P O V E R T Y & E C O N O M I C P O L I C Y R E S E A R C H N E T W O R K. 9th General Meeting Angkor Era Hotel Siem Reap, Cambodia December 3-9, 2011

Transcription

1 P O V E R T Y & E C O N O M I C P O L I C Y R E S E A R C H N E T W O R K 9th General Meeting Angkor Era Hotel Siem Reap, Cambodia December 3-9, 2011 Implementing a Community- Based Poverty Monitoring System in Argentina Institute of Economics UNICEN November 15 th 2011 IDRC photo: N. McKee

2 Implementing a Community-Based Poverty Monitoring System in Argentina Institute of Economics UNICEN November 15 th 2011 Abstract: The present article summarizes the experience learnt and some early results of the pilot CBMS administrated in Argentina in the cities of Urundel and Tandil during September of 2011.The institution in charge of the pilot was the Instistuto de Economía, UNICEN. 1

3 Table of Contents 1. Introduction Methodology Selection of places to census Questionnaire Strategy for Promotion Implementation Pre-census activities Selection and training of enumerators, supervisors and data enters Advertisement Time Window Problems and remedial actions in the fieldwork Preliminary Results Urundel. Some descriptive Statistics Tandil. Some descriptive Statistics Core Indicators Conclusions

4 1. Introduction The CBMS Pilot in Argentina was implemented in two very different places: Urundel, a rural town in one of the poorest provinces, and Tandil, a middle income urban city. These two places capture the heterogeneity of Argentina. We adapt the questionnaire to the reality of the two cities. The Census was implemented in September 2011, in both cases with the help and collaboration of the local authorities (municipal government). Enumerators were selected in the community. The rest of the article is organized as follows. Section 2 presents the design of the Pilot. Section 3 presents details of the implementation. Section 4 presents preliminary results, and section 5 concludes. 2. Methodology We closely followed CBMS methodology as proposed by Reyes and Due (2009). In particular we administrated a Census to two cities in Argentina, Urundel and Tandil. In the first case, it was a census to the entire city, in the second, we census a selected area which represents approximately 10% of the total population of the city. The questionnaires were adapted to the reality of each city. In particular, Urundel is a rural city in one of the poorest provinces of Argentina, with aborigine roots and an income per capita which is almost one fourth of the national average. We include in this case a module in the questionnaire asking about agriculture production and ethnicity. On the other hand, Tandil is a middle income and middle size city, with an income per capita which is 35% above the national average. As an urban city, the problems are different. We did not include the agriculture production and aborigine culture modules, and instead we include a module ofcrime. Both censuses were done with the assistance of local governments, who get involved in the process. The contact with the local government was established very early in the process, when we submit the proposal to PEP. In the case of Tandil, where due to the size of the city we had to restrict the area to census, the election of the zone was done jointly with the municipal government, as well as the strategy to select and train enumerators. Early involvement was a positive aspect of the experience, but on the other hand it tends to delay the implementation as the speed in making decision in local governments is usually low. Enumerators and data-enters were selected from the local communities. In the case of Tandil, the areas of the local government involved in the process were the Chief of Cabinet (leaded by Mr Matias Civale) and the Secretary of Social Development (leaded by Oscar Perugia). 3

5 In the case of Urundel, and given the distance to Tandil (more than 1700 kms), the University signed an agreement with a local entity located in Salta capital city (4 hours away from Urundel). We worked together in adapting the questionnaire and training modules. The local entity was in charge of designing the field work, selecting and training enumerators, validating and processing the data Selection of places to census The selection of the two places was strategic to capture the heterogeneity of Argentina. We contact municipal government directly, as the minor unit of government in Argentina, to be as close as possible to the community. Regarding the heterogeneity, Argentina has large variation in living condition across region and very high income inequality. The most salient characteristics of this country are: Very Urban: 90% of the population lives in urban areas (town or cities of at least 5000 people), and almost 75% lives in cities of more than inhabitants. Low Density: population density is 14 individuals by squared kilometers on average, but in some regions of the Patagonia (southern part of the country) the density is as low as 1 individual per squared kilometer. Economic Concentration: Buenos Aires city and the surrounding neighborhoods (the Great Buenos Aires (GBA) Region) has almost one third of the total population, and it produces almost 60% of total GDP. High income inequality and regional inequality. Large and relatively rich urban areas, such as GBA or Rosario, have entire neighborhoods (slums or favelas) of very poor people. On the other hand, half of the country, what sometimes is called the Great North, is poor and more rural. Rural areas are not poor. Contrary to what happens in many other countries, rural areas are mostly rich areas, of high technology agricultural production, the exception is the Northern part of the country (see Auguste (2007) 1. The northern part (known as Norte Grande) is the poorest and there are several communities living in self-subsistence. We selected the province of Salta first because it belongs to Norte Grande and second because we have already worked with its provincial government. This province has a large city called also Salta, and several small towns, most of them rural, many of them with high incidence of aborigine population. After evaluating several municipalities in the province and establishing contact with authorities we selected Urundel (provincial government helped us in the selection). 1 Auguste (2007) 4

