64276 POVERTY ASSESSMENT Social Protection Chapter Background paper 2 Social assistance – Cash transfers Potential impact of Decision 67 Pham Anh Tuyet, Paulette Castel 1 (Vietnamese Academy of Social Sciences - Centre for Analysis and Forecasting) This draft – February 2009 A wide range of public policies aims at improving the social welfare of the population or specific groups among the population. The use of the term social protection usually refers to the group of social policies that provide cash transfer to protect households from poverty. These policies are of two types. Social and health insurance policies are concerned about how to mitigate households’ lost of welfare in case of unexpected health expenditure or lost of labor income because of a member’s sickness, maternity leave, unemployment, disability, old-age or death. Social assistance policies intend to help poor households satisfy their basic needs and avoid deep or extreme poverty as a result. Following S. Sumarto, A. Suryahadi, L. Pritchett (2000 2) illustration, one can think that social protection policies are like ropes that help people reduce the welfare’s drop caused by the lost of income while social assistance policies are like rings that are thrown to the people so that they do not sink into poverty or extreme poverty. Social policies with different goals like for example: the social allowances paid to the meritorious persons, the subsidies on education fees granted to the poor students, preferential access to credit or the development of infrastructure in poor regions, do not belong to the sector of social protection. According to this definition, in Vietnam social assistance policies include programs that provide social allowances to the poor and the payment of emergency allowances in case of natural disaster. Since 2007, the government has indicated its willingness to develop and increase the impact of social allowances. The regulations have been changed in April 2007 and the budget in 2008 has been expanded to VND million 2100 from VND million 510 in 2007 (or approximately from 0.05 % to 0.15% of GDP). 1 Comments and questions are welcomed at paulettecastel@gmail.com 2 S. Sumarto, A. Suryahadi, L. Pritchett World Bank Policy Research Working Paper 2436, 2000 The goal of this study is to measure the possible impact of these new policies. Because the new regulations have been recently passed and few administrative data on the programs are available, the study cannot report on the effective impact of the new policy. The approach instead uses the 2006 VHLSS to estimate how efficient the new policy could be if it had enough resource to be fully implemented without any restriction. The first section measures how many of the poor could be eligible to social allowances if the decision 67 could be fully implemented. The second section describes the characteristics of the poor that would probably not benefit from this new policy. 1. New policy on social assistance allowances: Decision 67 The Decision 67/2007/ND-CP of April 2007 introduced changes in the categories of the beneficiaries of social assistance allowances as well as in the level of the allowances. In summary, the new group of beneficiaries includes: • the children who have nobody to rely on, children with AIDS/HIV living in poor households, • the poor elderly who have nobody to rely on, • the elderly who are 85 years and above and who are not pensioner or beneficiary from other special social monthly benefit • the mentally or physically seriously disabled persons living in poor household (including AIDS/HIV persons), • the poor people raising alone children, as well as, • the families that have adopted a child and the families that include at least two disabled persons. One important change in comparison to the previous regulations is that the decision 67 targets ‘the poor” instead of the persons that “have no source of income”. Some groups are also defined more broadly than before. The new regulation includes the elderly of 85 years and above while before it covered only those 90 years and above. All the households that include two or more seriously disabled persons are included whatever the cause of the disability while before only those affected by the “agent orange” 3 were included. Finally, children whose parents are in prison are possibly also considered now as children who have nobody to rely on. The amount of the allowance has been increased from between 65 and 260 000 VND per month to between 120 and 480 000 VND per month. Finally one “emergency” payment is realized in case of unexpected economic difficulties after events like: the death or the injury of a household member, hunger related to natural disaster or extreme poverty. 3 Toxic chemicals spread in Vietnam during the US – Vietnam war. All the beneficiaries are exempt from educational fees and benefit from free health insurance. Similar social allowances are also paid to the community-based social houses where some of the beneficiaries live. According to the administrative figures of MOLISA reported in Table 1, about 1.15 million people could potentially be eligible to social allowance in 2008. In order to achieve a higher coverage that in the preceding years, the budget assigned to the distribution of social allowances is 3.5 times higher than the budget of 2007. Table 1 Social Assistance administrative data Number of beneficiaries and budget 2000 2001 2002 2003 2004 2005 2006 2007 2008 Potential and effective number of beneficiaries in thousand and corresponding coverage in percent potential 480.0 488.7 489.0 546.2 843.0 831.0 860.0 1100.0 1150.0 effective 175.4 181.6 223.0 229.0 329.7 416.0 470.0 578.0 coverage 36.5 37.2 45.6 41.9 39.1 50.1 54.7 52.5 Budget VND billion 123.1 127.51 173.96 160.78 282.9 421.8 510.0 590.0 2100.0 in % of GDP 0.03 0.03 0.03 0.03 0.04 0.05 0.05 0.05 Source: Nguyen Hai Huu "Social assistance for the poor fact and measures" presentation made at the workshop on social protection policies for the poor Molisa - DFID Hanoi June 10, 2008. 2. Possible impact of Decision 67 on poverty: the approach Measuring the impact of the implementation of the Decision 67 is right now impossible. The data on poverty and households characteristics included in the most recent VHLSS were collected in 2006 before the Decision was adopted. Instead on analyzing who did receive social allowances, this study investigates, therefore, who in the VHLSS 2006 would have been a potential beneficiary of the Decision 67 if the policy had been implemented in 2006. Besides measuring the program’s coverage of the poor, the study investigates also what could have been the impact of the Decision 67 on these household’s poverty status and gives indicators of the efficiency of the program to push people out of poverty. The potential beneficiaries This section describes the approach used to estimate the potential coverage of the Decision 67. The study calculates the share of the poor in the VHLSS 2006 that could have been recipients of social allowance if the Decision 67 had been implemented in 2006 without any problems of targeting or funding. Based on the criteria defined in the Decision 67, the group of the potential beneficiaries in the VHLSS is defined as including: • All the children who live alone in households with not working-aged or elderly person • All the elderly aged 85 and above who are not pensioners • All the households that include at least 2 seriously