Guatemala Water Supply, Sanitation and Hygiene (WASH) Poverty Diagnostic Annexes 1 Annexes: Poverty Diagnostic for Water Supply, Sanitation, and Hygiene (WASH) This document summarises the analysis performed on the socio-economic profiles, poverty trends, consumption and human opportunities of Guatemalan households. Data come from the LHMS/ENCOVI survey, years 2000-2014, and are weighted to make the analysis representative at the national level. Summary Poverty and inequality A staggering 59 percent of the Guatemalan population currently lives in general poverty, and almost a quarter of the population lives in extreme poverty. These numbers are particularly worrisome as they represent a marked increase with respect to their value in 2000. Both general and extreme poverty are much more prevalent in rural areas, but urban areas saw a much more marked increase in poverty over time: extreme urban poverty in particular is now four times as high as it was in 2000. The distribution of general and extreme poverty across districts is quite similar, with particularly high concentrations in Alta Verapaz, Solola, Totonicapan, Quiche, Huehuetenango and Chiquimula, and lower levels in Guatemala, Sacatepequez and Escuintla. The Gini index for the country is 0.39: inequality is higher in urban (0.38) than in rural areas (0.32), and is highest in Zacapa and Quetzaltenango districts (0.42) and lowest in Escuintla (0.28). General and extreme poverty head-counts by area - percentages 2000 2014 Urban Rural National Urban Rural National General poverty 27.13 74.49 56.20 42.15 76.10 59.28 Extreme poverty 2.811 23.80 15.70 11.23 35.26 23.36 Source: own calculations using ENCOVI survey data, years 2000, 2014 2.2 Improved access to water 91 percent of the Guatemalan population have access to improved water. Overall, between 2000 and 2014, access to improved water increased by around 2 percentage points. The changes over time are mainly due to an increase by almost 8 percentage points in water piped directly to Access to water by year - percentages premises, and a reduction in access to other, still 2000 2014 improved, yet lower quality sources. As a result, Piped to premises 68.65 76.27 currently more than three quarters of the population Piped to neighobour/public 3.99 1.53 Tubewell/borehole 14.56 11.41 have access to piped water. At the same time, the Rainwater 0.57 2.21 usage of surface water dropped by more than a third Truck or barrel 0.44 0.47 with respect to its 2000 level. Improved access to Surface water 7.28 5.05 water is more prevalent in urban (95 percent) than in Other unimproved 4.52 3.07 urban areas (87 percent): the difference is even Source: own calculations using ENCOVI survey data, years 2000, 2014 sharper when it comes to access to piped water (89 vs. 64 percent) and especially use of surface water (1.5 vs. 9.3 percent). Among Guatemalan districts, Solola, Escuintla Suchitepequez and Sacatepequez are the ones with highest access to improved water (96-98 percent), Alta Verapaz, Chiquimula, Santa Rosa and Peten the lowest (80-87 percent). Alta Verapaz, which is by far the location with the lowest access (80.7 percent), is however the district that experienced the highest increase in access to improved water over time, moving by 18 percentage points from 63 percent in 2000. 2 2.3 Improved sanitation Only 58 percent of the Guatemalan population have Sanitation access by year - percentages access to improved sanitation. This however 2000 2014 Flush to sewer 30.31 42.94 represents a sizeable increase from very low levels in Flush to septic tank 3.18 7.76 2000, when more than half of the population lacked Pour flush/composting toilet/other improved 10.76 7.61 access to improved sanitation. At the same time, while Open defecation 13.18 5.02 more than 5 percent still perform open defecation, the Ambiguous 42.57 36.67 percentage has dropped by more than 60 percent Source: own calculations using ENCOVI survey data, years 2000, 2014 since 2000. Almost 43 percent of the population currently have a flush toilet. Access to sanitation is significantly better in urban than in rural areas: 83 percent of the urban population have access to improved sanitation, 72 percent have a flush toilet and 1.7 percent does open defecation; the corresponding figures for rural areas are 29, 24 and 9 percent, respectively. The percentage of population with access to improved sanitation varies widely by district: if on the one hand some locations like Sacatepequez and Guatemala have shares as high as 90 percent, others like Totonicapan and Peten barely reach 30 percent. In Alta Verapaz, roughly one in five individuals has access to improved sanitation. Socio-economic profiles While 96 and 79 percent of the non-poor have access to improved water and sanitation, respectively, the figures are 88 and 45 percent for the poor, and drop to 83 and 19 for the extremely poor. The most pronounced differences in improved access to water along socio-economic lines are due to age, education, employment sector, ethnicity and migrant status of the household head. When the household head is young (13-24 years old), improved access to water is 5-7 percentage points lower than for other age groups. The same applies when the household head has not completed primary education, and when they are employed in the primary sector. When the household head is an indigenous language speaker or when they have migrated to a new municipality in the last 5 years, the household’s access to improved water is 4 percentage points lower than when they do not speak an indigenous language or have not changed municipality. With the exception of age, the same socio-economic categories account for the main differences in access to improved sanitation as well, although in this case the gaps are much more pronounced. Households whose head migrated or did not complete primary education have respectively 20 and 30 percentage points lower access to improved sanitation than non-migrants or those with at least primary education, whereas primary sector workers and indigenous language speakers are less than half as likely to access improved sanitation as their respective counterparts. Education, employment sector and ethnicity also appear to be strong predictors of general and extreme poverty, whereas poverty headcounts are higher for non-migrants than for migrants. Perhaps counterintuitively, households headed by women appear to have a better access to both improved water and sanitation, as well as a lower likelihood of being poor. As regards the bottom 40 percent of the income distribution, it is predominantly composed of households living in rural areas, whose head is a male, aged 25-44, has not completed primary education, is a wage earner and is employed in the primary sector. Consumption model Simple regression analysis of household consumption on WASH variables and a set of covariates shows a positive, statistically significant and robust correlation between consumption and access to improved water and sanitation. When the household has access to improved water, their consumption increases by 19 percent; when they have access to improved sanitation, by 33-37 percent (depending on the actual specification used). Households that have access to both improved water and sanitation have 33 percent higher consumption. Other variables that are significantly positively correlated with consumption are gender (female) of the household head, living in an urban area, education, and being employed, non-wage earner. Age of the household head does not seem to matter much in determining household consumption, whereas households with higher dependency ratios (the ratio of the number of household members younger than 15 and older than 65 on the number of household members aged 15-65) are correlated with 20 percent lower consumption. 3 Human Opportunity Index (HOI) Children 0-16 The HOI for access to improved water in Guatemala was 85 in Human Opportunity Index: Children 0-16 years 2000 and increased to 89 in 2014. At the same time, the HOI for Opportunity HOI Coverage Dissimilarity access to piped water improved from 66 in 2000 to 72 in 2014. Improved water 2000 85.26 85.50 0.28 The HOI for access to improved sanitation experienced a more 2014 89.53 89.69 0.17 marked increase, almost doubling from 38 in 2000 to 75 in 2014. Piped water The HOI for access to any sanitation services and not using open 2000 66.17 67.92 2.58 defecation also improved substantially in the time span 2014 72.14 73.32 1.61 considered, from 84 to 93. In all cases, the increase in the HOI is Improved sanitation driven by both an increase in coverage and a decrease in 2000 17.62 34.12 48.36 inequality as measured by the dissimilarity index, although the 2014 35.1 46.88 25.12 reduction in inequality is much more substantial for access to Non-open defecation 2000 83.82 84.50 0.80 improved sanitation. 2014 93.42 93.54 0.14 Changes in the HOI for water were due almost entirely to a scale Own calculations using ENCOVI survey data, years 2000, 2014. effect, with only a moderate improvement in inequality (increase in equalisation). Composition effects, due to shifts in the group size circumstances, were also positive, and particularly significant for access to piped water, where they accounted for more than a fifth of the overall change over time. Patterns are very similar for sanitation, although again, the equalisation effect is twice as strong as the scale effect for access to improved sanitation. The HOI for water is highest in Solola district (98) and lowest in Chiquimula (75). Spatial differences are incredibly more pronounced for sanitation, where the HOI ranges from 86 in Sacatepequez to just 4 in Alta Verapaz. Inequality of opportunity for water and sanitation services is mainly attributable to wealth distribution and urban or rural location inequality, which account for around 75 percent of the dissimilarity index for both access to improved water and sanitation. The household head characteristic that matters the most in explaining access to water and sanitation services is education. Human Opportunity Index: Children 0-59 months Opportunity HOI Coverage Dissimilarity Children 0-59 months Improved water 2000 83.00 83.29 0.35 The HOI trends for younger children are very similar to those 2014 88.51 88.7 0.22 relating to the broader sample of children. The HOI for access to Piped water improved water changed from 83 in 2000 to 89 in 2014, the HOI 2000 63.68 65.64 3.00 for access to piped water from 64 in 2000 to 72 in 2014. The HOI 2014 71.68 72.89 1.66 Improved sanitation for access to improved sanitation more than doubled, but 2000 15.93 31.56 49.53 remains very low at 36 in 2014. The HOI for not using open 2014 35.59 47.05 24.37 defecation improved from 82 to 93. Again, in all cases both Non-open defecation coverage and inequality improve, and the reduction in 2000 82.00 82.77 0.93 inequality is particularly pronounced for access to improved 2014 92.62 92.75 0.14 sanitation. Own calculations using ENCOVI survey data, years 2000, 2014. Changes in the HOI for water were mainly due to a scale effect, although both equalisation and composition effects were now more relevant than for the broader sample of children. The same applies for sanitation, with again the exception of access to improved sanitation, where the equalisation effect accounts for 60 percent of the change. The HOI rankings across districts are also very similar to the broader sample case: the HOI for water is highest in Solola (97) and lowest in Alta Verapaz (71), for sanitation it is highest in Sacatepequez (86) and lowest in Alta Verapaz (7). Inequality of opportunity for water and sanitation services is overwhelmingly driven by inequality in the wealth distribution and urban or rural location, which account for around 80 percent of the dissimilarity index. The most important characteristic of the household head is again education. 