6 In the case of Tandil, the selection was done basically because the University to which the Institute belongs to is located in this city, and because the University has a history of collaboration with municipal authorities. Tandil is a city of inhabitants located in the middle of the province of Buenos Aires, 365 km south from Buenos Aires city, and Urundel a village of approximately 3000 inhabitants, located in the of the province of Salta, 200 km away from the city of Salta (capital of the province with the same name), and almost 2000 km away from Tandil. Urundel has in the 2001 Census less than 8 inhabitants per square kilometer, Tandil 22. In terms of regions, Tandil belongs to Region Pampeana, which is a plain with very fertile soil (Pampas, from Quechua, an aborigen languague, meaning "plain"). The Pampas are vast plains covering the provinces of Buenos Aires, La Pampa, Santa Fe, Entre Ríos and Córdoba, and only interrupted by the Ventania and Tandilia hills (with average height of 1,300 meters and 500 meters respectively). This region produces more than 70% of the Argentine GDP. Salta belongs to the Northwest Region (NOA), in the Andean mountains. This region was before the Spaniard conquest the Southernmost region of the Inca empire, belonging to the Collasuyu area (one of the four areas the empire was divided). It has a very rich aborigine culture and it was a strategic point connecting Lima with Buenos Aires. When Argentina obtained the independence from Spain, economic activity in the region refocused from Lima to Buenos Aires. Salta emerged from the War of Independence politically in disarray and financially bankrupt, a condition that lingered throughout much of the 19th century. However, in the late 19th and early 20th centuries, the arrival of Italian, Spanish and Arab immigrants, particularly Syrians and Lebanese, revived trade and agriculture all over the area while further enhancing the city's multicultural flavor. In the last two decades Salta has been growing fast, favored by oil production, tourism, and agriculture. Nevertheless is among the poorest provinces in the country. The first measure of poverty was obtained in the first national Census of 1869 (Basic Unsatisfied Needs Index). Salta was ranked as the fourth poorest province, in the last Census available (2001) was still among the poorest ranking third. The GDP per capita of Salta is one third of the national GDP per capita. Figure 1 shows poverty rates by income level obtained from Household Survey for the main regions of the country and for Salta city and Tandil city (there are not poverty measures by income for Urundel). As shown, Tandil has a poverty rate below the country average, and below its region (Pampeana), whereas the city of Salta is in the average of its region (NOA). 5

7 Figure 1. Urban Poverty (head count ratio, income definition of poverty, 2007) Figure 2. Tandil and Urundel 1753 km 6

8 Urundel is an area of "foothill forests". The main economic activity is agriculture, historically dominated by sugar, but currently more diversified. It has a conglomerate of farms doing premiere horticultural crops and export fruits such as: orange, grapefruit, lemon, mango, papaya, bananas, green peppers, tomatoes, watermelon, pumpkin, melon, strawberry, sweet potato, cassava, and coffee. There is also wood production from the forests. In terms of aborigine groups, there are currently two Guaraní native communities living in the town: the Iyiguapentirami and the Avaguaraní. They are not segregated in a particular neighborhood but rather distributed along the city. Figure 3. Satellite Picture of Urundel Tandil was founded in 1823 as a Fort in the frontier between what was Argentina at that point and aborigine territories (that later were controversially added to Argentina). Being victim of incursions of aborigines ( malones ), the city was several times attacked. In 1875 the native aborigines (Pampas and Ranqueles) jointed forces and performed the bloodiest attack to the forts of Tres Arroyos, Tandil, Azul, and other towns and farms in the region. The response of the argentine government was the Conquest of the Dessert, a military attack to all the communities living to the south (a genocide), that lasted almost 10 years (the last battle of the campaign was in 1884, which ended up shaping the current Argentina and destroying most of the aborigine communities in the Pampas and Patagonia. The new territories were populated (or colonized) with new European migrants. In the case of Tandil, most of the early migrants come from Spain (mostly Basks) and Denmark. Later, it come a large contingent of Italians that populated the region, being Italians descendents 7

