1

Does Self Help Group Participation ensure Women Employment? A Case Study

Tanmoyee Banerjee ( Chaterjee)a

Chandralekha Ghoshb

Abstract

Present paper theoretically investigates the factors responsible for employment decision

of woman self help group member as optimal solution to intra household utility

maximization process. The theoretical results have been empirically validated on the

basis of two rounds of primary surveys once during 2005 and again in 2009 of same set

of self help group members in North 24 Parganas, West Bengal, India. Empirical results

show past occupation of the members and the local politics significantly influence the

present employment status as well as loan use pattern of the members.

Keywords: Self Help Group, Employment Status, Loan use pattern

JEL classification: D13, G21

Corresponding Author Chandralekha Ghosh

Department of Ecopnomics

West Bengal State University, Barasat, North 24, Parganas, India

ghoshchandralekha@yahoo.com

Address for Communication B201,Prateeti, 165A,Prantik Pally,

Kolkata-700042.

a Department of Economics, Jadavpur University, Kolkata 700032, India. b Department of Economics , West Bengal State University, Barasat, North 24, Parganas

2

Introduction

“Unemployment is a scourge of every modern society”. Yunus(1998)

The ideology and practice of poverty alleviation has been deeply influenced by the idea

that access to credit can create employment opportunities. Different programs have been

developed around the world which covers millions of poor borrowers and they are

provided a variety of different financial arrangements to start their own projects. The

main focus of these programs is to generate self employment projects especially for the

rural poor woman by smoothing their financial needs. But there are criticisms regarding

this financial inclusion as the main objective of employment generation through self

employment projects may loose its inertia and the woman may turn into “just a

borrower”, catering to various financial needs of the family.

In India a self help group based micro credit programme, SGSY1 was initiated from

1st April 1999. This programme aimed at extending income generating self employment activities

among poor women in rural areas through financial inclusion2. It tried to link the rural women to

the formal micro finance institutions through providing joint liability loans to SHGs and covered

all aspects of self employments such as training, credit, technology, infrastructure and marketing

and enabling the rural poor to take decisions on all issues concerning poverty eradication. The

1 Swarnjaynti Gram Swarojgar Yojana (SGSY) which is a self-employment programme of Ministry of Rural Development t aims at providing assistance to the below poverty level (BPL) rural poor for establishing micro-enterprises through bank credit and government subsidy to acquire an income-generating asset. The Swarnjayanti Gram Swarozgar Yojana, a credit-based scheme sponsored by the Government of India for poverty alleviation, is perhaps the largest of its kind in the world. It was first announced in 1997 to commemorate fifty years of India's independence and

first allocations for it under the Ministry of Rural development (MORD) were included in the GOI budget of 1999/2000. The MORD was to fix the shares of each district in the country roughly on the basis of relative numbers of poor residing there. 75% of the funds would come from the Union Government and each state government would match it for the remaining 25%. Both parts of the funds are to go directly to the DRDA; in each district, the DRDA is in charge of executing the scheme. For this, it is to draw on the help of local governments, line agencies of superior governments and local branches of commercial banks. In case of SGSY programme it was directed by the Government of India that the groups will be formed by taking members from Below Poverty level (BPL) List. If above poverty level members are included in the group then that person cannot be an office bearer of the group. Under SGSY, the

individual beneficiaries and members of SHGs are called swarozgaris. Generally a self Help Group (SHG) consist of 10-20 persons are formed but in difficult areas like deserts, hills and areas with scattered and sparse population and disabled persons, this number may be 5-20. Focus is given on the on vulnerable groups- SC/STs should be at least 50% of swarozgaris; women-40% and disabled- 3% of swarozgaris. List of BPL households identified through BPL census and duly approved by Gram Sabha forms the basis for identification of families from which the members of Self Help Groups (SHGs) are drawn. DRDAs initiate and sustain the process of social mobilization for formation, development and strengthening of SHGs through facilitators viz. NGOs, CBOs, Banks, Community Coordinators, Animators and SHPIs. Source: http://rural.nic.in/FAQ_SGSY.pdf

2 Financial inclusion denotes delivery of financial services at a help groups are created across affordable cost to the vast sections of the disadvantaged and low-income groups.

