The vulnerability of household income to health shocks

Health daze could hold a assortment of inauspicious impacts on family public assistance, such as diminishing household income, cut downing family ingestions or even forcing a family into poorness. It is province that “ although there are many macroeconomic, climatic dazes and natural catastrophes could do families steal into poorness, wellness dazes are by and large considered one of the most of import triggers of poorness. ( Dercon 2003 and Hoogeveen et al.2004 ) ” So how to protect families from wellness dazes is ever a critical issue for policy shapers and have a significance in bettering societal public assistance.

There are many surveies which have been conducted on the subject, particularly for both developing and developed states. They examine different sorts of economic effects of wellness dazes and happen a relatively indistinguishable general decision that wellness dazes influence much family wealth and income ; partially via increased medical outgos ; nevertheless, the consequence on the family ability of smoothing ingestion is ill-defined. For developing states, the literature in this field is relatively little. the lower public assistance degree and deficiency of a wellness insurance system make research on exposure of wellness dazes more meaningful for a underdeveloped state, such as China, which has undertaken a reform in wellness system and wellness insurance.

Since the economic effects of wellness dazes include many facets, this paper focuses on the consequence of wellness dazes on family income. Among possible economic effects, income decrease may be an immediate consequence of inauspicious wellness dazes before wellness dazes affect ingestion and family wealth.

The measuring of wellness which I use in this paper is ADLs ( Activities of Daily Living ) , Stewart rt Al. ( 1990 ) , Strauss et Al. ( 1993 ) and Gertler and Gruber ( 2002 ) province “ that ADLs have been proven dependable and valid steps of physical operation ability for both developed and developing states and separate the type of serious exogenic wellness jobs that are correlated with alteration in labour market and ingestion chances ” .

The information from the China Health and Nutrition Survey ( CHNS ) in twelvemonth 2004 and 2006 are used to analyze the consequence of wellness dazes on Chinese household income.

Since China has transferred medical funding from the manus of the authorities to the market, it is really interesting to analyze the consequence of wellness dazes on family public assistance utilizing informations from 2004 and 2006, because family may hold troubles in covering with wellness dazes in absence of societal insurance which was in topographic point before the reforms.

Furthermore, in order to be more specific about which groups of people are more vulnerable to wellness dazes, I besides divide families into different groups, which are male and female ; rural and urban countries, different household income degrees at the basedline ( 2004 ) and working and non-working persons. We divide families into the above groups for the undermentioned grounds. First, because ADLs chiefly measure the physical disablements of people, ADLs are more suited for mensurating the ability in manual work than in sedentary work. For this ground, we expect that the relationships of alterations in ADLs and alterations in income are different between males and females and between rural and urban countries. Second, the consequence of wellness dazes on family income may be different across family with differing wealth degrees. Furthermore, because our mark family members are age 55 or elder, some of them are already retired and some of them are still at work. The sum of pensions does non depend on wellness position. Therefore, wellness dazes to working versus non-working members would ensue in different effects in income decrease.

The aims of this paper is to analyze the relationship between wellness and household income, and to look into which population sections are more prone to income decrease in response to wellness dazes.

The remainder of the paper is organized as follow. In the 2nd subdivision, review the most relevant literature in the undermentioned two Fieldss: wellness dazes measuring and economic effects of wellness dazes. In the 3rd subdivision, informations used in this paper are described. In the 4th subdivision, we describe the methodological analysis used in this paper. The 6th subdivision discusses the consequences. Finally, decisions and future research suggestions are provided in the last subdivision.

Literature reappraisal

Previous surveies which use family study informations to analyze the economic effects of wellness dazes can be divided into two types by utilizing different wellness daze measurings. One type uses self-reported wellness daze measurings and the other usage nonsubjective wellness daze measuring.

Self-reported wellness daze measurings include alteration of self-reported general wellness position, freshly added unwellness or disease, self-reported symptoms, wellness dazes. Decron, Hoddinott and Woldehanna ( 2005 ) utilize a household-level “ daze ” faculty. The faculty asks family asks family to see a list of inauspicious events and indicate whether the family was adversely affected by them.

