Orphans and discrimination in Mozambique.

This study is motivated by the high HIV prevalence in Mozambique, which, among adults aged 15-45 years in 2005, is estimated to be about 16.2% and is projected to climb (INE et al. 2004). By 2003 an estimated 400,000 Mozambicans had died of AIDS-related causes since 1991, and this number is projected to grow rapidly through the rest of the decade to double by 2010. Due to the tendency of the pandemic to strike young adults, AIDS-related deaths leave significant numbers of orphans in their wake. A demographic and health survey carried out in 2003 found that, for children under 15 years of age, approximately one child in ten had been orphaned (paternal, maternal, or dual) (INE 2004). Demographic projections based on a time series of HIV prevalence data point also to large numbers of orphans (INE et al. 2004). Furthermore, the number of orphans appears set to climb dramatically.

Mozambican national policy specifically favors the integration of orphans into substitute or extended families. This mirrors policy in other highly afflicted countries such as Botswana, Zimbabwe, Zambia, and Uganda (UNAIDS 1999). It has the advantage that orphans remain integrated within a family. This approach to coping with orphaning also implies that the resources available to families that accept orphans and the allocation of those resources within the household become of policy interest.

Generally, resources are tight within Mozambican households. In 2002-3, 58% of all children lived in households that were absolutely poor based on a consumption-based metric. Although nonbiological children tend to concentrate in households that are on average slightly better off (Nhate 2004), resource availability remains distinctly limited and difficult decisions regarding resource distribution have to be made. As noted by Hamilton (1964), biological bonds are important in the distribution of resources within the household implying the potential for discrimination against nonbiological children.

Intrahousehold resource allocations are difficult to measure directly; and household consumption surveys rarely attempt to do so. To partially counter this difficulty, Deaton, Ruiz-Castillo, and Thomas (1989) proposed a method, labeled "outlay equivalence," whereby spending on children is measured indirectly via spending on adult goods. The intuition is that the addition of a child should imply increased spending on goods for children. The budget constraint then implies reduced spending on adult goods. Since, particularly in developing countries, pure adult goods are much easier to identify than pure children's goods, the method has become popular.

Application has often focused on intrahousehold discrimination of girls relative to boys. Using Deaton's approach, evidence from Asia often shows that girls are at a disadvantage relative to boys in the allocation of family resources (Deaton 1989; Behrman 1990; Gibson and Rozelle 2004; Kingdon 2005). On the other hand, studies in African countries tend not to find statistically significant evidence of discrimination against girls (Deaton 1989; Haddad and Reardon 1993).

The present study employs the outlay equivalence approach to analyze potential discrimination in resource allocation within households against children who are not the biological descendant of the household head in Mozambique. Specifically, this study seeks to: (a) identify goods that are demographically separable from children (adult goods), and (b) test for discrimination against children who are not the biological descendant of the household head in the intrahousehold allocation of consumption.

Similar to Nhate (2004), this analysis compares children who are biological versus nonbiological descendants of the household head rather than orphans specifically. The available data base on consumption does not permit the separation of orphans. For the age group fifteen and under, about one child in four is not the biological descendant of the household head. For an unknown but likely substantial fraction of these children, the circumstance of being fostered reflects stress, such as the death of a parent, resulting in placement of the child with another family. We hypothesize that these children are at risk of being discriminated against. Nhate (2004) previously found that Mozambican children who are not biological descendants of the household head were less likely to attend school in both rural and urban areas.

Nevertheless, an important subset of children who are not the biological descendant of the household head is not likely to be at risk for discrimination. In particular, weak geographic coverage of complete primary school causes some families living in areas without access to primary school to send children to live with relatives or friends in areas where primary school is available. It may be plausibly assumed that children who are sent by their parents to live with another family in order to attend school are less likely to be discriminated against than children, such as orphans, who are forced into fostering due to some negative shock. As we are not capable of distinguishing between these two groups of children in our sample, we view our results as a lower bound on the degree of discrimination within families against the target group of interest.

