«D I R E C T I O N S I N D E V E LO P M E N T Human Development Public Disclosure Authorized The Elderly and Old Age Support in Rural China Challenges ...»
Vulnerability to Poverty Vulnerability to poverty over time is considerably higher than actual poverty at any one point in time. Following an approach used in the World Bank’s China poverty assessment (Chaudhuri and Datt 2009), this book determines a household to be vulnerable to poverty if it is poor in one of three survey years over a six-year period. Figure 2.3 shows the share of the rural population vulnerable to poverty during early and late periods of the CHNS panel (1991 to 1997, and 1997 to 2004).4 Because rural incomes are quite variable, the probability that a household experiences negative shocks and incomes below the poverty Poverty and Vulnerability among China’s Rural Elderly 33 Figure 2.2 Share of Rural Elderly Households Experiencing Chronic Poverty
Sources: CHNS 1991, 1993, 1997, 2000, 2004.
Note: A household is determined to be chronically poor if it has income below the poverty level in each of three survey rounds of the designated period. The incidence of vulnerability is calculated by the age of the oldest household member in 1993 for the 1991–97 period and in 2000 for the 1997–2004 period.
line in one year of a six-year period will be higher than annual income poverty rates.5 Rural households with older heads are more vulnerable to poverty, and declines in vulnerability between 1991 and 2004 have been more pronounced for households headed by working-age individuals and least pronounced for households headed by individuals in their seventies. Nearly 63 percent of households headed by an individual in his or her seventies in 2000 experienced income below the basic needs line in 1997, 2000, or 2004. This occurred for fewer than 40 percent of working-age adults and only 45 percent of households headed by 34 The Elderly and Old Age Support in Rural China Figure 2.3 Share of Rural Elderly Households Vulnerable to Poverty
Source: CHNS 1991, 1993, 1997, 2000, 2004.
Note: Vulnerability occurs if a household’s income falls below the poverty level in at least one of three years. The incidence of vulnerability is calculated by the age of the oldest household member in 1993 for the 1991–97 period and in 2000 for the 1997–2004 period.
someone in their sixties. Vulnerability was also fairly low among households with heads in their eighties, probably because rural residents in their eighties nearly always coreside with younger family members.
“Head of household” in these cases more likely reflects ownership of the dwelling and not primary income earner status in the household.
Poverty and Vulnerability among China’s Rural Elderly 35 The Geographic Dimension of Poverty The importance of location as a determinant of poverty has declined for all households since the 1990s, but the decline is higher in absolute terms for households headed by elderly people. Early in the reform era, spatial differences in incomes and poverty incidence were largely driven by restrictions on migration reinforced by differences in per capita land endowments, access to urban markets, and level of local development.
With China’s economic reform, the importance of geography has declined in explaining poverty and inequality (Benjamin, Brandt, and Giles 2005), while inequality within communities was increasing at an even faster rate as individuals differed in their ability to take advantage of wider off-farm labor markets with migration. Young adult migrants may provide support to elderly parents through remittances, so one might observe that geography becomes less important for explaining inequality and poverty among the elderly. In table 2.2, results show a decline in the share of variance in income that can be explained by village location from 1993 to 2000, with a slight increase after 2004 for all households and for households with elderly, respectively.6 The degree to which location explains variation in poverty status declined from 26 percent in 1991 to 12 percent in 2004.
The decline in importance of location in explaining poverty reflects the importance of increasing market integration as a means of raising incomes for households in villages that were previously isolated.
When one looks at households with elderly residents, location obviously was more important in earlier years for explaining both volatility of income and poverty status. Also, the importance of location in determining poverty status among the elderly declined even more sharply, from 40 percent to 20 percent, between 1993 and 2004. This finding suggests that even if the rural elderly are unable to directly benefit from
employment in distant markets, they enjoy some spillover benefits as family members remit incomes earned as migrant workers.
Nonetheless, location remains a stronger determinant of welfare among the rural elderly than among all rural households. The fact that 20 percent of the variation in elderly poverty status can be explained by village location as late as 2004 suggests that, relative to working-age adults, elderly persons remaining in home villages are less likely to have benefited from the increase in opportunities resulting from China’s economic reforms.7 Income Determinants of the Elderly Apart from geography, the incomes of the elderly may vary with common shocks experienced by the community, with their own health status or the health status of other household residents, and with the migration status of adult children. Indeed, a growing literature on migration in China shows how migration has contributed to raising the incomes of rural households and facilitated smoothing the effects of shocks to income.8 The rural elderly may not benefit from increases in rural-tourban migration if the elderly are members of poorer households that are less involved in migrant employment or if migrant children are not remitting a share of their income.
The results in table 2.3 highlight the effects of three policy variables on household income per capita: human capital (the average education of adult household members), formal social support (whether a family member receives a pension), and migration (whether the adult child of the household from a previous round is now working elsewhere as a migrant). The main findings in this regard follow.
• Increasing the average education of adult members of elderly households by one year is associated with a 7.6 percent increase in income per capita.
Thus, higher human capital among adults in a household may improve not only the well-being of working-age adults but also the well-being of their elderly parents.