Self-reported wellness dazes have been widely used but of class without jobs. But the major failing of self-reported wellness daze measuring is that replies given by study respondents are easy influenced by their subjective attitudes or ain backgrounds. John Bound ( 1991 ) , Strauss and Thomas ( 1996 ) province that respondents is likely to over-state the consequence of wellness position on labour supply, since persons is inclined to warrant their left occupation determination by describing a impairment of wellness even the true ground is non this. And Gertler and Gruber ( 2002 ) province that persons with different economic position and different instructions may hold different definitions of unwellness.

Comparing with self-reported wellness daze measurings, nonsubjective wellness daze measurings such as outgrowth of a new serious wellness status ( Smith 1999 ) , decease of a family member, alteration in BMI ( Body Mass Index ) , physician ‘s diagnosing with a waiter disease are less capable to mistakes which may do prejudices in gauging the relationship between wellness and socioeconomic position ( Wagstaff 2005 ) . The oncoming of damage and disablement ( Stephens ( 2001 ) ) and alteration in activities of day-to-day life ( Gertler and Gruber 2002 ) are typically considered as nonsubjective. Objective wellness steps seems less influenced by the respondents ” ain grounds descripted above. But the consequence still could be biased by the omitted variables job. Prescott and Pradhan ( 1999 ) points out that the possibility that a hapless families experience wellness dazes is much higher that rich 1s. And there is a correlativity between wealth and current outgo and besides between wealth and wellness dazes, which could make a specious reasoning backward consequences. In order to get the better of this, we use alterations of an nonsubjective wellness step as a wellness step in this paper. Specifically, we use actibities of day-to-day life ( ADLs ) as a wellness step. The cogency of ADLs have been discussed in the literatures. ( Anita, Stewart et al.1990 ; John strauss et Al. 1993, Gertler and Gruber ( 2002 ) and Bound et Al. ( 1991,1995 ) ) .

Hairi et Al. ( 2010 ) uses informations and shown that ADLs is valid mensurating wellness in developing states. ADLs has been shown to be a dependable and valid step of physical operation abilities to both developed and developing states, and separate the type of serious wellness jobs from less serious 1s. Gertler and Gruber ( 2002 ) usage Indonesia family panel informations and analyze the extent to which households are able to see ingestion against major unwellness utilizing alterations in ADLs as a wellness step. Furthermore, every bit far as I see, there seems no pervious paper which has of all time used alterations in ADLs to analyse the impact of wellness daze for China.

By utilizing these two types of measurings, the pervious literature chiefly examined the consequence of wellness dazes on labour supply incomes, wealth, ingestions and medical outgo.

Most of these surveies found that wealth is negatively influenced by wellness dazes. Levy ( 2002 ) finds that in American, a new diagnosing could cut down a family wealth every bit big as about of $ 28,000 for both insured and uninsured families. Smith, J, P ( 1999 ) besides shown that in center and at older age, a new wellness status could bring forth a marked effects on family wealth. Furthermore, Lee and Kim ( 2008 ) besides estimate the impact of wellness dazes on wealth non merely on the short-run but besides on the long-run. And they find that by and large new wellness events lower wealth in seniors when it is happened and the impact tend to vanish in long-run. However, if the wellness dazes happened later in life, the impact of lower the wealth will increase over clip.

Evidences has been accumulated that income reduced by the wellness daze. T.Riphahn ( 1998 ) finds that in German, if a on the job member in a family loses employment due to a terrible wellness job, benefits provided by the authorities can non do up the loss. And Levy ( 2002 ) states that wellness dazes could cut down family incomes by about $ 9,000, and no affair the family is insured or uninsured, wellness dazes cut down income by approximatelt the same sum. Furthermore, Smith ( 1999 ) utilizing American information, besides finds that the net incomes could diminish about $ 2,600 by a terrible new wellness oncoming. However, there are besides a few surveies which find merely a little diminution in household income in the period following a wellness daze ( Riphahn ( 1998 ) ) . The consequences besides vary between states. For illustration, Burkhauser and Daly ( 1994 ) after comparing the economic effect of work-related wellness damages of the US and German, argue that the household income is same no affair the Germans has or has non disablements, while more than 25 % income spread between disablements workers and non-disabilities worker in American.