3

present study will try to find out the incidence of employment among female members of SHGs

created under SGSY.

The present study will try to examine how far the SGSY has been successful in creating

and sustaining employment opportunities for the rural poor woman. The decision of the woman to

work outside the home is the outcome of several complex socio economic factors. The role of

local politics, the demographic dimensions play important role in this decision making of the

woman. This study is unique as it has tried to examine the decision of the woman to work outside

considering intra household utility analysis.

The unique feature of this study is that the same set of group members have been

interviewed twice with a gap of four years. Firstly a group of 290 female SGSY SHG members

were surveyed during 2005-2006 and again the same set of members was resurveyed during

2009-2010. Based on this unique survey present paper will try to identify the socio economic

political and demographic factors that are affecting the continuity of employment status of the

group members. Nature of this survey will help us to remove the impacts of fixed unobservable

on the decisions of the group members. As we are examining the same set of members over a

period of time, the concept of selection bias does not arise.

The existing studies on self help groups have shown that group participation has helped

to remove poverty; improved consumption level of the families of group members; improved rate

of asset creation of the families; finally it improves the women empowerment levels. Different

impact evaluation studies by Kandekar(2003)3, Kandekar(2005) , Swain and Floro (2007)

4,

Tedeschi(2008), Banerjee(2009)5 ,Imai et al(2010)

6 have shown different positive impact of

3 Study by Kandekar(2003) carried out an exercise by estimating the effects of micro-finance on consumption, poverty and non-land assets for participants, non-participants, and an average villager, assuming that micro-finance programs have spillover (externality) effects. The results are resounding: micro-finance matters a lot for the very poor borrowers and also for the local economy 4 Swain and Floro (2007) have developed a theoretical model to examine the mechanisms through the pecuniary and non pecuniary effects of the SHG program on the beneficiaries’ earnings and empowerment, influence their households’ ability to manage risk. Going beyond the traditional poverty estimates, they have used vulnerability measure which quantifies the welfare loss associated with poverty as well as different types of risks like aggregate and

idiosyncratic risks. Applying this measure to an Indian panel survey data for 2000 and 2003, they have found that SHG

members have lower vulnerability as compared to a group of non-SHG (control) members. Furthermore, they have

observed that the poverty contributes to about 80 percent of the vulnerability faced by the household followed by aggregate risk. This paper investigates whether or not SHG participation results in reducing poverty and vulnerability. 5 Banerjee(2009) has shown that income regenerations through group activities has improved the average income of group members but the inequality of distribution of income is high among the group members than that of the non-group members. Secondly, there has been a significant decline in the medical expenditure and school dropout rate in the families of group members than that of non-group members. 6 Imai et al(2010) examined whether household access to microfinance reduces poverty. Using national household data from India, treatment effects model is employed to estimate the poverty-reducing effects of Micro Finance Institutions (MFIs) loans for productive purposes, such as investment in agriculture or non-farm businesses on household poverty levels. These models take into account the endogenous binary treatment effects and sample selection bias associated

4

group joining on the families of the group member. Based on data from a primary survey from six

states in India (NCAER 2008) Parida and Sinha (2010) have showed that all females groups have

performed most efficiently compared to male SHGs.

Issues related to SHG participation and women empowerment have been dealt in papers

by Swain and Wallentin(2009)7, Garikipati(2008). The study by Garikipati(2008) shows an

interesting and paradoxical result where credit availability of the rural woman through self help

group formation benefits the households by reducing risk vulnerability and asset creation but it

does not really empower woman. They have found out that loans procured by women are often

diverted into enhancing household’s assets and incomes. This combined with woman’s lack of

co-ownership of family’s productive assets, results in her disempowerment

Study by Swain and Adel Varghese(2009) have shown that in case of Indian SHG

members with longer participation in SHGs, members move away from pure agriculture as an

income source towards other sources such as livestock income. Training by NGOs positively

affects asset creation but the type of SHG linkage per se has no effect.