Data and Methodology

Data

We use the national representative household survey on living conditions (IAF) undertaken by the National Institute of Statistics in 2002-2003 (INE 2004). The survey covered 8,700 households corresponding to about 44,000 individuals. Expenditure data were collected on 863 different goods (food and non-food). For our purpose, we are interested in identifying adult goods that children do not consume. The addition of a child (with the concomitant expenses necessary to support that child) reduces the income available to spend on adult goods. For normal goods, consumption should decline. Six candidate adult goods were identified; adult clothes; alcoholic beverages (inside and away from home); personal care (hair treatment, nail products, lipstick, "mulala," lotion, etc.); public and private transportation services; tobacco; and food and soft drinks away from home.

We conduct the analysis both at the national level and by rural and urban zones in order to capture differential characteristics of rural and urban families. Table 1 presents summary statistics (means) for the two subsamples and the national sample. The analysis is performed separately for poor and nonpoor households. Poor households are defined as those living below a poverty line that reflects basic needs (MPF, IFPRI, Purdue University 2004). Resource constraints for poor households are more severe and may influence intrahousehold resource allocation decisions. Finally, following standard practice, 1,046 households without any children and 538 households with only a single household member were excluded from the sample leaving a total of 7,116 households with at least one child present in the final sample. Sample weights are used throughout the analysis to take into account the stratified nature of the sample.

The average budget share of the candidate adult goods as a group is 13%. Tobacco and adult clothes are the goods that have the highest share among all adult goods. The "food and soft drinks" group and "personal care" represent small shares of total expenditures (0.2% and 0.6%, respectively). Generally, budget shares for adult goods are higher in urban than in rural areas (15 % vs. 11%). Overall, the shares for adult goods observed in Mozambique are similar to values found in other developing countries (Haddad and Reardon 1993; Gibson and Rozelle 2004).

Urban households consume on average more than rural households and also have slightly larger household sizes. The largest demographic child category is children aged 0-5 years. As one would expect, biological children represent a higher proportion on average compared to nonbiological children for each age group.

Method

We follow the method developed by Deaton, Ruiz-Castillo, and Thomas (1989)--except our objective is to study potential discrimination between children who are direct descendants of the household head (labeled "biological") and those who are not (labeled "nonbiological").

First, household members were categorized into one of ten demographic groups according to the groups shown in table 1. Children (less than sixteen years of age) were divided into six groups: three each for the biological and nonbiological categories. The remaining four categories consist of adults at different age levels. The next step consisted of the identification of adult goods. Adult goods are goods which have no relationship to a specific household demographic class namely children hence are referred to as demographically separable. To test whether good i is truly an adult good, we used the linear model of Deaton, Ruiz-Castillo, and Thomas (1989): a

(1) [p.sub.i][q.sub.i] = [[alpha].sub.0i] + [[alpha].sub.1i][X.sub.G] + [J.summation over (j=1)][c.sub.ij][n.sub.j] + [d.sub.i]z + [[epsilon].sub.i]

where [p.sub.i][q.sub.i] is expenditure on the candidate adult good, [X.sub.G] is total expenditures on adult goods, [n.sub.j] is the number of members in each demographic category j (with j = 1, ... J), z is a vector of other explanatory variables included in the model, and [[epsilon].sub.i] is the error term.

Given total expenditures on adult goods, children should not influence the distribution of spending across adult goods. If the goods included are really adult goods, children will not have any effect in equation (1). Therefore, the coefficients [c.sub.ij] should be insignificant, both individually and jointly, for demographic groups related to children in order for demographic separability to hold. Due to potential endogeneity of [X.sub.G], we performed two-stage least squares estimation with total household expenditure (logarithm) as instrument.