In developed states, it may be that wellness dazes could act upon the ingestion in a long-term ; while in a short-term, families could smooth their family ingestions. Stephens ( 2001 ) finds that there is a long-run diminution in ingestion if the caput of a family is disabled. Long-run ingestion is on norm reduced about 5 % by a disablement ; while there seems no instantly consequence on ingestion. Furthermore, levy ( 2002 ) besides find no consistent consequences about whether family ingestion remains smooth in the face of serious wellness dazes, and the whether they smoothing by cut downing wealth or by utilizing insurance.

For developing states, the literature on economic effects of wellness dazes is little ( one of such surveies from developing states is Wagstaff ( 2005 ) ) . Due to miss of formal and effectual insurance and public assistance systems, most groundss from the old surveies give the decision that families are unable to get by with inauspicious wellness dazes. ( Dercon, Hocldinott and Woldhanna ( 2005 ) ) by comparing a figure of wide classs of inauspicious dazes, such as climatic, economic ; political/social/legal, offense and wellness dazes, concludes that drouth dazes and unwellness dazes are the most of import dazes that are associated with lower degrees of per capita ingestion. The consequence of wellness daze seems more uncertainness in developing states than developed states. Wagstaff ( 2005 ) examined this issue in the context of post- reform Vietnam, and found that wellness dazes have a negative consequence on income, but they differ between earned and unearned incomes and between rural and urban families. In African, Emily, Wendy and Jacques ( 2011 ) examine HIV-related wellness dazes and happen that economic effects of wellness dazes are server for uninsured families, even they resort to a assortment of get bying schemes to cover with high medical disbursals and decreases in income. Besides in African, Simon ( 2010 ) find that illness merely has negative effects on ingestion in short-run. For ingestion, most consequences come from developing states are non statisticant, important. It seems that “ families are able to utilize precautional nest eggs, informal insurance and other establishments to smooth ingestion in the face of idiosyncratic dazes. “ ( Morduch ( 1995 ) , Townsend ( 1995 ) , Dercon ( 2002 ) ) .

Most old surveies in this field divide families into insured and uninsured groups, since this would be a good manner pull some policy deductions. For illustration, Smith ( 1999,2003 ) and Levy ( 2002 ) finds that “ the impact of wellness dazes on income and wealth are non significantly different between the insured and uninsured. ” As we know, the original wealth could assist families to defy against the unexpected inauspicious dazes, and higher incomes are ever related with higher instructions. There are besides some surveies which examine differing effects of wellness dazes on the rich and hapless. Dercon, Hocldinott and Woldhanna ( 2005 ) find that “ illness dazes appear more of import for richer families ( as measured by comparative landholdings ) and family where the caput has no schooling. ” And Hurst ( 1995 ) , O’Donnell and Propper ( 1991 ) , Benzeral ( 1997 ) , Benzeral and Judge ( 1998 ) province that “ In the early 1990s, working-age persons whose household income is in the bottom 20 per centum of the national distribution are over 3-5 times more likely to describe their wellness as hapless than those in the top fifth. ” Another manner dimension in which economic experts are interested is the difference bwtween males and females. Macran and Colleagues ( 1996 ) reference that “ Income is associated positively and significantly with adult females ‘s wellness. ” Charles ( 1997 ) finds that “ a family does non needfully depend wholly upon the income of a individual worker. Spouses and /or other family members can increase their work attempt to assist countervail the family caput ‘s earning losingss and multiple authorities plans are designed to assist households. ”

For China, empirical surveies on economic effects of wellness dazes are smaller in figure. One such article was written by Lindelow and Wagstaff ( 2005 ) . The writers focus on “ how alterations in self-assessed wellness of a family caput affect incomes, family labour supply and medical outgos ” . The consequences show that wellness dazes significantly cut down incomes and labour supply. However, the decrease of income is due to the lessening of earned income. Unearned incomes is addition when inauspicious wellness dazes occur. Furthermore, they strong grounds that wellness dazes increase out-of pocket medical outgos but the addition is smaller for the hapless. For ingestions, they found no systematic diminutions in response of wellness dazes.


