Presentation is loading. Please wait.

Presentation is loading. Please wait.

Paying the Piper: The High Cost of Funerals in South Africa Anne Case, Princeton University Anu Garrib, Africa Centre UKZN Alicia Menendez, University.

Similar presentations


Presentation on theme: "Paying the Piper: The High Cost of Funerals in South Africa Anne Case, Princeton University Anu Garrib, Africa Centre UKZN Alicia Menendez, University."— Presentation transcript:

1 Paying the Piper: The High Cost of Funerals in South Africa Anne Case, Princeton University Anu Garrib, Africa Centre UKZN Alicia Menendez, University of Chicago Analia Olgiati, Princeton University

2 Institutions that evolve over a long period of time often do so for many reason Institutions that evolve over a long period of time often do so for many reason Funerals, for example, Funerals, for example, Pay respect to those who have died Pay respect to those who have died Console the grieving Console the grieving Mark the social status of the dead and his or her household Mark the social status of the dead and his or her household Knit social fabric within extended families and within communities Knit social fabric within extended families and within communities Redistribute some of the deceased’s resources Redistribute some of the deceased’s resources

3 Large funerals may not put households at financial risk, when people die in old age, and so such funerals may be sustainable for a very long period Large funerals may not put households at financial risk, when people die in old age, and so such funerals may be sustainable for a very long period However, large funerals may have very different consequences when people begin to die in large numbers in prime age However, large funerals may have very different consequences when people begin to die in large numbers in prime age AIDS crisis has knock-on effects brought about by the fact that institutions developed largely to bury people in old age are now being applied to a great number of prime-aged deaths AIDS crisis has knock-on effects brought about by the fact that institutions developed largely to bury people in old age are now being applied to a great number of prime-aged deaths

4 The financing of burials will affect a household’s ability to improve their members’ life chances The financing of burials will affect a household’s ability to improve their members’ life chances to maintain a stock of productive assets to maintain a stock of productive assets to stake migrants in urban areas until they find work to stake migrants in urban areas until they find work to finance schooling, and more broadly to finance schooling, and more broadly to provide adequate nutrition and a healthy environment for children to provide adequate nutrition and a healthy environment for children

5 In this paper Document the cost of funerals in South Africa Document the cost of funerals in South Africa Explore how households make decisions on funeral spending Explore how households make decisions on funeral spending Discuss issues associated with reducing funeral spending, and possible ways forward Discuss issues associated with reducing funeral spending, and possible ways forward

6 This paper documents funeral costs and financing for deaths that occurred between 2003 and 2005 in the Africa Centre DSA This paper documents funeral costs and financing for deaths that occurred between 2003 and 2005 in the Africa Centre DSA Specifically, we analyze funeral arrangements following the deaths of 3,751 people who died between January 2003 and December 2005 Specifically, we analyze funeral arrangements following the deaths of 3,751 people who died between January 2003 and December 2005

7 What do funerals cost? On average, households spend the equivalent of a year’s total expenditure on food and groceries, measured at median household expenditure in the DSA On average, households spend the equivalent of a year’s total expenditure on food and groceries, measured at median household expenditure in the DSA Approximately one-quarter of all individuals had some form of insurance, which helped surviving household members defray some fraction of funeral expenses. Approximately one-quarter of all individuals had some form of insurance, which helped surviving household members defray some fraction of funeral expenses. However, an equal fraction of households borrowed money to pay for the funeral. However, an equal fraction of households borrowed money to pay for the funeral.

8 How do households determine appropriate spending for funerals? We build a model in which households respond to social pressure to bury their dead in a style consistent with the observed social status of the household and that of the deceased. We build a model in which households respond to social pressure to bury their dead in a style consistent with the observed social status of the household and that of the deceased. Households that cannot afford a funeral that meets social expectations must borrow money to pay for the funeral. Households that cannot afford a funeral that meets social expectations must borrow money to pay for the funeral. The model leads to empirical tests, and we find results consistent with our model of household decision-making. The model leads to empirical tests, and we find results consistent with our model of household decision-making.

