Abstract

The discussion on participation and assisting poor clients in the international development literature about NGOs is reviewed. The central issues studied are: a.Do Kenyan small enterprise development programs work with the poorest clients where unemployment is highest or not? b. Do staff believe that clients participate in design and management of the programs?

This research found that most NGOs in the near census of the 31 small enterprise development programs in Kenya are not working with the poorer members of their communities where unemployment is highest. While most (79%) of these NGOs report that most or all of their clients are poor, most NGO programs have no operational definition of the poor or poverty. When shown specific strata of poverty defined by Remenyi (1991), only two of these 31 NGOs claim to have 75% or more of their clients in the lower poverty strata. A regression analysis of data from 13 of these 31 NGOs suggests that a poverty focus in the corporate purpose statement is correlated with higher levels of financial sustainability and efficiency.

All 31 NGOs were also asked to what extent beneficiaries participated in the design and management of their small enterprise assistance program. Not surprisingly, a majority of NGOs claimed to use design methods that were considered empowering, (but no confirmatory research was conducted with clients themselves). Almost one third (31%) of NGOs admitted that clients had no input at all in the design of the program. Only 47% claimed program features that were deemed empowering. Less than 30% reported any significant client participation in governance.

This paper reports the results of one of the first virtual census surveys of all the NGO programs operating in a specific sector in a specific country. Multiple personal interviews with chief executive officers, small enterprise program managers, and their staff over a four year period (1990-1994) studying 31 (out of 33) NGOs in Kenya with field operations directly assisting small and micro-enterprises defined as having 50 employees or less. NGO's studied had at least two years of operational experience in the sector, over 120 clients during at least one year of the period, and were continuing their small enterprise assistance program.

1. Introduction

While historically many non-governmental organizations (NGOs) have been organized principally for charitable works and poverty alleviation,, economic development policy makers have seen industry, business and government as the principal actors in broadscale poverty alleviation in the past. However, that paradigm shifted during the late 1980's. Writing for the World Bank, Paul & Israel summarize well the paradigm shift that occurred between the 1960's and 1980's as NGOs were increasingly perceived as a more effective alternative to government programs in serving the poor.

Increased reliance on the state to intervene directly to reduce poverty began in the 1960s, after the market and the "trickle-down" approach failed to affect significantly the depth of poverty in many developing countries. Observers were struck by the worsening of income distribution that often accompanied economic growth. Some argued that the slow progress of poverty reduction was due not only to "bureaucratic failure" but also to the market catering primarily to elites and export enclaves and ignoring the basic needs of the poor (Kothari 1984; Rose-Ackerman 1980; and Drabek 1987). An alternative development model supported by NGOs advocated "bottom-up" projects that involved the poor directly in income-generating social activities. The purpose was to enhance their income-earning capacity and to "empower" them to develop on their own. Poverty reduction requires resources, the capacity to adapt programs to diverse and uncertain conditions, and a strong performance orientation supported by the demand pull of the poor themselves. Neither developing country governments nor NGOs possess all these attributes (1991:1)

The World Bank (as other donors)saw the need for a new strategy where governments, donors, NGOs, and the poor themselves cooperate together to produce development. Paul and Israel go on to elaborate the paradigm shift where the "poor" are now seen as important actors in the design, implementation and evaluation of projects:

Donors and recipients have given too little attention to socio-cultural and political factors and have not been sufficiently aware of the important role that the poor themselves can play in initiatives designed to assist them. Evidence supports the view that involving the poor in the design, implementation, and evaluation of projects in a range of sectors would make aid more effective. (1991: 133)

In the Kenyan context, non-governmental organizations have been asked by the Government of Kenya (GOK) as early as 1986 to be a major player in the development of new micro-enterprises and small businesses in order to increase employment. The GOK estimated then that the country needed an additional 6.2 million new jobs by the year 2000 in order to provide employment for the millions of new people entering the labor market during this period.

This research undertaken in 1992-94 studied in part the provision of services claimed by the major NGOs in Kenya which had small enterprise development programs that were ongoing and were significant with 120 or more clients in two or more years during the four year period 1990-1993. Thirty one of 33 NGOs identified were interviewed during 1993-94.

This paper focuses on participation and poverty issues and reports part of the larger study which was conducted for the purpose of better understanding the impact of an NGO's corporate purpose (as represented by corporate purpose or mission statements)and the financial efficiency of the NGO as represented by a series of financial ratios.

2. Poverty And NGOs

At the micro-level of analysis, Chambers (1986) raised the notion that the local people themselves should define poverty because poverty tends to be culturally specific. Participatory Rural Appraisal (PRA) and other techniques were developed to induce local people to define poverty in local terms and to develop their own "needs" analysis from a local perspective.

With the paradigm shift described by Paul and Israel, there was a recognition that often governmental poverty alleviation programs were not targeting nor reaching the "poorest of the poor." Remenyi (1991) delineated "strata" of poverty which could categorize levels of poverty. These strata were: a) "near poor" with full-time jobs in the formal sector but who lived in urban slums with poorer people; b) "entrepreneurial poor" who owned businesses employing others but were located in slums and otherwise associated with the poor; c) "self-employed poor" who were owner/operators of their own micro-enterprises but without other paid employees; d)"laboring poor" working on a casual or daily basis without permanent employment; and e)"vulnerable poor" who had no employment and were often mostly women--- especially, widows--- and children with no additional support. This typology was used in the study to help NGOs to categorize their target clients according to poverty status.

