.

China's move to measuring relative poverty: implications for social protection

Prof. Robert Walker

Dr. Yang Lichao

# Abstract

With rural extreme poverty officially eradicated in 2020, China is to move to measuring relative poverty in urban and rural areas. Relative poverty describes circumstances in which people cannot afford actively to participate in society and benefit from the activities and experiences that most people take for granted. It is conventionally defined as 40, 50 or 60 percent of national median disposable income. While a single poverty threshold symbolises national unity, separate poverty thresholds could be created for urban and rural areas or provinces. A unified national poverty standard for China based on 40 per cent of national median disposable income threshold implies income levels almost 4 times higher than the existing rural poverty line and 61 per cent higher than average social assistance (Dibao) payments in urban areas (often used as a surrogate urban poverty line). Relative poverty is more persistent than absolute poverty and less affected by economic growth. Research is required to determine the needs of peri-poor persons (i.e., those brought into poverty due to the new definition). Strategies needed to tackle relative poverty include: a comprehensive social protection system inclusive of floors; active policies to assist people out of poverty; poverty mainstreaming; and supportive, redistributive fiscal policies.

# Introduction

Despite the negative impact of the COVID-19, the government of China officially eradicated extreme poverty in 2020. This outcome, the result of a deliberate policy process is an achievement without global precedent. Believing that to create a ‘moderately prosperous’ (xiaokang) society, poverty needed to be eliminated, this goal was formally established in 2010 and revitalised in 2015 with the targeted poverty alleviation strategy.

Beyond 2020, the Sustainable Development Goals (SDGs) focus on 2030 by when much more is expected to be achieved: halving poverty in all its dimensions according to national definitions. Moreover, it is evident that China retains its commitment to an ambitious poverty alleviation and social protection agenda embracing urban as well as rural poverty.1 The Communique of the Fourth Plenary Session of the 19th Central Committee of the Communist Party of China issued in October 2019 proposes addressing relative poverty and commits to improving social assistance, and to providing higher quality employment, education offering life-long learning, health guarantees and comprehensive social security.2 Subsequently, in the context of the COVID-19 pandemic, the State Council has announced that the scope of unemployment insurance is to be expanded to urban and rural areas and that social assistance is to be made available to out of work migrant workers without unemployment insurance.3

With a view to contributing to the development of this ambitious agenda, this Working Paper focusses on relative poverty, first reflecting on the concept and its measurement, then considering how the new definition is likely to change the understanding of poverty and its dynamics in China, and finally, exploring the policy implications of incorporating relative poverty into policymaking.

