2 METHODOLOGY AND FINDINGS OF STUDIES TO ANALYSE POVERTY IN FISHING COMMUNITIES

2.1 SUMMARY OF REFERENCES

2.2 MAIN FINDINGS

2.2.1 What is poverty, why measure it and how has it been done?

A limited number of studies were found in published literature, which measure poverty in fishery communities; all are recent with most written since 2000 (see bibliography for this, and other, sections). There is however a more substantial volume of literature at a more general level dealing with poverty, how it has been measured in the rural development context, and discussions of the validity of different approaches. This body of work provides information that can certainly be useful in informing the measurement of poverty in fishing communities.

Until recently in fisheries, it has been a common assumption that i) small-scale fishers are poor (Smith, 1979 and World Bank, 1982 in Béné et al. 2000) and that ii) development initiatives in small-scale fisheries set in place by governments, donors and NGOs would implicitly contribute to the reduction of this poverty. The issue of poverty and the reduction of poverty in many cases has only recently become an explicit objective (pers comm., Gillet, 2002).

Why is poverty important? The interest of the international development community in poverty was renewed after the publication of the 1990 World Development Report, stimulating the international prioritisation of poverty reduction (Moser, 1998). This commitment is evident in the agreement of the International Development Targets, which arose from UN conferences in the early 1990s. An understanding of poverty is required at two levels. Firstly to be able to measure progress towards these ambitious targets, poverty needs to be measured. This has lead to a considerable evolution in the professional understanding of poverty, what it is, who is affected and how it can be most appropriately measured. Secondly, to be able to help the poor break the poverty cycle, one needs to understand the poverty cycle, who it affects, why and what opportunities can be used to plan interventions to improve their conditions (Pittaluga, et al., 2001 conf). The FAO has begun to consider poverty as a result of indications that although complex, there are links between poverty and the degradation of natural resources (Bellamy, 1995; Reardon and Vosti, 1995; Willmann 2001 conf.) and that to achieve their mandate of sustainable development, both natural resource and poverty issues need to be addressed.

Poverty has been measured using a range of indicators, and detailed descriptions and critiques of their relevance are well debated in the literature (Thorpe, 2001 conf.; Chambers, 1989). Figure 1 attempts to illustrate the evolution of poverty indicators showing the shift from simplified statistical/economic indicators based on nutritional inputs, income and consumption within the household, through an approach looking at basic needs requirements ("the deprivation of material requirements for minimally acceptable fulfilment of human needs, including food" (Cox et al., 1998)), to more recent attempts which try to embrace the multi-dimensional aspects of poverty.

Figure 1. Different indicators of poverty showing their evolution to embrace a more holistic understanding of poverty (based on, Thorpe, 2001 conf; Cox et. al., 1998)

The relevance of linear poverty indicators is questioned by many of the articles we have reviewed. It is argued that econometric terms are too narrow, and oversimplify poverty (Chambers, 1989; Cox et al., 1998; Bebbington, 1999; World Bank, 1999). It does not, for example allow elements of well-being or ill-being out of the cash economy to be included. (Carnagie and Goldman, 2001). The concept of poverty has wide implications, which must be addressed in order to understand links between rural poverty and livelihood sustainability (Bebbington, 1999). The use of participatory approaches in development with target communities is resulting in a change in the way professionals define poverty. It is the subject of extensive discussion that cannot be developed here. Some examples of poverty definitions based on, or influenced by, perceptions of the poor are presented in Box 1. This illustrates that when the poor are asked to define poverty, many of their criteria are based on assets, rather than income, some of which are intangible and at best difficult to quantify. Naryan et al. (2000) report five key elements that contribute to the concept of poverty according to the poor. These are:

1. Poverty is multidimensional and complex.
2. It is a lack of the assets needed for well being.
3. It includes psychological aspects.
4. It is an absence of basic infrastructure.
5. It focuses on assets and vulnerability to risk, not income.

Box 1

How poverty is defined, affects what should be measured, and how, as well as the ability to identify poverty reduction measures (Gillet, pers com., 2002). This has brought about a realization that new approaches are needed to cope with this more sophisticated understanding, in a move to narrow the gap between previous professional definitions and local perceptions. The main ideas and themes relevant to measuring poverty from a multidimensional perspective are:

1. Quantitative, qualitative or both? The subjective vs. the objective approaches to poverty are considered to be polarised (Moser, 1998, Bebbington, 1999). What is realistic? Quantitative indicators give an insight as to the scale of the problem; the qualitative indictors are more adept at addressing why people are poor. Addressing multi dimensionality is time consuming, and imposes practical restrictions. The appropriateness of how and what to measure depends on what the information is needed for, by whom, at what scale and the resources available (Thorpe, conf 2001). A study looking at results of qualitative (participatory poverty assessments) and quantitative poverty assessments (Household Surveys) in Uganda showed that these methods are neither contradictory nor complementary as they stand (McGee, 2000). The use of SL approaches allows space for creating opportunity for linkages in future work.