disabled members • The poor elderly (above standard retirement age) who do live alone or with other elderly persons. • The poor household that include one seriously disabled member or with long-term sickness • The poor households that include children and only one working-aged member or no working aged member Four criteria of poverty are used to define which households are poor: the GSO poor and extremely poor, the MOLISA poor and the registered poor. The numbers of GSO poor and extremely poor are computed according to the GSO poverty lines. These lines represent the cost of consumption baskets that include food and non-food items or only food items. The food spending is large enough to secure 2100 calories per day per person. A household is poor if its level of consumption per capita is lower than the combined food and non-food poverty line; it is extremely poor if its level of consumption per capita is lower than the food line. Instead of using consumption per capita, the number of MOLISA poor is obtained by comparing household per capita income to the official thresholds MOLISA uses in the field to define if an households is poor or not. MOLISA’s poverty income threshold was in 2006 equal to 200 000 VND per capita and per month in the rural areas and 260 000 VND per capita and per month in the urban areas. These limits are currently revised and could be increased to 300 and 390 thousand VND, respectively. The study uses, however, the thresholds used in 2006. The reason is that the use of the new thresholds would require estimate each households’ change in income in 2007 and 2008. This task is out of the scope of this study. Finally, the number of registered poor in 2006 is obtained directly from the answers reported in the VHLSS. Table 2 gives the number of poor households or individuals according to each criteria of poverty. The number of poor according to GSO methodology is about 2 times the number of the registered poor or the poor according to MOLISA criteria on reported income per capita. Based on GSO poverty lines, in 2006, there were roughly13 million of poor in Vietnam of which 5.5 million lived in extreme poverty. Only about 7 millions persons, however, live in households that report income per capita lower than the MOLISA’s poverty thresholds. The number of registered poor of about 6 millions is consistent with that observation. Table 2 Poverty rates and number of poor and poor households In the VHLSS 2006 Households Population Indicator of Survey Confidence poverty Number In percent Number In percent Interval1/ of total of total Lower Upper bound bound GSO Poor 2,637,468 13.4 13,047,757 15.8 14.8 16.8 GSO Extremely poor 1,013,592 5.2 5,448,373 6.6 5.9 7.3 MOLISA poor 1,388,157 7.1 6,772,450 8.2 7.5 8.9 Registered poor 1,302,239 6.6 6,106,525 7.4 6.7 8.1 Source: VHLSS2006 Table 3 and 4 compares the distribution of MOLISA poor and the registered poor in GSO categories of poor, extremely poor and non poor. The results indicate that a very few of the MOLISA poor (1.8% see Table 3) and no one of the registered poor (see Table 4) appear to be non poor according to GSO. As already observed in Table 2, the number of GSO is about 2 times the number of the registered poor or the number of the MOLISA poor. Therefore, at the opposite, 53.2 % of the poor GSO are not registered poor and, 57.8% of the poor GSO report levels of income per capita higher than the MOLISA poverty thresholds. More surprisingly, 46.1% of the GSO extremely poor are not registered poor and 37.1% report income higher than the MOLISA poverty thresholds. Table 3 The distribution of the poor and the non poor According to GSO and MOLISA criteria of poverty MOLISA Poverty - GSO Extreme poverty - GSO (income) Not poor poor Not extremely poor Poor not extremely poor Not poor 98.2 57.8 95.7 37.1 Poor 1.8 42.2 4.3 62.9 total 100.0 100.0 100.0 100.0 Source: VHLSS2006 Table 4 The distribution of the poor and the non poor According to GSO and MOLISA criteria of poverty Registered Poverty - GSO Extreme poverty - GSO poor Not poor poor Not extremely poor Poor not extremely poor Not poor 100.0 53.2 95.9 46.1 Poor 0.0 46.8 4.1 53.9 total 100.0 100.0 100.0 100.0 Source: VHLSS2006 It is interesting to observe (Table 5) that although the total number of registered poor is close to the total number of MOLISA poor, the two groups do not cover the same persons: 47.7 % of the registered poor report income per capita higher than the MOLISA poverty threshold. This result suggests that procedures used by MOLISA and the local authorities to screen who are poor include additional criteria beside the revenue threshold or that local authorities face difficulties in measuring the level of income of the households. Finally, only 3.4% of the MOLISA poor are not registered poor. This result suggests that in the procedure of defining who is poor, the detection and coverage of the people with little income is rather well done. Table 5 The distribution of the poor and the non poor according to MOLISA criteria of poverty and reported poverty registration MOLISA poor Registered poor Not registered Registered Not poor 96.6 47.7 Poor 3.4 52.3 total 100.0 100.0 Source: VHLSS2006 3. Potential coverage of the poor by the Decision 67 Table 6 presents the results regarding the potential coverage of the poor of the VHLSS 2006 if the Decision 67 had been applied in that year without problems for identifying the poor. Depending on the criteria of poverty used, between 30.3 and 35% of the poor households could have been eligible for social allowances. Table 6 Potential coverage of the poor by the Decision 67 Indicator of poverty Percent of poor households Percent of poor population potentially eligible potentially eligible GSO poor 29.3 25.7 GSO extremely poor 30.3 25.0 MOLISA income poor 27.8 24.1 Registered poor 34.1 29.8 Source: VHLS 2006 The Decision 67 is not totally oriented to the support of the poor households, children who live on their own, elderly no pensioners aged 85 and above as well as households that include two or more seriously disabled persons are entitled to receive social allowances even if they are not classified as poor. As a result, as the figures in Table 7 show, between 36.8% and 68.4% of the potential households recipients of social allowances under the Decision 67 are not poor. Table 7 Share of non poor recipients in the total potential beneficiaries of the Decision 67 Indicator of poverty Percent of beneficiaries Percent of beneficiaries households not poor not poor GSO poor 36.8 37.8 MOLISA income poor 68.4 69.8 Registered poor 63.6 66.3 Source: VHLS 2006 The figures presented in Table 6 indicate the levels of coverage of the poor the program could achieve if there was no problem of implementation. In the practice finding the poor and screening the poor from the non poor is difficult. Problems of finding the poor will result in lower coverage. Problems of screening will produce some leakage: non-poor who according the regulations are not eligible to social allowances could be among the beneficiaries. 