Rural poverty in China is currently defined by per capita income of less than 2,300 yuan/year in 2011 prices (about US$350/year). Because the income threshold has not been increased in real terms since it was fixed in 2011, this is generally taken to be a measure of absolute poverty. However, for most people, in most circumstances, this threshold probably exceeds the textbook definition of absolute poverty, the level of income necessary to sustain life. The rural poverty rate, measured according to this national standard, has fallen precipitously and was already below 2 per cent in 2018 (Figure 1). There is no national poverty standard for urban poverty, but poverty measured according to World Bank standards is lower than in rural areas although direct comparison belies the impact of higher living costs in major cities. Figure 1 Various absolute poverty rates, China, 1990-2018 Source: World Bank database The origins of the official measure date back to 2008 when two separate indices, one for ‘absolute poverty’ and one for other low-income families, were merged into a single poverty threshold set at 1,274 yuan/year. In 2011, this threshold was raised by more than 80 per cent to the current level which, as a result, increased the nominal poverty count by 128 million. The policy justification for the increase was to bring the threshold ‘closer to the international standard of$1.25 a day’ set by the World Bank in 2005 and which was subsequently increased to $1.90/day in 2015.4 The current ‘absolute’ measure was therefore established relative to the World Bank standard, itself an average of the national poverty lines set by the world poorest 15 countries.5 However, the level of dibao6, basic social assistance, which varies between provinces, across cities and between rural and urban areas, is generally fixed above the poverty line. Moreover, the government’s poverty alleviation strategy has, since 2011, embraced the concept of ‘two no worries’ (Liangbuchou) over food and clothing and three guarantees (Sanbaozhang) relating to housing, health care and education. This has caused some Chinese scholars to adopt higher poverty thresholds and also to express concern about the limited capacity of targeted policies to lift the living standards of recipients above that of their counterparts who are denied receipt of benefits because their incomes hover just above the eligibility threshold.7 8 Similar concerns have been expressed about high housing costs depressing the living standards of urban dwellers and pushing some into poverty.9 As a policy response, many cities have introduced low income lines higher than dibao as thresholds for financial and other support. For example, that in Guandong Province is 1.5 times dibao, while that in Wuhan is twice the amount of dibao.10 In Beijing, the low-income line has equalled the minimum wage since 2018.11 Therefore, when not focussed on the official poverty line, poverty alleviation policy and research in China tends primarily to examine the ‘near poor’ comparing their circumstances with those of the extremely poor rather than those with average living standards. To summarise, while the national policy goal is to eradicate extreme rural poverty defined in absolute terms, policy on the ground has been moving towards a more relative understanding of poverty, by making adjustments so as to make those who are near poor eligible for social assistance. # The concept of relative poverty As societies become more prosperous, poverty ceases to be primarily about hunger, destitution, and survival and more about whether people can afford to partake in the normal activities of daily life that are enjoyed by most people. The object of comparison therefore shifts from point-of-death impoverishment to average living standards. The underlying logic is that the nature of poverty changes as national wealth and living standards rise. When the current standard was established, China was transitioning from being a low middle-income to a high middle-income country and is now close to being designated as a high-income country by the World Bank (Figure 2).12 Figure 2 International poverty lines in 2011 Source: Ferreira, and Sánchez-Páramo (2017) In adopting a relative poverty standard, which most high-income countries have done, attention is directed to social norms regarding acceptable living standards, to the benefits of citizenship and the degree of deprivation that can be tolerated in a stable society. Relative poverty is the denial of what ordinary people can consume and do, a rejection of who they are and a barrier to what they could become.13 Such aspirations and social expectations are necessarily conditioned by culture and level of economic activity, making poverty inherently relative. Policy in response to relative poverty is not restricted to consideration of the conditions of people experiencing material poverty but is necessarily required to embrace social and community factors including solidarity, cohesion, and harmony. In the European context, for example, social inclusion is a strong policy concern and Member States have agreed to use the same benchmarks of within-country social cohesion and exclusion to gauge success in fighting poverty, social exclusion, and inequality.14 Another fundamental difference between absolute and relative poverty is that, whereas the former can be eradicated through economic growth, the latter cannot be. The prevalence of relative poverty is a product of the nature and extent of social inequality and must be addressed by various forms of distributional policies including social assistance, social security, and progressive taxation. Relative poverty embraces more than the necessities that are required to survive, additionally asking: what is needed for people to be active members of a moderately prosperous society?15 Society’s answer has necessarily to be expressed in terms of the degree of relative deprivation that is acceptable, a response that is inherently normative but which is informed by the desire to promote social cohesion and to avoid discord and social instability. # Fixing the threshold of relative poverty While poverty is an inherently complex, multifaceted phenomenon, it is usually simplified, for the purpose of measurement, to consumption or more usually income. Relative income poverty is usually defined and measured as some proportion of mean or median income, both simple indices of the general living standard. Median income, the mid-point of the income distribution with 50 per cent of people receiving more and 50 per cent receiving less, is generally preferred. While mean income is easier to calculate, it is susceptible to distortion when a few people have extremely high incomes. Moreover, a poverty threshold based on mean income would rise if poverty alleviation strategies boosted the incomes of the poorest individuals; this would have the effect of increasing the number of people counted as being poor, thereby reducing the apparent success of policy. Figure 3 Rates of ‘at risk of poverty and social exclusion’ and various thresholds in Europe 2010 -2018 (27 Member States) Source: Eurostat Income is usually measured at household level and adjusted to take account of differences in the consumption of adults and children.16 The choice of the precise threshold is generally based on convention rather than scientific analysis. But, unlike low-income lines in China’s cities, thresholds are set relative to middle incomes rather than to extremely low ones. For example, the European Union sets the ‘at risk of poverty threshold’ at 60 per cent of median equivalised household income after social transfers, but Eurostat also publishes estimates of proportions of the population with incomes less than 70, 50 and 40 per cent of median (and mean) income.17 In addition, statistics are available for poverty rates before social transfer, an index of market poverty, while ‘persistent at risk of poverty’ considers duration, counting persons with incomes below the threshold in the current year and in two of the previous three years. Moreover, since 2010, Europe has added persons who are materially deprived (currently indexed by lacking five or more items from a list of 13) or living in households with very low work to those poor on income grounds to create a headline measure of ‘people at risk of poverty or social exclusion’ (Figure 3). Because indices of material deprivation have remained unaltered this headline indicator mixes absolute and relative measures and has fallen a little due to increasing living standards. Beyond the European Union, analysis of the national poverty lines of 107 non-OECD countries indicates that most are a little above 50 per cent of median income or consumption (the median plus$US1/per day).18 Britain follows Europe in adopting the 60 per cent of median disposable income threshold, but employs additional measures including one requiring low income (70 per cent of current median equivalised net household income) to be accompanied by material deprivation (based on 21 indicators). Empirical research in Britain to determine acceptable living standards that is used to inform the level of the ‘living wage’ indicates that, for most household types, the basic threshold should be set at 70 per cent of median household income, adjusted to 60 per cent for retirement pensioner couples (who generally need to spend less) and to above 80 per cent for lone parents with two children who need to spend more.19

Most countries set a single relative poverty threshold for the entire country despite geographic variations in living costs. While pragmatic, simplifying measurement and policy structures, and strategic, emphasising national unity and solidarity, this inevitably creates distortion and territorial inequity. This is true even of small highly urbanised countries such as Britain, where consensually determined, minimum income thresholds are around 25 per cent higher in inner London than in other urban centres.20

China, of course, already makes distinctions between urban and rural areas reflecting the almost three-fold variation in average disposable incomes (36,396 yuan in urban areas, 13,432 yuan in rural ones in 2017).21 Incomes also vary markedly between provinces which further militates against setting a common poverty threshold for the whole of China. For example, per capita household urban income in Shanghai (62,596 yuan; 2017) is 2.25 times that in Gansu (27,763 yuan) and 7.75 times the per capita income of Gansu’s rural areas (8,076 yuan).

However, income differentials are much reduced when account is taken of living costs. Combining rural and urban areas, provincial real disposable income in 2015 varied by a factor of 3.9, and by just 2.8 if Shanghai and Beijing are excluded because of their exceptional nature.22 Furthermore, considering disposable incomes net of housing costs means that Beijing and Shanghai drop from top to respectively seventh and tenth place in terms of spending power with Zhejiang having the highest disposable incomes and Hainan and Gansu the lowest.