2. The use of participatory approaches is vital in allowing the inclusion of qualitative indicators, and considers the range of assets people have, and their ability to access these assets as a basis of analysis.

3. Triangulation is an essential process in the collection of qualitative information.

4. The application of an asset based framework (Reardon and Vosti, 1995), Moser (1998) describe frameworks slightly different from the SL framework supported by DFID. Recognising that people have a range of different assets that they manage to create their livelihood strategy is an important step in recognising the complexities and dynamics of poverty (Moser, 1998)

5. Bebbington (1999) emphasises the importance of access to assets as a means for the poor to build sustainable and poverty alleviating livelihoods. In addition, it is the proportion of different types of assets, not just the amount of assets an individual or household has access to, that affects poverty.

6. Assets are perceived by the poor as making visible one's poverty status. Needing to ask for help is a social sign of poverty, not being able to read is another. Similarly the lack of access to physical or natural assets is a sign of poverty (Cox et al., 1998).

7. The non-poor are not necessarily the least vulnerable (Moser, 1998). Those who are economically unstable can easily fall below the subsistence border line if conditions deteriorate (Jazairy et al., 1992 in Cox, Farrington and Gilling, 1998).

8. Describing communities as poor and non poor is an over simplification. Different types of poverty exist both within and between communities (Béné et al., 2000; Gillet, pers comm., 2002). In studies which disaggregate by socio-professional groups or gender in fisheries, or other rural sectors, these groups are also heterogeneous, with poorer and less poor individuals or households undertaking similar activities. It is the differences between these people and how they devise a strategy to survive that is important. The SLA is a holistic, dynamic, people-centred, sustainable approach which lays the framework on which different methodologies are being developed to address these questions. These approaches are being developed across the development sector in an attempt to look at who is poor and how their needs can be addressed based on strengths and opportunities.

2.2.2 Applied methods used in rural development

Participatory Poverty Assessments (as used by the World Bank) use trained field teams to apply participatory approaches. The teams use groups, focus groups, key informants and other tools, which are triangulated to increase validity. Such methods aim to understand poverty, how people deal with poverty, trends and dimensions, and the impact of policy (UPPA, Uganda, 1999). In the case of Uganda, the PPAs were useful in identifying the types of policy that could influence the factors that will allow the poor to move out of poverty (McGee, 2000).

Well being analysis uses PRA techniques to help village participants categorise poverty levels based on their assets, hunger period, income, and the amount they spent for an annual festival. This allowed the original target group to be desegregated into 4 sub groups for resource allocation purposes and reduced exclusion in a rural micro credit project (Kar, 2001). An important note from this example is that the poorest tended to be risk averse and the most difficult group to assist.

Poverty Mapping allows the location of identified determinants of poverty. It can be used from a global to a micro level. A range of methodologies can feed into poverty maps depending on the objectives of the exercise, definition of poverty, limits on the data, analytical capacities and costs. The extent to which these maps measure the broader aspects of poverty depends on which methods are applied. Participatory approaches solicit self-generated definitions of poverty, the determinants of which can be uncovered using the livelihoods approach (Davies, forthcoming). Davies highlights the importance of considering bias and potential errors in the data as limitations of poverty mapping. The robustness of the final output (including statistical error) should be scrutinised, as the attractive and clear presentation of information in the maps can easily be misleading.

Wealth Ranking is used as part of a participatory or livelihoods analysis (Khanya www, Carnagie and Goldman, 2001; ODG, 2001). This is a method taken from the PRA repertoire to desegregate the given community according to wealth/poverty criteria, which are defined by the participants. It is a very quick categorisation, and does not lead to understanding the broader aspects of poverty.

Poverty Profiling can be executed at the national, waterbody, or community level. National level profiles have been undertaken in countries preparing PRSPs in line with World Bank and IMF requirements (www.worldbank.org/poverty). Poverty profiles use secondary data and participatory methods to understand who the poor are, why they are poor, how many poor there are, and where they are concentrated.

2.2.3 Examples of methodologies and findings of studies in small-scale fishing communities

For each reference below, a short summary of the methods used and the key findings are given. From the documentation available and conversations with experts in the field, it is evident that there has been a recent upsurge in the amount of work undertaken to measure poverty in fishing and aquatic resource dependent communities. Much of the documentation on such work is not yet available.