Following the test of existence of adult goods using equation (1), we calculate the ratio of equivalent expenditures ([[pi].sub.ir]) for a normal adult good i and demographic category r:

(2) [[pi].sub.ir] = [partial derivative]([p.sub.i][q.sub.i])/[partial derivative][n.sub.r]/[partial derivative]([p.sub.i][q.sub.i])/[partial derivative]x n/x

where [[pi].sub.ir] measures the effect of the addition of a member of type r on total expenditure on good i measured in terms of the change in total expenditure that would be necessary to produce the same effect on demand with this change presented as a share of per capita expenditure. For adult goods, one would expect a reduction in expenditure given an additional child and hence a negative value for [[pi].sub.ir].

Following Deaton, Ruiz-Castillo, and Thomas (1989), the equivalent expenditure ratios in (2) can be calculated using the coefficients estimated from a standard Engle curve, specified in the following way:

(3) [w.sub.i] = [p.sub.i][q.sub.i]/x = [[alpha].sub.i] + [[beta].sub.i]ln(x/n) + [[eta].sub.i]ln n + [J-1.summation over (j=1)] [[gamma].sub.ij]([n.sub.j]/n) + [[delta].sub.i]z + [[mu].sub.i]

where [w.sub.i] is the budget share of the ith adult good, x is the value of total household expenditure, n is household size, [n.sub.j] is the number of people in demographic group j, z is a vector of control variables, and [[mu].sub.i] is the error term (heteroskedasticity-consistent standard errors are used throughout the analysis).

The estimated parameters in equation (3) are then used to calculate:

(4) [[pi].sub.ir] = ([[eta].sub.i] - [[beta].sub.i]) + [[gamma].sub.ir] - [[summation].sup.J-1.sub.j=1][[gamma].sub.ij](n.sub.j]/n)/[[beta].sub.i] + [w.sub.i]

These estimated ratios are obtained by substituting the parameters with their respective estimates--from (3)--and substituting for [w.sub.i] and the fraction [n.sub.j]/n by the mean values in the sample. After calculating the [pi]'s, we can test the hypothesis of equal treatment between the biological and nonbiological children in each age group and for all adult goods i, as shown below:

(5) [H.sub.0] : [[pi].sub.ij] = [[pi].sub.k]

where j refers to biological children and k to nonbiological children in the same age group. Using the calculated [pi]'s, a second test for demographic separability was performed providing a robustness check for the selection of adult goods using equation (1). If demographic separability holds, the values for the estimated [pi] ratios across goods for demographic group r should not differ. This test is implemented for a group of v goods by testing the following null hypothesis for each good:

(6) [H.sub.0] : [[DELTA].sub.ir] = [[pi].sub.ir] - [summation over j] [[pi].sub.jr]/v = 0 i = 1, 2, ... v

Alternative approaches to deriving standard errors for the [pi] ratios are described in Deaton, Ruiz-Castillo, and Thomas (1989). Here, the standard errors for the [pi] ratios were derived using a nonparametric bootstrap. The bootstrap method involves drawing synthetic samples of the same size as the original sample and according to the same stratification, by sampling with replacement from the original sample. Regressions using equation (3) were run on 1,000 synthetic samples and the w ratios were calculated in each instance. Standard errors are then calculated from this sample of 1,000[pi] ratios. The bootstrap approach has the advantage of accommodating the nonlinear nature of the [pi] ratios as a function of the estimated parameters. Standard errors were also calculated using the linear approximation method suggested by Deaton, Ruiz-Castillo, and Thomas (1989) with similar results.

Results

The analysis was performed at the national, rural, and urban levels for all households and with households further divided by socioeconomic status (poor and nonpoor households). We found no evidence of discrimination between biological and nonbiological children in the nonpoor sample and in the full sample, so we focus on results for poor households in the following exposition. Full results are available upon request.