1. Data beginning

The dataset used in this paper comes from the China Health and Nutrition Survey ( CHNS ) . The study is conducted by the Carolina Population Center and the National Institute of Nutrition and Food Safety at the Chinese Center. This is an on-going longitudinal study, which is used examine wellness, nutrition, and household planning policies and plans implemented by national and local authoritiess and how the wellness and nutritionary position of Chinese population are affected by the societal and economic transmutation of the society. ( CHNs Website ) Until now, there are eight moving ridges of study informations provided from 1989 to 2009, which includes a sample of about 4,400 family with 26,000 persons in nine states.

2. Sample used in analysis

This paper selects the most recent two old ages ‘ informations ( 2004 and 2006 ) to carry on empirical analysis. Since ADLs information is provided for family members whose ages are 55 or elder, the dataset merely include families where at least one family member ‘s age is above or equal to 55 old ages old. Different household members contribute otherwise to household incomes, but family caputs would give the largest part to the family entire income in most families. However, because both hubbies and married womans work to gain incomes in many families in China, this does non intend that hubbies are exclusive breadwinners. Further, some of Chinese households are large households, which consist of a male parent, a female parent, boies, girls or even grandsons, granddaughters. A family caput is typically the eldest individual, but major income subscribers may non the family caput. To turn to this instance, we examine the consequence of wellness dazes to both family caputs and all family members including household caputs.

We restrict our sample families as follows to decease, migration, separation and study mistakes,

a™¦ Households were interviewed in both 2004 and 2006 ;

Families whose family caputs did non alterations between 2004 and 2006.

The figure of sample persons whose ages are 55 or older and who have complete information on variables used in this survey is 2,530, which occupy about 9.2 % of the original longitudinal samples from 2004 and 2006. Out of 2530 sample persons, 1,302 persons are household caputs.

Measurement of wellness

Changes in ADLs is the chief independent variable used in this survey. Because of informations restriction, ADLs are merely available for family members of 55 twelvemonth or senior.

There are 20 ADLs inquiries in the CHNS inquiries and replies to these inquiries are used to cipher the ADL indices ( Table 1 ) . The replies picks are: ( 1 ) no trouble ; ( 2 ) have some trouble, but can still make it ; ( 3 ) demand aid to make it ; ( 4 ) can non make it at all ; ( 9 ) terra incognita. The “ unknown ” replies have been changed into losing value.

We normalize the replies to each ADLs inquiries to the standard divergences from the mean for each twelvemonth ; so the indices are the agencies of the normalized tonss over the 20 inquiries. The equation used to normalise reply pick is ( X-mean ) / ( standard divergence ) ( Ten: reply pick to an ADL inquiry ) . For illustration, the mean of inquiry “ running a kilometre ” of twelvemonth 2004 is 2.387417 and the standard divergence of this inquiry is 1.273882 in the dataset which merely uses family caputs. Suppose one family caput ‘s reply to this inquiry is “ 2 ” . Then the normalizing consequence is: ( 2-2.387 ) /1.273882 which about equals -0.3041. Since the “ no trouble ” answer figure is assigned the smallest figure 1, the larger the ADLs index, the worse of of wellness position of the indibidual.

Table 1 lists the mean and standard divergence of each activity. From this tabular array, we can see that there are some activities whose agencies are more closer to 1 and the standard divergences are smaller than some other activities ‘ , which indicates that a huge bulk of family members can finish these activities vary easy, so we divide these activities into two groups: ADLI and ADLII, based on the distance in the mean from 1. The ADLs activities whose agencies are the highest 13 are grouped into ADLI and whose agencies are the lowest 7 are grouped into be ADLII.