9 the most populous group today in Tandil. The new territories developed fast influenced by agriculture, what shaped the north-south dichotomy of Argentina. In the case of Tandil, the attractiveness of the Tandilia Hills allowed the early development of tourism and mining too. GDP per capita in Tandil is estimated in USD per capita. It has been growing well above the country average in the 2000s, with the development of new economic sectors (such as software industry) and the attractiveness of its good quality of life. The economic boom has pushed urban land prices up, particularly in the downtown area, what has segmented the city in a similar fashion as the larger cities of the country, expelling the medium and low income households to the periphery. It is estimated than in the north part of the city, called Tandil Norte (Northern Tandil) -developed outside the original urban zonecurrently has inhabitants. This part of the city is on the other side of the main route to Tandil (Route 226). We select this northern part for our pilot, which includes 3 of the 23 barrios (or neighborhoods) of Tandil: 1) San Cayetano, 2) Villa Aguirre and 3) Palermo. This area has approximately 10% of the total population of Tandil, and has large incidence of new migrant from nearby cities and other provinces that came to the city searching for a job. Figure 4. Pictures of Urundel Town 8

10 Figure 5. GIS Tandil Figure 6. Pictures of Tandil 9

11 But there is also the Other Tandil 10

12 2.2. Questionnaire We developed two different questionnaires for Tandil and Urundel, see Annex A. We grouped the questions in modules, and the same numeration was used for both places. Urundel includes a module of agriculture production and another module of aborigine culture. Tandil does not include these two modules but includes one of victimization (not included in Urundel). Based on the questionnaire we can have a set of common indicators. Both local governments were interested in asking specific questions about local public policies in the neighborhood, what was adapted to the programs and needs of each city Strategy for Promotion In Urundel, which is very small, the promotion was done by the Municipality. When we did the precensus all the population learnt of the Census. Similar diffusion effect had the call for enumerators. In Tandil promotion was done by the Municipality as well, and with posters in local shops. 3. Implementation 3.1. Pre-census activities In the case of Tandil, the municipal authorities had detailed information about the lots and population living in the area. We selected the Northern part together, because it has been a fast growing, low income area, which has many needs the local government was interested in better identify. The information provided by the municipality was used to determine the area to census and to allocate housed to each enumerator. In Urundel, we did a pre-census finding 633 inhabited lots that were divided into nine barrios or neighborhoods. This information was used to allocate enumerators Selection and training of enumerators, supervisors and data enters Recruitment of enumerators and supervisors In the case of Tandil, the Municipality on the one hand elaborated a list of interested people of the neighborhood contacting local churches and community and health centers (both institutions of the Municipality located in each neighborhood of Tandil). Some 11

13 organizations of neighborhood get involved too, as the Peñel Church and library La Esperanza. In addition, the University made a public call. In both cases the main requirement was that the enumerators were residents of the neighborhood. The first meeting with local interested was on August 16 th in La Florida (a municipal club in located in the area of the Census). In the first meeting both, representatives of the Municipality and of the Institute explained the census. After this initial meeting we did four meeting more, that helped to promote the census and select enumerators. Figure 7. Example of the First Presentation, selected slides translated For the training we distributed Enumerator Manuals, Supervisor Manuals and Questionnaires, that individuals took to their homes to study. We gave them one week between the initial training and the next one to read and practice. There were two meetings more, in one case they interviewed each other to practice, in the last one, we did a small precensus in a couple of selected blocks so the enumerator could face the real challenge of interviewing real people. After that, we evaluate the questionnaires taking into account the information of each person, and call a new meeting for the final training with those who were eligible. Among postulants we selected as supervisors those with experience in managing people or groups, surveys and accountable for working. On September 2 we did the final meeting with the selected enumerators and supervisors and we gave them credentials, cartography, forms, pencil and rubber. Most of enumerators were 12