4 Measurement and Data Issues in Guatemala Actual Population Size The population figures used in this report and in official statistics are projections, not actuals. The last Population and Housing Census dates from 2002 so some caution is needed when looking at per capita figures. It is expected that birth and death rates will have changed in the 14 years since the last census: certainly there is evidence of changes in migration patterns. Obviously, any errors in the population projections could affect the analysis of per capita GDP, and, if there are variations in the growth of different population groups, this could affect poverty measures as well, although there is no data to indicate that this particular problem exists. Poverty Measurement The data for the poverty measures come from the National Living Standards Measurement Study surveys (Encuesta Nacional de Condiciones de Vida or ENCOVI) for 2000, 2006, and 2014. There are also data for a 2011 round of the survey but there are some methodological concerns about these data in terms of comparability. The government is currently reviewing this issue and since this review may lead to revisions of the 2011 indicators, we have chosen to use only data from the 2000, 2006, and 2014 survey rounds in the present report. Guatemala uses a consumption-based monetary measure of poverty. This per capita measure covers: the consumption of purchased and non-purchased food (own production, gifts, and donations); (ii) transport and communications; (iii) consumer goods; (iv) household services and legal costs; (v) utilities, education, and health; and (vi) the annual use value of housing and durable goods. The poverty lines used reflect the costs of a minimum level of calories (extreme poverty line) and the cost of obtaining both food and other basic necessities (overall poverty line). We use this official measure of welfare throughout the report. However, the exception is when the focus of the analysis is on benchmarking Guatemala’s situation in comparison with other countries (as is in done in this chapter). In such cases, we use an income-based measure of poverty and an international overall poverty line of US$4 and an international extreme poverty line of US$2.5 per person per day in 2005 PPP terms. While there has been an update to the 2011 PPP for the US$1.25 line, there are no updates for the US$2.5 and US$4 lines. For this reason, the report continues to use the 2005 PPP and the global extreme poverty line of US$1.25 in 2005 PPP. Urban and Rural Definitions A new urban/rural classification was introduced in the 2002 Census by Guatemala’s National Statistics Institute (Instituto Nacional de Estadística, INE). The share of the urban population increased from 38.6 percent in 2000 to 48.1 percent in 2006. However, only 2.9 percentage points of this increase were due to migration and population growth; the remaining 6.6 percentage point increase was due to changes in the urban/rural classification (World Bank, 2009). This dramatic increase in the share of urban areas in the sample explains why extreme poverty decreased between 2000 and 2006, even though both urban and rural extreme poverty increased. The more recent evolution of poverty between 2006 and 2014 has not been affected by changes in the urban/rural classification. When the rural-urban definition is important for our analysis, the discussion will be restricted to changes between 2006 and 2014 only. 5 Poverty and Inequality Inequality in Guatemala is high, but has fallen much faster than in either Central America or Latin America as a whole. As measured by the Gini coefficient, inequality of income in Guatemala was 49 percent in 2014, higher than in most LAC and Central American countries, all aspirational peers, and all structural peers except for Honduras (Figures below). Guatemala’s inequality ranked in the 91st percentile worldwide (WDI, 2015). However, while there was no appreciable change between 2000 and 2006, the next eight years saw a very substantial decline in the Gini, which fell from 0.55 to 0.49 percent. This sharp fall in inequality occurred in both urban and rural areas, with urban inequality falling from 0.55 to 0.48 and rural inequality falling from 0.50 to 0.46. 1 As already indicated, these reductions in income inequality have mainly been due to a fall in the incomes of the rich rather than to a rise in the incomes of the poor. 2 And despite this convergence in the income distribution, the total share of income held by those in the bottom deciles remains extremely low. The richest one percent of the income distribution controls 13 percent of total income. The Rate of decrease in Income inequality Inequality continues to be higher in was higher in Guatemala than in Central Guatemala than in its peer Countries America or in LAC Poor and non-poor households have significantly different characteristics in terms of their household structure and employment status (Table below). The poor are more likely than the non-poor to live in male-headed households; in fact, overall and extreme poverty rates among female-headed households are lower than the 1 As might be expected, a consumption-based Gini shows much lower levels of inequality but the trend in the consumption-based Gini is similar to that seen with income. The decline in inequality between 2006 and 2014 is even more dramatic in consumption terms. The decline in both urban and rural inequality can be seen in both income and consumption terms. 2 One hypothesis about the change in the Gini that cannot be ignored is that there may have been a deterioration in the quality of data at the top end of the distribution over time. There is always some degree of under-reporting: studies of top earner databases from tax records in other countries have shown that the Gini is often under-estimated as a result of non-responses from households at the top end of the distribution. The concern is not so much that this phenomenon exists but that these non-responses may be increasing. Given the very sharp rise in crime and violence in Guatemala, particularly in the cities, it is quite plausible that households are not willing to provide information on their wealth to surveys. In Mexico the non-response rate in some areas of the country has increased to 25 percent in recent years and non- response is correlated to education levels, suggesting that it is correlated to income as well. No assessment has yet been carried out of the increase in non-responses in the latest round of the ENCOVI in Guatemala or of its implications, so until that happens, a degree of skepticism about the magnitude of the fall in the Gini might be warranted. 6 corresponding national figures and, in terms of overall poverty, the rates have fallen. 3 Having a younger head of household is also correlated with greater poverty, with all other characteristics held constant. The poor and extreme poor tend to belong to larger households than the non-poor. Households in extreme poverty have 6.8 members on average while non-poor households have an average of three fewer members. Having a head of household who is indigenous but not one of the four large Mayan groups is linked to a significantly greater risk of poverty, as is living in the North and having less education. Characteristics of poor and non-poor households (2014) Extreme Moderate Bottom 40 Top 60 Poor Poor Non-poor percent percent Household Characteristics Age of head 45.5 45.1 47.1 45.2 47.1 Female-headed household (percent) 12.9 19.5 25.5 17.3 25.7 Single earner without children 2.3 4.4 14.8 3.7 14.9 Single earner with children 44 45 33.8 44.7 33.6 Two or more adult earners with children 52.8 48.5 45.8 49.9 45.8 Head indigenous 65.5 43.1 21.2 50.5 20.9 Rural 75.7 56.5 29.3 62.6 29.2 Proportion age 0-14 44.9 36.8 22.7 39.4 22.6 Proportion age 15-64 51.3 56.9 67.9 55.2 68 Proportion age 65+ 4.9 6.4 9.8 5.9 9.9 Education of household head (years) 2 3.3 6.7 2.9 6.8 Dependency ratio on employed persons 35.7 39.5 50.1 38.3 50.2 Monthly per capita income 468 739 2102 650 2121 Monthly per capita consumption 363 660 1755 565 1768 Household size 6.8 5.3 3.8 5.8 3.8 Total households 544,458 1,084,735 1,722,837 1,653,516 1,698,514 Labor force (percent) Salaried 57.7 60.7 61.7 59.6 61.7 Self-employed 24.1 24.3 23.6 24.3 23.6 Employer 0.4 1 4.6 0.8 4.7 Unpaid worker 17 12.2 7.6 13.9 7.5 Unemployed 0.9 1.7 2.5 1.4 2.6 Informality (percent) Formal (with social security) 3.1 11.1 29.6 8.3 29.8 Informal (without social security) 96.9 88.9 70.4 91.7 70.2 Employment sector (percent) Agriculture, hunting 61.5 41.3 14 48.5 13.6 Manufacturing 9.8 12.2 13.9 11.4 13.9 Construction 5.2 7.9 5.3 7 5.2 Wholesale, retail trade 12.8 20.6 32.2 17.7 32.5 Source: Calculations based on the 2014 ENCOVI. 3 World Bank (2009). 7 Areas with a greater concentration of indigenous people tend to be poorer, but many of the ethnic groups are spread across departments with varying degrees of poverty: this is particularly true for the non-indigenous and the Mam (Figure 1.10). The Mam had the highest poverty rate in 2000 at 90 percent, and live in both departments with high poverty levels and in others with much lower levels. By 2014, the Q´eqchi had overtaken the Mam, having the highest poverty and extreme poverty rates (87 and 56 percent respectively) and representing 22 percent of the extreme poor. In contrast the Kaqchiquel have the lowest poverty rate among the indigenous groups despite this rate having increased in both 2006 and 2014 (Figure 1.11). As noted earlier, the Mam have the highest chronic poverty rate, and none of the indigenous groups have fewer chronically poor households than the non-indigenous. Ethnic Dispersion and Poverty Are Linked The Indigenous Are Much More Likely to Be Poor than the Non-indigenous Poverty is highest in rural areas, but there has been a striking increase in urban poverty. In 2006, rural poverty was more than two times the level of urban poverty overall (71 percent compared to 30 percent) and for extreme poverty, the ratio was 4.8. 