Table 2 presents results of the tests for identification of adult goods based on equation (1) for the subset of poor households. The results indicate that all six candidate adult goods qualify. The separability test across goods (equation (6)) gives similar results (not presented). Table 3 presents [pi] ratios for the analysis conducted at the national, urban, and rural levels, respectively for poor households (standard errors are not presented to save space). As stated above, negative [pi] ratios indicate compression of expenditure on the associated adult good due to the addition of a child in a given age group. There are seven goods (the six adult goods plus the results for all six goods combined) and three age classes resulting in 21 comparisons at each of the three levels of analysis (national, urban, rural) or 63 comparisons overall. However, the crucial comparison is with respect to the aggregate of all six adult goods. For this case, the relationship is as hypothesized (greater compression of expenditure on adult goods with respect to biological children) in eight of nine instances.

Table 4 presents the results of F-tests for equality of [pi] ratios between biological and nonbiological children. Again, the crucial tests are the ones for all six goods combined. For this aggregate, the greater compression of expenditure on adult goods with respect to biological children was statistically significant for four of the eight possible cases. Muddying the waters somewhat, the one case with an unexpected sign (more compression of household expenditures for nonbiological children than biological in the case of children from 0 to 5 years old in urban areas) is also statistically significant at the 10% level.

As a further robustness check, [chi square] tests were performed on the aggregate good to test the hypothesis that [pi] ratios are equal for each of the three age groups (e.g., three linear restrictions). The results reject the hypothesis of equal [pi] ratios between biological and nonbiological children at the rural, urban, and national levels. In the case of urban areas, the direction of the sign of the difference in [pi] ratios is counter to expectation rendering the joint test inadmissible.

Conclusions

The weight of evidence points to discrimination in the intrahousehold allocation of resources against children who are not direct biological descendants of the household head in poor households. Discrimination is significant for younger children (aged 0-10) in rural households and older children (aged 11-15) in urban households.

There is no evidence that nonpoor households discriminate against children who are not the biological descendant of the household head. There are two likely reasons underpinning the dichotomy of results between poor and nonpoor households. First, resources are more severely constrained in poor households forcing more difficult choices in resource allocation. Nonbiological children may experience discrimination under these harsher economic conditions. Second, our inability to identify the reason for the presence of a nonbiological child within a family may also play a role. The available evidence indicates that wealthier households are more likely to host children in order for them to attend school (Nhate, 2004). Hence, the bias from mixing together children who are likely to be discriminated against (AIDS orphans for example) with children who are not (those living with friends or relatives in order to attend school) under a single rubric "nonbiological children" may be substantially more profound in the nonpoor subset of the population. As indicated earlier, the results obtained are likely a lower bound on the discrimination against the target group of children.

Unfortunately, AIDS will almost surely increase the number of children requiring care from neighbors, friends, and/or relatives due to the death of one or more of their parents. As the overall burden on communities grows, the tendency for nonbiological children to reside in better off households may become less pronounced and the degree of discrimination against nonbiological children may accentuate itself.

If one wishes to target some assistance at particularly disadvantaged groups, then children living in poor households that are not the biological descendant of the household head, especially those that do not attend school or attend school only sporadically, would appear to be a logical choice. The results also indicate that the policy of placing orphans in families of neighbors, friends, or relatives likely functions less well, in terms of the interests of the orphans, than would occur in a world free of discrimination. Further, the policy may perform even more poorly as the burden grows. Nevertheless, the result does not necessarily imply that the policy should be abandoned. This decision can only be reached through comparison with potential substitute policies. While the analysis of potential substitute policies merits further attention, the available evidence indicates that attractive substitute policies are few to nonexistent. Despite discrimination, the current policy may be the best available alternative.

References

Behrman, J. 1990. Intrahousehold Allocation of Nutrients and Gender Effects: A Survey of Structural and Reduced-Form Estimates. Oxford: Oxford University Press.

Deaton, A. 1989. "Looking for Boy-Girl Discrimination in Household Expenditure Data." World Bank Economic Review 3(1): 1-15.