• Receipt of a pension is associated with 67 percent higher household income per capita. In observing the correlation between pension receipt and household income, however, one must remember that few households in rural areas received pensions at the time of the surveys and that receipt of a pension was likely to be correlated with other household characteristics that lead to higher incomes (for example, high human capital, prior military service, or experience as a village cadre).
Poverty and Vulnerability among China’s Rural Elderly 37
• The effect of having a migrant family member suggests a 5.7 percent increase in income, but this effect is statistically insignificant. This result is somewhat surprising, so one must examine further whether having a migrant family member is significant at times of communitywide shocks.
Although the effect of having a migrant household member on household income is inconclusive, a migrant child provides insurance against loss of income stemming from shocks to the local economy. As shown in table 2.3, the coefficient for the interaction between the occurrence of a community income shock and having a migrant adult child in the household is significantly negative, indicating that having a migrant child can mitigate the effects of a communitywide shock.
Another major source of shock to incomes comes from illness, either to oneself or to another elderly member of the household. Health status can be captured using objective measures, such as activities of daily living or body mass index, but a stronger indication of the likelihood that a 38 The Elderly and Old Age Support in Rural China household member has been incapacitated is subsequent mortality.
Typically, individuals who die experience periods of incapacitation a year or two before passing away. This analysis therefore uses death of a household member one year following a survey round as a proxy for the presence of a family member who is too ill to work and examines the impact on household income.9 The effects of ill health of the elderly on household income are substantial but are more than compensated for where households have absent migrant members, because of transfers from these migrant household members. Households experience a 37 percent drop in income per capita in the year before the death of an elderly resident. This drop is likely driven by the lost earnings of the incapacitated family member and other family members who provide care to that person. For households with a migrant family member, however, an increase in income of 65 percent occurs in such circumstances. When households have a migrant family member and a resident suffers from a serious illness, migrant children contribute sufficient funds to reduce lost earned income of the household and to support additional expenditures related to health care.
Although migrant children help smooth the effects of shocks to income and health, these effects might be predominantly experienced by households either below or above the poverty line. To examine whether insurance against shocks affects transitions into poverty, the analysis examined how shocks and migration status affect the probability that a household falls below the poverty line (table 2.4).10 The findings are as
• Raising the average education of adult members of the household by one year leads to a 6 percent decline in the probability that a household is in poverty.
• Those households in which a family member has pension income are 58 percent less likely to be in poverty.
• Elderly households with migrant family members are 26 percent less likely to have incomes below the poverty line. When this result is contrasted with the negligible effect of migrant family members on income levels shown in table 2.3, it suggests that migrant family members may play an important role in keeping households out of poverty even if they do not raise incomes much above this threshold. Although they may not boost income for the average household, they may play an important role in raising incomes of households that would otherwise fall below the poverty line.
Poverty and Vulnerability among China’s Rural Elderly 39
• Migrant children may help insure against poverty generally but not provide additional insurance against poverty that is caused by communitywide shocks. Negative shocks to the local economy raise the likelihood that a household with elderly residents will fall into poverty. This finding is evident from the results for model 2 in table 2.4, where adding a communitywide income shock suggests that a 1 percent negative income shock experienced by the community will raise the probability that a household with elderly residents will fall into poverty by 0.14 percent.
In contrast to results in table 2.3, one does not observe that migrant family members yield significant insurance against the effects of communitywide shocks. This result is reflected in the low and insignificant coefficient on the community shock–migrant family member interaction term in model 2, indicating that existence of migrant family members does not appear to insure against the possibility that communitywide shocks drive households with elderly into poverty.
• Elderly households with a resident who is seriously ill are 25 percent more likely to have incomes below the poverty line, but having migrant 40 The Elderly and Old Age Support in Rural China children removes this effect.11 This finding is evident from the coefficient on the variable “elder died in the following year” in model 2 of table 2.4.
• These estimates also suggest that migrant children eliminate the effect of serious illness on the probability of falling into poverty. Migrant family members either return to care for elderly parents who are infirm (Giles and Mu 2007), or a migrant child is able to provide transfers sufficient to keep the family income stable when an elderly resident becomes seriously ill.
On balance, the evidence raises concerns that shocks to the local economy and serious health shocks experienced by household members affect incomes of the elderly. Given the rural pension pilot, one potential value of this formal social insurance would be to provide a base level of support that would keep China’s rural elderly from falling into poverty.
In principle, the rural dibao already plays this role, but coverage is limited and the manner in which it is targeted and implemented may further limit its effectiveness. Local village leaders responsible for targeting households to receive dibao support may have less skill at identifying households that suffer from transient poverty rather than chronic poverty. For older workers, transient poverty caused by a shock may quickly develop into chronic poverty because aging workers are less able to find new income-earning opportunities after falling into poverty.
Conclusion The evidence in this chapter indicates that the rural elderly are on average poorer, more likely to remain poor, and more vulnerable than the younger population, as well as substantially poorer than the urban elderly. Location also matters more for elderly than younger adults as a determinant of poverty. In terms of the factors that affect the incomes of the rural elderly, those who have more education or a pension are likely to have higher incomes. The effect of having a migrant child is, however, more complex.
Although having a migrant child has an inconclusive impact on the income of elderly rural households, having migrant children has a clearer and positive effect in terms of ability to cope with shocks to household income, whether communitywide or household specific.