São Francisco River Basin Water and Poverty Relationship

The São Francisco River Basin of Brazil has been subject to multiple interventions to increase access to water for agricultural purposes. The analysis being conducted includes a comprehensive measurement of water availability and its impact on agricultural activities (and poverty) in the SFRB. It controls for variations in socio-economic and political variables. We use two types of measures of access to water, one traditional and one new, to try to identify the links (broadly defined) between water and poverty in the SFRB. Our analysis is done at município level (using 1990 boundaries, there are approximately 450 municípios that comprise the SFRB). We then take estimates of 2003 rural poverty, the only recent ones available for rural areas in Brazil, and have begun to examine interrelationships. Our results imply that overall explanatory power of water on poverty varies across rainfall regimes and between poverty and extreme poverty.

In particular, the results of this modeling effort seem to provide evidence that in areas with fewer natural constraints to large-scale agriculture production (areas with relatively more precipitation), the availability of water may play a more direct role in rural income generation. In areas with less rainfall, variabilty in precipitation may play a bigger role in determining income. The specific econometric results and the correlation between other factors and poverty rates will be analyzed in detail.

In general, the evidence we have on water access/ag poverty/links (distilled from this regression work and other research activities) suggests that where they do exist, the links between water and poverty may not be very strong, and will likely be ‘partial’ and increasingly ‘indirect.’ They will be partial in the sense that improving access to water may turn out to be a necessary condition for improving rural livelihoods in some areas and for some types of smallholders, but it is unlikely to be a sufficient condition for doing so. The links will be indirect in the sense that improving access to water for medium- and large-scale farming operations may help reduce rural poverty via on-farm employment generation, but this indirect link will vary quite substantially depending on product mix on these farms.

Specific outcomes and conclusions of this research are pending.

Water-related poverty in the Nile basin

We gain a better understanding of the linkage between water and poverty through:-

1. Identifying key drivers

― water access conditions in agricultural systems (determinants, physical and economic)

― changing water conditions in agricultural systems (variability in water access, water productivity)

2. Understanding the conceptual framework

—characterizing poverty hotspots using household income and expenditure data. Poverty levels are higher in rural agricultural areas compared to urban areas due to income access. In the Nile basin, with the exception of Egypt, rural poverty accounts for 90 percent of total poverty and close to 80 percent of the poor depend on agriculture and farm labor as their main source of income and livelihood. Related to type 2 through disparities in safe drinking water and sanitation

—characterizing hotspots of the “water poor” meaning people who are deprived due to physical and economic water scarcity. Most of the “water poor” are found in the degraded and deforested areas of rangelands and in mixed rainfed systems that have a poorly developed water access infrastructure. Related to type2 where infrastructure determines allocation to users

—a dynamic livelihood system characterizing hotspots of biophysical and social vulnerability, given that a weak asset capital base indicates a lack of capacity to adapt to water stress as manifested through changing water conditions in agricultural systems. Areas with high vulnerability scores in rangeland and mixed rainfed systems are associated with low crop and livestock water productivity. Related to type, 4 & 5 where loss of livelihood emanates from low adaptive capacity

—a characterization of hotspots of water-related hazards (droughts, floods, diseases), highlighting areas of high exposure in marginal land such as floodplains, where there is a high risk of outbreaks of water-borne human and cattle diseases Related to type 3 due to poor ability to cope with impacts from hazards

The four concepts of water-related poverty and vulnerability

Concept 1, Measures of well-being

Deprivation from consumption estimated from lack of financial ability to meet social needs commonly defined as absolute poverty. There are three approaches for this estimate; poverty line, poverty gap and poverty inequality. Although there are other types of deprivation, we focus on two of these since they are more relevant for our analyses approach with the Nile basin data sets.

Absolute poverty being the income measure which is linked to goods and market forces that determine the allocation of goods and services to meet a minimum level of consumption. This is a good indicator for developed economies and appropriate for the case of Egypt since relationship is direct where financial deprivation leads to poverty.

Physiological poverty refers to the inability of people to meet basic needs such as food, clothing and shelter. This form of deprivation is widespread and serves as a better indicator in developing economies as is the case with most countries in the basin. We show examples of cases of northern Uganda where high poverty manifests both low and high physiological deprivation.

Hotspots of poverty in agricultural systems are a rural phenomenon. They occur in both areas of high potential for crop and livestock productivity and in rangeland systems. This is surprising and we are currently testing whether this phenomenon is related to market access. For instance, we compare poverty levels in the Kenya highlands with better market access to those found in the Ethiopian highlands having relatively poorer access to markets. Also we show that high poverty levels in rangelands are generally less variable but in some cases they may be related to environmental insecurity from water and pasture conflicts. In general our case studies suggest that education level as a proxy indicator for poverty significantly contributes to the differences we find in the prevalence of water-related poverty in the Nile basin

Concept 2, Poverty and access to water

In this approach, access is viewed as deprivation from consumption which is a function of both physical and economic factors. Together they constitute the “water poor”. In our work, we emphasize variability arising from water access due to the development of water infrastructure rather than total water availability, therefore excluding rainfall which is only weakly correlated with poverty