Deaton, A., J. Ruiz-Castillo, and D. Thomas. 1989. "The Influence of Household Composition on Household Expenditure Patterns: Theory and Spanish Evidence." Journal of Political Economy 97(1):179-200.

Gibson, J., and S. Rozelle. 2004. "Is It Better to Be a Boy? A Disaggregated Outlay Equivalent Analysis of Gender Bias in Papua New Guinea." Journal of Development Studies 40(4):115-36.

Haddad, L., and T. Reardon, 1993. "Gender Bias in the Allocation of Resources within Households in Burkina Faso: A Disaggregated Outlay Equivalent Analysis." Journal of Development Studies 29(2):260-76.

Hamilton, W.D. 1964. "The Genetical Evolution of Social Biology." Journal of Theoretical Biology 7:1-16.

INE [National Institute of Statistics]. 2004. "'Inquerito Nacional ao Agregados Familiares Sobre Orcamento Familiar 2002/03." [National Household Budget Survey] Maputo.

INE [National Institute of Statistics], Ministry of Health, Ministry of Planning and Finance, National Council of Fighting HIV/AIDS, the Center for Population Studies and Faculty of Medicine of Eduardo Mondlane University. 2004. "Impacto Demografico do HIV/SIDA em Mocambique: Actualizacao." [Demographic Impact of HIV/AIDS in Mozambique: Update.] Maputo.

Kingdon, G.G. 2005. "Where Has the Bias Gone? Detecting Gender Bias in the Intrahousehold Allocation of Educational Expenditure." Economic Development and Cultural Change 53:409-51.

MPF [Ministry of Planning and Finance], IFPRI [International Food Policy Research Institute], and Purdue University. 2004. "Poverty and Well-Being in Mozambique: The Second National Assessment." Maputo.

Nhate, V. 2004. "Orphans in Mozambique: Vulnerability, Trends, Determinants, and Programme Responses." Ministry of Planning and Development, Mozambique, Maputo.

UNAIDS, 1999. "Criancas Orfaos Devido SIDA." [AIDS Orphans]. Mimeo.

Channing Arndt is associate professor in the Department of Agricultural Economics at Purdue University. Mikkel Barslund is a Ph.D. candidate in the Department of Economics at the University of Copenhagen. Virgulino Nhate is an analyst in the Ministry of Planning and Development, Mozambique. Katleen Van den Broeck is an assistant research professor in the Department of Economics at the University of Copenhagen.

The authors thank the World Food Program for financial support. Financial support to the Ministry of Planning and Development from British, Danish. Swedish. and Swiss development agencies is also gratefully recognized.

This article was presented in a principal paper session at the AAEA annual meeting (Long Beach, CA, July 2006). The articles in these sessions are not subjected to the journal's standard refereeing process. Table 1. Mean Values for the Data Variables National Urban Rural Proportion of candidates to adult goods 0.125 0.153 0.114 Proportion of alcohol in total

expenditure 0.010 0.011 0.010 Proportion of tobacco in total

expenditure 0.043 0.049 0.041 Proportion of adult clothes in total

expenditure 0.043 0.043 0.043 Proportion of transportation in total of

expenditures 0.022 0.035 0.016 Proportion of food and soft drinks

consumed away from home in total of

expenditures 0.002 0.004 0.001 Proportion of personal care in total of

expenditures 0.006 0.011 0.004 Log of total household expenditures 9.151 9.496 8.851 Log of household size 1.556 1.632 1.491 Proportion of biological children aged

