The 4th International Workshop on the Evaluation of Social Development, Oxford, UK. 3-7 April 2000.


Does empowerment start at home?

And if so how will we recognise it?


By Rick Davies[1], revised Tuesday, 18 April 2000

1. Where to look?


Oakley (1999) has provided a helpful overview of common approaches to empowerment in development, and the key issues involved in monitoring and evaluating empowerment. In the first section of this overview, which focuses upon the concept of empowerment, James (1999) refers to a growing scepticism about the use of the concept:


"Notions of sharing power, of stakeholders, of participation and representation and so on seem to refer increasingly to the self contained the world of projects themselves: the external structures of land holding and subsistence economy which have perhaps been disrupted, of political and military formations which have shaped and still shape the forms of social life in a region, tend to fade from view in the world of development speak."


Contrary to James I would like to argue in favour of such a focus, and even argue that perhaps as a trend it has not gone far enough. The reason for taking this line is however a concern about a similar type of myopia that James is concerned about. Organisations can easily become caught up in (and carried away by) their own rhetoric, regardless of which end of the political spectrum they place themselves. A more self-critical examination of organisational behaviour may be useful. Maybe we need to learn to crawl before we can learn to walk, and fly (e.g. address wider political and military formations).


I would suggest that the most appropriate way to assess achievements with empowerment is to start at home, with the organisation that wants to empower others. Referred to hereafter as an NGO, for convenience sake. We should start by examining an NGOs' immediate relationships, especially with the people it is trying to empower. We should also look at relationships between staff within that NGO and that NGOs' relationships with its own donors and supporters.


There are three reasons for taking this position. Firstly, economy of effort. All of this information should be close at hand, and thus much easier to access compared to information about their client's relationship with other individuals and organisations. Secondly, empowerment in this relationship should be the easiest to achieve. An NGO espousing empowerment should be able to act in those terms at least within its own span of management control. We should expect results here before anywhere else. Thirdly, the experience of empowerment in this relationship should provide a useful model of what can be done, effecting clients' expectations of other organisations. Ideally this experience should help them to be empowered in relationships with other significant people and institutions in their lives. On the other hand, if empowerment cannot be achieved here then what sort of confidence can clients have in their relationships with much less altruistic organisations?


We can take the same argument one step back, to the relationships between senior and junior staff within the assisting NGOs. We could make a second assumption that if field staff within an NGO are empowered in relationship to their superiors, then we might expect that that they in turn will be able to be empowering in their relationships with their immediate clients. A third and more provisional assumption can be made. That is, the state of an NGO's relationship with its donors and significant government authorities may reflect its capacity for empowerment of others. My caution here is due to the fact that NGOs are usually located in a network of relationships with other organisations, not in a simple dominating hierarchy.


These statements are assumptions, which should be treated as hypotheses to be tested, not obvious truths. They could form the basis of a useful research project. More immediately, they may stimulate a discussion in a meeting like this, which might suggest how useful such a research project would be. At best, their usefulness would become immediately apparent and people would go straight ahead and put these ideas to work in a practical way.


Proximate versus proxy indicators


A proxy indicator is simply a substitute indicator. What is being proposed above is something different, a proximate indicator. That is, evidence which is more immediately at hand - because it is closer to us in the chain of causation between what we do and what we hope will ultimately happen in the lives of people we are trying to empower.


In addition to the practical arguments put forward above, a proximate indicator has an additional "theoretical" advantage. In order to identify a proximate indicator, as above, we have to pay some attention to theorising how our intervention will have an effect. If we find our proximate indicator shows no evidence of empowerment, then there are two possibilities. One is that we failed, and did not manage to empower anyone. The other is that our theory of how empowerment comes about was incorrect. If we change the theory of expected cause-and-effect and come up with other proximate indicators, then we might still find some evidence of empowerment. 