McKernan (2002) has shown the role of micro credit programs in determining the

restricted self-employment profits of rural households. Using data from a special survey carried

out in 87 rural Bangladeshi villages during 1991-1992, this study then estimates the impact of

participation in three micro-credit programs (BRAC, BRDB's RD-12 program, and Grameen

Bank) on self-employment profits. Making use of the fact that credit from micro credit programs

is used to purchase capital, this paper also provides a measure of the noncredit effect of the

programs by estimating the effect of program participation while controlling for the level of

productive capital. The study found large positive effects of participation and the noncredit

aspects of participation on self-employment profits.

However none of the papers mentioned above has dealt with the employment status of the

SHG member which is the central focus of the present study. Our main objective is to find out the

effect of group participation on the continuity of employment status of the group member taking

household utility function into consideration. The study is based on two rounds of primary

surveys; one carried out during 2005-2006 and another during 2009-2010 on same set of SHG

with access to MFIs. Despite some limitations, such as those arising from potential unobservable important determinants of access to MFIs, significant positive effect of MFI productive loans on multidimensional welfare indicator has been confirmed. The significance of “treatment effects” coefficients has been verified by both Tobit and Propensity Score Matching (PSM) models. In addition, they have found that loans for productive purposes were more important for poverty reduction in rural than in urban areas. 7The study by Swain and Wallentin(2009) using household survey data on SHG from India adopted a general structural

model where the latent women empowerment and its latent components (economic factors and financial confidence, managerial control, behavioural changes, education and networking, communication and political participation and awareness) are measured using observed indicators. They show that for SHG members, economic factors, managerial control and behavioural changes are the most significant factors in empowering women.

5

members from North 24 parganas, West Bengal, India. The SHGs were formed under the SGSY

programme. During the first round of survey all the group members were employed. However

during the second round of survey we found that even if groups are in existence, a large number

of group members have become unemployed. Srinivasan (2009) has also observed that groups

created under SGSY do not seem to have a long existence. He has observed that 3.13 million

groups were created under SGSY programme up to October 2008 across different states of India.

However only 21% of the total number of groups created took up economic activities.

The present study will try to find theoretically and empirically the factors determining the

incidence of continuity of employment among the group members. Our study mainly tries to

identify the reasons for the members to continue their employment based on the intra household

utility analysis. In line with this, the study will also focus on the factors determining the loan use

pattern of the members for the loans taken from the group corpus. Apart from these the study will

also try to find out the determinants of loan size. In this context we must mention that we will

also try to find the impact of local political dynamics on the employment decision of the group

members. In developing economies due presence of electoral clientelism the amount of benefits

received by an individual from local self governments depend on his political contacts as well as

on political scenario of that particular area. This has been confirmed in the study by Bardhan et al

(2007). For this reason we will try to find the influence of the political factors on the

employment decision of the group members which has not been covered in the existing literature.

Rest of the paper is divided as follows: Section II presents a theoretical model that

highlights the employment decisions of the woman member of the family who can access the

easy credit from SHGs. Further this section focuses on the determination of loan use pattern of

the members while assuming maximization of household’s total utility. This section provides a

theoretical basis of the empirical relations that will be estimated in the study. Section III descries

the survey design and section IV will present the result of empirical analysis and Section V will

conclude the paper.

Section II: The Theoretical model

In this section of the paper, we present a simple theoretical model designed to derive the

determinants of continuity of employment status of the female member of the family along with

pattern of loan use that a representative group member can receive from the group fund.

The present model assumes that amount of labour to be given by the female member of the

household for income generation is determined by the house-holds’ utility maximization problem.

6

Tassel (2004)8 has considered a bargaining model where the female member of the

household can access loans from micro finance institutions and the aggregate consumption good

is divided between the husband and wife in a Nash bargaining framework. The model found that

a key motivation behind borrowing capital for a woman is to raise her threat point in the

household bargaining process, and this can sometimes lead to a decline in her husband's expected

payoff. We also find that in some circumstances, it is in the woman's personal interest to transfer

control over her loan to her household partner. However the assumption of household bargaining

model is not very appropriate for our context because we are concerned with female group

members of traditional India villages. Indian village family system is a patriarchal system where

the women usually work for overall wellbeing of the family instead of bargaining for household

consumption goods in her favour. For this reason we assume that the household maximizes the

total utility function to decide about the consumption levels and time allocation of the female

family member and the loan use pattern. Gatripati (2009) has considered a similar maximization

problem to determine the time allocation pattern of the female house hold member. In the present

model we will determine the time allocation of female family member along with the use of

pattern of the loan taken from the group corpus which is not done in the present literature.

