted out of poverty, and the arduous task of eradicating absolute poverty had been completed in China. At the same time, the two sessions proposed the “effective integration of poverty alleviation and rural revitalization” in 2021, emphasizing that the focus of future poverty alleviation efforts is to solve relative poverty and narrow the gap between urban and rural poverty and speed up the construction of rural welfare levels. The limitations of the poverty line based on income levels and the shift in the focus of poverty alleviation indicate that China’s poverty alleviation standards must be multidimensional in the future. Identifying the current state of poverty in China from a multidimensional perspective and studying poverty characteristics have important practical significance for poverty alleviation and provide theoretical support for the setting of multidimensional poverty standards in the future.
The essence of poverty is insufficiency, which is an individual’s lack of the ability to improve material levels and mental state. In 1976, Sen (1976) proposed capacity poverty, which pointed out that poverty is not only manifested as a low income or lack of income ability, but is also manifested as a person’s inability to avoid poverty. Ability poverty is the foundation of multidimensional poverty. Since then, scholars have begun to reflect on the insufficiency of defining poverty based on a lack of material goods and have introduced a multidimensional index system to measure the research object (Nussbaum, 2003; Chakravarty et al., 2008). Among them, the most representative is that of Alkire and Foster (2011) construct the MPI from the three dimensions of education, health, and living standards, which includes ten indicators and proposes a double critical value method to identify multidimensional poverty. The first is the threshold of deprivation of indicators, and the second is how wide the deprivation indicators of the poor households are. The index has a comprehensive and accurate description of the progress made in the human development of a country or region, and can reflect the real poverty of the research objects, so it is widely used in the current multidimensional poverty measurement (Wang et al., 2013; Chen et al., 2016; Alkire et al., 2017; Mohanty et al., 2017; Dong et al., 2021). The MPI and double critical value identification methods are used to measure the multidimensional poverty status of different countries (Wang et al., 2018; Pham et al., 2020; Cai et al., 2021; Zhang et al., 2021).
Drawing on the mature multidimensional poverty theories and identification methods abroad, domestic researchers have identified the multidimensional poverty status of the global, region, and special objects, and have analysed its influencing factors (Alkire and Fang, 2018; Ding and Leng, 2018; Qi and Wu, 2018; Zhang et al., 2019; Aguilar and Sumner, 2020). Based on the dual structure of China’s urban and rural areas, Wang and Alkire (2009) comparatively analysed the multidimensional poverty situation of urban and rural households in China in 2006, and found that 20% of urban and rural households in China have multidimensional poverty. This result is much higher than the poverty incidence rate based on income alone. A total of 36.11% of China’s population lives in rural areas1(1Data source: Main data bulletin of the seventh national population census in 2021.), which has always been a key area of poverty alleviation in China. Liu and Xu (2016) identified the multidimensional poverty situation in China’s rural areas in 2011 and found that 655 counties and 141 million people lived in multidimensional poverty, and 78.1% of the government-designated poor counties had multidimensional poverty, which was concentrated in poverty-stricken areas and mountainous areas with the harsh geographical environments.
The geographical environment of China is complex and diverse, resulting in regional variability and topographical variability in poverty, especially the concentrated contiguous poverty-stricken areas and the boundary between the second and third classes of terrain (Zhou and Xiong, 2018; Ding et al., 2020; Wang et al., 2020). Xu et al. (2018) had studied the state of the multidimensional poverty in China based on the regional level and found that the multidimensional development index (RMDI) is higher in the east coastal areas, in the middle in the central and northern border areas, and the southwest and western border areas are lower. Li et al. (2019) used the night light data from 1993 to 2013 to identify the temporal and spatial dynamics of poverty counties in China and found that they showed a trend of movement from the eastern to the central and western regions in a horizontal direction, and the number of poor counties in the central urban areas of the central and western areas decreased. Most studies on multidimensional poverty have considered regional differences and the specificity of poverty in mountainous areas (Mohanty et al., 2017; Wang et al., 2020; Yu et al., 2020; Jiao, 2020). However, what is the difference between multidimensional poverty in the plains, hills and mountainous areas? What is the difference in multidimensional poverty between households distributed in areas within 15 minutes to the nearest county seat and those located at a distance of 30 minutes or an hour? To clarify these differences, we explore them in this paper.
Poverty is not only multidimensional, regional, and diverse in topography and geographical location but is also dynamic, meaning that the time a family falls into poverty has an impact on the difficulty of getting rid of poverty (Wagle, 2016; Dang and Dabalen, 2018; Ren et al., 2018; Dong et al., 2021; Fang and Zhang, 2021; Yan and Qi, 2021). Glauben et al. (2012) calculated persistent poverty based on household data in three provinces and found that most households are in temporary poverty, and the duration of poverty has different effects on poverty alleviation of the people in different provinces. The continuous enrichment of data, it also provides data support for the dynamic research of poverty. The majority of studies mainly adopt the China Health and Nutrition Survey (CHNS), China Health and Retirement Longitudinal Study (CHARLS), Field Survey Data, DMSP/OLS, and Point of Interest (POI) data (Yu, 2013; Yang and Mukhopadhaya, 2018; Katumba et al., 2019; Shi et al., 2020; Xu et al., 2021; Zhang et al., 2021). The continuous accumulation of China Family Panel Studies (CFPS) data, it provides the possibility of measuring multidimensional poverty in China on a large scale, and the results are universal. Existing studies have primarily analysed the dynamics of multidimensional poverty based on the entire study area, including region, urban and rural areas, and the provinces; however, the differences that geographic factors bring to the dynamics of poverty from multiple perspectives, such as regional, topographical, and geographical locations, have rarely been analysed.
To explore these issues, a series of work has been done in this paper: (1) to clarify the impact of multiple geographic factors on multidimensional poverty, we analyse not only the regional nature of multidimensional poverty, but also the variability of multidimensional poverty among households under different topographical and geographical conditions within regions separately; and (2) to reveal the dynamics of multidimensional poverty in rural China, we explore changes in multidimensional poverty during 2010-2018 in terms of regional, topographical, and geographical locations. Specifically, we use the China Family Panel Studies (CFPS) data from 2010 to 2018, selecting 7 indicators from the 4 dimensions of economy, education, health, and living standards and adopting the Alkire-Foster double critical value method that measures the incidence of unidimensional poverty, the incidence of multidimensional poverty, average deprivation intensity, and the MPI of 3009 households. The multidimensional poverty and dynamic changes from different regions (east, central, west and northeast), terrains (plain, hill, and mountain) and geographical locations (expressed by the time taken by the family to the nearest county seat) are decomposed. Finally, at the regional level, the contribution of the indicator to MPI is calculated. Nationwide multidimensional poverty studies at large scales, over long periods of time, and from multiple perspectives can inform government efforts to address multidimensional poverty.