# The implications of relative poverty

It is well recognised in China that ‘relative poverty will persist for a long time, and work to reduce relative poverty will continue to be done after 2020’ (Liu Yongfu, 2018 cited by Lim [2018] and Huang [2019]). The persistence reflects the fact that poverty thresholds rise in line with national income enabling people in poverty to share in a nation’s growing prosperity. However, as already noted, this also means that relative poverty, unlike absolute poverty, generally cannot be reduced through economic growth alone but additionally requires deliberately redistributive policies.

The challenge is illustrated by the failure of Europe to reduce poverty between 1995 and 2018 during which time real per capita GDP increased by 44 per cent. This was despite an intergovernmental target to cut poverty by 20 million or 18 per cent (Figure 3).23 24 Over a similar period, during which absolute poverty in China fell substantially, relative poverty in urban China grew from below levels found in Europe to above them. This rise in urban relative poverty reflects a period of rapid urbanisation during which rural migration added to the incidence of low-income in urban areas, and disproportionate income growth increased median urban incomes and the proportion of people below the poverty threshold.

Figure 4 Relative poverty in Europe and urban China

Source: Eurostat and Gustafsson, B & Ding S. (2020).

The analysis reported for China in Figure 4 is directly comparable with Europe but relates only to urban areas. By adopting a somewhat different approach, it is possible to examine poverty trends for China as a whole (albeit for a somewhat earlier period) and to compare these with the rest of the world other than OECD countries.25 China’s success is evident from Figure 5. While global absolute poverty (measured as less than $1.25/day) declined from 40.5 to 25.2 per cent between 1981 and 2008, China begin the period with a much higher rate (84.0 per cent) and succeeded in reducing it to 13.1 per cent, little more than half the global figure. Moreover, China even substantially reduced relative poverty (defined as half of mean income26) from 85.2 to 41.0 per cent while the global rate scarcely changed. China’s success in reducing relative poverty was mostly the result of rural migrant workers gaining more benefit from economic growth than they would have done by remaining in their villages although, as noted above, this resulted in increased inequality in urban areas. With China now almost 60 per cent urban, the scope for lowering relative poverty further through rural-urban migration is much reduced. Figure 5 Relative and absolute poverty and inequality, China and other non-OECD countries Source: Adapted from Chen and Ravallion (2012) Moving to a relative definition of poverty will result in an increase in the nominal poverty rate, just as happened with the 2011 increase in the absolute poverty threshold. With a relative poverty line set at 40 per cent, 50 per cent or 60 per cent of median national household disposable income, the current rural poverty line would need to increase by factors of 3.8, 4.8 or 5.7, respectively. Clearly, this would result in a substantial rise in the number of persons counted as being poor; the 50 per cent relative poverty threshold is quite similar to the World Bank’s$5.50/day poverty standard which generated a poverty rate of 27.2 per cent for China with about 377 million people living beneath the poverty line in 2015. Subjective poverty, the income level at which people feel that they are poor, also points to increases of a similar order.27 People feel poor in rural areas when their income falls below three-quarters of the local median which means a poverty line set at around 40 per cent of national median disposable income.

Poverty is dynamic with many more people moving in and out of poverty than remain poor. However, raising the poverty threshold is likely to reverse the trend to shorter spells observed for the current measure.28 Extreme poverty is exceptional even in the lives of people prone to poverty whereas less severe forms are commonplace and persistent. While there is as yet little analysis of the duration of spells of relative poverty in China, spells of more severe poverty (below the 50 percent median threshold) are, on average, shorter than spells of less severe poverty (based on the 60 per cent threshold) in every European country.29

# What relative poverty means for policy

China’s moving to the measurement of relative poverty in both rural and urban areas reflects its economic success and place in the world and represents a contribution to the vertical development of its social protection floor. It is a statement that, as a citizen in the New China, one is not just guaranteed physical survival and subsistence but the ability to participate in the economic, political, social, and cultural life of the nation. To implement the guarantee requires decisions about the level of the poverty threshold, linkages between the threshold and welfare benefits and the necessary policy response in terms of the horizontal coverage of social protection.

## Level of the poverty threshold

Setting the level of the poverty threshold is a matter of political judgement informed by pragmatic considerations such as ease of measurement. Nevertheless, the threshold is only likely to attract social support and to be responsive to policy initiatives if it reflects actual ways of living and enables people to live in dignity.30 Further research on budget standards, that is determining what people need to have and do in modern China and how much it costs, would seem to be essential.31 This research might rely on expert opinion32 or, consistent with the principles of social governance, include use of consensual techniques in which lay participants from diverse backgrounds review evidence to derive minimum income standards. Consensual methods have recently been used in several countries including Japan, Singapore, Ireland, Britain, and France to assess the adequacy of social assistance thresholds.33 34

Adopting a single universal poverty line for all of China has the merit of promoting national cohesion and harmony and offering clarity in the implementation of policy. Initially, it would index markedly different living standards due to urban, rural and provincial differences in living costs and levels of economic development. Over time, it would be a stimulus for, and a measure of, the effectiveness the policies designed to enhance rural re-vitalisation and the spatial rebalancing of the economy.

The European Union, as noted above, has adopted the common standard for relative poverty of 60 per cent of national equivalised disposable income. In this basis, poverty rates in 2014 varied from 9.7 per cent in Czechia and 11.6 per cent in the Netherlands to 22.2 per cent in Spain and 25.1 per cent in Romania. However, when incomes across Europe are compared to 60 per cent of the EU-wide median income, national poverty rates fall to below 5 per cent in Luxembourg, Finland and Austria but rise to over 50 per cent in Greece, Lithuania, Latvia, Hungary, and Bulgaria, and to 93 per cent in Romania, the poorest country in the EU (Figure 6).35 This analysis illustrates the remaining challenges that confront Europe in building social cohesion within an open market economy.