Section II:1 Description of the basic Model

The model considers a two stage decision making process. In the first stage a

representative female group member will decide about her borrowing level along with here fellow

group members to increase the income of the group will in turn will improve the size of the group

corpus. Given the size of borrowing from the group the family maximizes the household utility

function to determine the household consumption level, time allocation of female member and

allocation of credit across different purposes: to meet sudden consumption need of the family, for

8 Tassel (2004) has considered a bargaining model where the female member of the household can access

loans from micro finance institutions and there is a two stage game with two parts. In the first part of the

game, two household members make a few non-cooperative production decisions regarding a credit

contract and risky business projects. In the second part of the game an aggregate household consumption

good is produced, which is then divided between the two members of the household. Under an assumption

that there are gains from being household partners, the division of the consumption good is modeled using

a Nash bargaining framework with an implicit assumption that the players cannot commit ex ante, to a

specific division of the consumption good. This implies that a player's payoff is sensitive not only to the

size of the aggregate consumption good, but also the player's outside option or bargaining strength. The model found that a key motivation behind borrowing capital for a woman is to raise her threat point in the

household bargaining process, and this can sometimes lead to a decline in her husband's expected payoff. It

is also find that in some circumstances, it is in the woman's personal interest to transfer control over her

loan to her household partner

7

household asset creation, for self employment of female family member or for self employment

of the male family member.

For any period, a well-behaved (quasi-concave with positive partial derivatives) household utility

function is assumed to exist. Let this function be given by

)()()(),,( hFlFlMhTp fdfdmdfccUU β−α−−= 1.

The total utility is assumed to be a function of total planned consumption expenditure of the

family (cp)9, transitory or unplanned consumption expenditure (cT)

10 and amount of time devoted

by the adult female member of the family for smooth functioning of the household activities (fh)

less of disutility of work where ml and fl be the labour endowment of the male and the female

member of the family for earning purpose, respectively. The model assumes that the planned

consumption is always financed through the income of the family and the transitory consumption

or (cT) is financed through borrowing. The nature of utility function of the family will depend on

demographic and socioeconomic factors like family size, religious identity of the family, land

holding status of the family and others.

The female member of the family is assumed to make distinction between disutility of labour

from household work and that of labour endowed for earning purpose.

If 1=β=α the female worker treats both types of activities equally and her inutilities of labor

from house hold activity is same as that of income earning activity for equal amount of labour

endowment. However if β<α , the female group member is more inclined to give labour for

income earning purpose and hence perceive it as less tiring. Alternatively we can say that the

representative female member of the family is more willing to work to be economically self

reliant.

Alternatively β>α implies that the female member is more willing to work at home and she

treats household activities as less tiring. The nature of disutility of labour of the female adult

member will again depend on her education level, family size, and number of other earning

members in the family.

Thus the labour constraints are as follows:

For female family member: : lehl ffff ++= ; 0fl≥ 2a

9 The planned family consumption expenditure implies the sum total of all planned expenditure like

expenditure on food or clothing, children related expenditure, planned health and education related expenditure and others. 10 The transitory or unplanned expenditures are different in nature. It will include sudden expenditures that

a family has to undertake for medical purpose, social obligation purpose like death, marriages, or

expenditure arising out of repair of house or other household accessories and other unforeseen

expenditures.

8

For male family member : lel

mmf += 2b

where that the leisure time of the family members ),(lele

fm are given and male member is

assumed to devote a fixed amount labour time as well.

For the simplicity of analysis we assume that the female SHG member does not have any

alternative source of income unless she takes loan from SHG and start any project.11

Let i be the

effective rate of interest that the group member pays for a loan taken from her group corpus and

the SHG member earns an income yh if her project is successful with probability p, otherwise she

gets no income. Probability of success increases with labour endowment ( lf ), amount of working

capital used( `fK ), the training received by the member (t) and her social capital endowment(s). 12

Equation 3 defines the expected income of the female family labour:

hFlFystKfpY ),,,(= 3

0yst0fpystK0p hlhF == ),,(),,,( , ; 0stKfpFlt

>),,,( ; 0stKfpFls

>),,,(

0stKfp FlFi>),,,( 0stKfp

FlFF<),,,( 0stKfp FlK F

>),,,( 0stKfp FlKK FF<),,,( 3a.