An ongoing debate exists in the development literature (Boomgard et al., 1992; Malhotra, 1994; Mutua, 1994) as to whether small enterprise development programs can focus on poverty alleviation and become self-sustaining programs. This discussion is limited mainly to credit programs. Serving the poorest is seen by some as very expensive in program costs while others believe it is possible to become cost efficient and self-sustaining.

3. Client Participation

There is a widely held belief (Brown and Korten, 1990; Cernea, 1986; Uphoff,1985 and 1986) that NGOs are more participatory in the design, implementation, and management of development work than are the government or commercial sectors, and this characteristic is often cited as an NGO "comparative advantage in development work. Bryant and White define well the importance of participation to the development process:

Development as a process of increasing people's capacity to determine their future means that people need to be included in the process - they need to participate. Participation, or empowerment is part of the process and definition of development. Managing participation is more than including the public in one stage of the design process or evaluation of the project. Rather participation informs the meaning and integrity of the entire process.(1982:205)

Clark (1991) also addresses the importance of participation by the client and makes clear that participation means also the ability of beneficiaries to participate in the management of the NGO as well as the NGO's participation in the poor's struggle for justice.

Other NGO observers (Tendler, 1982; Elliot, 1987; Fowler, 1985, 1988) question the degree to which client participation is realized by NGOs in field operations suggesting the very real need to test the idea that NGOs use participatory methods. Tendler's 1982 study and Sheperd's later 1986 study of 12 PSOs in India (discussed by Fowler, 1988) show little evidence of participation by clients in NGO development work. Fowler is much less sanguine about NGOs and their ability to include clients in meaningful ways.

Against these generalized positive findings based on indirect comparisons and not statistically representative sources, there are similar reviews and evaluations which point to the failure of NGOs to live up to their own articles of faith.(1988:12)

Similarly, in specifically discussing small enterprise development, Tendler (1982), Boomgard (1989) and others question how participative small enterprise development (SED) projects can be suggesting that they are less participative than other development interventions. However, Grameen Bank in Bangladesh is often advanced as one of the more successful organizations in combining commercial development among the poor and social development strategies. Most observers cite participation of the poor in the design, management and governance of the organization as a principal reason for its success of Grameen (Mann et al., 1989; Fugelsang & Chandler, 1988; Yunus & Morgan, l996). While many of these writers cite various pieces of research, very few sectoral program studies have been done of NGOs within one country in order to get a good idea of the common practice across the NGOs within one country. This study focused on doing that in Kenya.

In terms of measuring participation, Bryant and White discuss the degrees of empowerment compared to co-optation especially well. They note that often participation is merely co-optation of clients "to reinforce the influence of managers" (1982:210) and does not involve true granting of power to clients. They cite a listing of strategies that is particularly useful in evaluating the degree to which participatory processes are "empowering" target groups:

One listing of strategies is provided by Mary Hollnsteiner on the basis of her work in the Philippines. Reflecting on opportunities for the urban poor, she lists six different means of participation: (1)representation by a "solid citizens" group, (2)appointment of local leaders to official positions, (3)allowing the community to select one of several plans, (4)consultation throughout the planning process, (5)representing the public on decision-making boards, (6) control by the community over funds and expenditures. Only the last three modes Hollnsteiner observes actually constitute participation; the others are forms of co-optation in that they allow elites or bureaucrats to dominate the process. (1982:211)

From Hollnsteiner's list, a scale of "participatory measures" was developed based on the degree of participation and power granted to clients in the design of a development project (see endnote 1 for details).

Oakley (1991) uses a five point scale participation ranging from: 1.nothing/narrow; 2. restricted/ small; 3. Mean/fair; 4. Open/good; and 5. Wide/excellent. His matrix describes various conditions and/or behaviours for each indicator under each measure to give an analytical framework for evaluating participation. This scale was adapted to this study.

Four areas of participation were examined among the 31 NGOs studied:

(1) decision to enter the area (needs assessment);

(2) project or program design;

(3) project/program management (who makes the majority of management decisions); and

(4) governance (leadership and ownership) meaning representation or outright ownership of the organisation by the clients.

This part of the research was interested in five general questions about Kenyan NGOs working in SED as an "organizational class":

A. Do Kenyan NGOs' SED programs use participatory processes involving local people to choose the areas into which the NGO will start new work?

B. Do Kenyan NGOs use participatory processes with local people to design initial SED projects?

C. Do Kenyan NGOs use participatory project management processes involving local clients in the management of SED projects?

D. Do Kenyan NGOs with SED projects involve local project clients in the overall governance of the NGO?

E. Are Kenyan NGOs that focus on poverty alleviation in their purpose statements associated with higher levels of financial efficiency and sustainability and does client participation have any correlation with higher or lower costs?