2 Data and methodology

2.1 Data collection and processing

We use the China Family Panel Studies (CFPS) micro data organized by the Institute of Social Science Survey (ISSS), Peking University. The data has been tracked and cover 25 provinces, autonomous regions and municipalities from the three levels of individuals (adults and children), families, and communities, which account for approximately 95% of the total population in China (excluding Hong Kong, Macao and Taiwan). The survey has a scientific and rigorous sampling method, large sample size, strong representativeness and guaranteed data quality, and can more truly reflect the current basic situation of China’s society, economy, population, education, and health (Xie and Hu, 2014).
This study uses data of communities, households, and adults in rural China in 2010, 2012, 2014, 2016, and 2018. Filter data based on the following principles. The first step is to remove the blanks, refusal, and inapplicable samples in the community, family, and adult databases, and to retain the rural samples; the second step, based on the family code, is to merge the samples of the family and adult databases with stata15.0; in the third step, based on the second step and according to the family code, we take the average of the sample values of family members as the value of the indicator; and in the fourth step, according to the village code, we merge the village data and the family data processed in the third step. Finally, we match the data from the community, family, and adult databases, and obtained 3009 household samples, covering 24 provinces, autonomous regions and municipalities, 120 counties, and 285 villages, as detailed in Figure 1 and Table 1.
Figure 1 Map of regional divisions and locations in China
Table 1 Distribution of household samples in China Family Panel Studies from 2010 to 2018
Region Provinces, autonomous regions and municipalities County Village Household
East Tianjin 1 1 377
Hebei 7 21 23
Shanghai 3 3 13
Jiangsu 3 3 220
Zhejiang 3 5 567
Fujian 2 4 81
Shandong 7 17 212
Guangdong 10 25 32
Central Shanxi 6 17 192
Anhui 3 5 47
Jiangxi 3 9 62
Henan 13 34 97
Hubei 2 3 53
Hunan 4 7 46
West Guangxi 3 8 84
Chongqing 1 3 316
Sichuan 6 12 25
Guizhou 5 11 130
Yunnan 4 11 149
Shaanxi 2 6 63
Gansu 16 47 37
Northeast Liaoning 10 23 45
Jilin 3 5 116
Heilongjiang 3 5 22

2.2 Dimensional and weight

Dimension selection is crucial to the identification of multidimensional poverty. This study draws on the MPI dimension selection criteria (Alkire and Santos, 2014). The MPI measures poverty from the ability and welfare, which includes 10 indicators in three dimensions: education, health, and living standards. It is a relatively mature MPI that is widely used globally. Considering the current standards for defining rural poverty in China, we not only select education, health, and living standards dimensions, but also add per capita income to measure economic levels. Additionally, given the availability and uniformity of CFPS data, the indicators of MPI have been appropriately adjusted. The explanation of the dimensions and indicators is shown in Table 2.
Table 2 The components and indicators of the multidimensional poverty index
Dimension Indicator Deprivation threshold Weight
Economy Per capita income The family’s annual per capita income is less than 2300 yuan1(1Price adjusted calculation was made for family’s per capita income from 2012 to 2018, using 2010 as the base period), assigned a value of 1, otherwise it is 0. 1/4
Education Years of education The average years of education of the family over 16 years old is less than 9, assigned a value of 1, otherwise it is 0. 1/4
Health Chronic diseases If there is a chronic disease among family members, the value is 1, otherwise it is 0. 1/8
Self-rated health There are health self-assessments among family members, “Unhealthy”, “Relatively unhealthy” and “Very unhealthy” assigned a value of 1, otherwise it is 0. 1/8
Living
standard		 Cooking fuel Household cooking fuel is mainly non-clean energy such as firewood and coal, assigned a value of 1, otherwise it is 0. 1/12
Housing type Family housing is not a house type such as “bungalows”, “unit houses”, “small buildings”, “villas”, “townhouses and courtyard houses”, assigned the value is 1, otherwise it is 0 1/12
Drinking water Drinking water types are “river and lake water”, “rain water”, “pond water” and “cellar water”, etc. The clean water sources are not available, assigned a value of 1, otherwise it is 0. 1/12
First, considering China’s poverty standards, we still select the family’s per capita income to express the economic level. Second, education is people’s intellectual capital and the source of creativity. Since the implementation of nine-year compulsory education in 2006, the enrolment rate of school-age children has reached more than 99.94%2(2Data source: 2019 National Educational Development Statistical Bulletin.). Therefore, we only consider the average years of education for adults over 16 years old. Third, health is the basic guarantee for individuals to realize their self-worth and build a harmonious and happy family. We select whether individuals suffer from chronic diseases and self-assessments of physical health as indicators for evaluating health. Fourth, considering the indicators that exist in the CFPS data in terms of living standards from 2010 to 2018, we finally select three indicators of drinking water, cooking fuel, and housing type to measure the family’s living standards. The above three indicators reflect the basic guarantee of people’s lives.
The setting of weight is another important issue in identifying multidimensional poverty. It will affect the family’s multidimensional deprivation scores, thereby affecting the judgement results of multidimensional poverty. The United Nations “Human Development Report” uses the equal weight method to measure multidimensional poverty, so we also adopt the equal weight method, which is divided into two types: dimension equal weight and index equal weight. We adopt the method of equal weight between dimensions and equal weight of indicators within the dimension. The indicator of the weight and deprivation threshold is shown in Table 2.

2.3 Approaches

This study uses the Alkire-Foster double critical value method to identify multidimensional poverty, which includes two critical values. The first critical value is used to judge the poverty situation of the household in each indicator, and the second critical value is used to determine whether the family has multidimensional poverty. When the total deprivation score is greater than the critical value of 0.3 (Alkire and Foster, 2011), the family is considered to be in multidimensional poverty.

2.3.1 The identification of multidimensional poverty

Suppose that n represents the number of families and m (m≥2) presents the number of dimensions. Use the n×m matrix X = [xij] to represent the matrix composed of all household samples. Row vector xi represents the value of the family in the selected dimension, while column vector xj indicates the values of all households under dimension j, and xij corresponds to the value of family i under dimension j. The first step is to determine whether the household is deprived under the dimension j, and the row vector Z = [zj] expresses the poverty threshold of the selected dimension. The identification of multidimensional poverty is divided into two steps.
First, we judge the poverty situation of families in the selected indicators, and use the deprivation matrix g = [gij] to show the result, calculated as follows:
$g_{ij}^{{}}=\left\{ \begin{align} & 1,{{x}_{ij}}<{{z}_{j}} \\ & 0,{{x}_{ij}}\ge {{z}_{j}} \\ \end{align} \right.$
Second, we determine whether the family is in multidimensional poverty. Assume that the deprivation of family i in all dimensions is scored as ${{c}_{i(k)}}=\sum\limits_{j=1}^{m}{{{w}_{i}}{{g}_{ij}}}$, where the weight vector is wi = (w1, w2, …, wm), 0<wi≤1, and wi reflects the importance of the j dimension. Introducing the multidimensional poverty threshold k, then whether the family has multidimensional poverty is represented by Q = [qi(k)], calculated as follows:
${{q}_{i(k)}}=\left\{ \begin{align} & 1,{{c}_{i(k)}}>k \\ & 0,{{c}_{i(k)}}<k \\ \end{align} \right.$

2.3.2 The calculation of multidimensional poverty

It is necessary to add up the comprehensive index to reflect the overall poverty level after identifying multidimensional poverty.
(1) The incidence of multidimensional poverty (H). The incidence of multidimensional poverty represents the population ratio of multidimensional poverty. This indicator reflects the extent of poverty at a macro level, but it cannot describe the intensity of deprivation. The Alkire-Foster method is used to identify q multidimensional poverty in n families, calculated as follows:
$H=\frac{q}{n}\text{=}\frac{\sum\limits_{i=1}^{n}{{{q}_{i}}(k)}}{n}$
(2) The average deprivation intensity (A). The average deprivation intensity represents the average poverty width, which is the ratio of the index of deprivation of poor households to the number of multidimensional poverty families, calculated as follows:
$A=\frac{\sum\limits_{i=1}^{n}{{{c}_{i}}(k)}}{\sum\limits_{i=1}^{n}{{{q}_{i}}(k)}}$
(3) The multidimensional poverty index M0 is also called the adjusted incidence of multidimensional poverty, which is composed of the incidence of multidimensional poverty and the average deprivation intensity. The MPI can simultaneously reflect the distribution of poverty and the degree of deprivation, calculated as follows:
${{M}_{0}}=H\times A=\frac{\sum\limits_{i=1}^{n}{{{c}_{i}}\left(k \right)}}{n}$

2.3.3 The decomposition of multidimensional poverty

The multidimensional poverty index M0 is not only additive, but also decomposable. We will decompose the MPI based on different regions, terrains, and geographical locations. In addition, in order to analyse the contribution of each indicator to the MPI, we also decompose the MPI according to the indicators.
We divide n households into p categories, where ni represents the number of the i categories, so there is $\sum\limits_{i=1}^{p}{{{n}_{i}}=n}$. The calculation formula of M0 is as follows:
${{M}_{0}}=\sum\limits_{i=1}^{p}{\frac{{{n}_{i}}}{n}}{{M}_{i}}=\frac{{{n}_{1}}}{n}{{M}_{1}}+\frac{{{n}_{2}}}{n}{{M}_{2}}+...+\frac{{{n}_{p}}}{n}{{M}_{p}}$
where Mi refers to the MPI of i, and ni /n means the proportion of the sample of i to the total number of households.