14 selected to work part time, and compromised to do it for an entire month. We worked on average with 40 enumerators and 8 supervisors. Figure 8. Training in Tandil In Urundel, Delta Consultant was in charge of implementing the field work. This is a company advocated to empirical research (surveys) in Salta. A team of Delta did the precensus during a week in Urundel, and used that time to preselect and pre-train interested people. This group worked previously in a Social Census done in the city of Salta so they have a lot of experience, what was a key factor for the success in Urundel. One week before the census Delta team did the final training and final selection. Enumerators, supervisors and data enters worked full-time. In total we hired: 12 enumerators, 2 Supervisors and 3 data entry without previous experience with the exception of one, all natives from Urundel. Figure 9. Training in Urundel Recruitment and training of data enters In Tandil, data enters were selected and trained independently of enumerators and supervisors. The training was done in the University Campus. We asked for a different 13

15 profile, people with at least secondary education, with experience in computing. Most of the positions of data enters were filled by students of the university that live in the neighborhood, and for the data entering they had the option to do it in the University computer lab or at home. In Urundel, data enters were selected from the pool of people who applied to do the census, since there was just one call for the three positions (supervisors, enumerators and data enters) and individuals were selected into one category according to their profile. In the case of Urundel, the training and data entry was done with equipment facilitated by Secretary of Sports and Culture. For the data entry, Delta and the Institute worked together developing specific software programmed in Fox. This was done to avoid mistakes in the data entering process Advertisement In Tandil, the massive diffusion of the census was carried out by Institute of Economics by concurrent posters up in different places such as neighborhood community centers, schools, bakeries, kiosks, in local minimarkets and in the Universidad Barrial (A small branch of the university in Villa Aguirre, that provides training in different areas to the community), and delivering leaflets to residents, and on the other side by the Municipality, through radio and major newspapers of the city. Figure 10. Flyer and Poster. Tandil 14

16 In Urundel, it was easier because it is a village with few inhabitants. The diffusion of the census was done with equipment of the Municipality given for the Secretariat of Sports and Culture and through the radio of the village with a spot produced by Delta Time Window In Tandil the Census was officially launched on September 9. There was a public act in the Municipal Government House, in which participated Miguel Lunghi, Mayor of the town; Matías Civale, Secretary of Government; Natalia Martignoni, Director of Social Development; Alfredo Rebori, Dean of the Faculty of Economics and María Del Carmen Romero, member of the Institute of Economics and project coordinator. After this public act the census was launched on September 12 th. The first two weeks the process was slow because of the weather and because some enumerators deserted. We called new enumerators from the waiting list. The census was completed by October 31. It was also conducted on weekends. On total, almost 3000 units were visited. The process of data validation was done continuously. Every week (on Friday) supervisors met with the Institute team and updated the census evolution and worked together to resolve specific cases. In Urundel, the survey was carried out from Tuesday August 30 th to Monday October 17 th in two stages and with collaboration of the municipality of Urundel. The first stage, in the case of Urundel was from the day Tuesday August the 31 st to Sunday September 4 th. All the community was covered and rejections, absence and other nonresponses were registered. The following week the supervising team worked identifying partial deliveries of forms and errors. In the final stage supervisors and enumerators revisited to finish the census Problems and remedial actions in the fieldwork In Urundel, there were just a few problems. Worth to mention is the case of an aborigine enumerator rejected by individuals from a different aborigine community, which was resolved by reallocating enumerators. A couple of enumerators were not committed to fulfill the full time job and were replaced. Both problems cause a partial delay in the field work. In the first stage absence was high due to a local holiday, but in the second visit this was resolved. In general, minor problems were resolved by supervisors and also regular meetings conducted by the Field Work Coordinator. In total 670 houses were interviewed, with 92 nonresponses. In Tandil from the list facilitated by the Municipality, there were many empty lots, houses in construction or nonresidential buildings. From the identified households, around 2100 were 15