4 Both rural and urban poverty rose by 2014, and rural poverty remains much higher than in urban areas. The depth and severity of poverty levels are also worse in rural areas, although there has been some convergence in the second half of the decade as these indicators declined in rural areas while remaining unchanged in urban areas. 5 However, the increase in urban poverty between 2006 and 2014 was much greater (41 percent) than in rural areas (8 percent). This does not appear to have been due to rural-urban migration: a Huppi-Ravaillion decomposition shows that the poverty increases within each area are responsible for the changes. Movements between rural and urban areas are actually poverty-reducing, albeit on a very small scale. 6 The rise in urban poverty has narrowed the urban-rural gap and shifted the concentration of the poor quite dramatically. In 2000, 7 percent of all people in extreme poverty lived in urban areas, while by 2014 this had increased to 24 percent (Figure 1.8). While some of the change between 2000 and 4 A new urban/rural classification was introduced in the 2002 Census by Guatemala’s National Statistics Institute (Instituto Nacional de Estadística, INE). The share of the urban population increased from 38.6 percent in 2000 to 48.1 percent in 2006. However, only 2.9 percentage points of this increase were due to migration and population growth; the remaining 6.6 percentage point increase was due to changes in the urban/rural classification (World Bank, 2009). When the rural-urban definition is important for our analysis, the discussion will be restricted to changes between 2006 and 2014 only. 5 Changes in the definition of rural and urban areas between 2000 and 2006 make earlier comparisons suspect: see Annex. 6 Huppi-Ravallion decompositions look at intra-area and across areas changes in poverty. The within-area increase in poverty was 8 percentage points, but the across-area movements mitigated this effect, resulting in a 7.6 percentage point rise in overall poverty. For extreme poverty, the within-area increase in poverty was 4.1 percentage points, while movements from rural to urban areas led to a small decline in poverty for an overall rate of 3.7. 8 2006 can be attributed to a post-census reclassification of rural areas, this explains only a portion of the observed change and does not affect differences between 2006 and 2014. The larger change has been in terms of overall poverty. In eight years, the share of the poor living in urban areas has gone up from 28 percent to 57 percent. The regional change in the concentration of the poor is also quite striking. Country Comparators To benchmark Guatemala’s performance, this report uses six comparable groups of peers: Central American countries, Latin American peers, lower-middle-income countries, world average, structural peers, and aspirational peers. The structural peers were selected using the “Find your friends” tool. The group of structural peers includes countries that provide appropriate benchmarks for answering SCD-relevant questions such as whether or not certain conditions, policies, or economic performances in Guatemala are adequate. The criteria and filters for selection were the following: (i) lower-middle-income countries; (ii) population between 5 million and 25 million people; (iii) Agriculture, value added (% of GDP) less than 20 percent; and (iv) small island states were excluded. The use of these criteria resulted in the following set of countries: Bolivia, El Salvador, Honduras, Nicaragua, Paraguay, and Senegal. Under aspirational peers, we want to aggregate countries that may be used as good examples of development for Guatemala and that Guatemala may emulate. Thus, we set the following criteria for the period 2001 - 2013: (i) lower middle income and upper middle-income countries; (ii) GDP per capita growth higher than 3 percent; (iii) inflation below 5 percent; (iv) maternal mortality ratio (per 100,000 live births) less than 100; and (v) population below 35 million. Countries that are “natural intensive”, landlocked, or islands were excluded. This classification delivers the following group of countries: Structural Peers for Guatemala Nominal GDP per capita agriculture, value added ( Country Population (millions), 2014 (US $), 2013 of GDP), 2001-13 Guatemala 3,512 15.9 13.0 Bolivia 2,700 11.2 14.0 El Salvador 3,875 6.4 11.1 Honduras 2,323 8.3 13.4 Nicaragua 1,840 6.2 18.1 Paraguay 4,170 6.9 19.3 Senegal 1,073 14.5 16.3 Source: Find my Friends Tool 2014. Aspirational Peers for Guatemala Nominal GDP Population GDP per capita Maternal Inflation ( %), Country per capita (US (millions), growth (%), mortality ratio, 2001-13 $), 2013 2014 2001-13 latest data point Guatemala 3,512 15.9 0.9 6.3 140 Albania 4,610 2.8 5.4 2.9 21 Chile 15,776 17.7 3.3 3.1 22 Jordan 5,174 6.7 3.1 4.3 50 Latvia 15,205 2.0 5.5 4.7 13 Lithuania 16,003 3.0 5.9 3.0 11 Panama 10,839 3.8 5.3 3.2 85 Peru 6,674 31.4 4.1 2.6 89 Source: Find my Friends Tool 2014. 9 Rates of Stunting and GDP per country, circa 2013. Source: Estimates based on World Bank Data, 2015. Source: UN, 2015; World Bank, WDI. 2015. 10 Snapshot of Water and Sanitation Infrastructure and Nutrition in Indigenous Areas of Guatemala Data and Sampling Description This snapshot uses data from a survey collected to evaluate an early stimulation intervention from October 2016 to January 2017 in four departments: Huehuetenango, San Marcos, Quiche and Totonicapán. From a list of more than 100 communities, each of them was randomly selected to a treatment arm and within each community, 20 eligible families were randomly selected to our sample. Our sample consists of about 2,000 families with children under 24 months of age. The dataset contains information on basic demographic characteristics, WASH infrastructure and child health and nutritional outcomes. General Descriptive Statistics The majority of indigenous families in this sample obtained their water from a pipe located outside of the household (65%) and most purified the water by boiling it (89%). The most common sanitation service reported was latrine (65%), followed by toilet (21%). Fourteen percent of the families did not have any type of sanitation service. A few differences in water access and sanitation services were found across departments. Specifically, families from Huehuetenango had the easiest access to water as 82% reported having water piped inside or outside their household, while families in Quiche were less likely to have sanitation services (19% reported none). In terms of household and family characteristics, the majority of the houses had dirt floors (65%) and consisted of 2 rooms. In general, families were composed of 5 members of which 3 were children and the head of the household was mostly men (88%). The average number of years of education for the head of the household was 4. Across departments, San Marcos had the highest number of years of education with an average of 6. Regarding family income, the average monthly income per household was 999 GTQ. Using family income and number of family members, a Purchase Power Parity (PPP) index was calculated to identify households in which each member had more than $1.25 USD to spend daily—the majority of the families fell at or below the line (77%) and are thus considered to be poor based on this index. Of all the departments, Huehuetenango had the highest percentage of families at or under the PPP poverty line (84%). The number of children who had had diarrhea in the last two weeks was high (39%) and the percentage was higher in Huehuetenango (50%). On average, the closest health center was 37 minutes away from home. Approximately 46% of the respondents preferred Spanish only, while 54% preferred an indigenous language or a combination of Spanish and an indigenous language. Huehuetenango had the highest percentage of respondents who spoke Mam or a combination of Mam and Spanish (70%), while Quiche had the highest percentage of respondents who spoke Quiche or a combination of Quiche and Spanish (62%). The majority of respondents in Totonicapán (62%) and San Marcos (73%) preferred Spanish. We conducted additional comparisons across the three language groups, which are not included in the descriptive statistics table (see appendix). Respondents who preferred an indigenous language or a combination of an indigenous language and Spanish had a higher percentage of houses with dirt floors (76%). Additionally, respondents who preferred Mam or a combination of Mam and Spanish had a higher percentage of child diarrhea (51%). Water and Sanitation Infrastructure To consider the potential synergies between sanitation infrastructure and water source and our indicators, we compared the percentage of some indicators per type of sanitation and water source. The charts below depict the percentage of houses with dirt floors, PPP-based poverty, children with extreme stunting, and underweight children by sanitation infrastructure and water source. Compared to families who had a toilet, those without 11 sanitation infrastructure had a higher percentage of homes with dirt floor (86%), were poor (88%), had children with extreme stunting (40%), and underweight (21%). When considering our outcomes by water source, compared to piped water, families who obtained water from other sources had a higher percentage of homes with dirt floors (81%), were poor (82%), had children with extreme stunting (39%), and underweight (20%). Selected Indicators, by Sanitation Selected Indicators, by Water Source 100% 88% 100% 86% 82% 81% 80% 80% 70% 62% 60% 60% 40% 44% 35% 39% 40% 40% 23% 21% 25% 20% 20% 12% 20% 14% 0% 0% Dirt floor PPP Extreme stunting Underweight Dirt floor PPP Extreme stunting Underweight Toilet None Water piped inside Other water source (river, lake, rain) Anthropometric Measures Following the World Health Organization (WHO) guidelines, stunting or growth retardation is defined as -2 S.D. of the WHO Child Growth standard median for height-for-age and it reflects chronic malnutrition. We constructed an additional measure, extreme stunting, defined as -3 S.D. of the WHO Child Growth standard median for height-for-age and it measures extreme chronic malnutrition because it is the result of long-term effects of nutritional deprivation (cumulative effects of undernutrition) or recurrent infections. Underweight children are those with a weigh-for-age below -2 S.D. of the WHO