3 Results

3.1 Reduction of incidence of unidimensional poverty

We calculate the incidence of unidimensional poverty in rural households. In general, the incidence of poverty shows a dynamic weakening trend, with large differences between indicators from 2010 to 2018 (as shown in Figure 2). Specifically, the indicators with a higher incidence of poverty are years of education and cooking fuel. The moderate level incidence of poverty is self-rated health and per capita income. The lower incidence of poverty is associated with housing types, chronic diseases, and drinking water. It can show that poverty is not only related to income level, but also closely related to living standards, education, and health.
Figure 2 The incidence of unidimensional poverty in rural households from 2010 to 2018

3.2 Changes in multidimensional poverty

The critical value k is critical in the calculation of the MPI, which is between 0 and 1. Changes in k will result in changes in the average deprivation intensity and the MPI. We first measure the MPI with k between 0.3 and 0.8. Finally, we set the critical value k as 0.3.
Table 3 shows the calculation results of the incidence of multidimensional poverty (H), the average deprivation intensity (A), and the MPI (M0) of rural households in China from 2010 to 2018. It is worth noting the disparities between the incidence of multidimensional poverty and the incidence of poverty by income. In 2010, the incidence of per capita income poverty was 24.9%, while the incidence of multidimensional poverty was as high as 72.0%. In 2018, the incidence of per capita income poverty was 18.5%, and the incidence of multidimensional poverty was as high as 48.5%.
Table 3 The multidimensional poverty in rural households from 2010 to 2018
Year k 0.3 0.4 0.5 0.6 0.7 0.8
2010 H (%) 72.0 43.0 28.0 13.0 9.1 2.0
A 0.48 0.57 0.64 0.73 0.76 0.85
M0 0.34 0.25 0.18 0.09 0.07 0.02
2012 H (%) 65.9 38.9 24.2 10.5 6.8 0.9
A 0.47 0.56 0.62 0.72 0.75 0.87
M0 0.31 0.22 0.15 0.08 0.05 0.01
2014 H (%) 61.4 34.0 21.8 8.5 5.1 0.9
A 0.46 0.55 0.62 0.71 0.75 0.87
M0 0.28 0.19 0.13 0.06 0.04 0.01
2016 H (%) 58.9 31.4 20.5 7.5 4.6 1.1
A 0.46 0.56 0.61 0.72 0.76 0.86
M0 0.27 0.17 0.13 0.05 0.04 0.01
2018 H (%) 48.5 23.6 18.3 5.4 3.5 0.9
A 0.45 0.56 0.60 0.72 0.76 0.85
M0 0.22 0.13 0.11 0.04 0.03 0.01
Viewed from the rows of Table 3, as the critical value increases, the incidence of multidimensional poverty and the MPI gradually decrease, while the average intensity deprivation continues to increase (Figure 3). Specifically, in 2010, when k was 0.3, the incidence of multidimensional poverty was 72.0%, the MPI was 0.34, and the average deprivation intensity was 0.48; when k was 0.8, the values were 48.5%, 0.22, and 0.45, respectively. The incidence of poverty decreased by 0.24, the MPI decreased by 0.13, and the average intensity deprivation decreased by 0.03. With the increase in the critical value, the number of poor people almost decreases to zero, indicating that only a small number of households are in extreme poverty. With the help of the Chinese government, there are few extremely poor families in China.
Figure 3 The multidimensional poverty in rural households from 2010 to 2018
The columns of Table 3 show that as time increases, the incidence of multidimensional poverty, the MPI, and the average deprivation intensity decrease. When k is 0.3, the incidence of multidimensional poverty, average deprivation intensity, and the MPI decreased by 8.4%, 2.0%, and 10.3%, respectively, from 2010 to 2012; from 2012 to 2014, they decreased by 6.9%, 2.0%, and 8.7%, respectively; from 2014 to 2016, they decreased by 4.2%, 0.7%, and 4.8%, respectively; and from 2016 to 2018, they decreased by 17.6%, 1.5% and 18.9%, respectively. The incidence of multidimensional poverty and the MPI fell the greatest from 2016 to 2018. This demonstrates that the targeted poverty reduction policy is appropriate for China and has an essential effect on poverty eradication, and China has made great progress in poverty alleviation.

3.3 The decomposition of the multidimensional poverty index

In this section, we compare the multidimensional poverty in different regions (east, central, west, northeast), terrains (plain, hill, mountain), and geographical locations (the time taken by the family to reach the nearest county seat) in rural China and analyse their changes over time. The classification criteria for topography and geographical location are obtained based on the statistics from the 2010 village questionnaire1(1Plain areas are between 0-200 m above sea level, hilly areas are between 200 and 500 m above sea level, and mountainous areas are above 500 m above sea level, and the geographical location classification is based on the shortest time it takes to get to the nearest county seat as indicated by the villagers.), and the distribution of topographical and geographical location data for each region is shown in Table 4.
Table 4 Classification of data according to terrain and geographical location of China
Region Terrain Geographical location (mins)
Plain Hill Mountain 0-15 15-30 30-60 >60
East 375 286 58 75 183 358 103
Central 463 231 96 56 274 330 130
West 117 517 443 39 211 253 574
Northeast 180 192 51 41 63 220 99
China 1135 1226 648 211 731 1161 906

3.3.1 Decomposition by region

Table 5 shows the calculation results of the MPI across regions. Obviously, the incidence of multidimensional poverty, average deprivation intensity, and the MPI show a dynamic decreasing trend during 2010-2018, which is consistent with the national level (Figure 4). The disparity is particularly marked across regions, analysing the incidence of multidimensional poverty. The west is the largest, followed by the northeastern and central regions, the eastern region is the smallest, and the western region (74.3%) is more than 1.5 times that of the east (48.2%). The maximum of the average deprivation intensity is 0.47 in the west, and the minimum is 0.43 in the northeast. The central region is smaller than the west and larger than the east. The average MPI in each region is on the same order of magnitude as the incidence of multidimensional poverty. It is as high as 0.35 in the west and 0.22 in the east, with the former being 1.6 times higher than the latter.
Table 5 The multidimensional poverty of different regions from 2010 to 2018
Region 2010 2012 2014 2016 2018 Average
H (%) East 57.6 53.1 49.1 44.7 36.4 48.2
Central 69.2 58.6 56.1 53.0 40.8 55.6
West 84.5 78.7 74.4 72.6 61.4 74.3
Northeast 69.7 68.8 59.3 58.9 50.6 61.5
A East 0.46 0.46 0.46 0.45 0.44 0.45
Central 0.46 0.47 0.45 0.46 0.47 0.46
West 0.51 0.48 0.47 0.46 0.44 0.47
Northeast 0.44 0.42 0.42 0.44 0.44 0.43
M0 East 0.27 0.24 0.22 0.20 0.16 0.22
Central 0.32 0.28 0.25 0.24 0.19 0.26
West 0.43 0.38 0.35 0.33 0.27 0.35
Northeast 0.30 0.29 0.25 0.26 0.22 0.26
Figure 4 The multidimensional poverty index of rural households in different regions of China during 2010-2018
The comprehensive regional calculation results show that the indicators in the west are the largest, which implies that the west has a wide range of poor families, a deep average deprivation intensity, and a large MPI. In fact, the Chinese government has taken a variety of ways to help the western region out of poverty in the west, such as supporting key projects, railways, transportation, water conservancy, agriculture, forestry, information industry, and other departments in arranging construction funds, while continuing to increase the proportion of key projects for the western region.