Measurement of income

Income used in this paper is per capital income which has been inflated to 2006. We besides use income and log per capital income as the dependant variable. We replaced zero or negative per capita income by minimal positive value of per-capita household income in sample.

Descriptive Statisticss

Table 2 and table 3 shows the agencies, standard divergences, upper limit and minimal figure of all the variables which are used in the arrested development analysis for family caputs and all family members, severally. Because the dependant and many independent variables are used as alteration in our arrested development analysis, alterations of income, alterations in log income, alterations in wellness and alterations in some control variables are besides reported in the tabular arraies.

First, allow us see table 2 which merely includes the information of family caput. We can see that the mean household income increased by 422.237 Yuans during these two old ages. The standard divergences are all really big and it is larger in twelvemonth 2006. The maximal figure of household income is approximately 13,000 times of the minimal figure of household income.

The average alteration of entire ADLs is 0.0019929, connoting that ADLs became worse between the two old ages but the magnitude of the alteration is really little. Further, the standard divergence is acquiring big, connoting that differences in ADLs became larger across families between the two old ages. ADLI mean is larger, nevertheless, ADLII mean became smaller, so this once more proved that the ADLI is more exactly than ADLII.

Second, from table3, we can see that for the family all members whose ages are 55 old ages or senior, mean household income increased and standard divergence besides became big. The entire ADLs and ADLIs became worse on norm. However, the ADLIIs is became larger. The proportion of females in this dataset is about 52 % , while the rural proportion is still approximately 64 % , and working persons are about 30 % .


In order to analyze the consequence of wellness alteration on per capital income alteration, the paper adopted the method used by Nguyen and Mangyo ( 2010 ) . The gauging equation signifier is:

a?† = +* ++ ( 1 )

where the inferiors i indicate each person ; a?† is the per capital income alteration of family between 2004 and 2006 ; is the alteration of wellness position, in this paper we use the ADL index as a placeholder of wellness, so it is ADLs alteration between 2004 and 2006 ; is a set of household — degree control variables at the baseline ( in 2004 ) are used. Furthermore, we use other household-level control variables which are alterations between 2004 and 2006. Household-level control variables in degrees and alterations are follows:

a™¦ Household size in twelvemonth 2004 and household size alterations between twelvemonth 2004 and twelvemonth 2006 ; ( variable names used in the consequence tabular arraies are hhsize04 and differ_hhsize )

a™¦ Proportion of 18 to 55 old ages old family members in twelvemonth 2004 and the alteration in the proportion of 18 to 55 between twelvemonth 2004 and twelvemonth 2006. ( variable names used in the consequence tabular arraies are pro54 and differ_pro54 )

a™¦ Proportion of family members whose ages are above 55 old ages in twelvemonth 2004 and the alteration proportion of 55 and senior between twelvemonth 2004 and twelvemonth 2006. ( variable names used in the consequence tabular arraies are pro55 and differ_pro55 )

a™¦ Individual ‘s age in twelvemonth 2004.

Further, in order to happen different effects of wellness dazes on family incomes for different types of persons, we group the sample persons into the undermentioned classs:

a™¦ Dividing family members into males and females ( utilizing variable female=1 if the family member is female, and female=0, otherwise. )

a™¦ Dividing families into rural and urban families based on their life topographic points. ( Using variable rural=1, if the family lives in a rural country ; rural=0, otherwise )

a™¦ Dividing family members into working and non working based on the inquiry “ soon working? ” in twelvemonth 2004. ( work=1, if the family member is working in twelvemonth 2004, and work=0, otherwise. )

Then we use the undermentioned theoretical account to analyze differential wellness effects for different groups of family member:

a?† = +* +a?†++ ( 2 )

Where, is the indication of a different group, such as female, rural and work. This theoretical account is the same as the equation ( 1 ) , except for two extra independent variables: a silent person group variable and its interaction term with. After adding this silent person variables and its interaction with wellness, the consequence of wellness alterations for different groups can be examined.