An analysis of poverty and aquatic resource use - focusing especially on the livelihoods of the poor in Cambodia. DIFD-SEA Aquatic Resource Management Programme.2000.

Methods: The programme used a mix of (i) quantitative data, HDI, income based poverty lines based on the cost of providing minimum calories and per capita consumption, national and regional level census data and other secondary literature. In particular the article describes the development of a country specific HDI based on economic and social indicators, a broader indicator than the standard HDI. (ii) Key informant interviews. The analysis is based primarily on economic data.

Findings: The method identified who the rural poor are, defined on the basis of the poverty line and indicating differences between regions. Large provincial differences were found in human and social indicators. The importance of small-scale inland fisheries for the rural poor is emphasised, especially the poor aquatic resource users including the subsistence aquatic resource collection activities of the landless. Fish were highly significant as a source of dietary protein in the diet of the rural poor. Key issues identified to orientate change were: access rights to fisheries (varied according to the scale of fishing); communication difficulties between the poor and supporting agencies; and insufficient technical knowledge to support subsistence fishers.

Poverty and aquatic resources in Vietnam: an assessment of the role and potential of aquatic resource management in poor people's livelihoods. DIFD-SEA Aquatic Resource Management Programme.2000.

Methods: Involved secondary data collection for 7 regions and a validation workshop. It aimed to define different dimensions of poverty, and areas of high poverty of aquatic resource dependent people, to be used to target livelihoods analyses. There was insufficient information to achieve the aims using the secondary information.

Findings: Information was available on aquatic resources and on poverty, but was not linked together. Information was particularly lacking on subsistence or "wild" fisheries that are of importance to the poor. Regional reports identified some areas based on official poverty assessments, but no information as to the causes of poverty or where there is a high dependence on aquatic resources.

Evaluating livelihood strategies and the role of inland fisheries in rural development and poverty alleviation: The case of the Yaéré Floodplain in North Cameroon (Béné, C., Mindjimba, K., Belal, E., Jolley, T., 2000.)

Methods: The study used participatory wealth ranking in 21 villages. The structure of the approach used semi-structured group interviews with key people, participatory mapping of the natural and physical resources, and construction of a seasonal calendar. The criteria for wealth was defined by the respondents

Findings: Between 2 and 3 groups were identified, depending on the village. Income was not a major criteria, and herd size was the most sited measure as it acts as a fund bank in times of difficulty.

Access rights and ownership of rights to the water bodies was not cited as a cause of differentiation of wealth (it is noted by anthropologists that the ethnic majority in the area, the Mousgoum have a rather egalitarian culture). The poorest groups have the highest dependence on the fishery and therefore the equitable access to the water body is critical for the poor in an area where agricultural land is scarce. The number of different activities in poorer households was found to be significantly higher than in richer households.

Livelihood and Poverty Baseline, Integrated Lake Management Project, Uganda. September 2001.

Methods: Questionnaires were designed on the SL approach, translated into the appropriate language and administered by trained facilitators. Every 5^th house was sampled on a walk through of the parish. A target was set to sample 20% of the parish. Data was analysed using Microsoft Excel. The second method was wealth ranking based on physical, human and financial assets (including qualitative indicators). How questionnaires are delivered was seen to be critical to the outcome.

Findings: A clear demarcation of wealth groups was difficult as the communities are composed of a continuous spectrum of livelihood statuses.

The average number of dependants increases with the wealth group (raising the question: if groups with high levels of dependants are targeted will more people be reached?)

The majority of the poorest wealth group were not born in the parish (i.e. are non resident), however helping such groups is sensitive.

Female-headed households were disproportionately represented in poorer wealth groups and the poorer wealth groups have less opportunity to diversify livelihoods.

Handbook for Livelihoods Analysis and PRA. STREAM - Support to Regional Aquatic Resource Management. 2002. NACA, DFID, FAO, VSO learning and communication initiative.

Methods: A SL framework was used as a basis for analysis. Within the framework, secondary data sources (national, regional statistics, reports and indicators), sample surveys and existing PRA tools are used to collect the information. This provides the opportunity for the poor to express their own perceptions, remain at the centre of the investigation and be involved. The SL framework removes sectoral constraints. The method aims to understand poor livelihoods of aquatic resource users and the role of aquatic resource management in their livelihoods.

Pilot usage of the livelihoods analysis has been undertaken in 7 Vietnamese villages, with aims for replication at a larger scale.

Expected findings: The main contribution to household income; constraints to accessing assets; different risk factors of social groups in the villages; and the ability and constraints on o exercising individual and group choice.