Child Growth standard median. The following charts show statistics for our anthropometric indicators. Sixty eight percent of the children in our overall sample were stunted, 32% are extremely stunted and 15% are underweight. These results are similar for the departments of San Marcos, Huehuetenango and Quiche, but much higher for Totonicapán, as it can be observed in the bar chart. We also observed a disturbing trend for these indicators: as age increases, the percentage of stunted, extremely stunted and underweight children increases. We further examined Pearson correlations of age in months with each anthropometric indicator to confirm these findings. We observed that all correlations were statistically significant at the 1-percent level. While these findings are troublesome, they also indicate that policy makers could intervene as early as possible to prevent cumulative negative effects of nutrition deprivation that might impact children mortality, delayed mental development, poor cognitive capacity and school performance. Anthropometric Indicators Anthropometric Measures, by Age Groups 80% 73% 90% 68% 76% 60% 59% 60% 40% 32% 35% 42% 40% 19% 24% 20% 15% 30% 19% 12% 9% 11% 0% n=1,996 n=376 0% Stunting Extreme stunting Underweight Overall Totonicapán Stunting Extreme Stunting Underweight 0-5 months n=85 6-11 months n=817 12-23 months n=1,084 12 Are these associations significant? Results from Pearson Correlations Our descriptive statistics show differences in anthropometric measures and basic demographic characteristics for those with different type of sanitation in their households. In general, we observed worst indicators for households with no sanitation. Our correlation analysis showed a positive and statistically significant correlation between no sanitation and extreme stunting and underweight. At the same time, we found better indicators for those with toilet. Our correlation analysis confirmed these findings and showed a negative and significant correlations for toilet with stunting, extreme stunting and underweight. A similar pattern was found for those with water infrastructure in the home (negative correlation with stunting and extreme stunting) and those without it (e.g. water is obtained from river, creek, rain, lake; positive correlation with extreme stunting and underweight). Pearson correlations Water piped Water piped No Other water Latrine Toilet outside the inside the sanitation source residence residence Stunting 0.01 0.07 *** -0.09 *** 0.03 0.02 -0.06 ** Extreme Stunting 0.08 *** 0.03 -0.10 *** 0.05 * -0.01 -0.06 ** Underweight 0.07 ** -0.01 -0.05 * 0.04 * -0.02 -0.02 Diarrhea -0.02 -0.01 0.03 0.00 0.00 0.02 Note: ***p ≤ .001**p ≤ .01; *p ≤ .05. 13 Descriptive Statistics by Department Huehuetenango Totonicapán San Marcos Quiché Overall (n=284) (n=378) (n=292) (n=1,078) (n=2,032) Water Source Piped inside residence 15.2% 19.4% 10.7% 12.1% 13.7% Piped outside residence 67.1% 57.8% 66.0% 67.2% 65.3% Communal tap/well 3.5% 13.8% 10.0% 11.2% 10.5% Other water source 14.1% 9.0% 13.4% 9.5% 10.6% Water Purification Method Boiling 93.3% 89.4% 88.3% 88.2% 89.2% Other (chlorine, filter, buy purified) 4.9% 8.5% 10.0% 9.6% 8.8% None 1.8% 2.1% 1.7% 2.2% 2.0% Sanitation Service Toilet 30.0% 13.5% 18.9% 22.0% 21.1% Latrine 62.9% 77.2% 71.5% 59.2% 64.8% None 7.1% 9.3% 9.6% 18.8% 14.1% House Flooring Material (concrete, tile, etc) 36.4% 37.7% 37.8% 33.0% 35.0% Dirt 63.6% 62.3% 62.2% 67.0% 65.0% Head of Household Gender Male 85.2% 90.2% 89.4% 86.9% 87.6% Female 14.8% 9.8% 10.6% 13.1% 12.4% Purchasing Power Parity Above daily $1.25 line 15.9% 20.3% 24.2% 25.4% 23.0% At or below daily $1.25 line 84.1% 79.7% 75.8% 74.6% 77.0% Child Diarrhea No 50.0% 61.9% 62.3% 63.3% 61.0% Yes 50.0% 38.1% 37.7% 36.7% 39.0% Mother Preferred Language Spanish Only 29.9% 61.6% 73.3% 38.2% 46.5% Quiche and Spanish 0.0% 38.4% 0.0% 61.8% 39.9% Mam and Spanish 70.1% 0.0% 26.7% 0.0% 13.6% Household Rooms 1.82 (.98) 1.86 (1.06) 1.63 (.99) 1.79 (1.07) 1.78 (1.05) Household Members 4.76 (1.51) 5.46 (2.21) 5.30 (1.86) 5.25 (2.15) 5.23 (2.05) Children in Household 2.15 (1.20) 2.90 (1.74) 2.48 (1.46) 2.79 (1.64) 2.68 (1.60) *HH Years of Education 5.06 (3.42) 3.96 (3.10) 6.46 (4.08) 3.73 (3.23) 4.32 (3.50) Monthly Income in GTQ 841 (911) 967 (620) 1,119 (1333) 1,020 (650) 999 (805) Health Center Distance (min) 34.42 (22.82) 37.84 (25.21) 39.24 (27.69) 36.91 (25.78) 37.05 (25.60) Notes: *HH=Head of Household. Percentages shown for categorical variables; means and standard deviations shown for continuous variables. 14 Sector-based Data, Graphs and Maps Education and Literacy Guatemala’s substantial increases in education levels have exceeded those of its peers. Between 1970 and 2010, Guatemala’s rate of increase in years of schooling outstripped most of its peers, with only Bolivia, El Salvador, and Brazil having improved faster. Guatemala’s level of education is well below its Some progress has been made on closing the peers but has been rising fast large gaps in schooling Primary school enrolment became almost universal in 2009 (99 percent) but shows signs of reversal and, by 2014, had fallen to 82 percent, which was even lower than it had been in 2000 (Figure 1.17). To the contrary, between 2000 and 2006, the schooling gap due to ethnicity in Guatemala went from around 3.5 years to less than a half year, a significant achievement 7. 7 Ferreira et al., (2013). 15 Education Levels in Guatemala Remain Low and Unequal Literacy rates improved nationally from 69 percent in 2000 to 81 percent in 2014, however, this varies geographically, by poverty level and ethnicity. In 2014, 87 percent of people living in urban areas were literate in contrast to only 75 percent in rural areas. Similarly, 90 percent of all non-poor people were literate whereas only 75 percent of poor people and 68 percent of those living in extreme poverty were literate. The contrast in education and literacy between ethnic groups is notable. While 87 percent of non-indigenous people were literate, this contrasts with only 72 percent of indigenous people. In 2014, the Q’eqchi had the lowest literacy rates in any language at just under 67 percent (Figure 1.18). However, they have the highest literacy rates in a Mayan language. 8 The Kakchikel have the highest literacy rate of any indigenous group at 78 percent, which remains below the national average. 8 The sample size of the survey data used in this analysis is adequate at best for looking at these groups. The other groups are too small to be able to be examined separately. There are some concerns about the adequacy of the sample even for the larger groups. 16 Primary Enrolment Became Almost Universal But Literacy Rates in Both Spanish and Mayan Shows Signs of Reversal Languages Vary by Groups Further concerns exist regarding the levels of educational attainment at the secondary level. The share of children aged 15 to 19 years old who have not completed the nine years of schooling mandated in the 1985 Constitution is high (figure 1.19) 9 and only Nicaragua, among Guatemala’s peer countries, has a lower secondary school completion rate. In Guatemala, the drop-off in enrollment is fairly steep: there is almost universal enrollment in the first grade of primary school, yet only half of the enrolled children finish their first year of secondary school. Gross enrollment in secondary school is 65 percent, the lowest in the Central American region, and the net rate of 47 percent is well below both the LAC and Central American averages. Critically, the educational gaps after the primary level begin to divide along ethnic, geographic, and gender lines. By the age of 18 the enrollment gap between indigenous and non- indigenous youths is close to 15 percentage points, and there is a 20 percentage point difference between the share of rural girls who complete 9th grade and the share of urban boys who do so (Figure 1.20). 9There is a discrepancy between the survey data (from the 2014 ENCOVI) and the administrative data concerning the rate of completion of nine years of schooling. The survey indicates that two out of three children do not complete this level, while the administrative data indicate that only 55 percent do not. It is not clear what is driving the discrepancy. However, neither figure suggests that Guatemala has been able to fulfill its Constitutional obligation regarding education. 17 Guatemala has low levels of educational Levels of completion are much lower in rural attainment compared to the rest of Central areas, especially for girls America Percentage of Heads of Households without eductation Alta Verapaz*** Huehuetenango*** Solola Izabal* San Marcos*** Jutiapa** Santa Rosa Peten** Quetzaltenango* El Progreso** Guatemala 0 10 20 30 40 50 60 2014 2000 Note: Significance of difference between 2000 and 2014 percentages. p<0.05, ** p<0.01, *** p<0.001. Source: ENCOVI, 2000 and 2014 18 Profiles of Access to and Quality of WASH Coverage rates of WASH services Unimproved Improved Improved Open Socioeconomic Characteristics Piped water water Water* Sanitation** Defecation Gender Female-headed households 17.0% 11.8% 43.7% 8.5% 71.2% Age Group 15-24 27.9% 19.7% 16.5% 19.1% 52.4% 25-44 12.2% 21.4% 20.0% 8.3% 66.4% 45-64 12.1% 20.3% 15.9% 6.8% 67.6% 65 and older 29.7% 17.7% 12.6% 10.1% 52.6% Labor status Wage earners 13.5% 20.2% 32.0% 8.0% 66.3% Non-waged/unemployed 38.0% 15.7% 22.5% 15.1% 46.3% Subsistance Agriculture 35.6% 18.1% 20.0% 15.8% 46.1% Dependency Ratio Less than 2 7.6% 20.0% 29.0% 8.0% 72.5% More than 2 19.2% 23.1% 19.5% 11.4% 57.4% Household characteristics Dirt floors 26.2% 18.2% 14.7% 13.4% 55.6% Witout electricity 25.1% 30.4% 9.9% 15.9% 44.5% Overcrowding 16.3% 22.8% 24.3% 9.7% 60.9% * Excluding piped water **Excluding open defecation Source: ENCOVI, 2014/2015 19 Aggregate statistics mask some of the inequities and trends in the sector. 