Equation 3 defines a well-behaved (quasi-concave with positive partial derivatives and negative

second order partial derivative) production function of the male member of the

family: ),(lMM

mKFY = 4.

where KM be the working capital used by the male member and ml be the labour endowment.13

The model assumed that the family does not own any working capital at the beginning of the

production period and thus it borrows. Given the amount of group loan (B) received by the group

member, the family decides the size of L , the amount borrowed by the family from the market at

the rate of interest r and it is only used in the production of the male member of the family.

Secondly let k1 be the fraction of group loan B used for production activity of SHG member and

k2 be the fraction of this loan is used for the production purpose of the male member of the family

and rest is used to finance the transitory unplanned consumption expenditure.

11 The model can be easily modified to take account of the wage earning activities along with the self

employment project for female members.. 12 The social capital refers to refers to connections within and between social networks. In the present

context by social capital we imply the connection of the representative SHG member with the elected

member of local self government (village panchayat member) or village resource persons who can take

decisive roles in distribution of different types of government subsidized credits, seeds, fertilizers, or

arrange training programmes for SHG members, or recruit them in different government projects like

cooking of midday meals in village primary schools etc. This variable is included to take account of the presence of Electoral clientelism which is a major problem of economic development in developing

economies. 13 If the male member does not have any self employment, he will work for some fixed wage rate and earn

a fixed income per month or he can do a combination of both. But for simplicity it is not presently

considered.

9

Hence

BkLK 2M += 5

BkK 1F = 6

B)kk1(c 21T −−= ; 0cT ≥ 7.

1kk 1k0 1k0 2121 ≤+≤≤≤≤ 8

The income of the family is given as:

L)r1(B)i1(y)t,s,f,K(p)m,K(FY hlFlM +−+−+= 9.

It is assumed that the planned consumption is financed through the income.

Lr1Bi1ytsfKpmKFc hlFlMp )()(),,,(),( +−+−+= 10.

The model assumes that there is a two stage decision making process. In the first stage of the

process the loan amount of a representative group member is determined by joint interaction

among the group members. At this stage group members collectively maximize their group

objective function to optimally allocate loan among the group members from the group corpus.

The group objective function is optimized to increase the size of group corpus through interest

earning subject to constrains that utility received by the members must be higher than that of the

case when they are not receiving any loan.

Given the loan amount, we assume that in the second stage of the game, the family maximizes

the utility function given by (1) subject to 2a, 2b, (4), (5), (6),(7), (8) and (10) with respect to

21lTp k,k,f L, ,c ,c 14

The Lagrangean is given as follows:

11. kk1k1k1cBkk1cLr1

Bi1ytsfKpmKFffdfdmdffccU

2142413T212p

hlFlM1lwFlFlMlwTp

)()()())(())(

)(),,,(),(()()()(),,(

−−λ+−λ+−λ+−−−λ+−+−

+−+λ+−β+α−−−=ψ

From the first order conditions we get the following results

Proposition 1

i) The female group member will work for income earning purpose (fl>0 )for

F

h

FlFf

p

df

UdtsfKp

c

Ul

''),,,( β−δ

δ=α−

δ

δ 12.

ii) Female member would be unemployed if

F

h

FlFf

p

df

UdtsfKp

c

Ul

''),,,( β−δ

δ<α−

δ

δ 13.

14 See appendix I for first order conditions.

10

Proof: follows from first order conditions.

Eq (12) implies that when a female worker is involved in earning activity, she equates her net

marginal utility from working labour to net marginal utility that she receives if that labour is

applied in household activities. Otherwise she will remain unemployed as given in (13). This is

possible if probability of success is low for female worker and it increases at a lower rate with

respect to labour. A low probability of success is possible if the person did not receive any

training or she is not well connected to social network which help her to receive different

subsidised benefits or marketing of her products. Secondly ifβ is high relative to α , that is

perceived disutility of labour endowed for income earning purpose is higher than that of labour

endowed for household work purpose, then also the female member will not work.