0-5 years 0.150 0.128 0.170 Proportion of nonbiological children

aged 0-5 years 0.040 0.042 0.038 Proportion of biological children aged

6-10 years 0.104 0.098 0.109 Proportion of nonbiological children aged

6-10 years 0.032 0.032 0.032 Proportion of biological children aged

11-15 years 0.079 0.082 0.076 Proportion of nonbiological children aged

11-15 years 0.031 0.035 0.028 Proportion of people aged 16-20 years 0.110 0.130 0.092 Proportion of people aged 21-25 years 0.075 0.087 0.065 Proportion of people aged 26-59 years 0.320 0.320 0.319 Proportion of people with more than sixty

years of age 0.059 0.045 0.072 Proportion of households headed by women 0.252 0.266 0.239 Educational level of household head 1.106 1.884 0.432 The mean age of the household head 42.937 42.696 43.146 Proportion of people in agriculture and

fishing 0.756 0.503 0.976 Proportion of people in commerce 0.180 0.313 0.0065 Proportion of people in the services

sector 0.142 0.270 0.030 Table 2. p-Values for Tests of Demographic Separability between Children and Adult Goods--Poor Households

Biological Nonbiological Biological Adult Goods 0-5 0-5 6-10

National Alcohol 0.032 0.749 0.092 Tobacco 0.065 0.453 0.01 Adult cloth 0.807 0.378 0.592 Transportation 0.337 0.074 O.t36 Meal and soft drink away home 0.571 0.529 0.673 Personal care 0.108 0.488 0.892

Urban Alcohol 0.223 0.163 0.865 Tobacco 0.273 0.221 0.704 Adult cloth 0.452 0.370 0.510 Transportation 0.875 0.163 0.873 Meal and soft drink away home 0.192 0.494 0.468 Personal care 0.527 0.458 0.715

Rural Alcohol 0.058 0.501 0.079 Tobacco 0.185 0.839 0.008 Adult cloth 0.469 0.255 0.423 Transportation 0.261 0.202 0.108 Meal and soft drink away home 0.329 0.548 0.519 Personal care 0.21 0.177 0.86

Nonbiological Biological Adult Goods 6-10 11-15

National Alcohol 0.058 0.461 Tobacco 0.32 0.012 Adult cloth 0.761 0.781 Transportation 0.923 0.286 Meal and soft drink away home 0.689 0.012 Personal care 0.311 0.251

Urban Alcohol 0.418 0.411 Tobacco 0.477 0.243 Adult cloth 0.532 0.280 Transportation 0.05 0.954 Meal and soft drink away home 0.172 0.15 Personal care 0.539 0.876

Rural Alcohol 0.085 0.315 Tobacco 0.241 0.024 Adult cloth 0.518 0.997 Transportation 0.412 0.305 Meal and soft drink away home 0.443 0.038 Personal care 0.22 0.151

Nonbiological Joint Test: All Adult Goods 11-15 Children Groups

National Alcohol 0.022 0.284 Tobacco 0.027 0.055 Adult cloth 0.273 0.850 Transportation 0.056 0.219 Meal and soft drink away home 0.152 0.158 Personal care 0.901 0.135

Urban Alcohol 0.767 0.831 Tobacco 0.072 0.514 Adult cloth 0.670 0.827 Transportation 0.548 0.169 Meal and soft drink away home 0.036 0.207 Personal care 0.963 0.925

Rural Alcohol 0.016 0.379 Tobacco 0.099 0.060 Adult cloth 0.217 0.813 Transportation 0.069 0.624 Meal and soft drink away home 0.96 0.420 Personal care 0.585 0.113 Table 3. Outlay Equivalence Ratios--Poor Households

Biological Nonbiological Biological Adult Goods 0-5 0-5 6-10

National Alcohol -0.643 -0.991 -0.518 Tobacco 0.369 0.911 -0.070 Adult clothing -0.037 -0.035 -0.169 Transportation -0.318 -0.239 -0.465 Meal and drink away home 0.803 -0.565 0.005 Personal care -0.441 -0.155 -0.310 All six goods -0.054 0.065 -0.238

Urban Alcohol 0.093 0.027 0.206 Tobacco 0.715 0.325 -0.216 Adult clothing 0.028 -0.496 -0.68 Transportation 0.027 -0.195 -0.512 Meal and drink away home -0.153 -0.62 0.028 Personal care -0.234 -0.488 -0.213 All six goods 0.180 -0.179 -0.418