Poverty Profiles of Artisanal fishers: Methods Based on the SLA, examples of Benin, and Guatemala. (second case study in Guinea) (Corcoran, E, Working paper 2001; Pittaluga, F., Corcoran, E., Senahoun, J, in SFLP Workshop, 2001; Corcoran, E., Johnson, B., Senahoun, J., Theielun, C, working paper, 2002)

Methods: The methods for these poverty profiles are based on the SL framework, and use the SL approach. This enables the methodology to be flexible and adaptable to different objectives, depending on the objectives, time, country and the type of community (e.g. whether it is a small-scale fishing community set around an inland water body, or a maritime coastline). It can also be adapted to look at a national level profile, at the level of a particular water body, or at the community level.

The methodology has 4 main components, (i) a secondary data review (makes use of existing profiles, quantitative data, government or agency reports) (ii) participatory brainstorming sessions using multidisciplinary key resource people at the appropriate administrative level (iii) refining of livelihood groups with homogeneous focus groups of target beneficiaries identified in step ii, and (iv) validation of the findings in the field.

Existing participatory approaches and methods are used in components ii - iv. The role of facilitation in the brainstorming and in the focus groups is essential to gain the added benefit of group dynamics.

Findings: The main causes of vulnerability to poverty are identified by the participants; who the poor are and who are the most vulnerable to poverty; the location of the poor through mapping and using quantitative data supported by the participatory work; how many people are poor. This method is still being developed, but there are indications that it will be able to help understand the variation of poverty levels between people undertaking the same key livelihood activities.

Further work is needed on the quantification of the capital assets, the weighting of macro and micro variables on poverty and development in the understanding of poverty and vulnerability. The structure and presentation of the information needs further development to enable it to be informative to policy makers, and communities.

2.3 POLICY IMPLICATIONS

• Development of realistic participatory poverty measurement methodologies in the small-scale fisheries sector (although not exclusively) is a required pre-cursor to more effective policy formulation. Difficulties in defining poverty will affect the success of measurement and poverty reduction policies (Baulch, 1996 in Sporton, 1998). Baulch and McCulloch (1998) describe a high turnover of poverty in rural Pakistan suggesting that policies should be developed to try and increase the exit rate from poverty and at the same time reduce the entry rate.

• The level of policy development in inland fisheries is very low and often ignored (Béné et al, 2000), however, inland fisheries are important to the poor particularly in areas of inundation where agricultural land is scarce. This high dependence on inland fisheries means that policy change will have a proportionately larger effect on the poorest (Béné et al, 2000). It must be clear who these people are.

• Differences in long and short term poverty have big policy implications. Short term temporary or seasonal poverty might require safety nets or access to credit. Long term chronic poverty requires more fundamental policy adaptations in education for example (Sporton, 1998)

• Policy must be sensitive to gender issues in resource allocation and asset control and the reported feminisations of poverty (Sporton, 1998). It must recognise the economic and non economic roles of women in the household and the differing perceptions of poverty held by men and women.

• Without understanding the intra community heterogeneity of poverty differences there is a danger that poverty alleviation policies could have unintended negative impacts on some groups (Béné et al., 2000).

2.4 RESEARCH IMPLICATIONS

• Vulnerability is a separate issue to poverty (Chambers, 1989, Corcoran, 2001), and one that is linked to the asset-base of the household or individual. It refers to defencelessness, insecurity and exposure to risk, shocks and stresses. Many of these factors can be seen in the definitions of poverty. More research needs to go into identifying indicators for vulnerability. If poverty reduction is the priority of governments and aid agencies, then reducing the vulnerability of people to fall into poverty must be one important element of this task. As inferred in the Ugandan Lake Management Project, is it appropriate to assist the vulnerable rather than the poorest, in order to reach the largest number of people? (Also a development and policy issue).

• Development of a flexible methodology is required that will allow for site specific variations in the perceptions and relevant components of poverty as seen by those who are poor. Methods must make the best use of existing information and reflect the multi-dimensional, site specific and dynamic nature of poverty that is now being conceptualised as a result of working in partnership with target beneficiaries.

• Better understanding is required about the type of poverty prevalent in small-scale fisheries communities? What are the proportions of people in transitional and chronic poverty in fisheries? These differences also need to be distinguished in poverty measurement. How are different members of the community and the household affected? Where is the balance between having enough information and making timely decisions?

• More research is necessary to be able to ascertain the most significant contributing factors to poverty in the fisheries communities, and what type of intervention (not necessarily in the fisheries sector) is necessary to alleviate this poverty.