4. Potential impact on households’ poverty status of the Decision 67 Potential impact on households’ poverty status This section describes the approach used to estimate the potential impact on the poverty status of the VHLSS 2006 households of the receipt of social allowances according to the rules stipulated in the Decision 67. The study calculates how many recipients could have been “pushed out of poverty” thanks to the receipt of social allowances. The following methodology is used: The members of the household h are poor if the consumption per capita of the household h is lower than the poverty line l, that is: Ch (1) if = l Nh A precise evaluation of the possible welfare impact of the distribution of social allowances would require take into account that the increase of households’ total consumption can be lower than the amount of the received social allowance. The household can decide to save a portion of the allowance. Some members can decide to work less hours and thus finance lower level of consumption than in the absence of the transfer. Friends and relatives can decide to provide less support (gift, remittances, etc.). Having these caveats in mind, this study measures how much poverty could be reduced if social allowances were totally used to support households’ consumption. Table 8 presents the results regarding the potential impact of the distribution of social allowance on the VHLSS poor households’ poverty status if the Decision 67 had been applied in that year. Thanks to social allowances receipts between 17.9% and 26.2% of the poor households could cover their basic needs and “escape poverty” if they use all the transfers to buy consumption goods. As a result, the overall population poverty rate could be reduced from 1.0 percentage point to 2.1 percentage point depending on if the program manages to target poor populations selected like the registered poor in 2006 or populations who are poor according to the GSO criteria of poverty. For about three quarter of the poor households social allowances would, however, not be enough to cover their poverty gaps. The poverty gap is the distance between poor households’ level of consumption or income per capita and the poverty line. Table 8 Potential impact on poverty of the Decision 67 Impact on the poor households Indicator of poverty Percent of poor households Percent of poor population receiving social allowance receiving social allowance pushed out of poverty pushed out of poverty GSO poor1/ 17.9 13.2 1/ GSO extremely poor 24.0 16.6 2/ MOLISA income poor 26.2 19.2 Registered poor1/ 19.1 13.9 Impact on poverty rate Initial poverty rate Poverty rate after receipt 1/ GSO poor 15.8 13.7 1/ GSO extremely poor 6.6 5.5 MOLISA income poor2/ 8.2 6.6 1/ Registered poor 7.4 6.4 Note: 1/ if social allowances are entirely used to buy consumption goods. 2/ social allowances are added to households income per capita and compared to MOLISA poverty thresholds 5. Potential efficiency of the overall program This section describes the approach used to evaluate the potential overall program’s efficiency in terms of poverty reduction. Table 9 first compares the amount of budget that would be distributed to the poor and to the non poor. In 2006, if the Decision 67 had been fully implemented without problems of targeting, screening and funding, the total budget of social allowances would have ranged between VND (of 2006) 1.9 and 2.8 billion depending on the poverty criteria used to target the poor populations. The efficiency of these funds to reduce poverty would have been limited by the fact that the program does not target only poor populations. As the figures in Table 9 show a large portion of the budget from 33.8 to 60.7% could be distributed to non-poor. Table 9 Distribution of social allowances under the Decision 671/ Budget distributed in a year Indicator of Budget Total in % to the poor % to the % to the poverty needed to VND moving out of poor non poor eliminate billion poverty remaining poverty2/ poor GSO poor 2028 2828 40.6 25.6 33.8 GSO extremely poor 526 1878 30.6 8.7 60.7 MOLISA income poor 77 2262 38.4 13.2 48.4 Registered poor 1234 2199 28.9 21.1 50.0 Source: VHLSS 2006 Note: 1/ Difference between the initial and after receipts poverty rates reported in Table 8 2/ Sum of the poverty gaps of the poor households recipients of social allowance A second source of inefficiency is that the amount of the social allowance in the Decision 67 is determined by category but not in relation to the recipients’ poverty gap. There are risks, consequently, that the amount of social allowance is too short to help the family escape poverty or too high regarding its uncovered basic needs. In the later case, part of the funds could have been used to help other poor households. Regarding the budget distributed to the poor, Table 10 shows that 39% of the poor households (GSO criteria) and 43.9% of the registered poor households would receive allowance too small in relation to the poverty gap they cannot manage to cover. At the opposite, 31.8% of the poor households (GSO criteria) and 27.3% of the registered poor would receive allowances two times higher than their poverty gap. This situation is due to the fact that the amount of allowances is determined independently of the situation of the recipients. An extremely poor household with two seriously disabled persons receive exactly the same amount of social allowance than a poor or a better off household with also two seriously disabled members. Table 10 Distribution of poor households depending on the share of the poverty gap is covered by social allowances GSO poor households Registered poor households In percent Cumulative In percent Cumulative Number of total distribution Number of total distribution Below 25% 23855 3.1 3.1 12011 2.7 2.7 25% to less than 50% 98551 12.8 15.9 62164 14.0 16.7 50% to less than 75% 81997 10.6 26.5 48730 11.0 27.7 75% to less than 100% 96697 12.5 39.0 71985 16.2 43.9 100% to less than 125% 55023 7.1 46.1 34892 7.9 51.8 125% to less than 150% 80161 10.4 56.5 43658 9.8 61.6 150% to less than 175% 58667 7.6 64.1 33085 7.5 69.0 175% to less than 200% 31719 4.1 68.2 16200 3.7 72.7 200% and above 245545 31.8 100.0 121241 27.3 100.0 Total 772216 100.0 443968 100.0 6. Social allowances policies towards the extremely poor The previous sections show that if the Decision 67 had been implemented in 2006 without any targeting and financing problems, social allowance would have had covered only a small portion of the poor and would have had rather little impact on poverty in that year. Even if the goal had been to reduce extreme poverty, 4.2 million of the registered poor, and an additional 1.9 million of persons extremely poor according to the GSO criteria (but not registered as poor) would not have fit into one of the groups defined in the Decision 67 and, consequently, would have been excluded from social allowance receipts. This section, accordingly, investigates what changes or what new policies could be implemented in Vietnam so that social allowances policies play a significant role in the country’s strategy to eliminate extreme poverty. In that analysis, the targeted groups are the registered poor and the GSO extremely poor (who are not registered) who are excluded from social allowances under Decision 67. The aggregation of these two groups makes sense because the process of the poor registration appears to be very efficient at avoiding errors of inclusion. According to the results of Table 4, above, only 4.1% of the registered poor are not included in the group of the GSO extremely poor. The latter enclose the group of the registered poor. Table 12 Extremely poor excluded from social allowances under the Decision 67 Indicator of poverty Number of poor households Number of poor persons Registered poor 858 271 4 288 712 GSO extremely poor not