Figure 6 Relative poverty rates in Europe, 60 per cent of median disposable national and pan-European income, 2014

Source: Goedemé, et al. (2019)

The variation in the level of economic development between the European states is greater than between provinces in China.36 Nevertheless, despite the merits of a single poverty line, the marked discrepancies between the existing rural poverty line and urban dibao rates (often taken as urban poverty thresholds), and the low level of both in relation to conventional international indices of relative poverty, speak against this strategy. In 2017, the rural poverty standard represented just 10.5 per cent of median national disposable income or 26 per cent of national poverty threshold set at 40% of national disposable income; while average urban dibao equated to 61 per cent of this poverty standard. This would suggest an almost four-fold increase in the rural poverty threshold and a 61 per cent increase for the urban one.

An alternative approach would be to retain separate urban and rural poverty thresholds, setting both at 40 per cent of the corresponding median disposable income. This would mean increasing the rural poverty line and average urban poverty line (derived from dibao) by factors of 1.9 and 2.2 respectively. A further strategy might combine rural and urban populations and establish separate provincial poverty lines related to provincial per capita disposable income. While this might have merit as a means of determining payments of social assistance, it perhaps too much dilutes the concept of a national poverty line.

## Poverty and social assistance thresholds

There is a powerful administrative logic for equating social assistance and poverty thresholds, with much discussed precedents in the USA and the Netherlands. If poverty prevention is the policy goal and the principal policy instrument is social assistance, then 100 per cent take up of comprehensive social assistance assures the eradication of poverty. The poverty rate thus serves as a direct measure of policy effectiveness. The income threshold for receipt of rural dibao currently exceeds the rural poverty line by about 65 per cent, whereas a relative poverty threshold, set at conventional international levels, would be much higher than either existing rural or urban dibao.

If it was decided to set dibao equal to a national poverty line set at 40 per cent of the median personal disposable income, then urban and rural dibao would need to increase by factors of around 2.3 and 1.6 respectively. Urban dibao would require to be doubled and rural dibao almost tripled to reach the World Bank preferred standard of 50 per cent of median personal disposable income. These large multiples reflect the gap between China’s rural poverty threshold and conventional international measures of relative poverty.

If separate urban and rural thresholds were retained, rural dibao would be required to rise by 15 per cent to equal 40 per cent of median rural disposable income, while urban dibao would have to approximately double (increasing by a factor of 2.19) to reach the corresponding urban threshold. Choosing to adopt provincial thresholds that combine rural and urban populations would similarly require dibao thresholds to be doubled although with some variation. In Zhejiang and Jiangsu, dibao would need to increase by factors of 2.75 and 2.65 respectively, while, in the poorer provinces of Gansu and Guizhou, benefit rates would have to rise by 59 and 44 per cent respectively. It is striking and reassuring to note that the low-income lines adopted by certain cities and provinces are of the same order of magnitude, being approximately double rates of dibao. This suggests that policymakers generally feel that incomes below this level are no longer acceptable in today’s China. However, in the less developed Tibet Autonomous Region, the level of dibao is already 23 per cent higher than the 40 per cent of average disposable poverty threshold. This possibly reflects policymakers’ awareness of higher living standards elsewhere in China and may hint at future pressures towards adopting a single poverty standard as may be happening in Europe.37

There are, though, arguments against linking social assistance and relative poverty thresholds. A recession (which can reduce income inequality), or a new strongly progressive income tax system, might reduce median disposable incomes leading to a fall in the poverty threshold and hence a reduction in benefit levels which might prove administratively and politically difficult to implement. Global experience suggests that governments increase poverty thresholds as national income rises and the financial cost of moving from, say, 40 per cent to 50 per cent of median income might appear prohibitive especially as the recorded poverty rate would rise during the transition. Linkage of benefit levels to policy thresholds might also increase the chances of administrative gaming to exaggerate performance against the poverty target, thereby reducing the reliability of statistics. Furthermore, poverty, as noted above, is multifaceted and multidimensional and targeting policy on one measure may constrain overall policy effectiveness. This is because different income thresholds arguably index varying forms of poverty with unique dynamics that respond to distinct forms of policy intervention. Also, limiting the policy focus to income poverty means that the important social, human capital and relational dimensions of poverty are ignored.

## Vertical coverage of social protection

Despite the adoption of relative poverty standards by many governments, it is still often the case that minimum income benefits fall short of offering full protection against relative poverty (Figure 7). Sometimes the reasons are technical such as other systems and in-kind services being in place to bridge the gap between benefits and adequate living standards. On other occasions, the failing is due to political factors, concern about cost and, rightly or wrongly, fears of creating work disincentives. In response, Sustainable Development Goal, Target 1.3 is designed to ensure that all national governments ‘implement nationally appropriate social protection systems and measures for all, including floors, and by 2030 achieve substantial coverage of the poor and vulnerable’.

Figure 7 Adequacy of minimum income benefits latest year 2015-19, OECD38

Source: OECD (2020)

A first policy response to relative poverty is the implementation of a social assistance system sufficient to guarantee that beneficiaries can live with dignity. However, there needs additionally to be a comprehensive social protection floor to maintain incomes and to prevent people slipping into poverty for reasons such as illness, maternity, employment injury, death of a spouse or caregiver, funeral expenses, invalidity, and unemployment. Likewise, there will often need to be active provision to assist people to return to self-sufficiency, for example, measures to enhance human capital and to facilitate employment.