17 successfully interviewed, 336 refused to answered and 259 were temporarily empty houses. In Tandil we had a harder work to convince people to participate in the survey. Fear of unlawful behavior was common, what is understandable in the area because it has a very high crime rate. Another fear of many households was the use of the information, since social plans are very common in the area, and many houses have duplication of plans, what in principle is not allowed (they feared to lose some benefit). In cases in which people did not want to respond, enumerators explained the objective of the work so that they could observe the importance of his response. Enumerators also gave the option to call to a university telephone if the households wanted to verify information, and they revisited the house later, to give time to the household to verify the information. During the implementation of census, communication with the municipality was very fluid in both cases. The Institute of Economics reported at least once a week to the Municipality on the progress of the process with sufficient details as they were active participants in the implementation of the census. Information we reported to the Municipality was: Number of surveys that were made for each enumerator, Areas of each enumerator, Areas which were completed, Number of places where they did not find anyone or did not want to answer was an election year, what was not good for the census. In the case of Salta, the local election was done in the first semester of the year, and on October 23 rd it was only the presidential election, what did not introduced any noise in the pilot CBMS. In the case of Tandil there was some noise, since October 23 rd were both presidential and local elections. There were many political surveys that produced fatigue in the people we wanted to census. The local government, running for re-election, was at some stages of the process delaying the decisions and focusing more on the political debate than in implementing the census. On the other hand, there were some political problems and a very hard political battle between top contenders, which was mostly taken in the poorer areas such as the one we census. But in spite of this noise, we are in conditions of saying that it was a real success, local government did get involved, and it was interested in the process. 4. Preliminary Results 4.1. Urundel. Some descriptive Statistics In Urundel the census collected information about 670 households, or 2867 people. 16.5% of them self-reported to belong to an aborigine community: 41% to the Iyugua Pentirani community, 56% to Aba Guarani and 3% to other. Families are in general large for national standards, and augmented families (adults living with parents) is common. 16

18 Table 1. Urundel Male Female Mean Age Average family size Total Non Aborigine Aborigine Figure 11. Distribution of Families by family size. Urundel Density famsize Other demographic statistics are: 85% of the individuals were born in Salta, and only 0.17% were immigrants (all of them from Bolivia). 72% were catholic (30% of non-catholic were aborigines). 90% of the adult population did vote in the last election (which is an extremely high participation rate). Illiteracy rate is 13.3% (for population with more than 10 years old) Of those self identified as belonging to an aborigine community o 59% states that they do not speak and neither understand any aborigine language (with a much higher incidence among youngest). o 13% that is not important to preserve aborigine culture, and only 30% practice or respect his/her community customs and celebrations o But 38% participates in activities organized by his/her community or in community jobs 17

19 Regarding economic activity: 43 hours a week is the working time for an average worker 43% of workers are in the agriculture sector, but only 22% are farmers 3.8% of workers are professionals 40% of the employees are informal (not registered in Social Security) 50% of the workers states to work in the public sector Regarding the problems in the community, in the most important problem 38% choose insecurity, followed by health. As the second most important problem, 49% choose Education, and 29% Health. Table 2. The three most important problems of your community First Second Third Obs. Freq. Obs. Freq. Obs. Freq. 1. Insecurity Education Health Poor Infrastructure Lack of public space Other No problem We included in the questionnaire a set of questions to measure the perception of the families regarding their socio economic status and social mobility. The main results are presented in Table 3. Surprisingly the perceived situation is extremely favorable. 65% believe they are in much better socioeconomic situation now that their parents were when they had their age. 49% believe this is due his/her personal effort, but 36% believe this is due to the country offering better opportunities. The optimism projects also to the future, only 10% believe they will be worst in 20 years from now. Regarding the self-perceived socioeconomic status, 53% believe they are poor, and 46 that they are middle income class. Only 1.3% classify themselves as very poor, which is extremely low compared to other quantitative measures (as quality of housing or educational achievement), what shows that the perceptions is build up comparing with the local community and not the entire country. 18