20 Trends in access rates of different sources of water supply Differences between B40 and T60 in water access by type of water provision Sources: ENCOVI, 2000 and 2014 21 Percent distribution of mean days per month with water service, Guatemala Sources: ENCOVI, 2000 and 2014 22 23 Departmental Rates 2000-2014 Percent Households in Rural areas by Department (2000-2014) Guatemala Sacatepequez Quetzaltenango Solola Chimaltenango Escuintla Totonicapan Suchitepequez Zacapa Santa Rosa El Progreso Retalhuleu Izabal Jutiapa Jalapa Baja Verapaz Quiche Peten Huehuetenango San Marcos Chiquimula Alta Verapaz 0 20 40 60 80 100 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Percent Households in Urban areas by Department (2000-2014) Alta Verapaz Chiquimula San Marcos Huehuetenango Peten Quiche Baja Verapaz Jalapa Jutiapa Izabal Retalhuleu El Progreso Santa Rosa Zacapa Suchitepequez Totonicapan Escuintla Chimaltenango Solola Quetzaltenango Sacatepequez Guatemala 0 20 40 60 80 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 24 Percent Households in Extreme Poverty by Department (2000-2014) Guatemala Sacatepequez Escuintla Santa Rosa El Progreso Retalhuleu Quetzaltenango Suchitepequez Peten Zacapa San Marcos Jalapa Chimaltenango Jutiapa Baja Verapaz Huehuetenango Izabal Solola Totonicapan Chiquimula Quiche Alta Verapaz 0 20 40 60 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Percent of Female-Headed Households by Department (2000-2014) Alta Verapaz Solola Chimaltenango Quiche El Progreso Chiquimula Jalapa Huehuetenango Santa Rosa Peten Sacatepequez Totonicapan San Marcos Baja Verapaz Jutiapa Quetzaltenango Izabal Escuintla Suchitepequez Zacapa Retalhuleu Guatemala 0 5 10 15 20 25 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 25 Percent of Households Connected to Water Meter by Department (2000-2014) Alta Verapaz Quiche Huehuetenango San Marcos Baja Verapaz Santa Rosa Chiquimula Retalhuleu Totonicapan Solola Suchitepequez Jalapa Escuintla Izabal Peten Chimaltenango Jutiapa Zacapa El Progreso Quetzaltenango Guatemala Sacatepequez 0 20 40 60 80 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Percent of Households Connected to Public Provider by Department (2000-2014) San Marcos Huehuetenango Totonicapan Quiche Solola Baja Verapaz Jalapa Chiquimula Suchitepequez Jutiapa Alta Verapaz Izabal Quetzaltenango Chimaltenango Zacapa Santa Rosa Peten El Progreso Guatemala Escuintla Retalhuleu Sacatepequez 0 20 40 60 80 100 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 26 Percent of Households Connected to Private Provider by Department (2000-2014) Solola Baja Verapaz Totonicapan Quiche Jalapa Retalhuleu Zacapa Izabal Alta Verapaz Huehuetenango Chiquimula Chimaltenango El Progreso Jutiapa San Marcos Escuintla Santa Rosa Guatemala Suchitepequez Quetzaltenango Sacatepequez Peten 0 20 40 60 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Percent of Households with water service owned by water committee by Department (2000-2014) Sacatepequez Escuintla Guatemala Peten Retalhuleu El Progreso Santa Rosa Quetzaltenango Zacapa Chimaltenango Suchitepequez Jutiapa Izabal Alta Verapaz Chiquimula Jalapa Baja Verapaz Huehuetenango Solola San Marcos Quiche Totonicapan 0 20 40 60 80 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 27 Monthly Cost of Water (Q/M3) by Department (2000-2014) Solola Huehuetenango Baja Verapaz Quiche San Marcos Totonicapan Retalhuleu Alta Verapaz Izabal Suchitepequez Escuintla Chiquimula Jalapa Santa Rosa Zacapa Chimaltenango Jutiapa El Progreso Sacatepequez Quetzaltenango Peten Guatemala 0 10 20 30 40 50 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Number of days without water per month, by Department (2000-2014) Suchitepequez Solola Quiche Quetzaltenango Alta Verapaz Retalhuleu Totonicapan Huehuetenango Izabal Sacatepequez Peten San Marcos Escuintla Baja Verapaz Jalapa Chiquimula El Progreso Zacapa Guatemala Jutiapa Chimaltenango Santa Rosa 0 2 4 6 8 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 28 Hours per days without water, by Department (2000-2014) Chimaltenango Jutiapa El Progreso Guatemala Santa Rosa Zacapa Escuintla Jalapa Sacatepequez Peten Chiquimula Totonicapan Retalhuleu Solola San Marcos Quiche Alta Verapaz Quetzaltenango Huehuetenango Suchitepequez Baja Verapaz Izabal 0 5 10 15 20 25 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Distance to water supply (roundtrip in minutes) by Department (2000-2014) Retalhuleu Sacatepequez Solola El Progreso Guatemala Peten Izabal Alta Verapaz Suchitepequez Santa Rosa Baja Verapaz Quiche Chimaltenango Jalapa San Marcos Totonicapan Huehuetenango Escuintla Zacapa Jutiapa Quetzaltenango Chiquimula 0 20 40 60 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 29 Percent of Households that Boil Water by Department (2000-2014) Escuintla El Progreso Zacapa Jutiapa Retalhuleu Izabal Santa Rosa Guatemala Chiquimula Sacatepequez Peten Jalapa Suchitepequez Chimaltenango Baja Verapaz Quetzaltenango Quiche San Marcos Solola Alta Verapaz Huehuetenango Totonicapan 0 20 40 60 80 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Percent of Households that Filter Water by Department (2000-2014) Peten Huehuetenango Izabal El Progreso San Marcos Alta Verapaz Totonicapan Suchitepequez Santa Rosa Quetzaltenango Quiche Chiquimula Baja Verapaz Guatemala Retalhuleu Escuintla Jutiapa Jalapa Zacapa Chimaltenango Solola Sacatepequez 0 5 10 15 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 30 Percent of households that add chlorinate to water by Department (2000-2014) Huehuetenango Totonicapan Solola Quetzaltenango San Marcos Quiche Baja Verapaz Chimaltenango Izabal Alta Verapaz Sacatepequez Chiquimula Suchitepequez Guatemala Jalapa Peten El Progreso Zacapa Escuintla Santa Rosa Retalhuleu Jutiapa 0 10 20 30 40 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Percentage of Households Connected to drainage system by Department (2000-2014) Peten Alta Verapaz Solola Totonicapan Izabal San Marcos Retalhuleu Chiquimula Jalapa Quiche Baja Verapaz Huehuetenango Jutiapa Santa Rosa El Progreso Zacapa Suchitepequez Escuintla Chimaltenango Quetzaltenango Guatemala Sacatepequez 0 20 40 60 80 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 31 Percentage of Households practicing Open Defecation by Department (2000-2014) Guatemala Sacatepequez Chimaltenango Quetzaltenango San Marcos Solola Huehuetenango Retalhuleu Izabal Alta Verapaz El Progreso Peten Totonicapan Escuintla Baja Verapaz Quiche Santa Rosa Zacapa Suchitepequez Jalapa Jutiapa Chiquimula 0 20 40 60 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Percentage of Children with Reported Diarrhoea in the last month by Department (2000-2014) Izabal Totonicapan Quetzaltenango El Progreso Zacapa San Marcos Sacatepequez Jutiapa Jalapa Chimaltenango Guatemala Peten Solola Chiquimula Escuintla Quiche Baja Verapaz Santa Rosa Retalhuleu Suchitepequez Huehuetenango Alta Verapaz 0 20 40 60 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 32 Percentage of children receiving (Oral Rehydration Supplements) ORS by Department (2000-2014) El Progreso Izabal Quetzaltenango Solola Retalhuleu Suchitepequez San Marcos Chimaltenango Huehuetenango Chiquimula Alta Verapaz Sacatepequez Baja Verapaz Escuintla Jalapa Jutiapa Santa Rosa Peten Zacapa Quiche Totonicapan Guatemala 0 5 10 15 20 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. Percentage of children treating diarrhoea with medicines by Department (2000-2014) Santa Rosa Sacatepequez Izabal Chimaltenango Baja Verapaz Zacapa Jalapa Jutiapa Peten Quetzaltenango El Progreso Escuintla Quiche Chiquimula Guatemala Totonicapan Solola Huehuetenango San Marcos Alta Verapaz Retalhuleu Suchitepequez 0 20 40 60 80 100 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 33 Percentage of children without diarrhoea treatment by Department (2000-2014) Huehuetenango Solola Totonicapan Suchitepequez Chimaltenango San Marcos Jalapa Quiche Retalhuleu El Progreso Escuintla Santa Rosa Quetzaltenango Peten Baja Verapaz Guatemala Alta Verapaz Chiquimula Jutiapa Zacapa Sacatepequez Izabal 0 10 20 30 mean of Yr2000 mean of Yr2014 Source: ENCOVI, 2000 and 2014. 34 Maps and Spatial Distributions Source: ENCOVI, 2014 Own elaboration. Source: Own elaboration. Based on SEGEPLAN/Sinit data, 2014. 35 Access to continuous piped water supply by Department, 2006-2014 Source: ENCOVI, 2014 Own elaboration. Proportion of Households with intermittency of water (piped in dwelling or yard) Sources: ENCOVI, 2006 and 2014 36 Source: SEGEPLAN/Sinit, 2014. Data based on SINIT combines admin health records from SESAN, adjusted with CENSUS 2011. Health facilities and Index by MINSA. Social vulnerability combines the food insecurity index and the municipal marginality index that measures the levels of education, health and access to basic infrastructure (energy, WASH) and the households’ materials quality. Red colour denotes higher vulnerability, yellow colour denotes moderate vulnerability. Although the Government has implemented a decentralization process, not all local governments and institutions are prepared to execute budgets due to limited fiscal space and capacity. This could be an institutional constraint for WASH delivery and delivery of other social services. Darker brown colours with low fiscal capacity. Grey colours in both maps denote no data available. Source: SEGEPLAN, 2014. Orthophotos available at: http://ide.segeplan.gob.gt/geoportal/index.html Also see: http://web.maga.gob.gt/wp- content/blogs.dir/13/files/2013/widget/public/visan_2011.pdf 37 Source: Based on data from Segeplan, 2016. http://ide.segeplan.gob.gt/geoportal/index.html 38 Population density by district, 2012 Surface and underground water accumulation, 2014 Sources: UN Population and INE data, 2012. CONRED Guatemala. Right hand side map includes Belize’s water accumulation. Red colours in right- hand side map denote lower water accumulation levels. 39 Access Plus Framework Access to Drinking Water in Guatemala URBAN RURAL Improve JMP Improve JMP JMP d on Piped JMP d on Piped piped piped total premises on total premises on on on improve Improve Improve available premise Total improve Improve Improve available premise premise premise Total improved d d + 30 d on when s improve d d + 30 d on when s s s Year (T1) min (T2) premises needed d (T1) min (T2) premises needed 2000 94.5 95.6 93.7 87.9 53.9 87.3 90.0 82.1 83.8 77.6 54.7 40.8 53.8 60.6 2006 95.3 95.6 91.3 89.3 56.7 89.3 90.0 83.5 83.8 69.7 60.6 43.7 60.6 60.6 2011 94.9 94.5 92.0 90.0 57.7 90.0 89.9 79.8 78.4 66.4 57.3 42.0 57.3 56.1 2014 94.9 90.1 88.5 57.8 88.5 87.5 72.7 61.0 47.0 61.0 Access to Sanitation in Guatemala 40 Access to piped water, by selected characteristics of the household head - percentages 2000 2006 2011 2014 2014-2000 Male 71.26 77.77 74.00 76.41 5.15 ** Female 78.79 81.68 80.00 82.85 4.06 Age 13-24 65.54 74.12 62.84 69.14 3.6 Age 25-44 71.98 77.07 73.15 76.63 4.65 * Age 45-64 74.75 81.57 79.05 79.19 4.44 * Age 65 or more 73.52 78.97 78.91 81.48 7.96 *** Less than primary educ. 61.95 72.76 68.69 72.10 10.15 *** Completed primary or more 79.34 87.72 84.40 85.16 5.82 *** Wage earner 73.10 79.59 76.87 77.38 4.28 * Non-wage earner 71.01 76.24 71.15 76.62 5.61 * Unemployed or inactive 76.68 83.17 79.56 82.02 5.34 * Employed in other sectors 83.66 85.38 82.92 84.98 1.32 Employed in primary sector 57.29 65.62 61.98 64.71 7.42 ** Not indigenous language speaker 84.77 81.57 78.83 79.93 -4.84 Indigenous language speaker 70.01 72.44 68.02 73.22 3.21 No electricity 44.43 46.09 41.99 49.22 4.79 Dependency ratio: 0-1 77.09 83.53 78.91 80.45 3.36 Dependency ratio: 1-2 70.22 76.29 73.54 75.01 4.79 * Dependency ratio: 2+ 66.14 69.49 64.63 69.48 3.34 Non-migrant 72.66 78.70 75.17 77.62 4.96 ** Migrant 72.20 77.98 77.92 84.72 12.52 ** *p<0.05, ** p<0.01, *** p<0.001 Source: own calculations using ENCOVI survey data, years 2000, 2006, 2011, 2014 Use of surface water, by selected characteristics of the household head - percentages 2000 2006 2011 2014 2014-2000 Age 25-44 6.571 6.661 8.004 5.412 -1.2 Age 45-64 8.395 4.806 6.193 4.569 -3.8 *** Age 65 or more 6.461 5.645 6.841 4.513 -1.9 Less than primary educ. 12.29 8.605 10.58 7.246 -5.0 *** Completed primary or more 4.132 1.878 3.026 2.205 -1.9 *** Wage earner 6.026 4.775 5.544 4.964 -1.1 Non-wage earner 9.542 8.023 11.23 5.914 -3.6 ** Unemployed or inactive 3.801 3.259 4.653 3.283 -0.5 Employed in other sectors 2.122 2.031 2.690 2.431 0.3 Employed in primary sector 14.94 13.56 15.68 9.909 -5.0 ** Not indigenous language speaker 1.637 3.643 5.425 3.035 1.4 * Indigenous language speaker 12.10 10.85 11.48 9.310 -2.8 No electricity 23.00 22.29 24.35 16.83 -6.2 * Dependency ratio: 0-1 5.825 3.462 5.593 3.931 -1.9 * Dependency ratio: 1-2 8.443 7.431 7.790 5.918 -2.5 * Dependency ratio: 2+ 8.781 10.51 13.63 9.658 0.9 41 Non-migrant 7.413 5.866 7.470 5.125 -2.3 ** Migrant 5.102 7.250 5.989 2.006 -3.1 * *p<0.05, ** p<0.01, *** p<0.001 Source: own calculations using ENCOVI survey data, years 2000, 2006, 2011, 2014 Improved sanitation access, by district and year - percentages 2000 2006 2011 2014 2014-2000 Guatemala 82.02 88.71 87.40 89.74 7.7 * El Progreso 42.43 54.22 57.58 64.81 22.4 ** Sacatepequez 83.20 87.95 91.12 89.94 6.7 Chimaltenango 43.87 58.34 55.05 59.22 15.4 * Escuintla 44.69 58.82 71.07 71.61 26.9 *** Santa Rosa 34.60 45.46 50.13 64.84 30.2 *** Solola 8.354 26.04 34.60 38.28 29.9 *** Totonicapan 23.37 23.60 24.71 30.05 6.7 Quetzaltenango 50.35 56.18 52.89 61.69 11.3 Suchitepequez 32.72 50.78 66.76 61.11 28.4 *** Retalhuleu 34.86 45.88 42.23 44.24 9.4 San Marcos 19.72 30.34 39.03 35.59 15.9 Huehuetenango 19.99 36.33 38.86 37.30 17.3 ** Quiche 13.08 23.59 31.06 38.53 25.5 *** Baja Verapaz 24.98 27.03 43.39 46.02 21.0 ** Alta Verapaz 10.65 20.30 17.27 22.31 11.7 Peten 12.60 25.36 22.98 31.58 19.0 *** Izabal 43.03 57.67 69.40 62.58 19.6 Zacapa 37.79 55.81 59.90 68.94 31.2 ** Chiquimula 28.46 48.95 50.85 51.68 23.2 * Jalapa 27.82 32.88 41.33 50.15 22.3 ** Jutiapa 23.58 43.37 51.03 59.76 36.2 *** *p<0.05, ** p<0.01, *** p<0.001 Source: own calculations using ENCOVI survey data, years 2000, 2006, 2011, 2014 Improved access to water, by district and year - percentages 2000 2006 2011 2014 2014-2000 Guatemala 93.09 93.52 91.76 93.05 -0.04 El Progreso 86.94 88.19 87.42 88.6 1.66 Sacatepequez 97.96 94.52 94.22 96.12 -1.84 Chimaltenango 93.18 97.5 90.42 93.56 0.38 Escuintla 97.82 98.23 92.76 96.24 -1.58 Santa Rosa 89.02 92.5 81.31 86.32 -2.7 Solola 100 96.08 96.93 98.4 -1.6 ** Totonicapan 87.49 97.15 92.73 91.4 3.91 Quetzaltenango 95.64 93.88 85.21 93.83 -1.81 Suchitepequez 98.34 95.84 89.97 96.23 -2.11 Retalhuleu 99.24 95.76 95.02 95.48 -3.76 ** San Marcos 84.03 91.55 77.94 89.85 5.82 Huehuetenango 79.27 82.45 83.19 89.75 10.48 ** Quiche 84.98 84.55 82.58 92.56 7.58 42 Baja Verapaz 83.21 78.84 80.1 89.05 5.84 Alta Verapaz 63.08 69.3 81.3 80.7 17.62 ** Peten 76.24 79.13 76.4 87.99 11.75 ** Izabal 80.93 87.83 86.38 93.98 13.05 ** Zacapa 79.2 92.62 90.63 89.66 10.46 Chiquimula 54.53 86.82 81.88 83.12 28.59 * Jalapa 86.36 87.02 80.33 89.61 3.25 Jutiapa 91.65 89.57 84.52 90.24 -1.41 Source: own calculations using ENCOVI survey data, years 2000, 2006, 2011, 2014 Access to flush sanitation, by selected characteristics of the household head - percentages 2000 2006 2011 2014 2014-2000 Female 42.99 53.59 57.63 59.08 16.1 *** Age 13-24 27.78 45.54 40.53 46.61 18.8 *** Age 25-44 33.11 46.99 46.38 49.87 16.8 *** Age 45-64 34.22 48.93 51.97 52.05 17.8 *** Age 65 or more 36.30 45.17 54.24 51.70 15.4 *** Less than primary educ. 15.27 30.01 32.78 35.66 20.4 *** Completed primary or more 44.91 73.91 71.38 70.16 25.3 *** Wage earner 37.15 52.46 51.85 52.07 14.9 *** Non-wage earner 26.53 38.93 40.46 46.06 19.5 *** Unemployed or inactive 44.80 56.61 59.06 56.90 12.1 *** Employed in other sectors 50.04 63.76 66.27 68.47 18.4 *** Employed in primary sector 8.833 16.10 18.67 20.21 11.4 *** Not indigenous 57.33 57.18 60.29 60.53 3.2 Indigenous 12.34 26.23 26.43 29.90 17.6 *** No electricity 1.872 8.119 9.964 12.98 11.1 *** Electricity 44.99 54.92 56.36 57.10 12.1 *** Dependency ratio: 0-1 41.57 56.36 55.78 55.93 14.4 *** Dependency ratio: 1-2 29.59 42.02 43.65 46.69 17.1 *** Dependency ratio: 2+ 20.59 30.52 34.29 32.92 12.3 *** Non-migrant 32.41 46.72 48.69 50.18 17.8 *** Migrant 50.66 57.87 55.29 71.64 21.0 *** Source: own calculations using ENCOVI survey data, years 2000, 2006, 2011, 2014 43 Open defecation, by selected characteristics of the household head - percentages 2000 2006 2011 2014 2014-2000 Age 13-24 18.79 13.24 17.22 9.665 -9.1 *** Age 25-44 12.98 9.347 8.668 5.416 -7.6 *** Age 45-64 12.09 7.324 5.354 4.005 -8.1 *** Age 65 or more 13.80 10.01 6.616 4.402 -9.4 *** Less than primary educ. 20.19 13.49 11.27 7.141 -13.0 *** Completed primary or more 8.787 2.096 3.020 2.279 -6.5 *** Wage earner 11.46 8.353 7.346 5.179 -6.3 *** Non-wage earner 15.74 10.30 9.006 4.957 -10.8 *** Unemployed or inactive 10.33 7.115 6.540 4.673 -5.7 *** Employed in other sectors 6.571 3.848 3.779 2.505 -4.1 *** Employed in primary sector 22.50 18.36 14.41 9.126 -13.4 *** Not indigenous 7.202 7.505 6.850 5.139 -2.1 Indigenous 15.53 12.21 9.912 4.801 -10.7 *** No electricity 31.68 30.73 25.01 15.37 -16.3 *** Dependency ratio: 0-1 10.32 6.434 5.702 3.709 -6.6 *** Dependency ratio: 1-2 14.20 9.819 9.138 5.928 -8.3 *** Dependency ratio: 2+ 17.89 14.74 12.52 9.178 -8.7 *** Non-migrant 13.34 9.046 7.937 5.075 -8.3 *** Migrant 10.57 8.277 4.437 2.855 -7.7 *** Source: own calculations using ENCOVI survey data, years 2000, 2006, 2011, 2014 44 45 The effect of improved access to water and sanitation on household consumption, year 2000 (1) (2) (3) (4) (5) (6) (7) VARIABLES Consumption Consumption Consumption Consumption Consumption Consumption Consumption Improved water 0.559*** 0.213*** 0.195*** (0.024) (0.037) (0.040) Female 0.214*** 0.186*** 0.183*** 0.163*** 0.170*** (0.033) (0.034) (0.032) (0.034) (0.034) Age -0.005 -0.006 -0.007* -0.007** -0.006* (0.004) (0.004) (0.004) (0.004) (0.003) Age squared 0.000* 0.000** 0.000** 0.000** 0.000** (0.000) (0.000) (0.000) (0.000) (0.000) Dep.ratio -0.269*** -0.253*** -0.251*** -0.241*** -0.234*** (0.013) (0.011) (0.011) (0.011) (0.012) Employed, wage earner -0.058 -0.063 -0.053 -0.058 -0.062 (0.043) (0.044) (0.044) (0.045) (0.041) Employed, non- wage earner -0.056 -0.023 -0.043 -0.017 -0.017 (0.041) (0.042) (0.042) (0.043) (0.040) Urban 0.676*** 0.465*** 0.352*** 0.224*** 0.223*** (0.024) (0.021) (0.026) (0.025) (0.027) Primary or more 0.425*** 0.355*** 0.342*** 0.296*** 0.298*** (0.022) (0.021) (0.021) (0.021) (0.021) Improved sanitation 0.899*** 0.554*** 0.460*** (0.014) (0.026) (0.025) Improved water and sanitation 0.470*** (0.031) Constant 8.501*** 8.160*** 8.307*** 8.529*** 8.323*** 8.442*** 8.427*** (0.022) (0.089) (0.087) (0.009) (0.084) (0.080) (0.081) Observations 18,811 7,081 7,081 18,811 7,081 7,081 7,081 R-squared 0.041 0.451 0.516 0.290 0.497 0.545 0.548 District fixed effects NO NO YES NO NO YES YES Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 The outcome is the natural logarithm of household consumption. Data come from the 2000 wave of the National Household Living Standards Survey of Guatemala. 46 The effect of improved access to water and sanitation on household consumption, year 2014 (1) (2) (3) (4) (5) (6) (7) VARIABLES Consumption Consumption Consumption Consumption Consumption Consumption Consumption Improved water 0.559*** 0.186*** 0.186*** (0.024) (0.024) (0.023) Female 0.153*** 0.124*** 0.128*** 0.108*** 0.108*** (0.019) (0.019) (0.019) (0.019) (0.019) Age -0.005* -0.004 -0.006** -0.005 -0.004 (0.003) (0.003) (0.003) (0.003) (0.003) Age squared 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) depratio -0.234*** -0.223*** -0.217*** -0.212*** -0.210*** (0.010) (0.010) (0.010) (0.009) (0.010) Employed, wage earner -0.000 -0.004 -0.004 -0.006 -0.002 (0.029) (0.028) (0.028) (0.028) (0.028) Employed, non- wage earner 0.044 0.066** 0.044* 0.059** 0.062** (0.027) (0.027) (0.026) (0.026) (0.027) Urban 0.310*** 0.238*** 0.148*** 0.115*** 0.120*** (0.015) (0.015) (0.015) (0.015) (0.015) Primary or more 0.499*** 0.455*** 0.428*** 0.406*** 0.401*** (0.016) (0.016) (0.016) (0.016) (0.016) Improved sanitation 0.899*** 0.374*** 0.331*** (0.014) (0.015) (0.015) Improved water and sanitation 0.326*** (0.015) Constant 8.501*** 8.890*** 8.922*** 8.529*** 8.978*** 9.007*** 9.013*** (0.022) (0.077) (0.076) (0.009) (0.072) (0.072) (0.072) Observations 18,811 11,071 11,071 18,811 11,071 11,071 11,071 R-squared 0.041 0.370 0.410 0.290 0.412 0.438 0.439 District fixed effects NO NO YES NO NO YES YES Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 The outcome is the natural logarithm of household consumption. Data come from the 2000 wave of the National Household Living Standards Survey of Guatemala. 47 48 Source: ENCOVI survey, 2014 49 50 Water Access Plus - Improved water, by area and year (%) 2000 2006 2011 2014 Urban Tier 1 Improved water 94.5 95.3 94.9 94.9 Tier 2 Basic water 93.7 91.3 92 90.1 Water 4 Improved on premises 87.9 89.3 90 88.5 Water 5 Piped on premises 87.3 89.3 90 88.5 Rural Tier 1 Improved water 82.1 83.5 79.8 87.5 Tier 2 Basic water 77.6 69.7 66.4 72.7 Water 4 Improved on premises 54.7 60.6 57.3 61 Water 5 Piped on premises 53.8 60.6 57.3 61 National Tier 1 Improved water 86.9 89.2 87.1 91.2 Tier 2 Basic water 83.8 80.1 78.8 81.3 Water 4 Improved on premises 67.5 74.4 73.1 74.6 Water 5 Piped on premises 66.7 74.4 73.1 74.6 Source: own calculations using ENCOVI survey data, years 2000, 2006, 2011, 2014 51 Sanitation Access Plus - Improved sanitation, by area and year (%) 2000 2006 2011 2014 Urban Sanitation 1 Open defecation 2.8 1.9 1.9 1.6 Sanitation 2 Improved sanitation (including shared) 88.9 88.5 90.2 89.7 Tier 1 Improved sanitation 74.2 72 77.2 77.6 Sanitation 3 Flush to sewage 60 65.2 69.6 68.1 Rural Sanitation 1 Open defecation 20.8 17.2 13.5 8.9 Sanitation 2 Improved sanitation (including shared) 45.9 51.1 54.5 58.7 Tier 1 Improved sanitation 40.2 45.7 48.7 54.2 Sanitation 3 Flush to sewage 4.46 6.72 7.75 9.09 National Sanitation 1 Open defecation 13.8 9.9 7.9 5.3 Sanitation 2 Improved sanitation (including shared) 62.5 69.1 71.8 74.1 Tier 1 Improved sanitation 53.3 58.4 62.5 65.8 Sanitation 3 Flush to sewage 25.9 34.8 37.7 38.3 Source: own calculations using ENCOVI survey data, years 2000, 2006, 2011, 2014 Children 3 to 60 Months Old Quintile of Chronic Anaemia % Height Undernutrition below 145 Income prevalence cm (%) I 70.2 50.7 46.9 II 59.7 51.4 36.3 III 43.8 46.4 29.8 IV 25.5 43.2 18.6 V 14.1 39.8 10.3 Indigenous 65.9 49.5 48.3 Non- 36.2 42.1 19.0 Indigenous 52 Scatterplots with trends in access and key variables by Department 2000-2014 50 Children's diarrhea incidence 40 30 20 10 0 0 10 20 30 40 Access to water network Yr2006 Yr2014 Yr2000 Source: ENCOVI, 2000, 2006, 2014. Distance to water source (meters) extreme poor 0 200 400 600 800 0 200 400 600 800 Distance to water source (meters) bottom 40 Yr2000 Yr2014 Yr2006 Source: ENCOVI, 2000, 2006, 2014. 