3.3.2 Decomposition by terrain

The topography of regions has obvious differences, so the interior of the region is decomposed again according to the terrain. Table 6 displays the trends of in multidimensional poverty. Overall, the incidence of multidimensional poverty, the average deprivation intensity, and the MPI decrease over time in different terrains from 2010 to 2018.
Table 6 The multidimensional poverty of different terrains in 2010-2018
Terrain 2010 2012 2014 2016 2018 Average
H (%) Plain 63.1 55.6 50.7 49.6 38.5 51.5
Hill 73.1 67.8 64.1 59.8 49.0 62.8
Mountain 88.8 83.7 77.0 75.0 65.9 78.1
A Plain 0.45 0.45 0.45 0.44 0.45 0.45
Hill 0.48 0.46 0.45 0.46 0.44 0.46
Mountain 0.52 0.50 0.48 0.47 0.45 0.48
M0 Plain 0.28 0.25 0.23 0.22 0.17 0.23
Hill 0.35 0.31 0.29 0.27 0.22 0.29
Mountain 0.46 0.41 0.37 0.35 0.30 0.38
In the east, the MPI in mountainous areas is almost twice that in the plains, and the MPI in the hills is 1.5 times that in the plains (Figure 5a). For example, in 2010, the MPI was 0.21 in the plains and 0.30 in the hills, while it was as high as 0.41 in mountainous areas. In 2018, it was 0.13 in the plains and 0.17 in the hills, but in mountainous areas it was 0.35. From 2010 to 2018, there was a decrease of 26.1% in mountainous areas and 39.4% in the plains. In contrast, the hilly areas declined by up to 44.16%. For the east, the improvement of multidimensional poverty in the hills is greater than in the plains, while the improvement in mountainous areas is the smallest.
Figure 5 The multidimensional poverty index of different regions of China from 2010 to 2018
As reported in Figure 5b, in the central region, the MPI shows that the plains area was the largest, followed by the mountainous areas, with the hill areas being the smallest in 2010 and 2016; the mountainous area was the largest in 2012, 2014, and 2018, but the hills area was the smallest, and the plains was in between. Analysing the changes in the MPI, the mountain rate dropped from 0.31 to 0.21, a drop of 32.8%; the hill rate dropped from 0.29 to 0.18, a drop of 35.4%; and the plains rate dropped from 0.33 to 0.19, a drop of 42.3%. For the central region, the multidimensional poverty situation in the plains improved greatly, similar to the mountain and hill areas.
In the west, the MPI under different topographical conditions presents the largest in the mountains, the smallest in the plains, and the MPI of the hills is larger than the plains, which is similar to the east (Figure 5c). From 2010 to 2018, the hill areas decreased from 0.40 to 0.24, a decrease of 39.0%; the mountainous areas decreased from 0.49 to 0.32, a decrease of 35.5%; and the plains decreased from 0.33 to 0.23, a decrease of 29.2%. The result shows that the improvement of multidimensional poverty in the hills area is greater, and the plains and mountainous areas are similar to the west.
In Figure 5d, the MPI is the largest in the mountain areas, followed by the hills, and the plains are the smallest in the northeast. Analysing the changes in the MPI, mountainous areas dropped from 0.42 to 0.25, a drop of 40.5%; the plains dropped from 0.26 to 0.19, a drop of 27.1%; and the hills dropped from 0.32 to 0.24, a drop of 23.2% during 2010-2018. For the northeast, the multidimensional poverty situation in the mountainous areas improved greatly, while multidimensional poverty improved only minimally in the hills.
Based on the results of the comprehensive regional multidimensional poverty measurements, the characteristics of topographical differences are clearly visible, and mountainous areas have the largest proportion of multidimensional poor families, especially the western mountains. We analyse the relative changes in the MPI and find that the MPI of the hills fall the fastest in the east and west, and that of the plains fall the fastest in the central region, while that of mountainous areas fall the fastest in the northeast.

3.3.3 Decomposition by geographical location

The county seat is a village where employment, education, and medical facilities are highly concentrated. The geographical location of rural households is expressed in terms of the time it takes for a household to reach the nearest county seat. The time taken by the family to reach the nearest county seat is divided into four types: 0-15 minutes, 15-30 minutes, 30-60 minutes, and >60 minutes. Then, we decompose the multidimensional poverty in each region according to the above four types.
The results in Table 7 show that during 2010-2018, the incidence of multidimensional poverty, the average deprivation intensity, and the MPI are proportional to the time taken by the family to reach the nearest county seat. Families who spend more than 60 minutes traveling to the nearest county seat have the worst multidimensional poverty. At the national level, the areas with the fastest decline in the MPI are the households whose time to the county seat is 15-30 minutes, which dropped from 0.30 to 0.18, a decrease of 38.8%. The area with the slowest decline in the MPI is the type of household that takes more than one hour to reach the county seat, which dropped from 0.44 to 0.29, a decrease of 35.3%. This illustrates that households along the 15-30 minutes route are more likely to be lifted out of multidimensional poverty than those located over 60 minutes away and that improved access to transportation is also critical for poor households.
Table 7 The multidimensional poverty in different geographical locations during 2010-2018
Geographical location (mins) 2010 2012 2014 2016 2018 Average
H (%) 0-15 60.3 59.6 52.6 48.7 38.5 51.9
15-30 65.0 56.7 51.3 52.1 39.8 53.0
30-60 69.0 61.9 58.0 53.9 43.3 57.2
>60 85.8 81.8 76.5 73.7 64.5 76.5
A 0-15 0.45 0.45 0.46 0.44 0.44 0.45
15-30 0.46 0.47 0.45 0.45 0.46 0.46
30-60 0.46 0.45 0.45 0.44 0.45 0.45
>60 0.52 0.48 0.48 0.47 0.45 0.48
M0 0-15 0.27 0.27 0.24 0.21 0.17 0.23
15-30 0.30 0.27 0.23 0.24 0.18 0.24
30-60 0.31 0.28 0.26 0.24 0.19 0.26
>60 0.44 0.40 0.36 0.34 0.29 0.37
From the perspective of the region, Figure 6 shows that the MPI is the largest in the west, followed by the central region, the northeast and the east for all time periods except for 0-15 minutes. To be more specific, when the time to the county seat is 0-15 minutes, the MPI in the central region is the greatest, and the northeast is the smallest. The MPI of the eastern, central, western, and northeastern regions decreased by 51.8%, 34.8%, 41.7%, and 53.3%, respectively. The family takes 15-30 minutes to reach the county seat, and the MPI is the largest in the west and the smallest in the east; the values in the eastern, central, western, and northeastern region have decreased by 32.6%, 42.6%, 32.0%, and 26.1%, respectively. When it takes 30-60 minutes to reach the county seat, the MPI is still the largest in the west and the smallest in the east; the MPI in the eastern, central, western, and northeastern regions has dropped by 46.9%, 43.7%, 40.2%, and 32.6%, respectively. When the family takes more than one hour to reach the county seat, the MPI is still the largest in the west and the smallest in the east. The MPI in the eastern, central, western, and northeastern regions has dropped by 36.3%, 22.6%, 37.7% and 36.6%, respectively.
Figure 6 The multidimensional poverty index of geographical locations in different regions during 2010-2018
During the study period, the multidimensional poverty of families in the western region that took more than 30 minutes to get to the county seat was greatly alleviated compared with the families that took less than 30 minutes. The multidimensional poverty of households with less than 15 minutes distance from the county seat in the eastern and northeastern regions was greatly alleviated, and the households with 15-30 minutes in the central region achieved greater relief. In China’s poverty alleviation policies, the construction of traffic conditions has always been emphasized. In 1982-1993, 1994-2000, 2001-2010, 2011-2014, and 2015-2020, policies related to transportation were issued, including the construction of transportation infrastructure, road repair projects, and village access roads. Improving transportation accessibility is an important challenge for poverty alleviation in China, especially in ecologically fragile and geographically remote areas.