9 Data Data Funeral costs Funeral costs A model of household decision-making A model of household decision-making Structural estimates from our model, and additional tests using our data Structural estimates from our model, and additional tests using our data Concluding thoughts Concluding thoughts

10 Africa Centre for Health and Population Studies, UKZN Since 2000, approximately 11,000 households (~100,000 people) in the Umkhanyakude District in northern KwaZulu-Natal have been under demographic surveillance Since 2000, approximately 11,000 households (~100,000 people) in the Umkhanyakude District in northern KwaZulu-Natal have been under demographic surveillance The surveillance site includes both a township and a rural area administered by a tribal authority The surveillance site includes both a township and a rural area administered by a tribal authority At six month intervals, demographic and health information is collected on all household members At six month intervals, demographic and health information is collected on all household members Individuals may be resident in the Demographic Surveillance Area (DSA), or may be non-resident members of households that claim them as members Individuals may be resident in the Demographic Surveillance Area (DSA), or may be non-resident members of households that claim them as members Approximately two-thirds of all persons under demographic surveillance are resident in the DSA at any one time. Approximately two-thirds of all persons under demographic surveillance are resident in the DSA at any one time.

11

12

13

14 Data sets Verbal autopsies Verbal autopsies Illness and Death (IAD) Survey Illness and Death (IAD) Survey Covering deaths January 2003-December 2005 Covering deaths January 2003-December deaths 3751 deaths Household Socioeconomic Surveys (HSE1 2001) and (HSE2 2003/04) Household Socioeconomic Surveys (HSE1 2001) and (HSE2 2003/04)

15 Burial societies/funeral policies 28 percent of the deceased had some form of funeral policy, or belonged to a burial society 28 percent of the deceased had some form of funeral policy, or belonged to a burial society For Rand a month, these policies pay out upon death For Rand a month, these policies pay out upon death Participation highly correlated with old-age pension receipt Participation highly correlated with old-age pension receipt

16 Table 2. Burial Societies and Funeral Policies BURIAL SOCIETY AND FUNERAL POLICIES Fraction with a policy0.284 Fraction pension-eligible with a policy Fraction non-pension eligible with a policy Number of observations3668

17 FUNERAL POLICY PAIDfractionMean amount Money for the funeral Coffin0.230 Food0.232 Transport0.087 Tent0.134 Number of observations1007

18 Table 3. Costs of Funerals Funeral purchases Fraction making purchase Mean All deaths (Rand) Mean, Funeral policy holders (Rand) Coffin Meat Groceries Tent Clothing Blankets Transport Other Total Rands Number of observations

19 Number of assets owned at HSE Density HSE1:number of assets

20 Total funeral spending and number of assets owned at HSE1 (2001) 0 2,000 4,000 6,000 8,000 mean of total funeral spending

21 Table 4. Accounting for Funeral Costs CONTRIBUTIONS TO FUNERAL COSTS ( RAND) Fraction Contributing Mean amount Household members Other family Community Church Employer Other Total4228 Number of observations3747

22 MONEY BORROWEDFraction borrowing Mean conditional on borrowing (Rand) Conditional on borrowing, fraction borrowing from: Bank Money lender Employer of deceased Employer of another person Family outside the household Neighbor Other Number of observations862

23 Fraction borrowing money and number of assets owned at HSE1 (2001) Number of assets owned at HSE Fraction borrowing for funeral

24 Conditional on borrowing, fraction borrowing from a money lender Fraction borrowing from money lender

25 Summary of preliminary findings Funerals are expensive Funerals are expensive And often leave households vulnerable And often leave households vulnerable How are decisions made on funeral spending? How are decisions made on funeral spending?

26 Let = characteristics marking an individual’s status and = community and extended family perception of household “income” (resources) at the time of the death. The funeral expenses we observe in our data are the desired spending plus an idiosyncratic error:. The community and extended family form an opinion about the appropriate size of the funeral, F*, according to the deceased’s status and that of his household at the time of the death: A model of household decision-making

27 Community and extended family do not observe household income (resources). Instead, they observe a vector of household and individual characteristics X 2 that are correlated with income, which they use to form an expectation of household income. Households that experience an unobserved income shortfall will be less able to meet social expectations with respect to the size of the funeral, without borrowing money. The probability that the household will need to borrow (B=1) to finance a funeral of size F* can be written:.

28 This provides us with several checks, and a formal test, of our model. First, characteristics associated with lower individual status will have different predictions for spending and borrowing than do characteristics associated with lower household income. Characteristics of the deceased associated with lower individual status that is, with lower values of should reduce both the size of the funeral, as in (1), and the probability of borrowing, as in (4).

29 In contrast, any information available to the community that causes them to revise downward their estimate of household income, should reduce the size of the funeral, as in (1), but increase the probability of borrowing for the funeral. We examine these in turn.