4.0 Findings

4.1 Target Groups

A large majority (84 per cent) of the agencies interviewed claimed to have written descriptions of their target groups as shown in Table 1 below. This question was asked to ascertain agency efforts to clearly define target groups in anticipation of later questions about the "poor" as target groups.

Table 1: Description Of Target Groups

Do You Have A Written Description Of Your Target Groups For Sed?	Total Responses
Yes	26	84%
No	5	16%
Total Responses	31	100%
N=	31	100%

(Source: Kenyan Fieldwork Data)

An open ended probe question was then asked "Who are your target groups?" All respondents replied immediately without consulting plans or written statements suggesting that they knew very well who their target groups were even if their actual targets were different from their written target groups. Multiple responses were received from many NGOs. "Existing businesses in a poor area" and "women" received the highest number of responses (13 percent each of 31 NGOs). When the different replies about existing businesses are aggregated (single asterisked items in Table 2), existing businesses as a category account for fourteen responses or an unduplicated 13 NGOs (or 42 per cent of the 31 NGOs). This 42 per cent is much higher than the 10 aggregated responses (32 per cent of total) claiming to work with groups usually identified with the very poor in Kenya (represented by double asterisks below).

Table 2: Kenyan SEDManagers' Oral Descriptions Of Target Groups

Could you tell me what were your target groups?	Total Responses	Total SED NGOs
*Existing businesses in poor area	4	13% )
*Existing businesses	3	10% )
*Existing small and micro-enterprises	3	10% )
*Existing business: assets <50,000 KSH.	1	3% )
*Existing business: assets <100,000 KSH.	1	3% )
*Existing Women's Businesses	2	6% )
Women	4	13%
**Single Mothers	2	6%
**Street Children & Their Families	1	3%
Existing Project Families	2	6%
**Child Sponsorship Families	1	4%
Existing Women's Groups	2	6%
**School Dropouts	3	10%
**Low Income/ Poor in Target Area	3	10%
Members of NGO (MOs)	2	6%
Organisations requesting services	1	3%
Unclear	1	3%
No Oral Description Given	5	16%
Total Responses	42
Total Responding NGOs (N=)	31	100%

(Source: Kenyan Fieldwork Data)

4.3. Poor as Target Groups

When asked "to what extent are the poor your target group?", most respondents asked the interviewer for a definition of "poverty." They were then told that they should reply simply using their own subjective or agency definition of "poverty." Many respondents did not have a ready definition of poverty and tended to define their target groups in other terms such as "small and micro businesses" or simply "businesses."

As shown in Table 3, averagely, 79 per cent of the agencies claimed that "All" (defined as 86 per cent to 100 per cent) or "Most" (85 per cent to 56 per cent) of their beneficiaries were "poor." The Members Organisations (MO)(such as the Young Women's Christian Association) were clear that they responded to their members' needs rather than to other client selection criteria.

In the MO cases, the SED managers stated that most of their members were not among the very poor or even the poor.

Table 3: The Poor As Target Group

To What Extent Are The Poor Your Target Group?	Total Responses
All (86% -100%)	13	41%
Most (55%- 85%)	12	38%
About 50% (55%-45%)	1	3%
44%-25%	1	3%
Less Than 25%	1	3%
Not Asked	3	8%
Total Responses	31	100%
N=	31	100%

(Source: Kenyan Fieldwork Data)

Respondents were then shown a chart showing Remenyi's strata of poverty as they would seem to be dispersed from KREP's Kibera Slum Micro-enterprise study (Parker & Aleke-Dondo, 1992) using Remenyi's definitions of the strata of poverty as defined by the interviewer.

While most agencies say they focus attention on the poor, most of the respondents candidly admitted that they did not work much with the poorest groups needing the most help. The "poorest of the poor" for the most part seem to receive little attention from most responding agencies reportedly due to the difficulty in forming businesses with people who have little or no business experience. As shown in Table 4 below, almost three quarters of the responses claimed that the NGO worked with the entrepreneurial poor (having more than one paid employee) and the "self- employed poor" who are owners/operators of their own micro-enterprises but have no full-time additional paid employees. Most NGOs claimed to work with more than one category of clients. Five agencies claimed that 50% or more of their beneficiaries were in the bottom two Remenyi strata.

Table 4: Remenyi Poverty Strata And Target Groups

What Percentage Of Your Beneficiaries Fall In Each Group?	No. Of NGOs In Category	Average Percentage Of Clients In Stratum
Not Poor	5	3%
Near Poor	8	6%
Entrepreneurial Poor	19	23%
Self-employed Poor	25	48%
Laboring Poor	14	13%
Vulnerable Poor	8	9%
Total Responses	79	100%
Total Responding NGOs	28
Not Asked Due To Time/no Response	3
N =	31

(Source: Kenyan Fieldwork Data)

These multiple responses were then retabulated in Table 5 below to omit response duplication and to analyze how much attention was given to lower poverty strata. Only two NGOs (7 per cent) out of 28 respondents claimed that 75 per cent or more of their target groups were in the lower two categories of "Laboring Poor" and "Vulnerable Poor." Over one third (36 per cent) did not claim any of their target group in the lower two categories while almost two thirds (64 per cent) claimed that they targeted no one in the lowest poverty category which logic suggests is that category needing the most assistance in finding employment and increasing incomes.