The dummy work steps whether the family member worked in tear 2004. If he or she worked, work=1 ; if he or she did non worked, work=0.

Since the consequence of wellness dazes on family income universe be larger for working persons than for non-working persons, we expect that interaction term produces a negative coefficient estimation.

Using the gauging equation ( 2 ) , we examine the consequence of wellness dazes on family income individually for rural and urban families. Further, we divided the sample families into 3 groups harmonizing to the initial degree of per capita family income in 2004. Then, we examine whether the consequence of wellness dazes on income is different across families with differing income degrees in 2004.

There are two possible jobs which we need to aware of, one is omitted variable ; the other is simultaneousness. As Gertler and Gruber ( 2002 ) explain in their paper, “ first-differencing expanses out correlativity from omitted unseen single features ( such as penchants and wellness gifts ) . ” The simultaneousness seems to be most terrible job in degree arrested developments. However, even “ if wellness dazes observed in our informations themselves reflect feedback from income dazes, such ( simultaneousness ) prejudice is likely to be little as grounds indicates that income effects on wellness are normally dominated by more lasting income instead than short-run income dazes. “ ( Liu et al. , ( 2007 ) Smith, 1999 )


1. General reasoning backward consequence

We begin with gauging the impact of wellness dazes to household caputs on per-capital family income where family caputs are 55 old ages and older. We use the ADL index which was created by all 20 ADL inquiries ( stdadl ) . We besides present the impact of wellness dazes to any family member on per-capita family income where family members are 55 old ages old or older. As the dependant variable, we use both alterations in per-capita income and alterations in log per-capita income. The consequences are shown in tabular arraies 4. We can see that no affair whether wellness dazes are to household caputs or any household member with age 55 and elder, the estimated coefficients are non important at conventional degrees.

A possible ground for these undistinguished consequences could be that some ADL activities may non be sensitive plenty to reflect wellness alterations which would impact labor-market results. So we examine the effects of the alteration of ADLI and ADLII individually on income alteration. From table 5 and 6, we can see that the mark of the coefficient on ADLI is negative and that the absolute value is larger than the coefficient on STDADLTOT. However, the estimated coefficients on ADLI and ADLII are still non important at conventional degrees. For wellness dazes to household caputs, 1standard divergence ( SD ) addition in STDADLTOT reduces income by 38.73 RMB, while 1 SD addition in ADLI lessenings household per capital income by 244.61 RMB. However, the mark of the coefficient on ADLII is positive and is non important. For the remainder of analysis, we use ADLI as the step of wellness dazes.

2. Arrested development consequences with interaction parts

In table 7, we examine the consequence of wellness dazes to male and female family members, separetely. From the arrested development consequences, we find a statistically important association ( at the 10 % degree ) between alterations in male family caput ADL and the alterations in husehold income. In footings of magnitude, a 1 SD addition in ADLI of male family caputs reduces household income by 737.92 RMB, which is about 11 % of mean per capita income ( the norm per capital income of twelvemonth 2004 was 6426.03 RMB. ) . For wellness dazes to any family member, the coefficients are non statistically important. The magnitude of the coefficient estimates indicate that male wellness influence household income more than female wellness. Health dazes to female family member are captured by the amount of the estimated coefficients ( D.adlI+D.adlI_female, which is shown at the underside of table 7 ) . None of the estimated consequence is statistically important at conventional degrees.

Turning to the urban and rural difference in table 8, we do non happen statistically important differences in the consequence of wellness dazes on per-capita income between rural and urban countries. However, table 8 is shows that- the coefficient estimates on the interaction term ( D.adlI_rural ) are all negative, connoting that the consequence of wellness dazes on per-capita income would be more serious in rural countries than in urban countries, although the consequence of wellness dazes on per-capita income in rural countries is non statistically important at conventional degrees ( the underside of table 8 ) .