53 100 Improved Water Coverage Rate 2000 80 60 40 20 0 0 20 40 60 80 100 Improved Water Coverage Rate 2014 Indigenous Improved Water (2000) Indigenous Improved Water (2014) Non Indigenous Improved Water (2000) Non Indigenous Improved water (2014) Source: ENCOVI, 2000, 2006, 2014. 100 Improved sanitation Coverage rate 80 60 40 20 0 0 20 40 60 80 100 Indigenous (%) Yr2006 Yr2014 Yr2000 Source: ENCOVI, 2000, 2006, 2014. 54 60 Child Diarrhea Incidence (baseline) 40 20 0 0 10 20 30 40 Child Diarrhea Incidence (current) B40_2000 B40_2014 Rest of Groups_2000 Rest of Groups_2014 Source: ENCOVI, 2000, 2006, 2014. 500 Distance to Water (mtrs) baseline 100 200 300 0 400 0 100 200 300 400 500 Distance to Water (mtrs) current B40 2000 B40 2014 Rest of Groups 2000 Rest of Groups 2014 Source: ENCOVI, 2000, 2006, 2014. 55 100 80 Improved Sanitation (B40) Sacatepequez Guatemala 60 Quetzaltenango 40 Izabal Chimaltenango Escuintla El Progreso SuchitepequeZ Santa Rosa Retalhuleu Zacapa Jalapa Chiquimula 20 Baja Verapaz Totonicapan Jutiapa San Marcos Huehuetenango Quiche Peten Alta Verapaz Solola 0 20 40 60 80 100 Improved Sanitation (Extreme Poor) Yr2000 Yr2014 Yr2006 Source: ENCOVI, 2000, 2006, 2014. 100 Solola Retalhuleu SuchitepequeZ Sacatepequez Escuintla Quetzaltenango Improved water (Extreme Poor) Jutiapa Guatemala Chimaltenango 90 Santa Rosa Jalapa Totonicapan El Progreso Baja Verapaz Quiche San Marcos 80 Huehuetenango Izabal Zacapa Peten 70 Alta Verapaz 60 50 Chiquimula 60 70 80 90 100 Improved water (B40) Yr2000 Yr2014 Yr2006 Source: ENCOVI, 2000, 2006, 2014. 56 80 Acute respiratory infections among children (B40) Retalhuleu Chiquimula Alta Verapaz Jalapa 60 Baja Verapaz Izabal Suchitepequez San Marcos Chimaltenango Jutiapa Zacapa Quiche Peten Escuintla Solola Sacatepequez Quetzaltenango 40 Guatemala Santa Rosa Totonicapan Huehuetenango El Progreso 20 0 0 20 40 60 80 Acute respiratory infections among children (extreme poor) Yr2006 Yr2014 Yr2000 Source: ENCOVI, 2000, 2006, 2014. Source: ENCOVI, 2000, 2006, 2014. 57 Source: ENCOVI, 2000, 2006, 2014. 58 Other indicators and graphs 59 Statistical Analyses Chi-squared test for independence Lack of access to Lack of access to improved Attribute improved water sanitation Rural 2.7e+05*** 5.0e+06*** General poverty 3.7e+05*** 3.2e+06*** Extreme poverty 2.2e+05*** 2.3e+06*** Bottom 40 percent 3.5e+05*** 3.0e+06*** Household head is female and employed outside the home 1.0e+04*** 2.2e+05*** Indigenous 1.1e+05*** 2.0e+06*** Illiterate 9.7e+04*** 5.8e+05*** Household size 0.011*** 0.210*** Dependency ratio 0.103*** 0.192*** Lack of access to improved water 7.1e+05*** Lack of access to improved sanitation 7.1e+05*** *** p<0.01 Source: Authors’ calculations based on ENCOVI, 2014. Regression analyses by department Regression was run at the household level, restricting the sample period to the year 2014 only. Regressions are weighted using the sampling weights provided in the ENCOVI dataset. The analysis was performed both at the national level, with the inclusion of Department indicators, and for each Department separately. The outcomes that were investigated can be grouped into two categories: 1. Access to WASH services; 2. Child disease. Access to WASH services Outcome variables pertaining to group 1 were household access to improved water and access to improved sanitation. For these variables, the analysis was performed via Probit regression. Given a binary outcome taking either value 0 or 1, the probability that = 1 for household ℎ was therefore modelled as (ℎ = 1) = �ℎ �, where is a row vector of covariates (including a constant), is a vector of coefficients and is the distribution function of a Standard Normal random variable. The vector of covariates included the following regressors: • An indicator for living in a rural area 60 • An indicator for being poor using the Guatemalan poverty line as defined by the ENCOVI survey • An indicator for being extremely poor using the Guatemalan extreme poverty line as defined by the ENCOVI survey 10 • An indicator for belonging to the Bottom 40 of the consumption distribution at the national level • An indicator for female household head • An indicator for belonging to an indigenous ethnicity • An indicator for illiteracy • Household size • Household dependency ratio (measured as the ratio between number of household members younger than 15 and older than 65 and number of household members aged 15-65) • An indicator for access to improved sanitation (when the outcome was access to improved water) or for access to improved water (when the outcome was access to improved sanitation) • (A set of Department fixed effects when the analysis was performed at the national level). The output tables report the average marginal effects of the covariates. For a continuous variable 1 , the ̂1 is that a marginal increase in 1 raises the probability that interpretation of a marginal effect equal to ̂1 percentage points. For a binary variable 2 , the interpretation of a marginal effect = 1 by 100 × ̂2 is that when 2 switches from 0 to 1 the probability that = 1 increases by 100 × equal to ̂2 percentage points. Child disease Outcome variables pertaining to group 2 were household proportion of children with diarrheal disease and with respiratory infection in the month before the interview. For these variables, analysis followed an OLS approach. Given a continuous outcome , its realisation for household ℎ was therefore modelled as ℎ = ℎ + ℎ , where is a row vector of covariates (including a constant), is a vector of coefficients and is an individual error term which is assumed to follow a Normal distribution. The vector of covariates included: • An indicator for living in a rural area • An indicator for having water on premises • A set of mutually exclusive indicators for type of access to water, namely: o Piped to home o Piped to yard o Piped to neighbour/public o Tube-well/borehole o Rainwater o Truck or barrel o Surface water o (Other unimproved source is the omitted category) • An indicator for no treatment of drinking water • A set of mutually exclusive indicators for type of access to sanitation, namely: o Flush to sewer 10 The dummy variable for extremely poor therefore identifies a subset of those who also qualify as generally poor in terms of the dummy variable “poor” described above. 61 o Flush to septic tank o Pour flush/composting toilet/other improved type o Latrine o Open defecation The output tables report the coefficient of the covariates, which can be interpreted as their marginal effects. For a continuous variable 1 , the interpretation of a marginal effect equal to ̂1 is that a marginal ̂ increase in 1 raises the outcome by 1 . For a binary variable 2 , the interpretation of a marginal effect ̂2 is that increases by equal to ̂2 when 2 switches from 0 to 1. The regression model The analysis focusses on five child anthropometric outcomes: 1. height (measured in cm); 2. weight (measured in kg); 3. stunting (height-for-age Z-score ≤ 2); 4. wasting (weight-for-height Z-score ≤ 2); 5. underweight (weight-for-age Z-score ≤ 2). For items 1 and 2 of the list above, which are continuous in nature, OLS regression is per- formed, whereas items 3-5, which consist of binary variables, are studied through a Probit model. The set of controls included in the regressions does not however vary across specifications, and is composed of the following groups of variables: 1. Main covariates of interest (WASH variables): • Lack of access to improved water • Either: Lack of access to improved sanitation; Or: Open defecation • Unsafe disposal of child stools. 2. Location controls: • Rural-urban location (Rural=1) • 22 Department fixed effects. 3. Household socio-demographic characteristics: • Age of household head (linear and square) • Years of education of household head • Gender of household head (Female=1) • Household dependency ratio (ratio of number of household members aged 14- or 65+ over number of members aged 15-64) 1 The Z-score of an anthropometric index is computed as: (observed value − median of the reference population) Z-score = standard deviation of reference population . For a child to be stunted, their height-for-age ratio therefore has to be at least 2 standard deviations below the refer- ence median. The reference population is obtained from US CDC (Centre for Disease Control) Standard Deviation- derived Growth Reference Curves from the Reference Population recommended by the WHO, based on data from the National Centre for Health Statistics (NCHS). 62 • Household wealth index.2 4. Mother characteristics: • Mother’s age in years • Mother has primary education • Mother has secondary or higher education • Mother’s age at first birth • Mother’s height in cm • Mother’s BMI • Total number of children ever born to the mother. 5. Child characteristics: • Child age in months • Child gender (Male=1) • Child birth order • Child month of birth • Indicator for whether child is from a multiple birth • Child received iron supplementation in the last 7 days • Child has health and vaccination card (seen by enumerator) • Number of antenatal visits received by mother • Child was delivered in a healthcare institution • Child received Vitamin A supplements (on health card) • Child received Vitamin A supplements (recalled by mother). Regressions are weighted using the sampling weights provided in the DHS dataset, which make the analysis representative of the country’s population. Standard errors are clustered at the primary sampling unit level. The analysis is performed both on the total national sample of all children aged 0- 59 months and on the following sub-samples: • Urban vs. Rural location • Bottom 40 vs. Top 60 percent of the wealth distribution in the country3 • Children aged 0-24 vs. 25-59 months/boys and girls 2 The wealth index is built as the first principal component from principal component analysis on a set of household assets and dwelling characteristics (excluding WASH information), weighted by rural-urban location. 3 As mentioned in a previous footnote, the wealth index is built as the first principal component from principal component analysis on a set of household assets and dwelling characteristics (excluding WASH information), weighted by rural-urban location. 