3.3.4 Decomposition by indicator

We decompose the MPI in different regions according to the indicators to show their contribution to the MPI. Figure 6 shows that the contribution of each indicator changed little over the period 2010-2018, with the largest being years of schooling at approximately 50%, followed first by income per capita and cooking fuel at approximately 14% and 17%, respectively, and then by self-rated health at approximately 9% and all others at less than 6%.
As observed in Figure 7, the contribution of the indicator to the MPI varies between regions. In particular, the contribution rate of years of education is the largest in the northeast, with the maximum of 57.0% in 2014, and the smallest in the west between 2010 and 2016. The minimum is 47.7% in 2010, but there is no difference between 2016 and 2018, approximately 50%. The per capita income has a complex situation; the contribution rate is the smallest in the northeast in 2010-2016, while it is the largest in the northeast during 2016-2018. The contribution rate was the largest in the west in 2010-2012, and it was the largest in the east in 2012-2016, but smallest in the east from 2016 to 2018. In general, the living standard dimension, including drinking water, housing type, and cooking fuel indicators, have the largest contribution rate in the west, while they contribute less to other regions. China’s two “no worries, three guarantees” poverty alleviation standards mentioned that by 2020, the rural poor will be free from worries over food and clothing and will have access to compulsory education, basic medical services and safe housing. However, despite these successes, the western region still faces a number of challenges.
Figure 7 The contribution rate of indicators to the multidimensional poverty index in different regions of China during 2010-2018
Through the above analysis, we conclude that the overall structure of the contribution rate of each indicator of the MPI did not change during 2010-2018. The contribution rate of the years of education, per capita income, and cooking fuel to the MPI exceeds 80% from 2010 to 2018, which suggests that education, cooking fuel, and income level are still the key aspects of rural poverty alleviation, especially education.

4 Dynamic multidimensional poverty

China’s poverty alleviation policies have a significant impact on poverty eradication. 2001-2014 is the development-orientated poverty reduction phase, and 2015-2020 is the targeted poverty alleviation phase. The policy of targeted poverty alleviation was first proposed in 2013, and the real implementation started in 2015. The implementation of the targeted poverty alleviation policy has historic significance for poverty reduction in China. Therefore, we calculate the average annual relative changes in the MPI and divide the changes into two stages: the pre-targeted poverty alleviation period (2010-2014), and the post-period (2016-2018), analysing the impact of the policy on multidimensional poverty. Finally, according to the times that a family falls into multidimensional poverty, we divide the family into three types: those in persistent multidimensional poverty (more than 3 times), temporary multidimensional poverty (1-3 times), and no multidimensional poverty. The more times families fall into multidimensional poverty, the harder it is to lift out of poverty. We analyse the changes and the number of times families fall into poverty and identify the extent of poverty in order to better help families escape from poverty.

4.1 The relative changes in the multidimensional poverty index

Figure 8 reveals the average annual decline in the MPI from 2010-2014 and 2016-2018 in terms of region, terrain, and geographical location. Overall, Figure 7 clearly shows the different stages of decline in the MPI from 2010-2014 and 2016-2018; the MPI dropped dramatically in 2016-2018, which is more than twice that in 2010-2014.
Figure 8 The relative changes in the multidimensional poverty index from 2010-2014 and 2016-2018
Specifically, at the regional level, as seen in Figure 8a, the annual reduction in the MPI for households is essentially the same across regions from 2010-2014, averaging approximately 4%. From 2016-2018, there were differences in the improvement of MPI, with the west declining by only 7.36% per year, while the east declined by 10.50%. These results are consistent with the actual situation in China, where economic, educational, medical, and livelihood development are better in the east than in the west, and therefore multidimensional poor families are more likely to escape poverty in the east than in the west. Figure 8b depicts the changes in the MPI under different terrains. The difference in multidimensional poverty alleviation is obvious before and after targeted poverty alleviation, while the difference between different types of topography is not significant. The multidimensional poverty alleviation in mountainous areas is less than that in plains and hilly areas. Compared with the plains and hilly areas, it is more difficult to improve living conditions in mountainous areas, which indicates that the mountainous areas will remain a priority area for poverty alleviation for many years. In different geographical locations, what seems encouraging is that when the time taken by a family to reach the nearest country seat is 15-30 minutes, the decline in MPI is the largest in both 2010-2014 and 2016-2018, at 5.47% and 11.71%, respectively (Figure 8c). It is an astonishing thing that the MPI of a family located 15-30 minutes distance from the county seat dropped more than that of families located 0-15 minutes away. This may be related to the fact that 0-15 minutes distance households are less prone to multidimensional poverty. Thus, the rate of decline is less than those located 15-30 minutes away. The level of transportation development is also vital to the development of a region, especially for families who spend more than one hour traveling to the nearest county seat, which is also a challenge in China’s poverty elimination campaign.

4.2 The dynamic multidimensional poverty

We calculate the proportions of families with no multidimensional poverty, temporary multidimensional poverty, and persistent multidimensional poverty at the level of the region, terrain, and geographical location. An analysis of Figure 9 shows two important results. At the region, topography, and geographical location levels, except for the east and the plain regions, the proportion of households with persistent multidimensional poverty is higher than those with temporary multidimensional poverty, and the proportion of those in temporary multidimensional poverty is higher than that those without multidimensional poverty.
Figure 9 The proportion of dynamic multidimensional poverty in 2010-2018
Regionally, the proportion of households without multidimensional poverty is the smallest, at no more than 25%. It is 23.1% in the east and only 6.6% in the west (Figure 9a). Except for the east, the proportion of persistent multidimensional poverty households in other regions is generally greater than the proportion of temporary multidimensional poverty households. The number of temporary multidimensional poverty families is the largest in the east, at 42.1%, while the smallest is in the west at 25.5%. The number of persistent multidimensional poverty families in the west is 67.9%, while the smallest in the east is 34.8%. Overall, the proportion of households without multidimensional poverty and temporary poverty in the eastern and central regions is greater than that in the northeast and west, while the opposite is true for persistent multidimensional poverty. The disparity of persistent poverty in the eastern and western regions also coincides with China’s focus on poverty alleviation.
From the topographical level, the proportion of households without multidimensional poverty is less than 20%, and that of families in the plains is 2.98 times more than those in mountainous areas and 1.42 times more than those in the hills. Then, the proportion of temporary multidimensional poverty households is between 20% and 43%, and that of the plains is 2.03 times more than the rate in the mountainous areas and 1.23 times more than that of the hills. Except for the plains, families with persistent multidimensional poverty are the most numerous, up to 72.5% in mountainous areas and 51.63 in the hills, 1.90 and 1.35 times more than in the plains, respectively. Compared to the plains, what is remarkable is that there are more persistently multidimensional households and fewer households without multidimensional poverty in mountainous areas; thus, the government should pay more attention to the persistent multidimensional poverty of families in mountainous areas.
Analysing the differences in geographical location, we found that the proportion of households with persistent multidimensional poverty is greater than those experiencing temporary poverty and greater than that of families without multidimensional poverty (Figure 9c). When the time taken by the family to the nearest county seat is 0-15 minutes and 15-30 minutes, the percentage of families in temporary multidimensional poverty and persistent poverty is similar, distributed between 39% and 45%, respectively, which is more of its households without multidimensional poverty than those located 30-60 minutes and one hour away. However, when the time taken by the family to the nearest county seat is 30-60 minutes and more than one hour, the difference between persistent multidimensional poverty and temporary multidimensional poverty is evident; the former is approximately three times that of the latter, and the temporary multidimensional poverty is more than 10 times that of families without multidimensional poverty. Overall, the families that have not experienced multidimensional poverty and temporary multidimensional poverty are mainly concentrated within 30 minutes to the nearest county seat, and persistent multidimensional poverty is mostly distributed in the areas located more than 30 minutes away from the nearest county seat.
Based on the analysis, some interesting findings are that region, terrain, and geographical location have a great impact on the frequency and duration of a family falling into multidimensional poverty.