registered poor 318 257 1 964 413 Improving social allowance to reduce extreme poverty will, therefore, benefit the overall population in Vietnam. Expanding the coverage of social assistance allowances is not particularly related to problems of screening the extremely poor among the ethnic minorities groups. As the figures in Table 13 show, the households currently excluded from social allowances are as well Kinh and Chinese than from ethnic minorities: 66.8% and 78.2% of the extremely poor Kinh and Chinese and the extremely poor of ethnic minorities are excluded, respectively. The remark is still valid within each group of the registered group and of the GSO extremely poor who are registered. The extremely poor from ethnic minorities represent 54.1% of the excluded. Table 13 Extremely poor excluded from social allowances under the Decision 67 by ethnic groups Number of persons Excluded Distribution in total Potentially share in Excluded eligible Total category Excluded Eligible Total registered poor Kinh and Chinese 2 110 991 1 149 416 3 260 407 64.7 33.8 48.6 Other ethnic groups 2 177 721 668 397 2 846 118 76.5 34.8 28.3 GSO extremely poor not registered poor Kinh and Chinese 758 001 275 096 1 033 097 73.4 12.1 11.6 Other ethnic groups 1 206 412 272 698 1 479 110 81.6 19.3 11.5 Total registered poor and GSO extremely poor Kinh and Chinese 2 868 991 1 424 513 4 293 504 66.8 45.9 60.2 Other ethnic groups 3 384 134 941 094 4 325 228 78.2 54.1 39.8 Total 6 253 25 2 65 607 8 618 732 72.6 100.0 100.0 Demographic composition Most of the eligibility criteria in the Decision 67 are related to the structure of the households and the presence of members who are not able to work. Households with seriously disabled persons are effectively more likely to be poor than the rest of the households (see companion paper). More generally, individual income of the able bodies in poor households with large dependency ratio (that include very old persons or in which the household head raise children alone) are more likely not sufficient to help cover all the basic needs. The groups defined by the Decision 67 do not, however, include the families whose dependency ratio is high because they include many children or children and old people that are not very old (over age 65 but less than 85 years old). Resistance to provide social allowances to families with many children is often related to concerns that social assistance may reduce the direct costs of raising children and as a result encourage higher fertility (WB report, p. 39). In Vietnam such effect would work against the government’s policies that search to slow down the country’s demographic growth. Not supporting extremely poor children has, however, negative lasting consequences. Children living in extremely poor families are more likely to be working and among the poor children those working are much less likely to go to school (see Table 14). Children who are withdrawn early from school face a lifetime of lower earning (WB report p. 15). The PPA has collected evidence of these negative consequences. In the province of An Giang, the local officials not only observe that “many poor households have to stop their children’s education” but signal the negative long-term impacts on children’s physical and mental development as busy poor parent provide less care to their children. (PPA p.26). These children are more likely to continue to be poor during their life. As a result, the likelihood of poverty initially related to an unfavorable dependency ratio in the family is passed to the next generations that remain chronically poor even if they do not live in households with particularly unfavorable dependency ratio anymore. There are no easy solutions to solve the social problems related to child labor. Providing social allowances to the all the extremely poor families whatever their demographic structure is, however, likely part of the solutions. To avoid that social allowances promote fertility, certain programs, like the Bolsa Familia in Brazil cap the social assistance benefits set according to the number of children in the family to three (WB p. 134) . In Vietnam such cap could be viewed as discriminatory against the ethnic minorities who have currently larger families. As the WB report suggests the implementation of programs that require women to get minimal preventive health care and health education and increase the opportunities to deliver messages about the health benefits of birth spacing might also help reduce the unwanted impact on fertility of social allowances. Table 14 Rate of working children (age 6 to 14 years old) Average Confidence Interval Excluded with high dependency ratio over 0.5 6.2 5.5 7.0 Excluded with low dependency ratio equal or under 0.5 7.8 6.2 9.4 Total 6.7 5.9 7.4 Working poor Poverty is not only observed in the households with high dependency ratio. Table 15 shows that among the extremely poor households excluded from social allowances under the Decision 67 48.1% have a dependency ratio equal or lower than 0.5. In these households, there is one or more than one working age adult per dependent. In fact, 65.1 percent of the extremely poor households excluded from social allowance include 2 children or less. Table 15 Number of children in the poor families not eligibility to social allowances under the Decision 67 Registered poor and GSO extremely poor Share of households with 2 children or Number of households Distribution less Potential beneficiaries 552 800 32.0 78.8 Excluded with high dependency ratio over 0.5 344 599 19.9 13.6 Excluded with low dependency ratio equal or under 0.5 831 929 48.1 86.5 Total 1 729 327 100.0 69.5 Excluded with 2 children or less 766 372 65.1 This is true also among the households of ethnic minorities. The figures in Table16 show that more than half of the poor households of ethnic minorities who are currently excluded from social allowances are families with 2 or less children. As expected, this share is higher in the Kinh and Chinese group (71.1%). These figures show also that for most of the poor households excluded from social allowances the causes of poverty are not directly related to particularly high dependency ratio: 70.5 of the poor households from ethnic minorities are households with low dependency ratio, a share very similar to the one observed in the Kinh and Chinese group. Table 16 Number of children in the poor families not eligibility to social allowances under the Decision 67 Registered poor and GSO extremely poor Share with Share with 2 Share in 2 children children or total or less Share in total less Ethnic minorities Kinh and Chinese Excluded with high dependency ratio over 0.5 29.5 9.6 29.1 17.9 Excluded with low dependency ratio equal or under 0.5 70.5 80.3 70.9 92.9 Total 100.0 59.4 100.0 71.1 Based on the observation collected in the PPA, Table 16 to 18 present some other factors related to extreme poverty that might be important to take into account in the design of social policies that aim at eliminating extreme poverty. More exhaustive analysis can be found in the chapter of the poverty assessment that investigates the poverty profiles. Table 17 shows that the extremely poor households with low dependency ratio (the large majority of the extremely poor households not currently covered by social allowances according to Table 16) are in fact households with low income earners. Each able working age member earns on average 4.047 million