It is important to keep the level of contingency-related income maintenance provisions closely aligned with social assistance minima to sustain a social protection ecosystem that ensures levels of relative poverty are kept to a minimum, that insurance benefits deliver additional incomes commensurate with the level of contributions, and that work disincentives are avoided.Contingency benefits need to be such as to maintain the beneficiary’s family in health and decency. This could be achieved by, for example, providing earnings-related benefits, fixed to a prior wage or as a percentage of the minimum or living wage;39 this would ensure that beneficiaries’ normal living standards are not unduly disrupted and that their resources do not fall below the relative poverty threshold.40 Benefit levels would need to be reviewed following substantial changes in the general level of earnings where these resulted from substantial changes in the cost of living.41

Social insurance provisions, viewed in the context of relative poverty that must be addressed through distributional policies rather than economic growth, serve not only to redistribute income over people’s lifetimes but also between individuals within different age cohorts—intergenerational redistribution—and within the same age cohort—intragenerational redistribution. They serve to minimize the risk of reduction of living standards resulting from insured contingences (social assistance compensating in the event of the occurrence of uninsurable exigencies), while also protecting lifetime consumption levels. With social assistance supporting incomes at - or somewhat above - relative poverty, social insurance protects living standards above this level either by paying benefits in full, which is preferable on grounds of administrative efficiency and minimising beneficiaries compliance costs, or by supplementing social assistance payments. The proviso is that workers’ minimum wages are pitched at a level above the poverty threshold and that wage differentials and benefit replacement rates are arranged to ensure that insurance benefits generate incomes above the poverty threshold. With poverty defined in relative terms, the relativities between wages, welfare benefits and the poverty threshold constitute a partially self-regulating system, but one that requires careful monitoring and intervention to prevent differentials moving out of constructive alignment.

## Horizontal coverage of social protection

Moving to a relative poverty threshold consistent with China’s growing economic prosperity will inevitably change both the predominant sociodemographic characteristics of people experiencing poverty and the character of the poverty experienced.

Ongoing monitoring of relative poverty by age, group, sex, geographic area, origin, etc. can serve to identify those groups that are falling behind others in society.42 Initially, priority should be given to determining the life experiences of peri-poor persons, those brought into poverty due to the new definition. They are likely to be younger and better educated than people currently counted as poor, to live in towns or urban areas, to have children and to suffer recurrent spells of poverty. They may experience poverty for several reasons: because of low wages, unemployment, or sickness; because of the onset, or continuance, of prolonged, disability; or/and because of exclusion from social protection schemes. Such exclusion results from partial coverage of schemes due to design, institutional constraints such as hukou and fragmentation with diverse provincial and city level schemes and different modes of implementation, and because of a failure of people to enrol themselves or to be enrolled by employers.43

Recognition that relative poverty has to be addressed by a comprehensive social protection floor accords with the ILO concept of Universal Social Protection that encompasses income redistribution through social solidarity and a mix of instruments to provide income security to allow a life in dignity.44 Moreover, investment in such schemes brings macroeconomic benefits in addition to enhancing individual welfare and social cohesion. Such spending helps the economy to bridge recessions and confront unpredictable contingences such as the COVID-19 pandemic, while the strategic policy goal of transitioning to an economy more driven by consumption is facilitated by transfers to those on low incomes who tend to spend more of their income than other groups. China has greater capacity than most countries to benefit from such measures with government spending on social investment being less than in many comparable countries while the saving rate by the lowest income groups is exceptionally high.

Relative poverty needs also to be mainstreamed.45 46 This means that the prevention of relative poverty should be made an explicit objective of all policy initiatives or, exceptionally where this is not possible, treated as a constraint on the choice of policy options. Fiscal policies should correspondingly be designed to support movements out of poverty and to sustain the redistribution of market incomes necessary to fund adequate social protection. While relative poverty is unlikely to be eradicated, in Europe, over the last decade, it has been kept below 6.5 percent (using the 40 per cent of median disposable income standard, [Figure 2]). Persistent poverty, on this basis, has not risen above 4.6 per cent. The ILO Social Protection Calculator suggests that, in China, spending equivalent to 5.8 per cent of GDP would be sufficient to ensure that that everyone had an income at or above the 40 per cent median disposable income threshold.47

# Continuing improvement

It is important not to confound the concept of relative poverty with its measurement.48 Relative poverty is understood to describe the state of people unable to participate in the normal life due to lack of resources. However, there can never be a definitive point measure of normal life, merely estimates of a range of possibilities that, in the absence of empirical evidence, are only roughly ‘captured’ as proportions of median equivalised income or expenditure. Income is equivalised on the presumption that children consume less than adults and that there are economies of scale achieved by people living together, but all equivalence scales are inherently imperfect and fail to reflect society’s diversity.49 Resources are usually measured by income, which is notoriously difficult to assess, especially for persons working in the informal sector or when self-production, both in goods and services, substitutes for cash production. Most national poverty estimates are based on household surveys that do not require the same quality of evidence from respondents as welfare agencies would demand from social assistance beneficiaries.

While the measurement of relative poverty is imperfect, there has been a history of increasing refinement mainly in the detailed specification of resources but also in acknowledging the social construction of new needs, for example with respect of child-care.50 51 This can be drawn on and adapted to the Chinese context.

Relative poverty is sometimes criticised for its insensitivity to short-term changes in the economic well-being and the counter-intuitive movements in the index that can be caused by macro-economic cycles. However, it is possible, as a supplementary index, to take a relative measure and anchor it in real terms to a chosen date, much as China’s rural poverty line currently is (Figure 2). Indeed, a strong case can be made always to use a range of indicators when measuring poverty and assessing the effectiveness of policies to address it.52 All indicators have limitations and poverty is multifaceted.