20 Do you believe your current socioeconomic situation compared to what your parents had at your current age is: Table 3. Perceived socioeconomic situation What is the main factor that explains your relative situation? Do you think your personal situation in 20 years will be? In your opinion you are: Better 64.9 Personal effort 49 better 54.2 rich 0.2 Same 29.6 Educational achievement 12.4 same 35.6 middle income 45.7 Worst 5.5 The country offers better opportunities 36.1 worst 10.3 poor 52.9 Total Other 2.48 total very poor 1.3 Total 100 total % of the households have at least one member with some disability; 50% of them are due to birth defects, 22% to diseases and 16% to accidents (only 7% of the disabilities are due to advanced age). The disabilities are spread in several factors, none of them has an incidence rate above 7% Tandil. Some descriptive Statistics In Tandil the census collected information about approximately 2600 households. Some of demographic characteristics found in the Census are: while most of migrants come from cities close to Tandil (Juarez, Necochea, Mar del Plata, Rauch, Azul and Ayacucho). Only 3.26% of the surveyed population come from other provinces different than Buenos Aires and Capital Federal and 1.46% come from other countries, most of them from Paraguay, Dominican Republic and Chile. Almost 70% of the population is catholic, while approximately 10% is evangelist and 20% atheist. 82% did vote in the last elections. Illiteracy rate is 1.78% (for population with more than 10 years old) Regarding economic activity: 41 hours a week is the working time for an average worker The workers are distributed in the economic sectors as follows: 19

21 Figure 12. Workers by sector 4% of workers are professionals 50% of the employees are informal (not registered in Social Security) Regarding the problems in the community, 33% choose poor infrastructure and 32% insecurity. As the second most important problem, 41% choose other and 17% poor infrastructure. Table 4. The three most important problems of your community First Second Third Obs. Freq. Obs. Freq. Obs. Freq. 1. Insecurity , Education 19 3, Health 9 1, Poor infrastructure , Lack of public space 12 2, Other , No problem Total In the set of questions to measure the perception of the families regarding their socioeconomic status and social mobility, a general situation of optimism is perceived. More than half of the households surveyed think that their situation is better than their parents, and this is mainly because the country offers better opportunities and their personal effort. The optimism is projected towards the future: 60% think they will be better in 20 years, while only 15% think they will be worse. In addition, 61% of the surveyed people think they are part of the middle income class. 20

22 Do you believe your current socioeconomic situation compared to what your parents had at your current age is: Table 5. Perceived socioeconomic situation What is the main factor that explains your relative situation? Do you think your personal situation in 20 years will be? In your opinion you are: Better Personal effort better rich 0.19 Same Educational achievement same middle income Worst The country offers better opportunities worst poor Total Other 5.46 total very poor 0.93 Total 100 total Regarding disabilities, 14% of the households have a handicapped member. Among them, in 34% it is due to birth effects, in 21% it was caused by a disease and in 20% it was caused by an accident. The only cause of disability that stands out from the rest is regular intellectual disability, affecting 12% of the population Core Indicators To generate GIS maps we need a set of binomial indicators. From the dataset many can be constructed. In what follows we show the results for Tandil and Urundel for the core indicators Next table shows the main results for both Pilots. Table 6. Core indicators 21

23 INDICATOR Tandil Salta Proportion of female headed households Female headed households 26,96% 26,78% Households heads Proportion of young mothers households 2 # Young mothers 3 45,97% 63,57% # mothers Proportion of current young mothers households # Current young mothers 4 5,26% 11,94% # Young mothers Proportion of child deaths (0-5 years old) 5 # child deaths 0,3% 0,85% # children 6 Proportion of women who died due to pregnancy relatedcauses # deaths due to pregnancy related-causes 0,0% 0,19% # mothers Proportion of households whose members experienced hunger and not have anything to eat in the last 30 days at least once # Households whose members experienced hunger and not have anything to eat in the last 30 days at least once 3,19% 1,34% # Households 2 Women that are or were young when they gave birth. 3 It is considered young the mother that gave birth before being 21 years old. 4 Mothers that are nowadays young 5 Considers the child deaths ocurred in the last five years 6 Children today aged between 0 and 10 years 22

24 INDICATOR Tandil Salta Proportion of households whose members experienced hunger and not have anything to eat in the last 30 days more than 5 times # households whose members experienced hunger and not have anything to eat in the last 30 days more than 5 times 0,88% 1,04% # Households Proportion of households whose members experienced hunger and not have anything to eat in the last 30 days more than 15 times # households whose members experienced hunger and not have anything to eat in the last 30 days more than 15 times 0,00% 0,00% # Households Proportion of households without access to safe water supply # households supplied with unsafe water sources 7 0% 0,3% # households Proportion of households without access to sanitary toilet facilities # households only with precarious wastewater drain 8 18,00% 15,64% # households Proportion of households living in makeshift housing # households in makeshift housing 9 5,71% 3,61% # households 12 Proportion of households who are sqatters # households who are squatters 1,25% 0,45% # households 7 Households that extract water from a stream, rain water or a shallow well 8 Septic tank, in the open, indoor, or without bathroom 9 Households in which the walls were built with light and/or improvised materials or households in which the roof was built with improvised materials 23

