4. What are the Implications and Lessons Learned?

In the previous sections, the opportunity was taken to report on how the cases incorporated sustainable livelihoods principles (or not) and identify the impact these projects had on the rural poor. This section will provide a stock taking on what we think we know and what we still do not know given the findings of this study.

4.1 What do we think we know?

· Do livelihoods approaches lead to poverty reduction?

While further study is necessary, our initial analysis would suggest that livelihoods approaches can contribute to real poverty reduction if applied effectively. As discussed, all cases under review contained elements of the SLA to greater or lesser extents yet the outcomes differed significantly. Different levels of success appeared to be based on either the number and type of SL principles applied or the quality of their application.

The more successful cases appeared to be those which had applied the greater number of SLA principles e.g. the three principles specific to the SLA and a mixture of other essential principles, in an effective manner. Projects that experienced lower levels of success were affected less by an absence of some SL principles per se, but rather weaknesses in project diagnosis, design, implementation and/or monitoring and evaluation. As Diana Carney points out, ‘the usefulness of the SL framework, as with any tool, is set by the user’ (2002: 16), and so the question remains whether it is the user or the tool at fault^[21]? Is it the approach itself we should be looking at, or its execution in the field? Answering this question requires analysis of a project that has been specifically designed and implemented using the SLA.

· Which principles seem to positively influence poverty reduction outcomes?

We think that projects must address a minimum number of principles to set the stage for reducing poverty. These include those that are specific to sustainable livelihoods and depending on the context perhaps, a mix of those that are not SL-specific, but considered important principles of good development.

In their papers entitled "Goodbye to Projects", the authors conclude that SL principles provide a useful cross-checking framework and that attention to all SL principles is required to create sustainable impact (Toner et al., 2004; Franks et al., 2004). In order for this conclusion to be helpful in a practical sense however, we must be able to understand the sequence in which these principles are introduced, and the mechanisms (organisational elements and institutional linkages) associated with them. We have made a preliminary attempt to do this through timelines and implementation maps, but there is a need for more in-depth analysis here.

Equally, while there must be a minimum number of principles put in place in order for a project to achieve real poverty reduction, it is the way in which these principles are executed and not the number of principles per se, that is the key determinant of success. In other words, the quality of project design, implementation and monitoring and evaluation remains a critical factor in shaping the effectiveness of project outcomes and impacts. The DELIVERI project, widely considered as a positive example of the SLA in practice, did not address each and every SL principle^[22], illustrating that success can be achieved by addressing a minimum combination of principles.

The DELIVERI case also benefited from timely policy changes, amongst other factors, that put the project in a good position to succeed, suggesting that the immense value of having the right policies, practices or people in place to succeed, as reported in the Honduras and Nepal cases, cannot be discounted. The character, attitude, vision, and will of agents of change, project coordinators, government agents, development practitioners, elected or informal leaders associated with the project can move it to success or assist in its breakdown.

· What do we think we know about monitoring and evaluation in assessing the impact of SL approaches?

In general, the outcomes of M&E and PM&E strategies within the cases yielded mixed results. The constraints tended to cluster around issues of human resources and capacity (overworked or untrained facilitators) as well as communications and transportation issues. Tackling issues of social relevance continue to be difficult to grasp compared to data on more technical matters. In many cases, the impact on yield was recorded, but data to show the relationship with income, nutrition, household food security, or health was not. On a more positive note, where locally derived indicators served dual purposes (e.g. serving a wider use in the community as a guide for decision making as well as providing required indicators to serve outsiders) as was the case in Honduras, there is reason to believe that internalizing monitoring and evaluation and lesson learning can be achieved.

The widespread lack of good M&E data raises the question of whether this gap could be a result of inherent weaknesses of the SLA itself, of an incomplete application of the framework, or just the difficulties of carrying out effective (P)M&E. With an incomplete knowledge of how M&E processes were undertaken in each of the twelve cases, we cannot be certain of the reason. Further analysis towards answering this question might include an examination of the proportion of project budgets that have been allocated towards M&E in cases that managed to produce apparently high quality M&E data (Myanmar, Ethiopia, Honduras), and those that did not (Zambia) in order to understand whether poor M&E results from a lack of human and financial resources (and if so, how much should we be allocating in order to achieve good M&E), a lack of time, lack of good methods, or a lack of commitment?

What do we think we know about reaching the poorest of the poor?