# Beyond relative poverty

While the shift to measuring relative income poverty is enormously significant in supporting the moderately prosperous society that China has become, the UN’s Sustainable Development Goals, as already noted, require that by 2030 all countries ‘halve poverty in all its dimensions according to national definition’. China’s poverty alleviation strategy, as previously noted, already considers food, clothing, housing, health, and education, although these are not explicitly conceptualised in relative terms. Internationally the UNDP’s Multidimensional Poverty Index (MPI)53 measures material deprivation, poor health and limited education, dimensions of poverty that can be targeted through universal social protection. However, it is recognised that the MPI is severely constrained by data availability54 and proof of concept research conducted with people experiencing poverty in Guizhou55 suggests, in accord with other international research, that poverty may have as many as nine dimensions that need to be taken into account in the design and implementation of policy.56 These include, in addition to familiar elements such as insufficient income, decent work, and material deprivation, relational dimensions - social abuse and exclusion, institutional maltreatment and powerlessness - and response dimensions: physical and emotional suffering, struggle, and resistance. While further research across China is required to validate these dimensions, they already serve as a valuable template for understanding the complexity of poverty in a modern society and point to the importance of process, in policy design and implementation, as well as outcomes in determining the most appropriate government response.57

# Conclusion

The decision by the Fourth Plenary Session of the 19th Central Committee of the Chinese Communist Party to focus on relative poverty is of utmost significance. It reflects China’s economic success and growing status in the world, reinforcing China’s leading role in addressing global poverty within the framework of the Sustainable Development Goals. But it has major domestic implications too, in all probability revealing that many tens of millions of Chinese people and families are denied their full potential by persistent and repeated periods of relative poverty.

The decision signals a new approach to poverty alleviation based on different forms of measurement and demanding the development of a comprehensive system of social protection that prevents poverty and supports the most disadvantaged at or above the new poverty line. It offers the possibility of promoting national unity by instigating a unified poverty standard embracing the whole country, while supporting a sustainable economy through fostering consumption led growth and protecting China against the destructive impact of global economic events.

‘Win-win’ outcomes are rare in public policy but China’s decision to make poverty relative holds out the possibility of being one of them.

# Annex

Table A: Implied increase in poverty and dibao thresholds to equate with relative poverty standards

 Multiplication factors Current thresholds New poverty threshold as % of per capital disposable income 40% 50% 60% A national poverty line Urban dibao 1.61 2.02 2.42 Rural dibao 2.31 2.89 3.46 Rural poverty line 3.82 4.77 5.73 Separate urban & rural poverty lines Urban dibao 2.19 2.73 3.28 Rural dibao 1.15 1.44 1.73 Rural poverty line 1.91 2.39 2.86

Table B: Whether to make poverty and social assistance thresholds the same

 Poverty and social assistance thresholds Make the same Keep separate Pros (advantages) Logical to equate policy goals and measures Poverty rate is a direct measure of policy effectiveness Aids poverty mainstreaming Aids policy coordination Sends clear message to the public and to potential applicants May increase take-up of social assistance because of clear messaging May encourage comprehensive coverage of social assistance Pros (advantages) Emphasises multiple dimensions and types of poverty Facilitates observation and investigation of divergent trends in different measures and types of poverty May encourage cross departmental cooperation Separates of statistical from political presentation Cons (disadvantages) Poverty – and therefore linked social assistance thresholds – can fall in a recession (if income inequality is reduced) making some recipients ineligible (unless benefits protected) Progressive tax changes can have the same effect Can lead to policy gaming and distort statistical returns Social assistance may not be designed to reach all beneath poverty line Cons (disadvantages) Separate studies and monitoring required to measure the effectiveness of social assistance and other related policies

# Key References

Shidong Wang (Ed.) 2020. “Continuing the Successful Fight against Poverty in China: Challenges for Social Governance”, Modernisation of Education and Social Governance: 18-29. Oxford: Prospects and Global Development Institute, Regent’s Park College.

Yang, Lichao and Robert Walker. 2020. “Poverty and Anti-poverty in post 2020: Review, Prospect and Suggestions”, Guizhou Social Sciences 362(2): 148-154.

Deeming, C. (Ed.) 2020. Minimum Income Standards and Reference Budgets: International and Comparative Policy Perspectives. Bristol Policy Press.

Gustafsson, B. & Ding, S. 2020. “Growing into Relative Income Poverty: Urban China, 1988–2013”, Social Indicators Research 147(1):73–94.

Yang, L. et al. 2020. “Determining Dimensions of Poverty Applicable in China: A Qualitative Study in Guizhou”, Journal of Social Service Research, DOI:10.1080/01488376.2020.1734712

ILO. 2019. Universal Social Protection for Human Dignity, Social Justice and Sustainable Development: General Survey Concerning the Social Protection Floors. Geneva: International Labour Conference 108th Session, https://www.ilo.org/wcmsp5/groups/public/---ed_norm/---relconf/documents/meetingdocument/wcms_673680.pdf

# Acknowledgements

While as authors we take full responsibility for the content of this Working Paper, we wish to acknowledge invaluable assistance from Fran Bennett, Jonathan Bradshaw, Alan Deacon, Stephen Jenkins, Li Mianguan, Paul Spicker and Yao Jianping. We also thank Claire Courteille-Mulder, Director at the ILO Office in Beijing, Shahrashoub Razavi, Director at the Social Protection Department, ILO Headquarters in Geneva and Luis Frota, Chief Technical Advisor on Social Protection at the ILO Office in Beijing, for their valuable feedback.