5 Discussion and conclusion

This paper, based on the CFPS data, measures and analyses the unidimensional and multidimensional poverty of 3,009 rural households from 2010 to 2018. We also decompose the MPI of the families according to region, terrain, and geographical location and calculate the indicator contribution rate. Finally, the dynamics of multidimensional poverty are analysed from the perspective of the above three geographical factors during 2010-2018. We draw the following conclusions.
First, we find that the incidence of unidimensional poverty and multidimensional poverty has decreased in rural China during 2010-2018, which is consistent with other studies (Alkire and Fang, 2018; Zhang et al., 2021), and multidimensional poverty declined more slowly than unidimensional poverty. In addition, the highest contributor to the poverty measurements is the years of education, even exceeding 50%, and the lowest is the drinking water (Figure 1).
Second, we observe that rural multidimensional poverty shows obvious regional differences, which has also been recognized (Dong et al., 2021). The multidimensional poverty is most severe in the west and least severe in the east, while the situation in the central and northeastern regions is similar, between the east and west.
Multidimensional poverty is the result of the interaction of multiple geographical factors. We discover that there are topographical and geographical location differences of the multidimensional poverty. On the one hand, the multidimensional poverty situation of different terrains in the east and west is similar; both are highest in mountainous areas, followed by hills, and the smallest is found in plains. The multidimensional poverty in the central and northeastern regions is similar, with the largest in mountainous areas followed by the plains and then the hills. The results show that the effect of topography on multidimensional poverty varies across regions. On the other hand, the longer it takes for families to arrive in the nearest county seat in the region, the worse its multidimensional poverty. Within the region, the MPI in different geographical locations is the largest in the west, followed by the central region, the northeast, and the east with the smallest MPI. This implies that the impact of geographical location on multidimensional poverty also varies across regions. The above results demonstrate that region, topography, and geographical location all have an impact on multidimensional poverty variability.
In addition, the contribution rate of indicators to the MPI is varies. The factors with a high contribution rate are years of education and per capita income, while Yu (2013) found that the social security was the largest contributor during 2000-2009. These indicators are the main factors causing multidimensional poverty, especially the years of education. The indicators with the low contribution rate are the drinking water and housing types, meaning that the rural poor people have made greater progress in terms of food and housing security, but education security needs to be further strengthened.
Finally, there are spatial and temporal differences in the dynamics of multidimensional poverty. In terms of time differences, the implementation of the targeted poverty alleviation policy has enabled China to make great progress in poverty alleviation. The decline in the relative changes of the multidimensional poverty index in 2016-2018 is clearly greater than the decline in 2010-2016. Spatial variability is manifested in a higher proportion of households in persistent multidimensional poverty than in temporary multidimensional poverty in different regions, terrain, and geographical location, except in the eastern and plain regions, which in turn is higher than in households without multidimensional poverty.

6 Recommendations and future work

As the largest developing country in the world, China has achieved the goal of eradicating absolute poverty in 2020, lifting more than 700 million people out of poverty and contributing to more than 70% of the world’s poverty reduction. Multidimensional poverty analysis nominates directions for the future of poverty in China. Through the above analysis we make the following suggestions. First, education has the deepest impact on poverty, and improving education is an important measure for poverty eradication, especially in the west. Second, the construction of infrastructure plays a decisive role in poverty alleviation in remote areas, particularly the western mountainous areas, followed by the mountainous areas in the east. Moreover, enhancing transportation accessibility is another important measure for poverty alleviation in China, and the government should focus on areas where it takes more than an hour for families to reach the nearest county seat. Finally, chronically multidimensional poor families are the target of extra attention from the government, in addition to the larger proportion of temporarily poor families in the east and the plain regions.
The article has the following shortcomings. As the difficulty of the investigation increases, the indicators are gradually reduced, resulting in a loss of measurable indicators. Unfortunately, due to the requirement of confidentiality of panel data locations, this paper does not spatially display multidimensional poverty in rural China as other (Li et al., 2019; Vasishtha and Mohanty, 2021). Affected by the epidemic, individual and household databases for 2020 have not been published. The targeted poverty alleviation policy was implemented in 2015, and we plan to divide the data into two phases, 2010-2014, and 2016-2020, after the release of the 2020 data, to analyse the impact of the targeted poverty alleviation policy on the completion of the poverty eradication policy in China.