VND. Able working age members of the other group of the poor (the extremely poor with high dependency ratio, and the other GSO poor) earn significant higher levels. Table 17 Average labor income per able working age households members thousand VND per year Survey 95% Average Confidence interval Potential beneficiaries 5186 4843 5530 Excluded with high dependency ratio over 0.5 5735 5352 6117 Excluded with low dependency ratio equal or under 0.5 4037 3813 4261 Other GSO poor 5979 5674 6283 Table 18 indicates that this situation is not strongly related to the fact that the extremely poor families could not benefit from the recent economic growth. Only about one quarter of the households with high and low dependency ratio (currently excluded from social allowances) reports that living conditions have not improved since 2001. Table 18 Share of households in which living conditions have not improved since 2001 Percent in the category Survey 95% Average Confidence interval Potential beneficiaries 41.9 35.7 48.2 Excluded with high dependency ratio over 0.5 23.8 16.6 31.1 Excluded with low dependency ratio equal or under 0.5 25.2 20.8 29.6 Other GSO poor 14.5 11.0 18.0 Table 19 shows that these households, however, have been borrowing in the past year. The share is of 40.8 and 33.7% in the case of the extremely poor excluded from the current social allowances. Table 19 Share of households with borrowing Percent of the category Survey 95% Average Confidence interval Potential beneficiaries 26.7 21.0 32.3 Excluded with high dependency ratio over 0.5 40.8 33.1 48.6 Excluded with low dependency ratio equal or under 0.5 33.7 32.5 42.9 Other GSO poor 33.1 28.5 37.7 Low income can be associated to particularly low returns but also to income shocks. Most of the poor have low education and few assets. It is not surprising that they obtain, consequently, low return from their work. However, the fact that the households with relatively small families report particularly low income suggests that a non negligible share of these families have suffered some income shocks in the past. The data does not make strong conclusion but the high share of households with borrowing points out to the same interpretation. As the PPA observes many poor and near poor households receive loans from private lenders for their production. Natural disaster and epidemics are the most serious risks for the poor. Harvest or livestock losses leave the poor during several years with outstanding debts with private lenders and at the Social Policy Bank. As a result poor households accumulate private debts over many seasons with accumulated interests which make them increasingly poorer (PPA section 4.1.2 and 4.1.3). Similar situation was described after the massive flood of 1998 in Bangladesh. No major food crisis occurred. “The flood did exact a heavy cost in terms of increased private debt because of extensive borrowing in private markets, a major coping strategy of the poor. Even when the worst outcomes are avoided, more subtle issues may arise, such as increased debt from which households may take a long time to recover” (WB, p.). Resistance to provide social allowances to extremely poor households of low income earner and/or that have suffered income shocks and consequently are largely indebted is often related to concerns that social assistance may reduce these households work effort. Another way to express these concerns is that social allowance risk to create “a culture of dependency and expectancy”. These concerns are explicitly reported by the PPA participants. Officials complain that State supports discourage the poor from ethic minorities to be active in their livelihood or to cope with natural and social changes and as a result they fail to generate willingness to escape poverty. The project beneficiaries are said to still maintain their attitude of reliance and that they do not want to work (section 4.1.4) The evidence supports the view that, in developing countries, safety net programs do not often reduce labor effort substantially. Academic studies show virtually no effects on the household’s primary earner and a limited effect on the work effort of secondary earners in low income households. In Mexico, adult earners benefiting from the conditional cash transfer PROGRESA worked as much as those in a randomized control group. In Brazil, simulations of the impact of the program Bolsa Familia on adult work effort found that the transfer amounts would have very little impact. In Armenia, the employment rate and hours worked by adults in the Family Poverty Benefits Program were similar to those for a matched sample of non participants. On the other hand, in Sri Lanka, a rice subsidy program seems to have induced labor disincentives (WB, p). Concerns that social allowances reduce beneficiaries’ work effort should not, therefore, stop the government in implementing new social policies toward all the extremely poor. Concerns bout the emergence of culture of dependency should not be, however, discarded and have to be taken into account in the design of the new policies. 7. One example of Family Poverty Benefits There is not a single policy design to support the extremely poor. The reasons that families are extremely poor are numerous and vary depending on the country and the period. The design must, moreover, take into account how much concerns about the impact of the new policies on fertility and work effort are important in the country. Consequently, general advises how to proceed can only remain rather vague. Extensive advises about all sort of settings are presented in the WB report. To help the reader get a precise idea, however, of the type of policy that can be developed, this note has opted to present one example: the Family Poverty Benefit program of Armenia. Armenia is a transition country that moved in 1999 from a social allowances system based on categorical groups like the one currently implemented in Vietnam to a most inclusive approach. Armenia’s targeting performance is also mentioned as excellent by international standards: it is as good as or better than the targeting achieved by the much- praised conditional cash transfers in Latin America (SP paper). Because, through that process, all the documentation was made public, the example of Armenia gives a good opportunity to understand how new social policies can be developed and implemented to reduce extreme poverty. The following extracts are taken from the annex of the World Bank poverty assessment of Armenia, 200x. All the legal documents are presented in … The major component of the reform in Armenia was the consolidation of 26 small, uncoordinated categorical benefits in cash paid to individuals into a single monthly cash benefit that would be targeted at the most needy families (the family benefit). Various targeting options were considered, including using a regression calculated from the 1996 Household Budget Survey. However, consumption correlates were not sufficiently strong and the Government decided to target the new family benefit using a somewhat adjusted proxy means-testing targeting mechanism developed for humanitarian assistance distribution. The new system was introduced in 1999 but all the new regulations were approved in 2000. The Regional Social Service Centers were given the responsibilities of implementing the policy. In order to create local dialogues about the way the policy was implemented, local Social