2.  What to look for?


In this section I will outline three types of indicators of empowerment, where evidence should be close at hand. They are all qualitative indicators, and they are all based on the concept of difference. The base assumption is that for a person to be able to say or do something differently involves some degree of choice. Having more choice, compared to the past, implies empowerment. This view can be related to Jo Rowlands (1997) concept of "power to", as mentioned in Oakley's review.


Three types of difference will be considered:

·        Differences of opinions between individuals

·        Differences between the activities undertaken by individuals

·        Differences in organisational structures


2.1 Differences of opinion. 


In a healthy relationship between two individuals we might expect that both are able to freely express their opinions, including their differences, without putting the future of that relationship at risk. Similarly, if clients are empowered in their relationship with an NGO then we might expect them to be able to freely and openly express their differences of opinion with that NGO. The larger the NGO and the number of clients it is working with, the more differences of opinion we might expect to find, if those clients are empowered in relationship to that NGO.


In the mid 1990's I was working with a large Bangladeshi NGO which had offices in many different parts of the country. Each of these offices was working through a number of large structures known as "people's organisations". There were more than 70 such organisations spread across the country. Given the number and scale of the organisations involved it seemed inevitable that many kinds of differences of opinion would exist between that NGO and the various people's organisations they were working with. However, when I asked some individual senior staff  about the existence of such differences I was told, rather surprisingly, that there were none. The same opinion was repeated again more strenuously in a large meeting between this NGO and its donors. If this reported lack of difference was really true, then the implications were very serious. Either the organisations concerned were either failing to communicate their members views, or worse, they were not even recognising their members' views. If the difference of opinion actually existed, but were not being reported by staff, this would also be dis-empowering of the people's organisations and their members. At the very least it would discourage any macro-level differentiation of assistance being provided to those people's organisations. This argument could be tested.


This unwillingness to disclose difference could of course be seen as primarily a reflection of that staff member's perception of his relationship with me as an outside consultant hired by one of their major donors. At the field level there may well have been many areas of disagreement which most of the staff were all too well aware of. However, if this was the case, then it suggests even more serious problems. In what should have been one of the most supportive relationships the organisation has with various organisations in its institutional environment these staff members still felt defensive and were unable to assert the truth as they knew it, without fear of any repercussions. If senior staff were not empowered to do so, then to what extent should we be expecting them to be empowering their own field workers?


In other more recent settings, in the Cameroon in 1999, I have found field staff of a development project ready and able to talk about the most significant differences of opinion between their team and the communities they were working with (as represented through various committees). I have also found they were willing to talk about the most significant disagreements within their team. The senior management of this project has subsequently taken steps to institutionalise the reporting of such differences in the team's six monthly reports, along with a number of other changes to the project's monitoring system. I have since learned that one of the four field staff teams were now unwilling to report differences of opinion within their own team. In my judgement this team was one of the less effective of the four area team's of field staff. The other area teams were willing to report on internal differences and appeared to be more capable in their work. The actions of the non-disclosing team suggested vulnerability rather than empowerment.


2.2 Differences in behaviour


At the population level, diversity of behaviour can be seen as a gross indicator of agency (of the ability to make choices), relative to homogenous behaviour by the same set of people. Diversity of behaviour suggests there is a range of possibilities which individuals can pursue[2]. At the other extreme is standardisation of behaviour, which we often associate with limited choice. The most notable example being perhaps that of an army. An army is a highly organised structure where individuality is not encouraged, and where standardised and predictable behaviour is very important.


Like the term "NGO" or "non-profit", diversity is defined by something that it is not -  a condition where there is no common constraint, which would otherwise lead to a homogeneity of response.


Homogeneity of behaviour may arise from various sources of constraint. A flood may force all farmers in a large area to move their animals to the high ground. Everybody's responses are the same, when compared to what they would be doing on normal day. At a certain time of the year all farmers may be planting the same crop. Here homogeneity of practice may reflect common constraints arising from a combination of sources: the nature of the physical environment, and the nature of particular local economies.