30 Table 5. Individual Status, Funeral Spending and Borrowing Dependent variable: Funeral spending (Rand) =1 if borrowed money for funeral Female– (107.06) – (107.70) –0.026 (0.014) –0.038 (0.015) Household characteristics? NoYesNoYes Number of observations

31 Table 5. Individual Status, Funeral Spending and Borrowing Dependent variable: Funeral spending (Rand) =1 if borrowed money for funeral Relation of deceased to current head is ‘other’ – (110.78) – (113.21) –0.037 (0.016) –0.038 (0.017) Household characteristics? NoYesNoYes Number of observations

32 Table 6. Household Income, Funeral Spending and Borrowing Funeral spending (Rand) Household asset holdings (21.38) -- Max education of any household member HSE (23.68) -- Ind: payments med treatment before death -- – (131.89) -- Indicator: cause of death was AIDS -- – (119.52) -- Education of the deceased (23.32) -- Deceased had a funeral policy -- – (315.30) Funeral policy paid money (342.03) Money was borrowed for the funeral -- –41.78 (109.87) Number of observations

33 =1 if borrowed money for funeral Household asset holdings–0.007 (0.002) -- Max education of any household member HSE1 --–0.005 (0.003) -- Ind: payments for med treatment before death (0.016) -- Indicator: cause of death was AIDS (0.017) -- Education of the deceased-- –0.007 (0.002) -- Deceased had a funeral policy-- –0.044 (0.043) Funeral policy paid money-- –0.074 (0.043) Number of observations Table 6. Household Income, Funeral Spending and Borrowing

34 Formal tests of the model The model suggests patterns of coefficients that should hold between the spending and borrowing equations. Writing We see

35 Table 8. Testing Predictions of the Model Dependent Variable:Ratio: Total spending on funeral (1') Borrowed money for the funeral (4') coefficient from (1')/(4') Indicator: female– ( ) –0.034 (0.015) Indicator: relationship to head of household is ‘other’ – ( ) –0.039 (0.016) Chi-square test: X 1 coefficients (p-value) 0.00 (0.994) Household assets (22.002) –0.006 (0.002) – Indicator: funeral policy paid money ( ) –0.106 (0.018) – AIDS death– ( ) (0.017) – Chi-square test: X 2 coefficients (p-value) 0.07 (0.966) Estimate of Gamma0.557

36 Maximum Likelihood Estimation where

37 Table 9. Maximum Likelihood Estimates coefficient (standard error) z- score : Individual characteristics Female– (104.59) 5.91 Indicator: relation to head is ‘other’– (112.11) : Predictors of household income Household assets (204.29) 3.98 AIDS death– (605.58) 3.54 Indicator: funeral policy paid money ( ) 4.00 : Fraction of household income to be used for the funeral (0.086) 4.04 Observations3381

38 Recent economic work has suggested, if the crisis results in lower population growth, that AIDS could “endow the economy with extra resources which … [will] raise the per capita welfare of future generations.” (Young, 2005). Recent economic work has suggested, if the crisis results in lower population growth, that AIDS could “endow the economy with extra resources which … [will] raise the per capita welfare of future generations.” (Young, 2005). This earlier research, however, assumes a constant savings rate over the life of the crisis, in order to focus on the effect of a potential fertility decline. This earlier research, however, assumes a constant savings rate over the life of the crisis, in order to focus on the effect of a potential fertility decline. To the extent that productive resources are diverted into expensive funeral celebrations, earlier predictions that the pandemic will benefit future generations economically are less likely to come to pass. To the extent that productive resources are diverted into expensive funeral celebrations, earlier predictions that the pandemic will benefit future generations economically are less likely to come to pass.

39 Conclusions Whether due to honoring an ancestor, or to the desire to please extended family or community, funerals in South Africa are elaborate and expensive Whether due to honoring an ancestor, or to the desire to please extended family or community, funerals in South Africa are elaborate and expensive Cost appears to be dictated by the status of the deceased and observable household resources Cost appears to be dictated by the status of the deceased and observable household resources Changes to this institution may be difficult without social coordination that everyone will spend less (e.g. Swaziland) Changes to this institution may be difficult without social coordination that everyone will spend less (e.g. Swaziland) Given that this entails agreement not only between those living locally, but also extended family coming from afar for the funeral, this may be difficult Given that this entails agreement not only between those living locally, but also extended family coming from afar for the funeral, this may be difficult


Download ppt "Paying the Piper: The High Cost of Funerals in South Africa Anne Case, Princeton University Anu Garrib, Africa Centre UKZN Alicia Menendez, University."

Similar presentations


Ads by Google