We are still not getting there. Most cases failed to address the needs of the most poor. It seems that the livelihoods approach, while a means to diagnose who the poor are, does not necessarily build in the ability to work with the most poor. In Zambia for example, the characteristics of nutritional vulnerability and food insecurity had been identified during the preparatory phase of the project, and yet it was slow in addressing the direct nutritional and health needs household most at risk of food insecurity.

The main explanation for this would appear to be that many development interventions, by their very nature, exclude the poorest/most vulnerable groups, who do not possess the necessary physical, natural and financial asset base (i.e. land, labour, livestock, seeds, farming implements, agricultural inputs, capital) to participate in production-related and income-generating activities. In order to have positive impacts on the poorest and most vulnerable individuals/households/groups, it is necessary to ‘recapitalise’ those that have no disposable income or assets to invest in enterprise development. In other words, initial investments must be made to provide the very poor with a minimum asset base if they are to be brought back into the development process, investments that can be recovered, at least in part, after an appropriate grace period.

This was done successfully in Myanmar, where 251 destitute individuals were targeted through the distribution of small-scale livestock (sows, does, ewes and scavenging poultry), an intervention that proved to be an effective way of generating a rapid change in income for low-risk landless persons/households, and where project inputs costs could be recovered relatively quickly.

· What do we think we know about building assets?

There were very few cases that addressed all five assets (human, social, financial, natural, physical), although all cases addressed human, social and financial assets. In terms of building human and social assets, we found that successful cases tend to be a result of combined technical and organizational skills. While building human, social and financial assets are important, vulnerability and long-term sustainability may not be addressed unless the natural resources upon which livelihoods depend are managed and sustained.

· What do we think we know about governance?

From the vantage point of enhancing good governance, the investment of time and resources seems to pay off. Many of the projects reviewed invested in various aspects of governance, but those that simultaneously worked with local and national government seemed to enhance ownership, build awareness, increase responsiveness, and influence policy effectively. In cases where central government was brought in during the planning stages, there was greater ownership and readiness for support of changes. When broad spectrum awareness raising and organizational and leadership skills are built at the local level (including representation of community groups in government), there is less chance that political shifts (mayoral elections) will paralyze implementation. This also puts government in the position of negotiating with donors for additional or extended projects. Norton and Foster (2001) looked at the potential for SLA to be used in the development of Poverty Reduction Strategy Papers (PRSP).

· What do we think we know about long-term sustainability?

Not enough is being done about it. There were few cases that put in place measures that would indicate long term sustainability, particularly financial, food, and environmental sustainability. Either little attention is being paid to this OR we do not yet know what to put in place or what to look for in terms of evidence. Although not included in this study, there are examples in which communities have built detailed visions of long term sustainability up front allowing for planning and decision making (relative to community activities or internal and external projects) to take into account whether the action will lead in this direction (Savory, 1999). Another question one could ask is, if the principle of long term sustainability is fully employed, do exit strategies become obsolete?

· What do we think we know about livelihoods approaches?

There were cases in which the livelihoods were strengthened by incorporating a natural resource component (making existing practices ecologically sustainable or building in incentives not to employ destructive practices). In the case of Yemen, existing livelihoods strategies were improved upon and new strategies were introduced through vocational training aligned with credit and support of community organizations. A clear case is also made in Honduras for building on existing livelihoods that can reduce and/or eliminate risk (increase or stabilize beans and corn yields) before addressing income generation through different livelihoods strategies or the facilitation of new enterprises.

· What do we think we know about partnerships?

Projects that were successful tended to have a mix of partners and links to the local, national and international level. Two important aspects of partnerships are a) the complementarity for advancing progress in a cost effective way with each partner working from their comparative advantage, and b) ownership by actors whose support is needed for success.

The projects under review tended to have a mix of local grass roots organizations, internationally affiliated organizations, and governments at the local level an in some cases at the national level. Two cases revealed partnerships with business and industry, although these relationships were abandoned at an early stage. A number of projects collaborated with other international or local organizations’ projects to advance their work.

· What do we think we know about multi-sectoral approaches and the engagement of ministries?