Assistance Boards were also put in place with representatives (on a voluntary basis) of social and employment services, heads of social insurance centers, local administrators, local officers dealing with adolescent’s problems as well as representatives from NGOs. Families who regard themselves as poor or extremely poor and seek state assistance submit, first, their applications and all the required documents to the Regional Social Service Center. Inspectors process the application and eventually visit the family in order to check the accuracy of the information. This set of information constitutes the family Social Passport on which the Center calculates a ranking score. Families that score within the eligible ranges can benefit from the state support for 12 months. In order to continue receiving the benefit, the household has to apply for the benefit again and go through the same eligibility testing procedure. The benefit administration is computerized, and the centralized data base is managed by the Ministry of Social Security. The data base is linked to other data bases relevant for cross-checking of the economic and social situation of the applicant families (electricity consumption, telephone bills, customs records, the list of pensioners, etc.). The benefits are delivered to the beneficiaries through the post offices. The family score is obtained with the use of a proxy means testing formulas that includes the following variables: - social category of each of the family members, or individual “social risk” (Pk) and related average “social risk” for the family (Pm); - number of the family members not capable of working (Pc); - place of residence (Pr); - housing situation (Ph); - car ownership (Pa); - private business (Pb); - the document issued by respective territorial center for social services verifying the social and economic situation of the applicant family (Pf) and its eligibility for the benefit (“eligible” and “not eligible”); - family income (Pi). The particularity of the Armenian system is that whereas in other countries regression analysis or similar statistical techniques are used to determine the PMT scores (see example in Box 1), Armenia’s weights and point system was determined based on expert opinion of what would be the likely predictors of poverty 4. The score of the family need (P) is calculated using the following formula: P = Pm * Pc * Pr * Ph * Pa * Pb * Pf * Pi (a) Socio-economic category of each of the family members (individual social risk)—Pk and related average “social risk” for the family (Pm). Each family member is screened for a certain social category, e.g. for whether she/he belongs to a certain category of “social risk”. Each category brings a certain number of points. The number 4 This aspect was apparently driven by the fact that households with informal income can underreport income to become eligible reflects the assumed level of need of each category. The list of categories and corresponding number of points is presented in the following table. These categories are similar to those considered in Vietnam under the decision 67. Social category Points 1 Biological orphan (no parents) 50 2 First category disabled 48 3 Child invalid (up to 16) 45 4 Biological orphan (one parent deceased) 43 5 Second category disabled 39 6 Pensioner (75+) 39 7 Single pensioner 36 8 Child below 2 year of age 35 9 Pensioner 34 10 Child 2-18 33 11 Pregnant women (20+ weeks) 30 12 Third category disabled (below pension age) 28 13 Unemployed 27 14 Single mother child 26 15 Child of divorced parents 26 16 Public university student 22 17 No social category 20 20 In cases where one person belongs to several social categories, a weighted average is calculated. The weights are as follows: for the category with the highest number of points 1.0; for the second highest 0.3 and for the third and all the rest 0.1. For instance, a person can be a 17-year old child (category 10), with divorced parents (category 15), a student (category 16) and a third category disabled (category 12). Her/his individual social category (or “social risk”) score is calculated in the following way: Pk(ind) = P10 + 0.3*P12 + 0.1*(P15+P16) = 33+0.3*28+0.1(26+22) = 46.2 The average “social risk” score for a family is calculated as arithmetic mean of the family members’ scores. (b) Number of family members not capable of working (Pc). The value of this factor is calculated in the following way: Pc = 1.0 + 0.02*m, where m is the number of the family members incapable of working, namely children up to 16, women over 63, men over 65 and first and second category disabled. (c) Place of residence (Pr). For most of the settlements in Armenia, the value of this factor is one. However, there is a list of 173 settlements (in the earthquake zone and border territories) for which the coefficient ranges between 1.03 and 1.05. (d) Housing situation (Ph). Housing situation is classified into 6 categories with the following coefficients: “domik” (temporary shelter such as a carriage, a barrack, etc., in particular in the earthquake zone) – 1.2; homeless – 1.07; unsafe dwelling – 1.05; collective center – 1.03; other – 1.02; permanent dwelling – 1. (e) Filter variables (0 or 1). The following factors are used as filters: a car ownership (Pa), private business (Pb), and a document issued by respective territorial center for social services verifying the social and economic situation of the applicant family and its eligibility for the benefit (Pf). Their value can be either 1 or 0. Obviously, 0 for any of the three (the family has a car and uses it, the family or its members are running private business and the social services center has assessed the family as ineligible for the benefit) eliminates the family from the list of beneficiaries. (f) Family income. The family income coefficient is calculated using the following formula: Pi = 1.2 – 0.04*(ΣSj/m*M) (j=1...n) Where n is the number of the household members, sj is the income of the j-th household member, m is the number of the present household members, and M is the minimum wage (regulated by the Government). The income includes wages and salaries, income from self-employment, pensions, stipends and unemployment compensation. Income from farming is estimated based on cadastral income, while income from cattle breeding is estimated separately using a methodology regulated by the Government. The score that qualifies the household for the benefit is 36.01. It has remained unchanged since the family benefit was introduced. As it was mentioned already above, the Armenia’s targeting performance is mentioned as excellent by international standards. One advantage of the ranking score is that the system can differentiate the amount of aids according to it. For example in Georgia, families with very low score that are considered extremely poor receive monetary aid for subsistence while the others receive subsidized access to health insurance and electricity (xxx). Evaluation reports mention, however, that the policy does not reach all the poor families because many poor families do not apply. The reasons are not clear: maybe they do not understand well the process, or it seems too complicated; maybe the do not want to ask for the State support (xxx). Currently the screening process in Vietnam is very efficient: only a few registered poor are not extremely poor. Many extremely poor are excluded from registration, however. Some studies suggest