Appendix

The appendix provides detailed statistics and further illustrates the multidimensional poverty situation of Chinese rural households under different terrain conditions and different geographical locations conditions in various regions. Appendix Tables 1-4 show the impact of terrain on the multidimensional poverty of households when the regions are different, and Appendix Tables 5-8 show the impact of geographic location on multidimensional poverty in different regions. The detailed contribution rate of indicator to the multidimensional poverty index is presented in Appendix Table 9.
Appendix Appendix Table 1 The multidimensional poverty of different terrains in eastern China during 2010-2018
Terrain 2010 2012 2014 2016 2018
H (%) Plain 47.7 42.9 40.8 40.0 30.4
Hill 64.0 62.2 54.6 46.5 38.8
Mountain 89.7 74.1 75.9 65.5 63.8
East 57.6 53.1 49.1 44.7 36.4
A Plain 0.45 0.45 0.46 0.43 0.43
Hill 0.48 0.47 0.45 0.48 0.44
Mountain 0.46 0.45 0.47 0.44 0.48
East 0.46 0.46 0.46 0.45 0.44
M0 Plain 0.21 0.19 0.19 0.17 0.13
Hill 0.30 0.29 0.25 0.22 0.17
Mountain 0.41 0.34 0.36 0.29 0.31
East 0.27 0.24 0.22 0.20 0.16
Appendix Appendix Table 2 The multidimensional poverty of different terrains in central China during 2010-2018
Terrain 2010 2012 2014 2016 2018
H (%) Plain 73.2 60.0 57.0 53.8 40.0
Hill 62.3 54.1 55.4 52.0 41.1
Mountain 66.7 62.5 53.1 52.1 43.8
Central 69.2 58.6 56.1 53.0 40.8
A Plain 0.46 0.47 0.45 0.46 0.48
Hill 0.46 0.46 0.45 0.45 0.45
Mountain 0.47 0.48 0.48 0.46 0.48
Central 0.46 0.47 0.45 0.46 0.47
M0 Plain 0.33 0.28 0.25 0.25 0.19
Hill 0.29 0.25 0.25 0.23 0.18
Mountain 0.31 0.30 0.26 0.24 0.21
Central 0.32 0.28 0.25 0.24 0.19
Appendix Appendix Table 3 The multidimensional poverty of different terrains in western China during 2010-2018
Terrain 2010 2012 2014 2016 2018
H (%) Plain 73.5 65.0 64.1 59.0 52.1
Hill 81.0 74.1 70.6 68.7 55.1
Mountain 91.4 87.8 81.5 80.8 71.1
West 84.5 78.7 74.4 72.6 61.4
A Plain 0.45 0.44 0.47 0.43 0.45
Hill 0.49 0.47 0.47 0.45 0.44
Mountain 0.54 0.51 0.49 0.47 0.45
West 0.51 0.48 0.47 0.46 0.44
M0 Plain 0.33 0.28 0.30 0.25 0.23
Hill 0.40 0.35 0.33 0.31 0.24
Mountain 0.49 0.45 0.40 0.38 0.32
West 0.43 0.38 0.35 0.33 0.27
Appendix Appendix Table 4 The multidimensional poverty of different terrains in northeastern China during 2010-2018
Terrain 2010 2012 2014 2016 2018
H (%) Plain 62.8 65.0 47.2 53.9 45.0
Hill 70.8 68.2 67.2 61.5 54.7
Mountain 90.2 84.3 72.6 66.7 54.9
Northeast 69.7 68.8 59.3 58.9 50.6
A Plain 0.42 0.41 0.43 0.42 0.42
Hill 0.45 0.43 0.42 0.45 0.44
Mountain 0.46 0.44 0.41 0.47 0.45
Northeast 0.44 0.42 0.42 0.44 0.44
M0 Plain 0.26 0.27 0.20 0.22 0.19
Hill 0.32 0.29 0.28 0.28 0.24
Mountain 0.42 0.37 0.30 0.31 0.25
Northeast 0.30 0.29 0.25 0.26 0.22
Appendix Appendix Table 5 The multidimensional poverty of different geographical location in eastern China during 2010-2018
Geographical location (mins) 2010 2012 2014 2016 2018
H (%) 0-15 52.2 47.9 41.7 36.8 25.8
15-30 56.7 53.6 48.5 43.7 38.8
30-60 68.9 54.4 51.7 46.9 39.7
>60 70.7 66.7 62.7 61.3 48.0
A 0-15 0.44 0.44 0.44 0.44 0.43
15-30 0.43 0.46 0.45 0.45 0.42
30-60 0.48 0.46 0.46 0.46 0.44
>60 0.48 0.48 0.48 0.48 0.45
M0 0-15 0.23 0.21 0.18 0.16 0.11
15-30 0.24 0.25 0.22 0.20 0.16
30-60 0.33 0.25 0.24 0.21 0.17
>60 0.34 0.32 0.30 0.29 0.21
Appendix Appendix Table 6 The multidimensional poverty of different geographical locations in central China in 2010-2018
Geographical location (mins) 2010 2012 2014 2016 2018
H (%) 0-15 55.2 58.6 65.5 44.8 34.5
15-30 61.7 49.6 44.9 48.2 34.3
30-60 75.5 63.0 61.5 54.6 41.5
>60 78.5 75.4 67.7 64.6 58.5
A 0-15 0.47 0.48 0.47 0.46 0.49
15-30 0.46 0.47 0.44 0.46 0.47
30-60 0.46 0.46 0.45 0.44 0.47
>60 0.46 0.49 0.47 0.47 0.48
M0 0-15 0.26 0.28 0.31 0.20 0.17
15-30 0.28 0.23 0.20 0.22 0.16
30-60 0.35 0.29 0.28 0.24 0.19
>60 0.36 0.37 0.32 0.30 0.28
Appendix Appendix Table 7 The multidimensional poverty of different geographical locations in western China in 2010-2018
Geographical location(mins) 2010 2012 2014 2016 2018
H (%) 0-15 48.7 53.9 38.5 30.8 28.2
15-30 80.6 72.0 67.3 66.4 56.9
30-60 76.3 67.9 63.3 63.3 47.5
>60 93.1 88.2 84.4 82.0 71.1
A 0-15 0.46 0.43 0.48 0.42 0.46
15-30 0.47 0.50 0.47 0.45 0.46
30-60 0.45 0.45 0.45 0.44 0.43
>60 0.54 0.50 0.49 0.47 0.44
M0 0-15 0.22 0.23 0.18 0.13 0.13
15-30 0.38 0.36 0.32 0.30 0.26
30-60 0.34 0.30 0.28 0.28 0.21
>60 0.51 0.44 0.41 0.39 0.32
Appendix Appendix Table 8 The multidimensional poverty of different geographical locations in northeastern China in 2010-2018
Geographical location (mins) 2010 2012 2014 2016 2018
H (%) 0-15 46.2 38.5 15.4 30.8 23.1
15-30 60.7 59.0 50.8 60.7 42.6
30-60 71.4 67.3 57.3 54.1 47.3
>60 72.7 82.8 73.7 70.7 62.6
A 0-15 0.43 0.45 0.65 0.34 0.40
15-30 0.41 0.44 0.42 0.43 0.44
30-60 0.43 0.43 0.41 0.44 0.44
>60 0.47 0.40 0.43 0.44 0.43
M0 0-15 0.20 0.17 0.10 0.11 0.09
15-30 0.25 0.26 0.22 0.26 0.19
30-60 0.31 0.29 0.24 0.24 0.21
>60 0.34 0.34 0.31 0.31 0.27
Appendix Appendix Table 9 The contribution of indicator to the MPI during 2010-2018 (%)
Year Region Drinking water Cooking
fuel		 Housing
type		 Per capita income Years of education Self-rated health Chronic diseases
2010 China 1.5 15.6 3.2 17.7 51.0 8.6 2.5
East 0.2 14.8 0.4 18.6 52.5 9.9 3.6
Central 0.6 16.3 2.9 16.4 53.2 8.0 2.7
West 2.7 14.9 4.7 19.6 47.7 8.2 2.2
Northeast 0.1 16.9 1.2 13.6 56.5 9.9 1.8
2012 China 1.4 13.9 3.7 17.7 51.4 10.2 1.7
East 0.3 12.6 0.9 19.3 51.3 12.7 2.8
Central 0.7 12.8 4.7 18.9 51.2 9.9 1.9
West 2.7 14.5 4.7 17.2 50.1 9.4 1.5
Northeast 0.1 15.7 2.5 14.6 56.1 10.3 0.7
2014 China 1.5 14.4 2.7 16.4 52.8 9.2 2.9
East 0.3 12.0 1.3 18.9 52.4 11.8 3.3
Central 0.7 14.2 3.4 16.1 53.4 9.5 2.9
West 2.9 15.0 3.5 16.8 51.4 8.0 2.5
Northeast 0.1 16.5 0.9 12.0 57.1 9.4 4.0
2016 China 1.5 14.5 1.6 16.5 53.3 9.6 3.1
East 0.8 12.0 0.1 17.4 53.4 13.2 3.2
Central 1.1 13.6 2.2 17.4 53.4 8.9 3.4
West 2.3 15.6 2.2 15.6 52.9 8.4 3.1
Northeast 0.2 15.6 0.9 16.9 54.0 9.8 2.5
2018 China 1.5 13.7 1.4 19.8 53.4 6.3 3.8
East 0.7 12.0 0.1 20.0 54.1 8.8 4.2
Central 1.5 11.5 1.7 21.0 50.8 7.9 5.6
West 2.4 15.0 2.1 17.8 54.7 4.9 3.1
Northeast 0.0 15.3 0.2 24.2 52.6 5.1 2.7

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Aguilar R G, Sumner A, 2020. Who are the world’s poor? A new profile of global multidimensional poverty. World Development, 126: 104716.

DOI

[2]
Alkire S, Fang Y F, 2018. Dynamics of multidimensional poverty and uni-dimensional income poverty: An evidence of stability analysis from China. Social Indicators Research, 142(1): 25-64.

DOI

[3]
Alkire S, Foster J, 2011. Counting and multidimensional poverty measurement. Journal of Public Economics, 95(7/8): 476-487.

DOI

[4]
Alkire S, Roche J M, Vaz A, 2017. Changes over time in multidimensional poverty: Methodology and results for 34 countries. World Development, 94: 232-249.

DOI

[5]
Alkire S, Santos M E, 2014. Measuring acute poverty in the developing world: Robustness and scope of the multidimensional poverty index. World Development, 59: 251-274.