Rural Alcohol -0.745 -1.014 -0.591 Tobacco -0.157 0.919 0.070 Adult clothing 0.059 0.211 -0.032 Transportation -0.388 -0.011 -0.425 Meal and drink away home 1.991 -0.264 0.057 Personal care -0.555 0.043 -0.387 All six goods -0.168 0.223 -0.144

Nonbiological Biological Nonbiological Adult Goods 6-10 11-15 11-15

National Alcohol 0.268 -0.414 -0.952 Tobacco 0.212 -0.251 0.342 Adult clothing -0.037 -0.194 0.301 Transportation 0.324 -0.268 -0.514 Meal and drink away home 0.020 -1.069 -0.565 Personal care 0.418 -0.596 -0.125 All six goods 0.142 -0.281 0.002

Urban Alcohol -0.216 1.199 0.094 Tobacco -0.638 -0.723 0.287 Adult clothing 0.144 -0_886 0.279 Transportation 0.074 -0.401 -0.136 Meal and drink away home -0.472 -0.382 -1.408 Personal care -0.103 -0.461 0.079 All six goods -0.168 -0.58 0.096

Rural Alcohol 0.560 -0.715 -1.129 Tobacco 0.369 -0.092 0.485 Adult clothing 0.053 0.032 0.260 Transportation 0.662 -0.111 -0.786 Meal and drink away home 0.416 -1.887 -0.028 Personal care 0.793 -0.674 -0.339 All six goods 0.300 -0.152 -0.025

General General General Adult Goods 16-20 21-24 25-59

National Alcohol -0.480 -0.931 0.350 Tobacco 0.913 -0.246 -0.110 Adult clothing -0.106 -0.474 -0.384 Transportation 0.096 0.095 -0.035 Meal and drink away home -0.927 0.132 -0.213 Personal care 0.558 1.333 0.777 All six goods 0.172 -0.246 -0.131

Urban Alcohol -0.498 -1.168 -0.514 Tobacco 1.184 0.109 0.164 Adult clothing -0.036 -0.516 -0.345 Transportation 0.414 -0.013 0.418 Meal and drink away home -0.456 0.236 0.695 Personal care 0.472 0.968 0.458 All six goods 0.393 -0.145 0.038

Rural Alcohol -0.431 -0.906 0.412 Tobacco 0.391 -0.186 -0.146 Adult clothing -0.134 -0.516 -0.466 Transportation -0.171 0.068 -0.434 Meal and drink away home -2.075 -1.198 -2.163 Personal care 0.551 1.697 1.083 All six goods -0.028 -0.282 -0.254 Table 4. p-Values for T-tests. Equality of [pi] ratios by Child Status--Poor Households

Children Adult Goods Children 0-5 Children 6-10 11-15

National Alcohol 0.38 0.35 0.33 Tobacco 0.22 0.59 0.35 Adult clothing 0.99 0.63 0.17 Transportation 0.86 0.05 ** 0.51 Meal and drink away home 0.13 0.98 0.33 Personal care 0.45 0.33 0.25 All six goods 0.53 0.05 ** 0.24

Urban Alcohol 0.92 0.59 0.32 Tobacco 0.41 0.68 0.26 Adult clothing 0.22 0.08 * 0.10 * Transportation 0.63 0.12 0.56 Meal and drink away home 0.44 0.52 0.18 Personal care 0.54 0.73 0.28 All six goods 0.06 * 0.40 0.05 **

Rural Alcohol 0.56 0.26 0.50 Tobacco 0.04 ** 0.60 0.38 Adult clothing 0.67 0.83 0.63 Transportation 0.45 0.08 * 0.23 Meal and drink away home 0.22 0.62 0.08 * Personal care 0.29 0.37 0.57 All six goods 0.09 * 0.06 * 0.71 Note: single (*) and double (**) asterisks denote significance at 10% and 5%, respectively.