Constraints on diversity can also arise within the assisting organisation. Credit programs can impose rules on loan use, specific repayment schedules and loan terms, as well as limiting when access to credit is available, or how quickly approval will be give.


Ideally, the diversity of economic activities undertaken by an NGO's clients, with the assistance of NGO credit, will be greater than that which existed before or without the NGO credit. We could then theorise as to which of the NGO imposed constraints make the biggest difference to the diversity of activities undertaken with that credit. Timing of credit availability may, for example, make a bigger difference than the absolute amount of credit available for any one loan. The removal of that constraint would then become one proximate indicator of empowerment in that setting.


What we then need are ways of separating out the sources of constraint, at different scales of analysis. At the more macro level there are constraints arising the assisting organisation, versus the economic and physical environment. At the more micro-level there are the constraints arising from different requirements associated with agency services. One way is to look for differences in diversity across different settings. This issue will be returned to later in this paper. 


2.3 Differences in organisational structures.


As Fritz Wills (2000: 6) has pointed out, "the subjects of empowerment - the ultimate beneficiaries - can be both groups and individuals". One way of identifying the absence of empowerment in a set of organisations is to look for homogeneity in their structures and processes.  One interesting example are the peoples' organisations of the kind that have been promoted by the large NGOs in Bangladesh. If these organisations have evolved over time in response to local needs and pressures then we should expected a fairly high degree of diversity in their structure and process. Diversity should increase over time, unless there are some common dominating constraints. On the other hand, if those organisations were developed primarily to meet the needs of the NGO assisting them then we might expect more homogeneity in structure and process. Standardised structures are a more efficient and noiseless channel through which to pass commands and to receive information. My impression of one large NGO in 1995 was the peoples' organisations were remarkably homogeneous, despite being scattered across the country, and in some cases being quite old. They appeared as unpaid extensions of the NGO's own organisational structure.


One candidate proximate indicator of organisational diversity is the degree to which the control over financial resources is decentralised, or not. To what extent do they have control, if any over grant funds, and to what extent do they make final decisions over loan authorisations? In the past, the evidence for significant devolution of financial control has not been very impressive. In 1991 IIED sent a questionnaire out to more than 1,000 rural development organisations in 50 countries.  The main conclusion of their research was that "local community participation in problems assessment and analysis is fairly common...[but] substantially less so for the monitoring and evaluation phases and more notably, there is very limited complete financial control given to the local community in all four phases of the work" (Guijt, 1991).



3. How to monitor change?

3.1 Differences in opinions


In the Cameroon project mentioned above we also explored the dynamic nature of differences between people. In normal life some differences of opinion are resolved, others remain a continuing concern. The resolution of significant differences of opinion in relationships between an organisation and its clients can be seen as an indication of empowerment of both parties, if both parties agree that the result is a successful resolution. Conversely, the persistence over time of unresolved differences can be seen as an indication of ineffectiveness, or lack of empowerment. In the Cameroon project one of the area teams used a simple table format (below) to document two types of resolved and unresolved differences of opinion. One was of opinions within the membership of a particular stakeholder group. The other was of differences in that group's relationship with others.


This table was constructed by the members of a Natural Resources Management Committee. This is an apex group representing a wide range of different stakeholders interested in the fate of a nearby forest area called Bimbia Bonadikombo, in southern Cameroon. They were assisted in this exercise by field staff who had already done a similar kind of analysis in respect to the functioning of their own team.




Most Significant Disagreement in the last 6 months






Internal (within





Definition of who is a stakeholder: This affects who is represented in the NRMC, and thus how their interests are addressed.

When to conduct a farm survey in northern Bimbia Bonadikombo:

If neglected, people in that area may feel marginalised




(with others)



Location of Inner Core boundary of the forest reserve. (northern part):

(Consequences not detailed)

Low incentives paid to Management Committee members during field work: 

(Consequences not detailed)



Each cell was expected to contain information - in the form of a "difference that makes a difference" The difference of opinion that was identified as most significant is in bold. The consequences of that difference are given in italics. It is expected that this information will be collected on a six-monthly basis. At the end of a six-monthly reporting period three different forms of change are possible:


·        The worst possibility will be that the contents of all the cells are the same. Unresolved issues remain unresolved. In the face of this problem, no other important issues have emerged and been resolved.