25 INDICATOR Tandil Salta Proportion of children 6-12 years old not in elementary school # children 6-12 years old not in elementary school 1,15% 1,80% # children 6-12 years old Proportion of children years old not in elementary school # children 6-12 years old not in elementary school 7,91% 9,80% # children 6-12 years old Proportion of children who are repeaters # repeater children 6-18 years old attending school 22,69% 37,90% # children 6-18 years old attending school Proportion of youth aged years who are not literate (cannot read and write) # youth aged years who are not literate (cannot read and write) 0,31% 1,16% # youth aged years Proportion of people in the labor force who are unemployed # Unemployed people 12,93% 26,26% # Economically active population 10 Proportion of Households that receive Social Assistance # households that receive social assistance 11 39,58% 52,61% # households While in some aspects the situation between Tandil and Urundel is alike or does not present significant differences (proportion of female headed households (1) or the sanitary conditions (9 and 10), for instance), in others, substantial differences arise. The economic situation is much worse in Urundel, with more than half of the population receiving social assistance and the unemployment reaching a quarter of the economically active population. In Tandil, on the other hand, while less people receive social assistance (40%, still a significant amount), the unemployment is half than Urundel s (17 and 18). 10 People searching for a job or willing to start a job if offered 11 Do not include the case of resources provided by relatives 24

26 If we look at the educational level, we can also see a better situation in Tandil. While in Urundel 5% of the children do not go to elementary school, in Tandil the percentage reaches 3.5% (Indicators 13 and 14 combined). Analphabetism reaches 0.31%, in Tandil and 1.16% in Urundel, and the proportion of repeaters is 22.69% and 37.9% respectively (16 and 15). Surprinsignly, the housing situation is a bit better in Urundel (11 and 12). However, if we consider as makeshift housing only the houses built with improvised materials, rather than considering both improvised and light materials, the housing situation looks worse in Urundel. Regarding the sanitary conditions, the big majority of inhabitants have access to the water mains (9), but a considerable amount of households have precarious wastewater drains (10). 5. Conclusions We consider the argentine Pilot CBMS as a full success. The strategy to implement it in two very different places significantly enriched the experience. In addition, as in each place the experiment was not exactly the same, we were able to extract some conclusions regarding to which alternative might be better to implement in the future. We learned that working very close with local authorities is good, but it will delay the process. Election years or implementing the census close to an election might complicate the field work substantially. We learned that there are many local institutions willing to help, and they have a positive effect on the quality of the job. In the case of Tandil, for instance, where we have the largest set of enumerators and coming from different sources, we identify as enhancing factors: i) persons with attachment to some local institution (who might have some reputation to lose, such as those who come from the local church), and ii) supervisors with previous experience (supervising activity is much more difficult than the enumerator activity, and enumerators with bad supervisors performed homogeneously worse). The option of implementing the census at fast pace, as it was the case of Urundel, it seems to have worked better than the option to have a more flexible schedule as in Tandil, where the effect of advertisement seems to have diluted fast. Although, it seems to be also true that the type of problems that arise in a marginal urban area seems more difficult to solve; fear of violence or to lose a social plan might affect negatively the willingness of the individuals to participate in the Census. In a very small community as Urundel this is not the case because everybody knows each other. The involvement of the University which is a very respected institution in Tandil was a positive factor that help us to overcome the before mentioned fears. A toll free number where individuals can call to evacuate their doubts is another good idea to implement in the future. 25

27 References Reyes, Celia and Evan Due (2009), Fighting Poverty with Facts: community-based monitoring systems, International Development Research Centre (Canada) Auguste Sebastian (2007) Pobreza Urbana (Urban Poverty)" in El Desorden Urbano. Los Problemas Locales de la Calidad de Vida y el Crecimiento, FIEL,