―measured as contribution of location of water sources, relative to the spatial demand of people and livestock in agricultural systems. For the Uganda country data set we use spatial variability in the location of water sources to explain differences in the poverty measures at the parish level. A number of water sources associated with high coefficients show unique positive as well as negative attributes in the overall model. This overrides the contribution of human and livestock population factors although we know these are major drivers of water use. (1)We provide more examples from regression analyses of poverty indicators and water access variables in Uganda and their implications for the entire basin

― by evaluating the role of water technologies and water policies in significantly influencing access to water for poor people in rural agricultural systems. There is little or no adoption of technology to adapt to changing water conditions, exposing poor people to water-related stress. (2)We provide examples from “cattle corridor” household survey of water-related poverty in 2 districts, Nakasongola (0⁰ 55N to 1⁰ 40N and longitude 31E 55E and 32 50E) and Kiruhura (00⁰15” to 00⁰ 24” and longitude 031⁰ 34” to 031 ⁰ 4)

Concept 3, Biophysical and social vulnerability

Vulnerability = f (sensitivity, adaptive capacity).

Sensitivity (S) and adaptive capacity (AC) [High S, Low AC = Vulnerable; Low S, High AC = Resilient]

The key point is to combine biophysical and social indicators with understanding of water-related hazards in concept 4. As opposed to poverty which is static, vulnerability is dynamic and is a characteristic of the sensitivity of the human and natural environment. In this starting point approach, poor people are sensitive to water stress since their livelihoods are dependent on rainfed agriculture. Their ability to adapt to changing water demands for agriculture is a function of how well endowed they are from natural, physical and social capital assets.

We determine vulnerability as the ability of people to deploy these assets in order to adapt to water stress while improving water productivity, given changing water quantity and quality conditions. (1) Biophysical vulnerability is assessed through scoring natural asset indicators such as water, land suitability and physical assets such as market access infrastructure. (2) Social vulnerability is assessed through scoring social asset indicators of human conditions such as agricultural labor as well as financial assets for investing in water technologies. A poor asset indicator (or index) score renders one vulnerable to impacts from variable water conditions in the agricultural systems. The linkage to poverty is in a livelihood system that modifies access to food through agricultural water management while accounting for risks associated with declining water resources, loss of pasture and cropland (or crop yields) in degraded areas. Other than rainfall, severe land degradation/deforestation and environmental insecurity are factors that cause variation in the availability agricultural water in the basin.

Concept 4, Poverty and water-related hazards

This concept is viewed as an end point assessment where poor people living on marginal land are exposed to water-related hazards. The most common events occurring in the Nile basin are associated with loss of livestock assets and crop failure from drought and flooding. A second category is in variability related to extreme events in the distribution of rainfall and temperature, for example prolonged intense dry spells and flooding events lead to crop failure and pasture loss in the rangelands as well as in floodplain regions of mixed rainfed systems. Also, extreme rainfall events may cause structural damage to infrastructure, disrupting agricultural production, transport and market functions particularly in mixed rainfed systems. The third category is water-related disease risk due to livestock and human exposure to poor quality agricultural water particularly where open water bodies are located as the main source of Water.

Outcomes of the 4 concepts

From concept 1 on poverty. Dependency on rainfed agriculture is a key driver of the high prevalence of rural poverty in agricultural systems

From concept 2 on water poverty. Poverty is exacerbated by low water access and poorly developed water infrastructure in agricultural systems

From concept 3 on vulnerability: Loss of crop and livestock productivity which occurs as a result of high exposure and low capacity to adapt to water stress under changing conditions of water access.

From concept 4 on water-related hazards: Loss of livelihoods which occurs as a result of high exposure and low ability to cope with impacts from water-related hazards

The Five types of Water / Poverty in the Andes

Location, Magnitude, Causes.

The Andean region is a very diverse landscape. There are regions on which lack of water is a key characteristic such as in the western side of Peru and North Chile (Atacama Desert), on which there is virtually no rain. However, people who rely on agriculture for their livelihoods is located near main rivers from which some irrigation schemes are in place. How water is managed inside these is another issue. Absolute water scarcity is not the cause of poverty rather the way in which the one available is distributed and managed. Migrating population near those green oasis becomes marginalized since their access is restricted or limited to the existing conditions. Similar situation happens in many other locations in the Andes on which topography, infrastructure and organizations restrict access to water for easy use in farming. Even in the most humid locations is easy to find limited access to water for productive purposes. Access and rights to use water is a growing concern.

Plains and valleys are frequently affected by floods. Rainfall patterns in the highlands are uncertain and information systems and early alerts are limited to prevent these. Many people is affected every year by floods near the Atlantic coasts of Colombia, the Pacific coast of Ecuador and wetland areas in the eastern Andes on which human settlements are rapidly appearing. Risk and vulnerability schemes do work in the region and local resources are easily responding to these chronic problem. Type 3 of water - poverty does exist but in some degree society is prepared to cope with this.