Fifteen NGOs (54 per cent of 28 responding NGOs) said that 75 per cent or more of their target group were in the lower three poverty strata ("Self employed," "Laboring," and "Vulnerable Poor"). However, most of these were "self-employed poor."

Table 5: NGO Aggregated Responses About Target Categories

Aggregated Target Groups Categories	Total Responses
Over 75% is Self-employed & below	15	54%
Over 75% is Laboring Poor & Below	2	7%
50% or more Laboring Poor & Below	8	29%
No Laboring Poor/Vulnerable Poor	10	32%
No Vulnerable Poor	18	64%
Responding NGOs N =	28	100%
No Response	3
Total NGOs in Census	31

(Source: Kenyan Fieldwork Data)

Two of the NGOs not responding to this question were large international NGOs which targeted at the project level, had many diverse projects, and, therefore, did not have aggregated data or an "educated" guess. However, these two are "traditional" relief and development NGOs, one might speculate that they target high percentages of the lower poverty strata due to their rural emphasis. One other NGO declined to answer. These results are certainly not definitive due to the subjective nature of the questions and the answers which were often admittedly "guesses" on the part of respondents. But, the responses do reveal that most NGOs' SED managers do not really believe that they are reaching the "poorest" people in society.

In summarising this section, then, one can conclude that very few (only two responding in this study) Kenyan NGOs involved in SED are specifically targeting the poorest groups of the communities where they work. For the most part, Kenyan SED NGOs target people already having businesses who are not unemployed. This finding is not unexpected and correlates with some of the SED literature. However, this finding does raise questions about the values and purposes of NGOs which do claim NGOs generally are more responsive to the poorer elements of society.

Because very few of these NGOs have done impact assessments focused on target groups, there is little data internally generated to confirm the "educated guesses" made by NGOs' SED managers here. The Remenyi strata used in this study were very effective in raising this poverty issue with the NGO SED program managers. This analytical tool could be very helpful if used more widely in other studies to gain comparative and validating data.

5. Participation

5.1 Participation of Beneficiaries in Area Selection

Most agencies studied use more than one method for deciding where to put programs suggesting that agency decision-making is complex and uses multiple methods as demonstrated in Table 6 below. Fifteen NGOs claimed to do "needs assessment" of some sort in the area before starting SED work. Almost one third claimed that local churches requested them to come and work in the area. (Many of these are the VOC NGOs that are part of the church.)

Table 6: Methods For Choosing New Program Areas

How Did You Decide Which Areas To Go To?	Total Responses	Per Cent Of Total Respondents
Needs Assessment	15	48%
Church Request	10	32%
Internal Agency Criteria	7	23%
Existing Project Request	6	19%
Group Req./Meetings	5	16%
Informal Business Density	3	10%
Govt	3	10%
Call	3	10%
Local Leaders	2	6%
Other:	2	6%
Donors	2	6%
Board Membership	1	3%
Capacity Of NGO/Diocese	1	3%
NGO Requested	1	3%
Membership	1	3%
Total Responses	62	194%
N=	31	100%

(Source: Kenyan Fieldwork Data)

In Table 7 below, these responses are categorised further into five participatory categories--- wide, open, fair, low, and no participation. Most (58 per cent) survey respondents cited methods that were rated as "wide" or "open" participation for clients in their NGO's SED program.

Table 7: Methods For Choosing New Program Areas

Participatory Rating	Total Responses
Wide Participation	10	32%
Open Participation	8	26%
Fair Participation	1	3%
Low Participation	5	16%
No Participation	6	19%
Total Responses	31
Not Applicable/Not Asked	1
Total Interviews	31	100%

(Source: Kenyan Fieldwork Data)

This data leads to the conclusion that a majority of NGO SED programs in Kenya were participatory in choosing areas as sites for their programs but it is also clear that this participation by local people is by no means a universal NGO SED practice in Kenya. As will be demonstrated below, choice of new areas for the program may be the most participatory aspect in project design, management and governance.

5.2. Client Participation in Program Design

In a follow-up question reported in Table 8 below, agencies were asked how they had designed their SED program model. Again most agencies showed complex processes and gave more than one response. Most agencies (61 per cent) used "Experts" (either internal or external to the agency) to design their SED programs, and an allied 55 per cent freely admitted to "copying other" programs. Only 45 per cent involved clients in the design through "group meetings" (35 per cent) or "community meetings" (10 per cent).