Multi-sectoral approaches are definitely being carried out. The cases that were most successful in addressing multiple sectors at once invested in building coherent multidisciplinary teams with government agency individuals and raising awareness with local groups about relationships between the sectoral aspects of the project (for example, water and nutrition). They tended to engage multiple ministries in the form of national steering committees or were attached to a level of bureaucracy that supersedes the ministries. Where central government was involved in the planning processes, it was easier to transcend from a single sector to additional sectors according to project needs (agriculture to education). A key element that must be ensured is keeping the higher levels of government informed/engaged. The engagement of different ministries and/or partners can allow for multi-sectoral approaches to be carried out in a collaborative way even if the project has a single focus or limited resources.

· What do we think we know about exit strategies?

While the SL approach lends itself to building community capacity so that "external" projects can exit at some point, there is a strong need to explicitly build in exit strategies to the design process. (see the paragraph on sustainability above).

· What do we think we know about empowerment?

There were a number of cases in which dramatic changes were made in women’s capacities and confidence. These included women taking on entrepreneurial endeavours (replacing middle-persons), becoming social mobilizers, forming or joining committees, and demanding services, among others. Other aspects of empowerment centred primarily around rights to access natural resources (Cambodia, Gambia), and the improved representation of community groups at district and national levels (Yemen). Empowerment was strongly linked to both social inclusivity, and the nature and quality of participation.

· What do we think we know about flexibility and embracing errors?

The iterative nature of projects is a plus for continued learning. Building in the flexibility to change directions or activities based on new knowledge can both enhance the learning of the project participants and project managers. In the case of the Gambia, the strategy has changed dramatically based on lessons from the earlier phases. In the case of Honduras, the iterative learning has empowered the local community and leadership. In the case of Nepal, there was some inflexibility among team members, possibly based in an over ambitious design, which created tension among local implementers.

While there is a theoretical consensus on the importance of flexibility and/or embracing errors, this is sometimes difficult to achieve in practice. There was a willingness to improve projects along the way, and participatory monitoring and evaluation provided a means for mid-course corrections, however, many of the projects had ineffective PM&E in place due to lack of human or financial resources to carry them out. This often kept projects from seeing areas of improvement until too late.

· What do we think we know about furthering the SL approach?

The SL approach is a logical evolutionary iteration in participatory development. Because purely SL cases are difficult to find, it seems that application of principles follows an "include and transcend" approach. Development approaches will continue to build on experience and lessons learned with or without a specific terminology.

4.2 What do we not know?

63. We have already mentioned some of the things we know less about in previous sections. Here we add a few more, principally at the "conceptual" level.

• We do not know how an additional set of cases would influence the findings in this study, but we are keen to find out. There are also additional cases that we wish to incorporate as well as a number of cases that are in early stages of implementation, or are in need of an evaluation which we will keep on a "need to watch" list.

• We do not know if a truly SL "start to finish" effort that has evidence of reducing poverty actually exists. It seems more likely that the cases we reviewed were intended to be participatory efforts with a livelihood perspective and brought in the principles that seemed appropriate.

• While we find that these cases showed evidence of reduced poverty and incorporated livelihoods principles, we are unable to say, unequivocally that the SL approach is better at reducing poverty than other approaches.

• While we can easily speculate, we still do not have a recipe for which mixture of principles lead to more effective poverty reduction. As mentioned earlier, the Bradford study (Franks et al, 2004) noted that all principles were needed for positive impact.

• We cannot say if SL approaches are more cost effective, however in the case of Nepal, the government reported that the approach used was effective given limited resources.

• We do not have enough evidence to draw a connection regarding scales of impact and principles applied. The cases covered a wide range of impacts, but there was not consistent data on numbers and types of beneficiaries.

• We do not have data to indicate if, having positively influenced the poor, the principles applied will keep the poor from falling back into poverty when exposed to new shocks.

• While we know that community organization and the sense of solidarity were key components of successful cases, we do not know which community organizations built upon existing or endogenous networks and which were created anew through project activities. In Honduras, effectively "new" organizations were built out of existing ones, which process catalysed some significant evolution in previously very unequal power structures. Along those same lines, we did not have a case that is not project driven, so we did not compare these cases with ones of endogenous community efforts and activities that led to poverty reduction without external influences. It would be useful to carry out such a study in the future.

• We do not know where they are now. Most projects were keen to have post project data collected but it is rare that resources are made available. In several cases, the capacity to take stock of continued progress by community members is present but it was not clear whether data collection would continue.

• We do not know how communities would respond to the common pattern outlined in the operational map (see Section 2.4).