Understanding the Nature of MFI Participation:Evidence from Rural Bangladesh

3.1 Preambles

While ex-post theorization on microfinance focused on group lending and group liability [1] , empirical work considered households as sampling units and often resorted to program-control comparison, spending much effort on identifying appropriate econometric techniques to control for selectivity bias. Since money borrowed from the MFIs is fungible, economists were happy with the choice of households as their units of analyses for impact assessment, while sociologists and gender specialists had their primary focus on member-specific participation. In both cases, however, relative stability in participation – whether it was for members or for the households – was taken for granted. The only variable factor, qualifying participation, was the length of participation. Such a premise was consistent with a growing industry, especially since the stated statistics suggested annual growth in membership at almost two-digit levels. [2] It is therefore no wonder that “drop-outs” from the system, which were being captured with micro-level snapshots, generated fuzzy views. At one extreme, this was interpreted in terms of “graduation”, while others expressed concerns for the MFIs’ inability to retain rural poor in their service-net. In the light of such confusion, it is important that the nature of participation in MFI programs be better understood, for better monitoring and interactive policy designs, as well as for meaningful interpretation of findings from impact assessment studies.
During the recent past, one aspect of participation has attracted much attention – that of multiple participation. The concern of the MFIs with such participation has been addressed by coining the term “overlapping”, which captures only one dimension of multiple-participation. This aspect will also be addressed in this chapter, within the broader perspective of understanding the nature of MFI participation in rural Bangladesh.

3.2 Defining Participation and related Conceptual Issues

Before engaging with numbers in subsequent sections, it is worth clarifying some definitions and conceptual issues. For any empirical work, it is important to identify, at the foremost, the level at which participation is being addressed – whether member, household, or both. In a strict sense, participation in MFI programs is by individual members; and household-level participation is a derived category. If one or more members of a household participate in one or more MFI programs, the particular household is identified as “participant”. A more practical problem arises in defining participation itself. A loose definition would include all individuals who report to be members of MFI groups. A stricter version may take into cognizance of the regularity in attending group meetings and depositing the required savings – thus, distinguishing between “active” and “passive” participation. Since group-level interactions are largely timed with financial transactions pertaining to borrowing and repayment, one may alternatively define participation in terms of borrowing status of an individual. Thus, active members may include those who are active borrowers – that is, either regularly repaying loan, or, waiting to receive a loan. In such categorization, one would have to separately identify the defaulters. While the findings presented in this chapter mention of the proportion of members borrowing from MFI sources, participation is operationally defined to include those who are active group members.

Within the above categorization, which is a static one, participation by individual members may be grouped into two: participation with a single MFI, and participation in more than one MFI. Accordingly, participant households may be grouped into the followings [3] :

Only one member participates, which may either of the two forms:
(a) in a single MFI, or,
(b) in two or more MFIs.
Two or more members of the household participate, which may take one of the following forms:
(a) all members participate in the same MFI;
(b) members participate in two or more MFIs. The latter may or may not include participation by an individual in two or more MFIs.
The above static classification is further complicated due to instability in participation itself. The BIDS survey of a matched sample over three rounds shows that a large number of individuals change their affiliation, or temporarily disassociate themselves from the MFI groups, every year. Thus, one needs further classification across “never participated” and “ever participants”; and within the latter group, across “regular participants”, “occasional participants” and “drop-outs”. Such classifications require some reasonable choice on period of affiliation (for identifying a regular case) and on period of disassociation (for identifying a case of drop-out). As will be shown in the subsequent sections, all these categories have emerged only during the recent past; and therefore, it will take a while to arrive at conclusive operational definitions. We bring these aspects from our data to provide an alternative perspective to view the dynamics that are underway in the microfinance industry.

3.3 Evidence on Participation in MFI Programs

We will initially present some evidence in support of the arguments made in the previous section. Estimates on current participation are then presented, following the classification suggested. The last part of this section presents findings on “current participation” varies on an annual basis. A fuller explanation of the observed pattern of participation is addressed in Section 3.4.

Presence of Passive Participation

Of all the MFI members captured in our sample survey during 1998, 12 percent reported of previous affiliation (see Table 3.1). When asked, if they were still involved with those MFIs, one-third reported of no involvement, while 9 percent reported of only passive involvement. In subsequent rounds in 1999 and 2000, extents of passive involvements were found to be respectively, 7.7 % and 15.2 %. Normally, borrowers’ perspective on involvement varies widely with those of MFIs. In many instances, MFIs may continue to consider an individual to be its member even though the latter reports of not being involved.