Type 4 form, low water productivity, is restricted to isolated and very specific areas. In particular to those places on which infrastructure, services, access to markets and technology is also limited. This is not a extensive form in the Andes but improvements are feasible since many producers, while relying on the abundance of water do not take care of it polluting and spoiling it.

Land and water degradation is a silence but growing problem. This is causing rural displacement, changes in land use, migration to cities and empowering population. Causes are linked to the history of colonization patters. Extensive cattle ranching in the hillsides and previous agricultural areas are an example. The fight for local power control associated with competition for the best land results in social violence. Areas near big plantations or remote areas devote to ilegal crops are some examples of this type of problem.

In the suburban settlements lack of sanitation and water provision is a common problem. Main capital cities are plenty of unplanned neighbourhoods where poor people from rural and secondaty cities live. This increases the demand of water from the sources and povision areas of the highlands.

How to cope with each particular situation is what Andes BFP is trying to contribute to. Using the three selected subbasins we will address some of these issues without loosing the whole picture analyzing the region as a whole.

Manifestation of WP types in the Niger Basin

S. Cook suggested that finding examples of the five classes would help in their communication and refinement (see previous posts). This post details a few of our findings (some formal, some informal and subjective) regarding the manifestation of the five water poverty categories in the Niger Basin, West Africa. We are particularly interested in alignment between our examples and those presented by the Volta team, given the immediate proximity of these rivers (see previous post)

Where people are deprived of water for basic needs of consumption or sanitation as a result of water scarcity. (Insufficient assets to compensate for physical scarcity)

There is little correlation between physical availability of bulk water and poverty evident at a landscape level of statistical analysis, as is generally understood. Weak evidence exists in North West Nigeria and East Burkina Faso. Availability of protected water (for example, piped sources) however is often associated with poverty, not surprising given its health implications. This relationship is quite widespread. Particular incidences we found include near the east Nigeria/north Cameroon border, and north west Nigeria. A 1% decrease in the number of people accessing water from unprotected surface or well sources is associated with a 0.1 – 0.15% decrease in child mortality rates.

Where people lack equitable access to water. (Political environment & institutions that lead to inequitable access)

There’s a large disparity in the average time taken to access water which may represent differences due inequality. In some regions average time to reach the primary water source is < st="on">Niger). Whether this represents inequity depends on its definition (either as a social constraint or a physical resource scarcity issue also). Also, such geographic differentiation does not well describe inequality due to different social strata in one area. It is difficult to capture this variety of water poverty in landscape statistical models (given the complexity of representing institutions) however can be assessed using case studies.

Where people are vulnerable to water-related hazards such as floods, droughts or disease. (Physical variability & lack of assets to buffer against natural variability)

The Sahel region encompasses a large area of the Niger Basin and is particularly prone to drought. Impact of this risk on poverty is assessed by Hyman et al (2008) (see good discussion of this by the Volta team’s posting below). Parts of Sahel have seen a 30% decrease in rainfall in past 40 years – may return or may represent a new drier reality. This has had a dramatic impact on landscape of north Burkina Faso for instance, where brush described as ‘too dense to walk through’ has now considerably thinned (as shown in photo below). The combination of drought, grazing and cropping pressure has led to soil degradation (due to wind erosion) in some parts of this region also.

Brush in north Burkina Faso

Where people suffer loss of livelihood as a consequence of change.

An emerging issue is conflict between water uses, which have until now been largely avoided simply due to minimal development of the Niger’s water resources. Large dams provide water to those in a particular area who have legal rights to its provision. However, they potentially deprive downstream users. A proposed large dam in Guinea may reduce water flows across the Niger Delta, a large wetland that supports over 1 million people by the natural irrigation of rice. A decrease in flooding extent directly reduces the number of people who can be supported by this agro-ecosystem.

Niger Delta flood plain - a reduction in extent due to upstream water diversions will likely entail a loss of livelihood.

Using retrospective analysis to estimate the poverty alleviation impacts of change

I just re-read an interesting report from Sahal Darghouth and others of the World Bank about the poverty alleviation impacts of its agricultural water projects. Although it focuses strongly on blue agricultural water projects, the report makes an interesting observation – probably true of all similar projects – that ‘despite the likely poverty reduction impact….there was rarely a sense of agricultural water projects as part of a coherent poverty reduction strategy ’ (my italics).

A cause, the reports states, is the lack of clarity about agricultural water’s role in poverty reduction. The report proceeds to call for better analysis to understand the more complex (distributional) effects that can arise as water distribution and productivity changes.

My question is this: Are we making the opposite mistake? In addition to understanding the water-related causes of poverty, should we also be thinking about methods to predict the effects of ‘improvements’ (whatever we mean by this) in water distribution and productivity?

I do not want to distract teams from the excellent work they are doing of analyzing water-related poverty in basins. But I am wondering how to make the forward link with, for example, analysis that shows the impact of changes in water productivity. Who has ideas on a conceptual model for this?