DOI

[6]
Cai P R, Wang X, Yan Y H et al., 2021. Spatio-temporal pattern and its influencing factors of children multidimensional poverty in China. Acta Geographica Sinica, 76(10): 2551-2567. (in Chinese)

DOI

[7]
Chakravarty S R, Deutsch J, Silber J, 2008. On the watts multidimensional poverty index and its decomposition. World Development, 36(6): 1067-1077.

DOI

[8]
Chen Y F, Wang Y H, Wang X L, 2016. Measurement and spatial analysis of poverty-stricken villages in China. Geographical Research, 35(12): 2298-2308. (in Chinese)

[9]
Dang H-A H, Dabalen A L, 2018. Is poverty in Africa mostly chronic or transient? Evidence from synthetic panel data. The Journal of Development Studies, 55(7): 1527-1547.

DOI

[10]
Ding J J, Leng Z M, 2018. Regional poverty analysis in a view of geography science. Acta Geographica Sinica, 73(2): 232-247. (in Chinese)

DOI

[11]
Ding J J, Wang Z, Liu Y H et al., 2020. Rural households' livelihood responses to industry-based poverty alleviation as a sustainable route out of poverty. Regional Sustainability, 1(1): 68-81.

DOI

[12]
Dong Y, Jin G, Deng X Z et al., 2021. Multidimensional measurement of poverty and its spatio-temporal dynamics in China from the perspective of development geography. Journal of Geographical Sciences, 31(1): 130-148.

DOI

[13]
Fang Y F, Zhang F, 2021. The future path to China’s poverty reduction: Dynamic decomposition analysis with the evolution of China’s poverty reduction policies. Social Indicator Research, 158(2): 507-538.

DOI

[14]
Gava O, Ardakani Z, Delalić A et al., 2021. Agricultural cooperatives contributing to the alleviation of rural poverty: The case of Konjic (Bosnia and Herzegovina). Journal of Rural Studies, 82: 328-339.

DOI

[15]
Glauben T, Herzfeld T, Rozelle S et al., 2012. Persistent poverty in rural China: Where, why, and how to escape? World Development, 40(4): 784-795.

DOI

[16]
Haushofer J, Fehr E, 2014. On the psychology of poverty. Science, 344(6186): 862-867.

DOI PMID

[17]
Jiao W X, 2020. Analyzing multidimensional measures of poverty and their influences in China’s Qinba Mountains. Chinese Journal of Population, Resources and Environment, 18(3): 214-221.

[18]
Katumba S, Cheruiyot K, Mushongera D, 2019. Spatial change in the concentration of multidimensional poverty in Gauteng, South Africa: Evidence from quality of life survey data. Social Indicators Research, 145(1): 95-115.

DOI

[19]
Li G E, Chang L Y, Liu X J et al., 2019. Monitoring the spatiotemporal dynamics of poor counties in China: Implications for global sustainable development goals. Journal of Cleaner Production, 227: 392-404.

DOI

[20]
Liu Y H, Xu Y, 2016. A geographic identification of multidimensional poverty in rural China under the framework of sustainable livelihoods analysis. Applied Geography, 73: 62-76.

DOI

[21]
Mohanty S K, Rasul G, Mahapatra B et al., 2017. Multidimensional poverty in mountainous regions: Shan and Chin in Myanmar. Social Indicators Research, 138(1): 23-44.

DOI

[22]
Nussbaum M, 2003. Capabilities as fundamental entitlements: Sen and social justice. Feminist Economics, 9(2/3): 33-59.

DOI

[23]
Padda I U H, Hameed A, 2018. Estimating multidimensional poverty levels in rural Pakistan: A contribution to sustainable development policies. Journal of Cleaner Production, 197: 435-442.

DOI

[24]
Pham A T Q, Mukhopadhaya P, Vu H, 2020. Targeting administrative regions for multidimensional poverty alleviation: A study on vietnam. Social Indicators Research, 150(1): 143-189.

DOI

[25]
Qi D, Wu Y C, 2018. Comparing the extent and levels of child poverty by the income and multidimensional deprivation approach in China. Child Indicators Research, 12(2): 627-645.

DOI

[26]
Ren Q, Huang Q X, He C Y et al., 2018. The poverty dynamics in rural China during 2000-2014: A multi-scale analysis based on the poverty gap index. Journal of Geographical Sciences, 28(10): 1427-1443.

DOI

[27]
Sen A, 1976. Poverty: An ordinal approach to measurement. Econometrica, 44(2): 219-231.

DOI

[28]
Shi K F, Chang Z J, Chen Z Q et al., 2020. Identifying and evaluating poverty using multisource remote sensing and point of interest (POI) data: A case study of Chongqing, China. Journal of Cleaner Production, 255: 120245.

DOI

[29]
Vasishtha G, Mohanty S K, 2021. Spatial pattern of multidimensional and consumption poverty in districts of india. Spatial Demography, 9: 213-240.

DOI

[30]
Wagle U R, 2016. The caste/ethnic bases of poverty dynamics: A longitudinal analysis of chronic and structural poverty in nepal. The Journal of Development Studies, 53(9): 1430-1451.

DOI

[31]
Wang X L, Alkire S, 2009. Multidimensional poverty measurement in China estimates and policy implications. Chinese Rural Economy, 12: 4-10. (in Chinese)

[32]
Wang C C, Wang Y, Fang H Y et al., 2020. Determinants of rural poverty in remote mountains of southeast China from the household perspective. Social Indicators Research, 150(3): 793-810.

DOI

[33]
Wang Y H, Chen Y F, Chi Y et al., 2018. Village-level multidimensional poverty measurement in China: Where and how. Journal of Geographical Sciences, 28(10): 1444-1466.

DOI

[34]
Wang Y H, Qian L Y, Duan F Z, 2013. Multidimensional poverty measurement and spatial distribution pattern at the country scale: A case study on key country from national contiguous special poverty-stricken areas. Scientia Geographica Sinica, 33(12): 1489-1497. (in Chinese)

[35]
Xie Y, Hu J W, 2014. An Introduction to the China Family Panel Studies (CFPS). Chinese Sociological Review, 47(1): 3-29.

[36]
Xu L D, Deng X Z, Jiang Q O et al., 2021. Identification and alleviation pathways of multidimensional poverty and relative poverty in counties of China. Journal of Geographical Sciences, 31(12): 1715-1736.

DOI

[37]
Xu Y, Duan J, Xu X R, 2018. Comprehensive methods for measuring regional multidimensional development and their applications in China. Journal of Geographical Sciences, 28(8): 1182-1196.

DOI

[38]
Yan X Y, Qi X H, 2021. The measurement method and evolution mechanism of poverty dynamics. Acta Geographica Sinica, 76(10): 2425-2438. (in Chinese)

DOI

[39]
Yang J, Mukhopadhaya P, 2018. Is the adb’s conjecture on upward trend in poverty for China right? An analysis of income and multidimensional poverty in China. Social Indicators Research, 143(2): 451-477.

DOI

[40]
Yu J T, 2013. Multidimensional poverty in China: Findings based on the chns. Social Indicators Research, 112(2): 315-336.

DOI

[41]
Yu R Y, Liang L K, Su X Y et al., 2020. A driver based framework for vulnerability assessment of the poverty stricken areas of Funiu Mountain, China. Ecological Indicators, 113: 106209.

DOI

[42]
Zhang D Y, Li J J, Han P, 2019. A multidimensional measure of energy poverty in China and its impacts on health: An empirical study based on the China family panel studies. Energy Policy, 131: 72-81.

DOI

[43]
Zhang Z, Ma C Y, Wang A P, 2021. A longitudinal study of multidimensional poverty in rural China from 2010 to 2018. Economics Letters, 204: 109912.

DOI

[44]
Zhou L, Xiong L Y, 2018. Natural topographic controls on the spatial distribution of poverty-stricken counties in China. Applied Geography, 90: 282-292.

DOI

Options
Outlines

/

〈 〉