Table 8: SED Program Design Methods

How Did You Design The Sed Program Model?	Total Responses	Total Respondents
Expert Design	19	61%
Copied Others	17	55%
Group Meetings	11	35%
Evolution	6	19%
Baseline Needs Assessment Survey	7	23%
Expert Surveys	4	13%
Other	3	10%
Community Meetings	3	10%
Focus Groups	3	10%
Community Leader Survey	2	6%
Key Informant Panels	1	3%
NGO Requested	1	3%
Internal Agency Criteria	1	3%
Membership Meetings	1	3%
Total Responses	79
Total Respondents N=	31	100%

(Source: Kenyan Fieldwork Data)

In a follow-up probe question, respondents who had not volunteered some specific beneficiary related design method in the preceding question were specifically asked how the beneficiaries might have participated in the overall design of their SED program. Ten NGOs (32 per cent) admitted that beneficiaries had no input in the overall SED design. This seems surprising when three of these ten agencies refer to the Grameen Bank in Bangladesh as their ideal program model. As noted earlier, the Grameen Bank literature stresses beneficiary participation in every aspect of organisational design and management as a key feature of Grameen's long term success. Perhaps even more surprising is that one of these programs is a training program which one would expect needs to understand what target groups know or do NOT know before designing training.

Table 9: Target Group Participation In Project Design

How Did The Target Group Participate In The Overall SSE Design?	Total Responses	Total Respondents
No Input	10	32%
Group Meetings	8	26%
Meetings With Members	4	13%
Existing Program	1	3%
Evolution	3	10%
Trial Program	1	3%
Participatory Rural Appraisal (PRA)	1	3%
Low Input	1	3%
Not Asked (Note)	7	23%
Total Responses	36
N=	31	100%

(Source: Kenyan Fieldwork Data)

NOTE: "Not Asked" Due To Participatory Response(s) In Table 8.

When these NGOS responses were retabulated and categorised according to "wide participation," "open", "fair," "small" and "no participation," there is a surprising bimodal distribution with NGOs tending to be "widely participatory" (52 per cent) or not participatory at all (35 per cent).

Table 10: NGO SED Project Design Participatory Rating

Participatory Rating	Total Responses
Wide	16	52%
Open	1	3%
Fair	1	3%
Small	1	3%
No Participation	11	35%
Not Asked Due To Time	1	3%
Total Responses	31	100%
Total Interviews=	31

(Source: Kenyan Fieldwork Data)

5.3. Client Participation in SED Project Management

This issue is defined as whether or not clients have any formal power within the project management structure. SED program managers were asked, "How do your target groups participate in management of the program?" This issue is difficult to judge from respondents' answers. However, respondents' comments throughout the interviews tend to reveal if NGO staff were attempting to make project management truly participatory. In Table 11 below, most of the responding NGO SED programs cited informal roles for clients to participate in management of the project. Of the 31 surveyed NGOs, 45 per cent claimed that senior NGO staff or lower level SED program staff meet with client groups periodically or simply claim "staff contact" (58 per cent of NGOs).

Another large group (45 per cent) in Table 11 claim that clients manage solidarity credit groups which are a real formal part of the project structure but which have very limited formal power to make important project decisions such as setting interest rates or loan terms. For this reason, "credit group management" was considered at best only a "fair" participation in project management. Smaller number of SED programs reported client participation in formal structures like "Board of Directors membership" (16 per cent), "project management" (16 per cent) and "advisory committees" (13 per cent).

Table 11: Client Participation In Management

How Did The Target Groups Participate In The Management Of Your SED Program?	Total Responses	Total Respondents
Staff Contact (Small/none)	18	58%
Credit Group Management (Fair)	14	45%
Staff Meetings With Client Groups (Small/none)	14	45%
Church/NGO Structure (Wide/open)	8	26%
Other	4	13%
Board Membership (Wide/open)	5	16%
Project Management (Wide/open)	5	16%
Advisory Committee (Open)	4	13%
Monthly Reports (Small/none)	3	10%
No Participation	2	6%
Not Asked	1	3%
Total Responses	77
N=	31	100%

(Source: Kenyan Fieldwork Data)

Of the 31 surveyed NGOs, an unduplicated 9 (29 per cent) claimed that SED clients had formal positions in parent NGO (or church) structures (26 per cent) and/or as members of Boards of Directors (16 per cent).

When surveyed NGOs were again recategorised according to the five point participatory scale 46 per cent were seen as having a "wide" or "open" participatory management style. A smaller group (35 per cent) were categorised as having a "fair participatory" rating. (This "fair" group was smaller than the percentage of credit group management responses in Table 10 above because of multiple responses that were more participatory). However, most of this 35 per cent would have a lower participatory management rating if they did not use a group credit self-management methodology. The distribution of NGOs in this Table is quite different from the "bimodal" patterns seen in Table 10.

Table 12: Project Participatory Management Ratings

NGO SED Project Management Rating	Total Respondents
Wide	7	23%
Open	7	23%
Fair	11	35%
Small	3	8%
None	2	5%
Not Applicable/Not Asked	1	3%
Total Responses/Total Interviews	31	100%

(Source: Kenyan Fieldwork Data)

The "Fair" ratings are from "solidarity" group credit programs which some consider an "empowering" intervention, and, therefore, it is difficult to judge the issue of how to assign them. While it is true that clients manage the small credit groups approving loans and collecting payments, the power of the group is very limited in its ability to change the major elements of the program like the interest rate charged, the term for loans, and maximum size loans.