that the main reason is that the lack of funding of social policy leads local authorities to reduce the list of the potential beneficiaries. It could be also that the current screening method has some limitations. Based on interviews conducted in 3 provinces at province, district and commune level, a GTZ mission (2008) signals that “potential beneficiaries (even those registered) display a general lack of information ….concerning [the] services and benefits delivered by social assistance … together with a low awareness of having actually a (claimable) right to receive social assistance and proper services” . Eliminating extreme poverty requires, therefore besides the systematic collection of information, the implementation of more resolute actions to reach and include all the extremely poor (public campaigns, systematic visits of social workers in potentially extremely poor households etc.). Box 1 The following table gives an example of the weights derived from poverty regression used in means testing procedures. (A. Posarac The system of Social Assistance based on proxy means testing, presentation in Sarajevo) Coef sd Coef Sd lhhsize -0.562*** 0.036 Cooking Oven 0.054*** 0.020 lhhsize2 -0.014 0.011 Cooking Stove -0.001 0.019 Share of children 0-6 -0.534*** 0.037 Microwave Oven 0.067*** 0.013 Share of children 7-14 -0.520*** 0.035 Dishwasher -0.007 0.035 Share of children 15-18 -0.121*** 0.046 Vacuum Cleaner 0.060*** 0.009 Share of adult males (19-60) (dropped) TV Set 0.112*** 0.029 Share of adult females (19-60) -0.166*** 0.035 Satellite Receiver 0.042*** 0.010 Share of elderly -0.144*** 0.038 Video 0.045*** 0.012 Share of job holders -0.010 0.030 Video Camera 0.067*** 0.020 Share of employed in public sectror 0.038*** 0.012 Computer 0.063*** 0.010 Recieves any subsidy -0.106*** 0.017 Telephone 0.075*** 0.009 Characteristics of the head Mobile Phone 0.128*** 0.011 Female headed hh 0.037** 0.015 Air Conditioner 0.099*** 0.018 Age of the head 0.001* 0.001 Solar Heater 0.008 0.012 Education of the head Private Car 0.210*** 0.009 Illiterate (dropped) Radio 0.005 0.008 Read/Write 0.048*** 0.017 Sewing Machine 0.004 0.011 Primary 0.056*** 0.016 Characteristics of dwelling Elementary 0.043*** 0.016 Number of rooms 0.050*** 0.005 Basic 0.021 0.024 Villa (dropped) Secondary 0.083*** 0.018 House 0.008 0.041 Intermidiate 0.088*** 0.020 Flat 0.029 0.041 University 0.156*** 0.021 Dwelling area m2 0.001*** 0.000 Occupation of the head Type of ownership hocc_1 -0.031 0.059 Own 0.040*** 0.015 hocc_2 0.101* 0.059 Rented 0.018 0.016 hocc_3 0.007 0.057 For work -0.096* 0.049 hocc_4 0.026 0.058 Free (dropped) hocc_5 -0.013 0.058 Source of heating hocc_6 0.038 0.056 Kerosine 0.056 0.054 hocc_7 (dropped) Gas 0.102* 0.055 hocc_8 -0.005 0.056 Central 0.249*** 0.057 hocc_9 0.006 0.056 Electricity 0.169** 0.070 hocc_10 -0.033 0.055 Wood or coal 0.039 0.058 hind_1 0.030 0.040 Other (dropped) hind_2 -0.010 0.060 No heating 0.005 0.065 hind_3 0.033 0.058 Source of drinking water hind_4 0.068 0.067 Piped water -0.237*** 0.068 hind_5 0.024 0.058 Well -0.217*** 0.070 hind_6 -0.010 0.056 Water tank -0.180** 0.071 hind_7 0.051 0.057 Mineral/filtered water -0.185*** 0.069 hind_8 0.027 0.057 Spring water (dropped) hind_9 -0.011 0.057 Public sewage 0.017 0.011 hind_10 -0.000 0.060 Construction material Household assets Clean Stone 0.209*** 0.065 WASHING MACHINE 0.095*** 0.024 Stone & concr 0.136** 0.064 REFRIGERATOR 0.070*** 0.024 Concreate 0.113* 0.064 FREEZER 0.044*** 0.014 Cement block 0.096 0.063 8. Conclusion This study investigated the possible impact of the new social allowance policy designed in the Decision 67 of 2007. Because there are still no statistics collected about its implementation, the study uses the data collected in the VHLSS 2006 to measure how many people and families could benefit from such policy if its implementation did not encounter problems of screening or funding. It shows that a large share of the extremely poor would remain excluded. References Bender Katja , Bui Xuan Du and Eichler Philip (2008) Strengthening public service delivery in social assistance. Fact finding mission GTZ and ILSSA March-April 2008.Processed Government of the Republic of Armenia, Ministry of Social Security “The poverty family benefit and lump sum assistance. Enforcement provisions” Decree 350. Yerevan July 2000 Grosh M., del Ninno C., Tesliuc E., Ouerghi A, Milazzo A.and Wiegand C. (2008) “For Protection and Promotion: the design and implementation of effective safety nets” . The World Bank World Bank (2003) Armenia poverty assessment Report No. 27192-AM Annex on households with people with severe disability With the participation of Tam Tran Ngo Minh (VASS/CAF) Disabled people face many difficulties and deprivations that are far beyond economic problems and, the economic difficulties they face are multidimensional. Studying the situation of the disabled person and design policies to support them are very complex tasks. This note only investigates one aspect: household’s welfare. Several reasons justify that having a disabled person at home can reduce household’s welfare. First, some disabled persons have lower productivity than not disabled people so they earn lower income. Some cannot work at all. Second, some disabled persons need special care and spending. This reduces other household members’ consumption level and limits their possible dedication to work, study and other direct or indirect income opportunities. All disabled people do not such have negative impact on the households they live in. More likely, people with severe disability do. In order to alleviate this impact, the Decision 67 target poor people with severe disability who live in poor households. This notes looks at this issue and shows that among the households that are severely disabled those including children are much more likely to be poor than the others. Households that include a working age person with severe disability are also more likely to be poor. These results suggest that particular effort should be developed during the implementation of the Decision 67 to effectively cover these households and design specific interventions to support better these populations. A. Data and definitions The analysis is based on the results of the VHLSS of 2006. The section muc3B reports people self-rating (little difficult or very difficult) upon 6 types of disability: seeing, hearing, concentrating & understanding, taking care of themselves, mobility, communicating with others. The study focus only on the people with severe disability that is the persons who have answered that at least one of these tasks was very difficult. According to these criteria, 1219 respondents of the VHLSS 2006 were with severe disability. The small number of the sample makes difficult analysis by sub-groups, the study, accordingly, searched to regroup the 6 types of disability into 3 groups: Group 3: includes all the persons who report that is very difficult to understand something new or, to understand what somebody else is talking or, to start and keep a conversation going Group 2: includes all the persons who report that is very difficult without using tools to walk about 400 meters long or, to walk up 10 steps without stopping or, that report that they need somebody else to help in the everyday life. Group 1: includes the other severely disabled to whom it is very difficult to see even when wearing glasses or, to listen even with