Proposition 2

i) Entire amount of loan taken from the SHG will used for female family production if

ypK

FhK

M

F<

δ

δand Byp

c

U

c

UhK

pT

Fδ

δ<

δ

δ 14.

ii) B is allocated between family expenditures, a part is used as capital of male member and the

rest is used as working capital of the female member if

K

F

c

U Byp

c

U

c

U

Mp

hK

pTF δ

δ

δ

δ=

δ

δ=

δ

δ 15.

iii) Entire B is used as the working capital of the male member of the family if

TMp

hK

pMp c

U

K

F

c

U and Byp

c

U

K

F

c

UF δ

δ>

δ

δ

δ

δ

δ

δ>

δ

δ

δ

δ 16.

iv) The following conditions determine the optimal loan amount of the male family member

)( r1K

F

c

U0L

Mp

+=δ

δ

δ

δ⇒> 17.

Proof: See Appendix 1

Thus family will allocate the loan taken by the SHG member by comparing the marginal utility of

the loan according to different purposes.

Solving the first order conditions optimal values for 21lTp k,k,f L, ,c ,c can be defined as

functions of variables corresponding to socioeconomic and demographic features of the family,

social capital and training related variable of the representative SHG member, amount of

borrowing of the family from the SHG.

Given the following variables as function of B, the equilibrium value of B is determined from the

optimization of the group objective function of the SHG subject to reservation constrains of the

11

agents. Thus optimal borrowing level of a representative agent is determined by the joint

interaction of the agents. Thus optimal value of B for a representative agent is given as follows:

.),.,,(* stFGBB =

Here G be the vector of group related variables like size of the group, age of the group, SHG

member’s status in the group. Thus given the equilibrium value of B as B* the sub-game perfect

values of 21lTp k,k,f L, ,c ,c are given as follows:

),,,(* GstFCcTT

=

),,,(* GstFLL =

),,,(* GstFffll

=

),,,(* GstFkk11

= 18.

),,,(* GstFkk22

=

Here F be the vector containing variables corresponding to socioeconomic and demographic

features of the family including the parameters βα, and s be the social capital and t be the

training related variable of the representative SHG member.

Section II: 2 Empirical Implication of the Theory

The theory presented has several interesting empirical implications. A group member will

undertake income earning activity if and only if it satisfies the utility maximizing condition of the

family. A group member with higher probability of success of self employment project is more

likely to be employed in income earning activity. This probability of success is likely to increase

with training received by them in the micro finance programme or her connection with the social

network which helps her to get subsidized benefits. This in turn implies that these variables have

significant impact on the continuity of employment status of the group member along with

socioeconomic and demographic characteristics of the family and group related variables. An

objective of the subsequent empirical analysis is to identify the set of variables that have

significant impact on the continuity of employment status of the group member. Secondly the

theory also shows that there is always a possibility that loan received by the group member may

be used for family consumption purpose or for business purpose of the male member of the

family or for family business purpose and this allocation decision of loan obtained from the

group will depend on member’s household utility function which in turn depends socioeconomic

and demographic characteristics of the family of the member, training of the member and group

related variables and also on local politics of the region. In the subsequent empirical analysis we

12

will try to isolate the significant set of variables that affects the allocation of group loan for

different purposes and size of such loans.

The socio economic demographic variables that we will include in the empirical analysis

are age of the group member, her education level, religion, family size, her past occupation, land

holding status of the family. Age of the group member, her education level will affect the

parameters corresponding to her perception regarding disutility of labour required for different

purposes or βα, . An educated female member is more likely to be more inclined to work for

income generating activity rather than staying at home. Alternatively an aged female member

may prefer to stay at home due to low productivity. Family size can affect the employment status

of the female family member. A large family size implies more work at home and female worker

has to devote more time for smooth functioning of family affairs than work for income earning

purpose. Religion of the family also affects the employment status of women and the loan use

pattern due to access to social network which in turn can affect the ability to find work, access to

informal child care and business network (Gatripati (2009)). Variables related to economic status

of the family such as land holding mostly affect the loan use pattern. If the husband of the group

member is a farmer, then it is more likely that loan will be taken to meet the agricultural working

capital purpose. Gatripati (2009) points out that land holding may act as a proxy for returns on

self-employment loan which in turn influences the individual’s time use.