1

Research Report of the Standing Committee of the Standing Committee of the National People’s Congress on the Work of Poverty Alleviation — The Ninth Meeting of the Standing Committee of the 13th National People’s Congress on 26 February 2019].’ http://www.npc.gov.cn/ npc/xinwen/2019-02/26/content_2072766.htm

2

Xinhua (2019) Communique of the Fourth Plenary Session of the 19th Central Committee of the Communist Party of China, Xinhuanet, 31st October. http://news.xmnn.cn/xmnn/2019/10/31/100620623.shtml

3

Speech by Li Keqiang at the State Council Executive Meeting, April 21, 2020.

4

EPRCUS (2011) Xinhua: China raises poverty line by 80 pct to benefit over 100 mln, Washington, DC.: Embassy of the People’s Republic of China, 29th November.

5

Ferreira, F. et al. (2016) A global count of the extreme poor in 2012: data issues, methodology and initial results, Journal of Economic Inequality, 14:141–172.

6

First introduced in Shanghai in 1993 and implemented in all cities by 1999, Dibao was consolidated and extended to rural areas in 2007. Arguably the world’s largest cash social assistance system, national in scope, benefit thresholds are fixed at county level, with decisions on individuals’ applications nominally taken at town level following recommendations from villagers’ committees.

7

Lu, S. and Lu H. (2013) Can government aid help low-income groups out of poverty? Empirical Research Based on CHNS Data 1989-2009. Journal of Finance and Economics, 39(1):4-27 (In Chinese)

8

Zhu, D.(2019).Situation and Policy Solutions of Marginalized Poor Households (pinkun “bianyuanhu”de xiangdui pinkun chujing yu shizhi). People’s Tribune (renmin luntan) 07:58-60.

9

Wang, P.& Hu, F.(2007). Problems and Strategies of the Current Housing Welfare Policy of the Urban Low-income Families (dangqian woguo chengzhen dishouru jiating zhufang fuli zhengce de wenti yu duice tantao). Reform of Economic System (jingji tizhi gaige) 03:44-48.

★ Chen, L., Lie, Y., Tan, J., Zhou, Y., &Wu, K. (2010) Analysis of Housing Needs and Housing Security for Low-income Groups – Empirical Study from Guangzhou (dishouru jiating zhufang xuqiu tezheng yu zhufang baozhang yanjiu – laizi Guangzhou de shizheng fenxi). China Soft Science (zhongguo ruan kexue) 10:133-143.

10

Civil Affairs Bureau of Wuhan (2017) Implementation Measures to Identify Low-income Family in Wuhan (wuhanshi dishouru jiating rending shishi banfa), http://mzj.wuhan.gov.cn/zwgk_918/fdzdgk/ggfw/shjz/201710/t20171031_157952.shtml

11

Beijing Daily (2018) Income Threshold has been Moved up to 2120 yuan in Identifying Low-income Households in Beijing (Beijing dishouru jiating rending biaozhun tiaozhi 2120yuan), http://www.xinhuanet.com/local/2018-08/28/c_1123338088.htm

12

Ferreira, F. and Sánchez-Páramo, C. (2017) A richer array of international poverty lines. Washington: World Bank blog. 13th October. https://blogs.worldbank.org/developmenttalk/richer-array-international-poverty-lines.

13

Sen, A. (1985). Commodities and capabilities. Amsterdam: North–Holland.

14

SPC (2015) Portfolio of EU Social Indicators for the Monitoring of Progress Towards the EU Objectives for Social Protection and Social Inclusion. Brussels: European Union, Social Protection Committee Indicators Sub-group

15

Ravallion, M. (2016) The Economics of Poverty: History, Measurement, and Policy. New York: Oxford University Press.

16

Eurostat use the ‘OECD-modified scale’ which assigns a value of 1 to the household head, of 0.5 to each additional adult member and of 0.3 to each child. There is evidence that this generally understates the needs of children and overstates those of elders. See: Bradshaw, J. (2020) From normative budget standards to consensual minimum income standards in the UK Chapter 2 in Deeming, C. (ed.) Minimum Income Standards and Reference Budgets. Bristol: Policy Press.

17

Eurostat (2020) EU statistics on income and living conditions (EU-SILC) methodology -monetary poverty. Brussels: Eurostat. - 14/04/2020. https://ec.europa.eu/eurostat/statisticsexplained/

18

Prydz E. & Jolliffe, D. (2019) Societal Poverty: A global measure of relative poverty. Washington DC: The World Bank. http://datatopics.worldbank.org/world-development-indicators/stories/societal-poverty-a-global-measure-of-relative-poverty.html

19

Hirsch, D. (2019) A Minimum Income Standard for the United Kingdom in 2019. York: Joseph Rowntree foundation. https://www.jrf.org.uk/report/minimum-income-standard-uk-2019

20

Padley, M. (2020) A Minimum Income Standard for London 2019. London: Trust for London

21

NSBC (2019) National Data, Beijing: National Bureau of Statistics China, http://data.stats.gov.cn/english/index.htm

22

Wen, Y & Reinbold, B. (2018) Income and Living Standards across China, St Louis: Federal Bank of St Louis, 8th January. https://www.stlouisfed.org/on-the-economy/2018/january/income-living-standards-china

23

EC (2010) Europe 2020: A strategy for smart, sustainable and inclusive growth. Brussels: European Commission: Communication from the Commission, COM(2010) 2020 final, 3.3.2010

24

Gustafsson, B & Ding S. (2020). Growing into Relative Income Poverty: Urban China, 1988–2013. Social Indicators Research 147(1):73–94

25

Chen, S.& Ravallion, M. (2012) More Relatively Poor People in a Less Absolutely-Poor World. Washington DC: World Bank Policy Research Working Paper 6114.