hearing aid or, to bow and bend the body or to stand on the knees or, to raise hands over the head or to use the fingers to button the shirt or, to put on socks or, to tie shoe laces or, to comb the hair, or to eat without aid. Group 0: includes all the persons who are not severely disabled and included in the above groups The groups were built based on the type of disability and the poverty rate among the disabled of each specific type when compared to the poverty rate of the population that is not with severe disability. Because the poverty rate is the highest among the people with disabilities of the group 3, all the people with such disability and other disabilities were included in the group 3. Similarly, because the poverty rate of the people with the disabilities of the group 2 are also very high in comparison to the people with other type of disabilities, all the persons with such types of disabilities and other types were included in the group 2. The survey reports also the reasons that caused the current disability. There are 12 possible causes and several reasons can be reported for one kind of disability. The study selected only the first cause reported by the person. The causes are organized into 4 groups: War and birth: when the disability is a consequence of birth defect, war, agent orange or low education level (any person with no birth defect is able to realize the proposed tasks without any education) Accident: when the disability is a consequence of road traffic accident, work injury or, natural disaster Illness: when the disability is a consequence of illness or behavior related to social evil, old-age (because most of the disabilities are related to developing illness), problems related to other disability, and other. The cause is reported if the person is heavily disabled and the cause is cited as one of the first cause for one type of disability. This means that one person can be associated to several causes if these persons have several disabilities (even at a mild level included because the question is presented whatever the level of difficulty). In fact only this happens in the case of 2 observations; 41 persons do not report the cause. The survey, finally, reports the length of time the person has been having the disability. B. Frequency and general characteristics According to the results of the VHLSS 2006 (and this methodology), there are currently in Vietnam about 2.6 million people severely disabled, that is 3.1% of the population. As the figure in Table 1 show most of them, 77.8%, are disabled as a result of illness or old-age; 17.4% are disabled as a consequence of the war and/or because of birth defects. Finally 4.8% of the persons with severe disability report that the cause was accidental. Table 1 People with severe disability: Frequency and Reasons by type of disability Frequency Reason reported for the disability Number of In % of the War or Accident Illness Total people total birth population defects Other population 79 885 013 96.9 - - - - Severely disabled 2 595 840 3.1 17.4 4.8 77.8 100.0 Mentally 870 330 1.1 28.2 2.9 68.9 100.0 Walking and taking care 1 261 638 1.5 11.4 4.6 84.0 100.0 Others 463 872 0.6 14.0 8.8 77.2 100.0 Total 82 480 853 100.0 Source: VHLSS 2006 The people with important problems that affect their concentration, understanding and communication with the other represent 1.1 % of the total population. The causes of these problems are more likely related to the war or birth defects that in the other cases. People with important mobility problems or that cannot take care of themselves represent 1.5% of the population. Table 2 shows the same results by age groups. Because health deteriorates with age the incidence of severe disability among the elderly is very high (15.4%). About 1.2 percent of the children and 1.5% of the working aged people are severely disabled. Table 2 People with severe disability: Frequency and Reasons by age group Frequency Reason reported for the disability Number of In % of the War or Accident Illness Total people corresponding birth population defects Children 242 738 1.2 25.7 0.8 73.6 100.0 Working age 795 704 1.5 37.6 12.3 50.2 100.0 Elderly 1 557 398 15.4 5.7 1.6 92.7 100.0 Total 2 595 840 3.1 17.4 4.8 77.8 100.0 Source: VHLSS 2006 The causes related to accidents are more important among the working age than in the other groups. There is a rather large share of children disabled because of illness. The graphs in Figure 1 compare the frequency of severe disability among men and women by age groups. The shapes of the curbs are similar for both sexes. Before the age of 50, the incidence of disability slowly increases with age then it accelerates, particularly after the age of 70. Among the elderly 70 and above, the frequency of disability appears to be higher among women than among men. Because women have longer longevity, this result probably reflects that contrary to the men, women at advanced ages survive more likely health shocks than men. Some of them, however, remain severly disabled. Figure 1 percent of heavy disabled 70% 60% 50% 40% w omen 30% men 20% 10% 0% er rs rs rs rs rs rs rs rs rs ov 4y 4y 4y 4y 4y 4y 4y 4y 4y 0- -1 -2 -3 -4 -5 -6 -7 -8 & 10 20 30 40 50 60 70 80 90 Finally, about 73% of the population in Vietnam lives in rural areas. It is not surprising, therefore, that most of the severely disabled live in such areas. However, the incidence of severe disability among the working aged people is more important in rural areas: 1.6% of the working aged populations in rural areas are severely disabled against 1.2% in urban areas (the difference is statistically significant). One reason could be that people working in agriculture are particularly more exposed to work injury or disease than other people. Another reason could be that disabled people are less likely to migrate to urban areas. C. Poverty rates About 12% of the households in Vietnam include one or several persons with severe disability. These households are statistically more likely to be poor than the rest of the population. As the figures reported in Table 4 show, 20.8% of the people living in household with one or several severely disabled persons are poor while the rate is 15.1% among the rest of the population. The measure of poverty is based on GSO evaluation of household expenditure per capita and estimate of the poverty line. Table 4 Severe disability: poverty rates Number of In percent Poverty rates Survey Confidence Interval1/ Type of household household of total in % of Lower bound Upper bound households population Do not include severely disabled 17 324 857 88.3 15.1 14.1 16.2 Include severely disabled persons 2 305 016 11.7 20.8 17.8 23.8 Total population 19 629 872 100.0 15.8 14.8 16.8 With severely disabled children 232 836 1.2 35.3 25.3 45.2 With severely disabled working aged persons 821 378 4.2 23.4 18.1 28.6 With severely disabled elderly 1 340 384 6.8 16.6 13.1 20.2 Note: 1/ poverty rates are statistically different if the corresponding survey’s confidence intervals do not overlap. Source: VHLSS 2006. There are about 1 million households that include children or working aged severely disabled. These households are significantly more likely to be poor than the rest of the population. The poverty rate of the households that include children severely disabled is particularly higher: 35.3 %. By contrast the households that include severely disabled elderly are not particularly poor.