In the present model we have assumed that female group member’s connection to local

government authorities via social network may help her to increase productivity. This assumption

is based in the existence of political clientelism in developing economies. The study by Bardhan

et al (2007) based on a survey of 85 villages in 15 districts of West Bengal has shown that the

electoral support for incumbent parties was related to reported benefits derived from gram

Panchayats or local self governments. Given this observation we have tried to find out the impact

of political change at the village level on the continuity of employment status and loan use pattern

of the group member. This aspect has not been covered in the existing literature.

Finally the impact of group related variables will be analysed on the employment status

of the group member and on the loan use pattern. The theoretical model shows that the group

related variables will affect the continuity of employment status through its impact on size of

loan. A larger group may posses a large group corpus, but the number of loan takers may be high

in each period resulting in small the loan size of the each member. This can affect the

productivity of group member and her employment status.

13

In the analysis that follows we first provide the description of the survey and the data and

then we estimate the empirical relations of the employment status, loan use pattern and loan size

of the group members on the socio economic political variables.

Section III: The Survey

The survey was conducted in two phases in the district of North 24 Parganas, West

Bengal. The first round of survey was carried out during October 2005 to March 2006 and again

the same set of members was resurveyed during Octobr 2009 to March 2010.

For the first round of survey data has been collected by simple random sampling in

three stages. The district of North 24 parganas is divided in five subdivisions namely Bongaon,

Barasat, Barackpur, Basirhat and Bidhannagar where Bidhannagar is urban area and was not

covered in our survey. From the rest of the four subdivisions we randomly selected six blocks15

namely Bongaon, Barasat I, Barackpur I, Basirhat I, Hasnabad and Hingalgang. Each block

consists of some Gram Panchayat (GP) areas. So from each block one Gram panchayat area has

been randomly selected. The selected gram Panchayats are Palla(Bongaon), Kashimpur

(BarasatI), Panpur-Keutia (Barackpur I), Nimdaria-Kodalia (Basirhat I), Makalgacha(Hasnabad)

and Dulduli(Hingalgang).

From each gram panchayat area four groups were selected randomly from the list of existing

SGSY groups as provided by District Rural Development Cell (DRDC) of North 24 Parganas

(only seven groups were selected from Hingalgang as the size of three of the selected groups were

less than 8). In this way we had 27 selected groups with 300 members. However during survey

we found that one group in Hasnabad block was male group so that group was not included in the

empirical analysis. So from first round of survey we had data on 290 female group members

covering all four subdivisions of North 24 parganas. Then in the second round of survey during

2009 we were able to trace out 272 female group members who were initially selected. Of these

272 female group members four were single woman and not included in the present analysis.16

Hence the subsequent analysis will be carried out on 268 female married group members based

on data collected during two surveys. .

In the surveys we have collected socio-economic information about the SHG members That

includes information about their religion, age, sex, education, social category (SC/ST), number of

family members, number of dependents, occupation category, husband’s occupation category

15 One block from each of the subdivisions is selected randomly and rest of the two is selected from the rest

of all blocks by simple random sampling method. 16 The single women are excluded from the empirical analysis because the employment decision and loan

use pattern of the single women will be different from that of married family women.

14

their average family income, savings, SHG member’s income, savings, family wealth level ,

level of agricultural land holding, family consumption expenditure, food expenditure, medical

expenditure, number of school drop out. The second survey further collected information about

the size of borrowing of the member from the group corpus; the training received by the member;

the loan use pattern and other loan related information.

It has been observed that functioning of SHGs formed under SGSY in West Bengal depends on

its co-ordination with the elected member of the village panchayat17

and two gram panchayat

employed resource persons( GPRP) who in turn co-ordinate with the panchayat Prodhan18

.