63 In what follows, Sections report results for each outcome in turn. Within Sections, each plot reports the coefficient/marginal effect of one WASH variable of interest at a time. Different rows in the same plot identify the sample on which the regression is performed. Child´s height Regression Coefficients for lack of access to improved water (dependent variable) Regression Coefficients for lack of access to improved sanitation (dependent variable) 64 Regression Coefficients for lack of access to open defecation (dependent variable) Regression Coefficients for lack of access unsafe disposal of child faeces (dependent variable) 65 Regression Plots for Access to Improved Water by Department and Basic Characteristics, DHS 2014 66 67 Regression Plots for Access to Improved Sanitation by Department and Basic Characteristics, 2015 68 69 Regression coefficients and confidence intervals of determinants of children´s diarrhea 2015 (By department, only significance of coefficients at 10 and 5 percent reported) 70 71 Conceptual Framework of Institutional Analysis MAPAS Methodology The MAPAS initiative seeks to overcome the limitations of the traditional monitoring processes, which are focused solely on determining progress in the sector based on the expansion of water and sanitation coverage, and analyzes other factors that contribute to the sustainability of the service related to the quality and management of the service provided, equitable distribution, and user satisfaction. The usefulness of MAPAS lies in its capacity to identify, through two instruments—the scorecard and financial analysis—the sectoral strengths and challenges for the provision of high- quality water and sanitation services in urban and rural areas. In addition, the MAPAS initiative stimulates broad sectoral dialogue, facilitating a suitable environment for national governments to expedite sectoral reform processes and encourage the transformation of the processes of economic and technical resources into sustainable services. The scorecard evaluates the service delivery pathway through which the country transforms funding in to sustainable water and sanitation services in each of the four subsectors: rural water, urban water, rural sanitation, urban sanitation (Table 1). The pathway is evaluated through nine building blocks, which are grouped into three main pillars, analyzed through: (1) the effectiveness of the institutional framework, (2) the concrete results obtained through the process of developing plans and sector activities, and (3) the sustainability of the services provided. Each building block is scored from 0 to 3 depending on the responses to the three to five indicators per block. The scorecard uses a traffic-light color code. A green block means a highly satisfactory score (higher than 2); a yellow block means a neutral score (between 1 and 2); and a red block means an unsatisfactory score. The development of the MAPAS initiative included the following steps: 1. Collection and organization of the sectoral technical and financial information in each country. 2. Analysis and validation of the sectoral information and assessment of subsector capacity to convert available financial resources into sustainable water and sanitation services. 3. National dialogue to establish a consensus on the initial analysis. Results of Balanced Scorecard of the WSS sector in Guatemala The MAPAS analysis suggests that there is a framework of policies in the sector that has favored the water subsector. However, despite the existence of a long-term National Plan, planning is “unsatisfactory” in all the subsectors; there are no comprehensive strategies that would address in an integral manner the problems affecting the sector at different levels of intervention. Table 1: MAPAS Scorecard Guatemala 2016 Source: MAPAS. Legend: Green fill means a highly satisfactory score (higher than 2.0); yellow fill means a fairly satisfactory score (1.0 ≤ x ≤ 2.0); and red fill means an unsatisfactory score (less than 1.0). 72 Because of the lack of strategic planning and investment in the sector, it is difficult to allocate resources, which is reflected in a budget that does not cover expectations and lacks criteria of equity in its allocation, combined with low budgetary execution in recent years. This situation is reflected in the results at the level of the four subsectors, with low performance in the quality of services, access, and protection of water resources. Finally, there is low operational and managerial capacity in the providers, both rural and urban, which makes the services provided unsustainable because of the limited technical and managerial capacity of the municipal or rural providers and the virtually nonexistent monitoring by the central government, especially in the rural area, where scores of “0” in the Scorecard 11 reveal that the rural sanitation situation is one of the most important challenges confronting the sector in Guatemala. WDR Framework Following the WDR 2004 framework, this lens examines the set of oversight and accountability relationships between service users, providers and the state. In this subsection we assess the strength of both the short and long route of accountability. The short route of accountability relies on consumers’ direct engagement with service providers, while the long route relies on citizens’ exercising their ‘voice’ through elections and the government’s ‘compact’ with service providers (see Figure 6.3). In the Guatemalan context, the extent of decentralization requires particular attention be paid in the long route to the municipal authorities who hold responsibility for provision. The nature of these relationships, including the interests, incentives and power of the actors involved, help to shape behavior in ways that contribute to the identified service delivery outcomes. It should be noted that rural water provision in Guatemala does not fit neatly into the WDR 2004 framework, since water users (through community drinking water providers, CAAP) play an important role in service provision. CAAPs are thus simultaneously agents of the community and agents of the providers. Central Government and Municipalities Long route Short route Service Providers: Citizens/Clients: Client Power Municipalities, CAAPs, B40 and T60 private companies Source: World Bank based on WDR (2004). Note: Includes short and long routes of accountability in the WASH Sector in Guatemala A generic PEA framework (Harris 2012) has been adapted in a revised framework, presented below, to draw on thinking in and since the World Development Report (WDR) 2004, with a focus on: • Problems are focused on the bottom 40%, including disaggregation and comparisons between different groups • The specific sector and sub-sector (‘service’) characteristics relevant to these problem(s) are identified, with a focus on the implications for monitoring, accountability and other relationships between core actors 11 The MAPAS Scorecard is generated based on secondary information, bibliography, interviews, and a broad process of consultation with representatives of the urban and rural sector, in this case in Guatemala. 73 • Core sets of relationships are explored as part of actor mapping, with a focus on citizens – policymakers – service providers • Heterogeneity within these groups will also be explored – including the potential for multiple, overlapping ‘triangles’ (following the WDR 2004 framework) • A wider set of motivations and networks will be taken into account • The understanding that the bottom 40% are more likely to use informal service providers than the broader population. The framework below is structured around: • Step 1: Initial problem identification, with a focus on the bottom 40% • Step 2: Analysis of relevant institutional arrangements, structural factors and stakeholder mapping, that explain why this problem persists, exploring both sub-sector and sector level factors, and factors in the wider context, structured around the sets of relationships identified in WDR 2004 ‘triangle’ (the ‘Key relationships of power’) • Step 3: Identifying potential reform entry points – developing up options for interventions, processes for reform and potential progress markers (e.g. how to measure whether these interventions are contributing to the institutional changes envisaged). For all countries undertaking CQ#4 analysis the first phase involves completing steps 1 and 2 as a desk-based exercise, leading to identification of hypotheses about the key political economy problem (rather than identification of specific reform entry points). This is the focus of this note and constitutes a ‘B40-focused WASH institutional diagnostic’. The Global Firm refers to this level of analysis ‘Tier 1’. For those countries, such as Mozambique, which progress to the full problem-driven political economy analysis, referred to as ‘Tier 2’ the research team will deepen step 2 and extend to step 3 through analysis in-country. For Indonesia the detailed approach to this Tier 2 work will be agreed through a separate Task Order and Approach Paper. MAPAS contains some elements of Tier 1 and 2. Overarching PEA framework for WASH Poverty Diagnostics (Tier 1 and Tier 2) 74 In Guatemala, clients do not have significant scope for exercising client power, particularly in rural areas. As mentioned previously, in rural areas community-managed systems dominate. Though municipalities are legally responsible for providing WASH services in rural areas, communities generally build, operate, and maintain their own systems through drinking-water committees. This makes it difficult to distinguish between “citizens/clients” and “service providers” in the rural sector, and represents a constraint in itself: The fact that water users are responsible for operations and maintenance (O&M) makes it difficult for water users to express “client power” and make use of the short route of accountability. In terms of institutional aspects available to ensure water quality, the Municipal Code requires municipalities to provide chlorinated water to their communities, a requirement with which few municipalities currently comply. Under the Peace Accords of 1996 and the Decentralization Law of 2002, municipal governments are guaranteed a “Constitutional Assistance” (Aporte Constitucional) of 10 percent of government ordinary revenue plus 1.5 percent of the 10 percent value added tax. Municipalities are required to spend a minimum of 90 percent of the Aporte Constitucional on infrastructure, including – but not limited to – potable water and sanitation. While MSPAS oversees compliance with water and sanitation laws and promotes improvements to water and sanitation infrastructure, it is primarily the responsibility of the municipalities to undertake both their construction and financing 75