Most agencies do not provide for the direct oversight of programs and staff by beneficiaries in any kind of day to day management relationship or position other than through client group credit self-management. NGO literature suggests that NGOs are more responsive to local beneficiaries and "the people" due to their close relationship including management structures that provide for the beneficiaries to actually hold NGO staff accountable. For the most part, these SED programs are not accountable to local beneficiaries in a direct way.

When surveyed NGOs answers to the previous open ended question, "How were clients involved in the management of the SED program?" were recategorised according to answers about NGO governance in which clients actually had an official accountability role in the NGO either through ownership and election of Directors (Wide), membership on Board of Directors (Open), or membership on official advisory boards (Fair), Kenyan SED NGOs fell into a more bimodal distribution where 10 (38 percent of respondents) had no participation in governance while 6 NGOs (23 per cent) had a "wide" participation in governance.

TABLE 13: SED NGOs Participatory Governance Ratings

SED NGO Governance Rating	Total Responses
Wide	6	23%
Open	3	1%
Fair	4	13%
Small	3	1%
None	10	38%
Total Responses	26	100%
Not Applicable/Not Asked Due To Time	5
Total Interviews	31

(Source: Kenyan Fieldwork Data)

None of the responding international NGOs with Kenyan SED programs had any beneficiaries in their formal organizational structures although as "foreign" agencies one might expect them to want to be more sensitive to empowering beneficiaries. This lack of representation in management does not mean that these agencies are "ipso facto" less responsive to beneficiary concerns but more research might be done in this area.

None of the ten agencies that cite Grameen Bank as their model incorporate beneficiaries in their formal organisational structures although an important feature of Grameen organisation is the partial ownership, management, and governance by the "poor" who are also Grameen clients. The lack of client participation in governance does suggest that the Kenyan organisations may be less responsive to clients needs and desires than Grameen.

It is important to point out that there was no specific question focusing on governance and Boards of Directors specifically in the interviews. However, these responses did come out in the open ended questions and the researcher believes that they are accurate categorisations of the overall NGO governance based on interviews and his personal knowledge.

6. Step-Wise Regression and Operational Self-Sufficiency Ratios

Stepwise regression was used with five dependent variables and 25 independent variables. Using stepwise regression analysis to find the variables that better fitted a regression equation explaining Operational Self-Sustainability (0PSS) ("Y_S") defined as Operational Income (non-grant) divided by Operational Expenses (omitting Financial costs such as capital asset costs, foreign exchange expenses, loan interest) (Christen, 1990), the strongest statistically significant relationship was found using the following independent variables in order of their significance within the equation:

(1) Frequency of communicating the Corporate Purpose Statement to staff ("X_F");

(2) Total number of clients assisted by the programme ("X_C");

(3) Programme goals focused on poverty objectives ("X_P");

(4) A negative correlation with bureaucratic management ("X_B"); and

(5) A negative correlation with the number of different ways ("X_W") that respondents claimed to use their purpose statements with staff.

The equation has an adjusted R² of 0.634 and an "F" value of 15.52 at a significance level of 0.01 with a D.W. statistic of 2.12. The regression equation is:

Y_S = 6.77X_F + 0.02X_C + 45.16X_P -49.69X_B -7.03X_W + 9.410 (8.1)

Note: Y_S = Operational Sustainability Ratio

X_B = Bureaucratic Management (Nominal)

X_C = Number of Clients

X_F = Frequency of Communicating Purpose Statement

X_P = Poverty Focus in CPS

X_W = Ways of Communicating Purpose Statement

TABLE 14: OPSS Regression Equation "t" Statistics, P-Values and Partial R²

Independent Variables	t(23)	P-Value	Partial R2
XF	6.38537	1.89204E-7	0.524256
XC	6.74624	6.18936E-8	0.551579
XP	4.05672	2.46655E-4	0.307855
XB	-3.94369	3.44011E-4	0.295945
XW	-2.59896	1.33510E-2	0.154374

The poverty focus of an SED program's goals and objectives is closely allied to the purported normative "power" of the corporate purpose statement and may be even more important in providing program direction. The positive correlation of the Poverty focus with organisational sustainability supports Remenyi's contention that the poor have resources and are good business partners. The evidence here suggests that:

NGO SED programs that are focused on the poor and unemployed are capable of providing fair returns for NGOs leading to higher operational sustainability ratios.

This is an important finding in supporting NGOs that are pursuing poverty alleviation strategies rather than trickle down strategies at the micro-level by helping he less poor in relatively poor communities.

7. Client Participation in Design and Management

Using the data above, participation in design and management was assigned a nominal rating of "1" for "widespread/open" or "0". These data were used in stepwise regression analysis to test the explanatory importance of participation as a factor in financial efficiency in thirteen of the NGOs which reported additional operational and financial data for two or more years. "Widespread/open" participation by clients in program design and participation by clients in program management was found to have no part in any regression equation with explanatory power with an R² of .5 or above.

Although caution needs to be exercised to recognize the limitations of this data to this group of NGOs at this time, this finding suggests that while participation may have important empowering impact on clients, widespread participation by clients in program design and in program management is neither positively nor negatively correlated with operational efficiency measures (operational sustainability, total income per client, or total cost per client) demonstrating that client participation in program design or management are not important factors in considering operational financial efficiency.