·        A better result will be that the contents of half the cells remain the same and half have changed. Either: (a) The unresolved issues remain the same, but other important issues have arisen and been resolved. These will be visible in the first column. Or: (b) The most important unresolved issues from the previous 6 months have been resolved, but other important issues that have arisen have not yet been resolved. In this case the contents of the second column would have moved to the first column, and there would be new contents in the second column.


·        In the best case the contents of all cells will have changed completely. Previously unresolved issues have been resolved and new more important issues have arisen and been resolved.


These three grades of evidence could be made more substantial by providing more cases. Additional rows representing differences between other types of stakeholders could be included. The relative significance of resolved and unresolved differences in any row could in turn be weighted, by ranking rows such that the most important relationships were at the top, and the least important at the bottom of the table.


3.2 Diversity of behaviour and organisational structures


Some changes in diversity are self-evident. In a savings and credit groups new economic activities may be undertaken for the first time, and easily be reported on by the members themselves, or others. The range of activities undertaken by one group compared to another may also be easy to identify in many circumstances. This aspect of diversity has been described as "variety", by Stirling (1998), in his analysis of the significance of diversity within economics.


A more sophisticated assessment might look at the degree of concentration, in terms of numbers of people involved, or amount of money invested, in one category of activity versus others. Stirling has described this aspect of diversity as "balance". His view is that for a particular system of given variety, the more equal are the fractions (e.g. of money, people), the more even is the balance, the greater is the diversity[3].


Stirling also proposes a third aspect, called "disparity". "This relates to the nature and degree to which the categories themselves are different from each other. It is notions of disparity which determine when a particular type of option is recognised as falling into one category and when it is judged to be two..." (Stirling, 1998: 40). Disparity can be seen in the proximity of different branches in a tree structure representing the evolution of species. In the diagram below (a) and (b) could be people and chimpanzees, whereas (c) could be insects. Human genealogies can show this type of difference (disparity) in the same way.




In the case of human activities Stirling makes the important comment that "…such disparity is an intrinsically qualitative, subjective and context-dependent aspect of diversity. Notions of disparity will vary, depending on the particular frame of reference which is adopted for any given purpose" (Stirling, 1998: 40). Disparity seems much more difficult to measure, or even observe, than variety or balance.


However, in an earlier unpublished paper I have spelled out a "tree mapping" method which can be used to elicit peoples' classification of other people, places or events, (Davies, 1997). This method leads to a nested categorisation structure, of the same kind as shown above. 


Furthermore, the subjectivity of this method is not necessarily a problem, if it is applied in an appropriate context. At this stage it is useful to return back to the idea of proximate indicators, discussed in section 1 of this paper. Diversity may exist to some degree in reality, but what matters also is the extent to which that diversity is recognised or not, especially by key actors within the assisting NGOs. If field staff, or middle managers, do not recognise important differences between their clients (in their behaviour or opinions, for example), they cannot respond differentially. Their denial of difference is effectively dis-empowering.


On the other hand, the ability to differentiate clients' opinions and behaviours to a great degree of detail (i.e. using many categories and sub-categories) holds out the potential to be empowering. Clients different needs and opinions may be responded to in an appropriate manner. If differences are recognised in great detail, then we can go on to check less proximate indicators of empowerment.