26

In China, income was assessed against either mean rural or mean urban income as appropriate.

27

Wang, et al., (2020) Poverty and Subjective Poverty in Rural China, Social Indicators Research, https://doi.org/10.1007/s11205-020-02303-0.

28

Ward, P. (2016) Transient Poverty, Poverty Dynamics, and Vulnerability to Poverty: An Empirical Analysis Using a Balanced Panel from Rural China. World Development, 78: 541–553.

29

Vaalavuo, M. (2015) Poverty Dynamics in Europe: From What to Why. Brussels: European Commission Working Paper 03/.

30

ILO Social Protection Floors Recommendation, 2012 (No. 202)

31

Deeming, C. (2020) (Ed.) Minimum Income Standards and Reference Budgets International and Comparative Policy Perspectives. Bristol Policy Press.

32

Saunders. P. (2018) Using a budget standards approach to assess the adequacy of newstart allowance. Australian Journal of Social Issues, 53:4–17

33

Deeming, op cit

34

CRSP (2020) MIS in other countries. Loughborough: Centre for Research In Social Policy.

35

Goedemé, T., Zardo Trindade, L., and Vandenbroucke, F. (2019). A pan-European perspective on low-income dynamics in the EU Chapter 3 in B. Cantillon, T. Goedemé, and J. Hills (eds), Decent Incomes for All. Improving Policies in Europe. Oxford: Oxford University Press.

36

Disposable income varies by a factor of 4.7 between the richest (Luxembourg) and poorest country in Europe before adjusting for price differentials and by 3.9 between Shanghai and Tibet Autonomous Region

37

Dauderstädt, M. (2019) Addressing poverty and inequality in Europe , Social Europe, 15th January 2019. https://www.socialeurope.eu/poverty-and-inequality-in-europe

38

OECD (2020). Adequacy of minimum income benefits (indicator). doi: 10.1787/dcb819cd-en (Accessed on 12 June 2020) https://data.oecd.org/benwage/adequacy-of-minimum-income-benefits.htm

39

ILO Social Security (Minimum Standards) Convention, 1952 (No. 102)

40

Yang, L. (2018) The social assistance reform in China: towards a fair and inclusive social safety net addressing inequalities and challenges to social inclusion through fiscal, wage and social protection policies. New York: United Nations Headquarters, 25 - 27 June. https://www.un.org/development/desa/dspd/wp-content/uploads/sites/22/2018/06/The-Social-Assistance-Reform-in-China.pdf

41

ILO Social Security (Minimum Standards) Convention, 1952 (No. 102), (Article 65. Para. 10). https://www.ilo.org/dyn/normlex/en/f?p=NORMLEXPUB:12100:0::NO::P12100_ILO_CODE:C102

42

Eurostat (2020) Income poverty statistics. https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Income_poverty_statistics

43

Jiang, J., Qian, J & Wen Z. (2018) Social protection for the informal sector in urban China: institutional constraints and self-selection behaviour. Journal of Social Policy, 47(2): 335–357.

44

ILO: (2019) Universal Social Protection: Key concepts and international framework. Geneva: Social Protection for All Issue Brief ,April.

45

HABITAT (1998) Proceedings of the regional workshop on mainstreaming urban poverty reduction in Sub-Saharan Africa. Nairobi: African Forum om Urban Poverty, United Nations Centre for Human Settlements

46

ILO (2001) Mainstreaming Poverty Alleviation Strategies through Sustainable Rural Infrastructure Development. Geneva: Employment Intensive Investment Programme

47

Social Protection Floors Cost Calculator, setting target income to the national poverty rate times 3.82, 140 per cent of national median disposable income. https://www.social-protection.org/gimi/SPFCalculReport.action. 12th June 2020. China’s total spending on social protection was estimated at 7.7 per cent of GDP in 2015 according to the Asian Development Bank (https://spi.adb.org/spidmz/)

48

Lister, R. (2004) Poverty. Cambridge: Polity Press.

49

CMC (2019a) Equivalisation in poverty measures: can we do Better? A technical paper of the Social Metrics Commission. London: Social Metrics Commission. https://socialmetricscommission.org.uk/wp-content/uploads/2019/12/SMC-Equivalisation-Report-2020-01-03-Web.pdf

50

Jenkins, P. (2020) Perspectives on Poverty in Europe. Following in Tony Atkinson’s Footsteps. Italian Economic Journal (2020) 6:129–155.

51

CMC (2019b) Measuring poverty 2019: A report of the Social Metrics Commission. London: Social Metrics Commission.

52

Atkinson, A. (2019) Measuring Poverty around the World. Princeton: Princeton University Press.

53

Alkire, S. et al., (2019) Global Multidimensional Poverty Index 2019: Illuminating Inequalities. New York: UNDP and OPHI.

54

Alkire, S., & Santos, M. (2014). Measuring acute poverty in the developing world: Robustness and scope of the Multidimensional Poverty Index. World Development, 59: 251–274.

55

Yang, L. et al., (2020): Determining Dimensions of Poverty Applicable in China: A Qualitative Study in Guizhou, Journal of Social Service Research, DOI: 10.1080/01488376.2020.1734712

56

Bray, R. et al., (2019). The hidden dimensions of poverty. Paris: International Movement ATD Fourth World; Bray, R. et al. (2020) Realising poverty in all its dimensions: A six-country participatory study, World Development. In press

57

Yang, L. Walker R. with Chen, J. and Zhang, R. (2019). Universal child benefit and dignity and shame. New York: UNICEF.