This is a surprising finding in that Bryant and White (1982) suggest that participation has real costs. However, this may be explained in that participation can also reduce costs as clients' knowledge is put to practice in the local situation.

Summary

Prior to this study, there were no systematic surveys of NGOs with SED operations in Kenya that researched NGOs sense of purpose as well as the extent of their operations especially with the poor and marginalised. While NGOs are reported to be involved with the poor more than government or business, the evidence from these data supports the claim that NGOs SED programs are involved with the poor in general but not with the poorest members of the community who most need assistance.

There is further evidence to suggest that these NGOs are not as participatory as some of the literature suggests that NGOs are. While this is not surprising for SED operations, there is clear empirical evidence here that should cause one to question broad claims about the participatory nature of NGOs in general. Obviously, more studies need to be done on NGOs in other countries and in other sectors to ascertain whether NGOs in general are any more participatory than business or government. Although Boomgard, Tendler and others have clearly stated that SED NGOs are not particularly focused on the poor nor are they particularly participatory, their generalisations have been based on collections of studies from a wide array of NGOs across geographic, cultural and economic lines. This study has given clear data for a group of NGOs in one country that broadly supports their generalisations.

At the same time, many writers about NGOs and many of the managers of these Kenyan NGOs look to the Grameen Bank as a particularly good example of a SED NGO that is poverty focused and participatory in its operations. This study demonstrates that none of the Kenyan NGOs seem to be as poverty focused nor as participatory as Grameen Bank.

This last factor suggests that these NGOs may become even less participatory and poverty focused as they grow into the larger organisations that many of them aspire to develop.

Finally, the strong association of a Poverty focus in the Corporate Purpose Statement with higher operational self-sufficiency ratings is important to encourage NGOs and donors to support poverty alleviation strategies--- a longtime characteristic of many NGOs throughout the world. Nevertheless, more research needs to be done with broader groups of NGOs in many countries before accepting that this Kenyan finding is a universal condition.

Endnotes

1. From Hollnsteiner's list, a scale of "participatory measures" was developed based on the degree of participation and power granted to clients in the design of a development project

(1) No Client Input (Experts or NGO staff design project; or project copied from elsewhere).

(2) NGO Field Staff contact with clients

(3) NGO Management and Field Staff meet with clients

(4) Community Surveys (Clients included)

(5) Surveys of Needs Focused only on Client Groups

(6) Focus Groups with Clients Discussing Plans

(7) Consultation with Clients Throughout Process

(8) Client approval of staff written plans

(9) Client Involvement in Writing/Making Plans

(10) Client Planning/Designing Project with NGO as Catalyst

(11) Clients Forming Own New Organisations Without External Assistance

Within such a scheme, levels one, two, and three above are essentially non-participatory in that clients input is totally filtered through the NGO staff's own preconceptions and "baggage." The same may also be true of levels four to nine to some extent. It is certainly difficult to assess what really happens in the design of a project. If NGO staff are real "listeners," then, levels one to three may in fact be more responsive to clients needs and desires than more participatory methods used with insensitive or authoritarian NGO staff who appear to "listen" but never really "hear."

2. Specific design methods that are viewed as participatory in an empowering sense in this study are:

(1) Focus Groups Among Prospective Clients to ascertain Needs/Desires

(2) Community Meetings to Ascertain Needs/Desires

(3) Involvement of Prospective Client Groups in Project Design Meetings

(4) Action learning. Evolutionary Approaches in which Clients' Feedback has an important Impact on Design

Methods judged to be "Non-Participatory" are specifically:

(1) Expert or Consultants' Design;

(2) Copying other Projects Design; and

(3) NGO Staff Design with No Prospective Client Input.

References

BOOMGARD, J., D. DE SANTIS, and A.TZAVARAS (1989) Taking Stock of A.I.D.'s Microenterprise Stocktaking Portfolio: Background and Conceptual Overview. Washington D.C.: U.S. Agency for International Development.

BROWN, L.D., and D.C. KORTEN (1990) Understanding Voluntary Organisations: Guidelines for Donors, World Bank Staff Working Papers WPS 258. Washington, D.C. : World Bank.

BRYANT, C. and L.G. WHITE (1982) Managing Development in the Third World. Denver: Westview Press.

BURBANK JR., K. (1994) A Survey of NGO's As Small Business Development Agencies in Kenya, Institute for Development Studies, University of Nairobi, IDS Working Paper No. 493.

CALAVAN, K.L.,Latifees, H.I., and M. YUNUS (1989) "Cases 10, 11, and 12: The Grameen Bank Project (A, B and C)" in MANN,C., M.S.GRINDLE and P. SHIPTON, Seeking Solutions. Hartford: Kumarian Press.

CERNEA, M. (1988) Nongovernmental Organizations and Local Development, World Bank Discussion Paper No.40. Washington, D.C.: World Bank.

CHAMBERS, R. (1987) Rural Development: Putting the Last First. London: Harper & Rowe.

CHAMBERS, R. (1995) "Primacy of The Personal", pp. 207- 218 in EDWARDS, M. and D. HULME (1995) Non-Governmental Organisations - Performance and Accountability Beyond the Magic Bullet. London: Earthscan Publications Ltd.