The tree mapping method was designed to elicit informal and almost tacit knowledge. We can also go on to explore whether NGO staff recognise client differences in other ways. We can look at an organisation's formal representations of its world, including its client population. We can look at progress reports and annual reports and examine the type of differences that are prioritised in the form of the report structure: its headings and sub-headings, and even the structuring of paragraphs within those sub-sections. To what extent are major differences between clients the basis of this structure, or are they even visible at all? My own experience with this sort of analysis over the past few years is that clients differences are rarely the basis of report structures, and instead the focus is on differences between the various activities the NGO is implementing[4]. This experience reinforces my view that James' concern about organisational myopia on power issues was not misplaced, but rather that he was seriously underestimating the scale of the problem. The lack of recognition of differences between clients suggests that many organisations have serious difficulty with the most basic of empowerment tasks, that is simply listening to people.


4. Wider issues: The limits to empowerment


There are clearly limits to the scale on which empowerment can take place. Some differences of opinion between individuals cannot be expressed without undermining the future of their relationship. Some differences in needs between clients cannot be responded to by an NGO because those responses are seen by itself, or its donors, as being outside the NGO's ambit, its raison d'être. At the level of larger groups and societies the extension of some forms of choice (e.g. gun ownership) may diminish other forms of choice, and even threaten a larger scale collapse of choice (e.g. via civil war). In the biosphere, the diversity of ecosystems can be degraded, rather than extended, through the introduction of new species (e.g. cats in the Australian bush). The relationship between individual and collective empowerment is obviously a complex one.


When we are talking about empowerment we are, by definition, not able to specify particular desirable outcomes at the level of individuals. Such an approach contradicts the very notion of choice. But we can define desirable outcomes in their collective form, at a population level. In the paragraph above, the implied ideal is a sustainable expansion of diversity. Such a view can even be applied as an ideal beyond the human sphere, to the wider biosphere. Doing so is a salutary exercise, to say the least, since homo sapiens is believed to be responsible for the most recent "great extinction" of species (Wilson, 1992).












Proximate indicators may be helpful in advocacy work, where the chain of causation is often long and intermeshed with the influence of many other actors. For example, NGOs could start by looking at those who have been the immediate recipients of their advocacy communications and ask what types of changes would they would expect to find in that persons knowledge and attitudes if advocacy messages were beginning to have an effect. Do the people concerned know more than they have been originally told? In what areas have they developed more knowledge and what does that signify?




Davies, R.J. (1997) Tree Maps: A Tool for Structuring, Exploring and Summarising Qualitative Information. Unpublished paper.


Guijt, I. (1991) Perspectives on Participation: Views From Africa, An Inventory of Rural Development Institutions and Their Uses of Participatory Methods. London. IIED.


James, W. (1999) "Empowering Ambiguities" in Cheater, A. (1999) The Anthropology of Power: Empowerment and Disempowerment in Changing Societies. London. Routledge.


Oakley, P. (1999) The Monitoring and Evaluation of Empowerment. Resource Document for the 4th International Workshop on the Evaluation of Social Development, Oxford, UK. 3rd to 7th April, 2000.


Rowlands, J. (1997) Questioning Empowerment. Oxford. Oxfam.


Stirling, A. (1998) On the Economics and Analysis of Diversity. Science Policy Research Unit. Electronic Working Papers Series. No. 28. Univ. of Sussex. Brighton.


Wills, F. (2000) Empowerment and its Evaluation: A Framework for Analysis and Application. Prepared for the 4th International Workshop on the Evaluation of Social Development, Oxford, UK. 3rd to 7th April, 2000


Wilson, E.O. The Diversity of Life. Allen Lane. The Penguin Press. London.



[1] Cambridge, UK


[2] As noted by some workshop participants, diversity in the behaviour of a set of individuals does not necessarily mean that all have equal choice. Inequalities of power (defined as choice) may still exist. Where we do find diversity in the set as a whole we could then do a more-micro-level analysis and examine the amount of diversity in the behaviour of one individual compared to another.

[3] However, while ethically preferable in the case of human incomes or assets, it seems questionable whether a homogenous distribution should be seen as being more indicative of diversity.

[4] I have seen annual reports describing development activities undertaken by tribal people in three adjacent states in India described as though they were all one single group, although the economic and political constraints they each faced were quite different from state to state.