Relationship between Membership and Borrowing

Of all current participant households, as reported during 1997 census of 91 villages, only 85 percent had ever borrowed from MFI sources, and 77 percent were found to be current borrowers (Table 3.2). Of all the current members, reported in 1998 sample survey, 90 percent were found to have borrowed at least once from an MFI. [4] The incidence of borrowing had reduced drastically during the last year of the survey (1999-2000) - of all the MFI members, only 62 percent had reported of taking loan. Of the latter, 98 percent were active borrowers while the rest had repaid. Note however that 84 percent of the 1551 MFI members in 1999-2000 knew the number of group meetings held during the month preceding the survey, which means that 58 percent of the non-borrowing members are involved in group activity; and may borrow at a later period. More than 90 percent of this latter group attended the group meetings.

Table 3.1
Distribution of Current MFI Members in terms of their Past
Affiliation and Nature of Involvement, by Current Primary Affiliation

Name of MFI	% of current memberswith no previous affiliation	% of ever members whosought new affiliation	Nature of affiliation with previous MFI
Name of MFI	% of current memberswith no previous affiliation	% of ever members whosought new affiliation	Not involved	Passive involvement	Actively involved
Grameen Bank	92.7	14.3	12.7	5.5	81.8
BRAC	80.7	14.6	38.3	12.8	48.9
ASA	84.7	3.0	66.7	0.0	33.3
Proshika	90.2	6.6	23.1	23.1	53.8
Anubhab	84.0	35.9	64.3	7.1	28.6
Solidarity	75.8	31.3	53.3	0.0	46.7
PPD	78.0	6.3	50.0	0.0	50.0
PRP	81.7	9.1	100.0	0.0	0.0
GUP	91.7	14.3	12.5	12.5	75.0
SSS	70.6	6.8	20.0	20.0	60.0
Sabalamby	96.1	1.9	50.0	0.0	50.0
Noabeky	98.6	0.0	-	-	-
OSDER	85.5	1.6	0.0	100.0	0.0
TMSS	82.1	11.8	55.6	11.1	33.3
RDRS	85.4	21.3	38.5	15.4	46.2
Prottashi	97.5	2.4	50.0	0.0	50.0
Others	84.4	16.4	32.6	7.0	60.5
All sample members	87.4	11.2	34.6	9.2	56.3

Source: Sample Survey on MFI Members, 1998.

Table 3.2
Extent of Borrowing among Program Participant Households by Study Area

Study Area	Percentage of participant household
Study Area	Availed at least one loan	Current borrower
Panchagarh	92.9	79.4
Kurigram	94.1	88.9
Sirajganj	80.1	72.5
Meherpur	80.1	69.7
Madaripur	91.6	78.1
Barisal	89.7	77.8
Tangail	74.0	66.5
Netrokona	85.1	83.3
Satkhira	72.7	64.2
Munshiganj	93.3	85.2
Chittagong	82.3	80.0
Feni	60.2	57.2
Bogra	96.4	93.0
Total	84.6 (6913)	77.2 (6303)

Source: Household Census in 91 Villages, BIDS Study on PKSF-MES.

Member-level Participation

Of those reporting current participation during 1998, more than 6 percent reported of their affiliation with two or more MFIs. Extent of multiple-participation increased to more than 10 percent during the later years (Table 3.3) [5]

Table 3.3
Estimate on Multiple Membership – at Member level

Regions	Number of MFI members			% of members who participate in more than one MFI
Regions	1997-98	1998-99	1999-00	1997-98	1998-99	1999-00
Panchagar	136	128	112	2.21	11.72	14.29
Kurigram	131	131	118	12.98	13.74	13.56
Sirajganj	148	134	79	7.43	14.93	11.39
Meherpur	144	116	93	2.08	12.07	7.53
Madaripur	137	135	131	9.49	16.30	16.03
Barisal	149	152	134	8.72	14.47	12.69
Tangail	169	239	191	15.38	18.83	15.18
Netrokona	139	133	113	0.72	5.26	8.85
Satkhira	168	168	167	0.60	4.17	2.99
Munshiganj	143	118	94	9.09	10.17	13.83
Chittagong	115	97	80	0.87	4.12	3.75
Feni	143	118	94	2.80	5.93	4.26
Bogra	194

Percentage of participant household

Member-level Participation