CHANDLER, M. & A. FUGELSANG (1988) Participation As Progress. Dacca: Grameen Bank.

CHRISTEN, R.P. (1990) Financial Management of Micro-Credit Programs: A Guidebook For NGOs, Boston: ACCION.

CLARK, J. (1991) The Role of Voluntary Organizations: Democratizing Development. Hartford: Kumarian Press.

ELLIOTT, C.(1987)"Some Aspects of Relations Between the North and South in the NGO Sector," pp. 57-68,in DRABEK, A. (Ed.)(1987)Development Alternatives: The Challenge for NGOs, World Development Special Issue, Vol. 15, Autumn Supplement, Oxford: Pergammon Press.

FOWLER, A. (1985) "NGOs In Africa: Naming Them By What They Are" in Non-Governmental Organisations (NGOs): Contributions to Development, Proceedings of a Seminar Held at the Institute of Development Studies, Nairobi, September 19, 1985, Working Paper No. 50. Nairobi: University of Nairobi.

FOWLER, A. (1988) Non-Governmental Organisations in Africa: Achieving Comparative Advantage in Relief and Development, Working Paper, IDS, DP 249. Brighton: IDS, University of Sussex.

FOWLER, A. (1994) The Growth of Non-Governmental Organisations and Democracy in Kenya Ph.D. Dissertation. Brighton: IDS University of Sussex.

KORTEN, D.C. and R. KLAUSS (ed.) (1984)People Centered Development: Contributions Toward Theory and Planning Frameworks. Hartford: Kumarian Press.

KORTEN, D.C. (1987) "Third Generation NGO Strategies: A Key To People-centered Development," pp. 145-160 in DRABEK, A. (Ed.)(1987)Development Alternatives: The Challenge for NGOs, World Development Special Issue, Vol. 15, Autumn Supplement, Oxford: Pergammon Press.

KORTEN, D.C (1991) "The Role of Nongovernmental Organizations in Development: Changing Patterns and Perspectives," pp. 20- 43 in PAUL, S. and A. ISRAEL (1991) Nongovernmental Organizations and the World Bank: Cooperation for Development. Washington D.C.: World Bank.

MALHOTRA, E. (l992) "Poverty Alleviation" and Microenterprise. Development Working Paper, Washington,D.C.: U.S. Agency for International Development.

OAKLEY, P. et al. (1991) Projects with People. The Practice of Participation in Rural Development. Geneva: International Labor Organization.

PAUL, S. and A. ISRAEL (1991) Nongovernmental Organizations and the World Bank: Cooperation for Development. Washington D.C.: World Bank.

REMENYI, J. (1991) Where Credit Is Due. Income-Generating Programs for the Poor in Developing Countries. London: IT Publications.

TENDLER, J. (1982) Turning Private Voluntary Organizations into Development Agencies, Questions for Evaluation, A.I.D. Program Evaluation Discussion Paper No. 12. Washington, D.C.: U.S. Agency for International Development.

UPHOFF, N. (1985) "Fitting Projects to People" in M.M. CERNEA (ed.) Putting People First: Sociological Variables in Rural Development pp. 479-486. New York: Oxford University Press.

UPHOFF, N. (1986) Local Institutional Development: An Analytical Source Book with Cases. Hartford: Kumarian Press.

YUNUS, M. and MORGAN, E.(1996) Mohammed Yunus Video Interview with Dr. Elizabeth Morgan, St.Davids, PA: Eastern College.

Catchment

Size

Interviewed Agency's Name Class

Action Aid- Kariobangi Project LOCAL

Adventists Development & Relief Agency REGIONAL

Ambassador's Development Agency REGIONAL

CARE- Women Income Generating LOCAL

Christian Children's Fund REGIONAL

Christian Industrial Training Centre NATIONAL

Christian Self Development Fund LOCAL

(now known as Village Enterprise Fund)

Dandora Community Credit Project LOCAL

Daraja Trust LOCAL

Diocese of Eldoret, C.P.K. REGIONAL

Diocese of Machakos, Catholic LOCAL

Food for Hungry- Faulu Loan Scheme LOCAL

Imani LOCAL

Improve Your Business (IYB) REGIONAL

Jisaidie Cottage Industries NATIONAL

KREP- Juhudi & Chikola LOCAL

Kenya Institute of Management(KIM) NATIONAL

Kenya Management Assistance Program NATIONAL

Kenya Small Traders Society (KSTS) NATIONAL

Kenya Women Finance Trust LOCAL

Kolping Society Financial Assistance REGIONAL

National Council of Churches in Kenya REGIONAL

OXFAM LOCAL

PCEA Chogoria Hospital Enterprise Development Program LOCAL

PRIDE- Kenya REGIONAL

Partnership for Productivity NATIONAL

St. John's Church, Revolving Loan Fund LOCAL

Tototo Home Industries Savings & Credit REGIONAL

Undugu Society of Kenya LOCAL

World Vision NATIONAL

Young Women's Christian Association-SED REGIONAL

Total Agencies In Study 31