Table 10 and table 11 show that the estimated impacts of wellness dazes on per-capita income for different families with differing initial per-capita income degrees in 2004. Table 10 is for wellness dazes to household caputs and table 11 is for wellness dazes to any family member with age 55 and elder. In table 10, we can see that rural families in the lowest initial income group are badly affected by wellness dazes to household caputs ( the underside of table 10 ) . 1 SD addition in ADLI reduces household income in 2004. However, for wellness dazes to any family member with age 55 or older, we obtained an unexpected consequences: Households in the richest initial income group suffered the largest from wellness dazes. A 1 SD addition in ADLI reduces household income by 18 % .

From the descriptive statistics, we already find that more than half of family members do non work any longer, which means that their incomes by and large do non depend on labour supply. This may be another ground why we saw many undistinguished consequences in our analysis above. To research this, we divided families into two groups, depending on whether family caputs work or non. The consequence is shown in table 12. For working family caput, 1 SD addition in ADLI reduces family per capital income by 2,232.75 RMB ( the underside of table 12 ) , and for any on the job family member, 1 SD addition in ADLI reduces family per capital income by 1,091.64 RMB. These estimated effects are statistically important at the 1 % and 10 % degrees, severally.

Consequences treatment

All in all, we do non happen strong grounds that wellness dazes cut down per-capita income, which is different from the decision by Lindelow and Wagstaff ( 2005 ) in which the writers utilizing the same information as this paper, declining of self-assessed wellness had s important and sometimes big negative impact on income. A possible ground of these inconsistent consequences of my paper and Lindelow and Wagstaff ( 2005 ) is that they use a subjective wellness step while I use an nonsubjective wellness step.

Mentioning to the interaction parts, we can see that wellness dazes on family income are different for different groups of families. For some groups the impact is visiable, while for other groups the impact may be negligible. By and large talking, males are still the chief income subscriber to household income. Health dazes to males cut down household income more than wellness dazes to females.

The most important consequences we get on this research is that wellness dazes to working family members cut down per-capita income much more than wellness dazes to working family members cut down per-capita income much more than wellness dazes to non-working family members. This is an expected consequence since the sum of pension reception is non influenced by wellness dazes.

Furthermore, in this paper, we do non see household income is earned or unearned. Earned income is supposed to be more associated with wellness dazes ; while the relationship between wellness dazes and unearned income is much loosen. Wagstaff ( 2005 ) states that in China, unearned income may well increase for many families due to gifts and remittals, from distant households and friends if wellness jobs hit household members.


In this paper, we examine the exposure of household income to wellness dazes. We explore this issue through gauging correlativities between alterations in ADLs and alteration family per capita income between twelvemonth 2004 and twelvemonth 2006 utilizing informations from the CHNS. The sample we used is limited to household members whose ages are 55 and elder. Although there are econometric designation jobs which bias our consequences, such as omitted variables jobs, we find some grounds of exposure of per-capita income to wellness dazes for some Chinese subpopulations.

First, we find that alterations in family income are significantly sensitive to alterations in ADLI and ATDADLTOT of working family members, while this consequence is non observed for non-working family members. This could be likely the grounds why the impact of wellness dazes can non be seen for the whole sample.

Second, the family exposure of income to wellness dazes is different across groups. Household income is more sensitive to alterations in ADLIs of makes than of females. For rural low income families, the consequence of wellness dazes to household caputs would mostly cut down household income.

Our findings imply that there may be subpopulations of China which are peculiarly vulnerable to wellness dazes. In contemplating policies against wellness dazes, it is of import to place which subpopulations are peculiarly vulnerable to wellness dazes.

Future research way

Overall, we find undistinguished correlativities between alterations in ADL and alterations in per-capital income. This paper does non turn to how families cope with wellness dazes. There are many possible methods to cover with wellness dazes, such as authorities public assistance plans, remittal and fiscal aid from distant households and friends, and alternate family members working alternatively of sick-fallen breadwinners. Depending on what get bying schemes are taken, household public assistance may diminish significantly although household income may non alter much. A farther way of research to inform policy devising in this field is to analyze how families cope with wellness dazes.