Different information are communicated to SHG members through village panchayat member

that includes information related to works distributed through gram panchayats like cooking of

midday meals at schools, 100 days work projects, distribution of government grants, training of

SHG member at the district or state level, different technical help related to self employment

projects, information related to distribution of inputs like seed, pesticides, poultry and sale and

marketing of SHG products. Under this context due to electoral clientilism political identification

of group members become significant in receiving benefits from gram panchayats. Bardhan et al

(2007) have also observed that in West Bengal the electoral support for incumbent parties was

related to short term benefits derived from gram panchayats. In this context we must mention

that the state of West Bengal had a Panchayat Election in May 2008 that is the period after our

first survey. For this reason during second time survey in 2009 we have collected data about the

political affiliation of village panchayat member and corresponding gram panchayat in the survey

area pre and post panchayat election of 2008 to study the impact of political change at village

level on the activities of the group members.

17 In West Bengal a three tier Pamchayat system of local self Government is in operation. In the state level,

Panchayats & Rural Development Department of the Government of West Bengal is the nodal agency for

implementation, supervision & monitoring of the major poverty alleviation programmes in the rural areas

and at the district level, “Zilla Parishad” is the implementing agency for the same. Under a three-tier system

of democratic decentralization, “Zilla Parishad” is the apex body at the district level; followed by

“Panchayat Samitis” at the block level as the second tier; and “gram panchayats (GP),” the third tier and

members of which will be elected by the voters who elect members to the State Legislative Assembly for

four years. The number of members in a GP varies between 7 and 25. Basically a Gram Panchayat is

formed of a whole Mouja or a part of it or more than one mouja together. Three to four villages build up

one gram panchayat. 18 For SGSY groups at district level DRDC acts as the nodal authority at the District level. At the Block

level, the Block Development Officer, along with SHG supervisor and block resource persons maintain the link between the DRDC and the gram panchayats who directly co ordinates with group members at the

village level through elected village panchayat member and gram panchayat resource persons. Sometimes

at the village level along with the gram panchayat member, NGOs also take important role in nurturing of

the groups. From 2006 onwards at the state level in West Bengal the ministry of Self Help Group and Self

Employment acts as the central coordinator of SHGs operating across different districts.

15

Using information on occupation of female group member we have divided the sample into three

occupational groups: self-employed (women with occupations like basket making, bidi binding,

sewing, handicraft work, vegetable grower small sellers of fish, flowers, etc) that is respondents

who work for themselves, wage labourer, (people working in agricultural and non agricultural

activity against wage payment) and animal husbandry (like poultry farming, cattle breeding etc.)

Table 1 gives the occupational pattern of the respondents for 2005 and 2009.

Table 1: Distribution of the Group Members Across Different Occupational Categories.

Source: As reported during two surveys.

Note: Column totals give numbers for 2005 and row totals provide 2009 numbers.

Table 1 represents the distribution of occupational category for two surveys. The distinct feature

is that during first survey all the respondents were employed, however during the second survey

we found that 55% of the respondents have turned into housewife without any employment.

Table 2 provides the descriptive statistics of the socio-economic variables:

Table 2: Descriptive statistics

N Mean Std. Deviation

Annual Price Adjusted Family income( in Rs) of

2009*

268 6802.70 3327.970019

Annual price Adjusted Family Income(in Rs.) for

2005*

268 4851.70080 1653.09038

Number of Family Members 268 3.72 1.01

AGE_OF_SHG Member as on second survey 268 36.22 8.98

LOAN AMOUNT IN 2009 (In Rs) 268 2317.54 2508.2

Source: As reported in the survey.

*At Base: July 1986 – June 1987

19 Further it has been observed that family income has significantly increased at constant prices during the

second survey.(Table A2-I of Appendix 2)

Occupation Categories Labour 05 Self employed 05 Animal Husbandry 05 Total

labour09 19 11 10 40(16%)

Self employed 09 1 53 11 64 (24%)

Animal Husbandry 09 2 2 11 15(5%)

Unemployed housewife09 46 47 55 148(55%)

Total 68 (35%) 113 (42%) 87(33%) 268(100%)

16

Thus from table 1 and table 2 it has been observed that even if the mean family income level of

the group members has significantly increased, a large number of group member has became

unemployed. From the survey it has been observed that of 268 group member income has

increased for 195 group members. But among them 50% members are currently unemployed.

Among 73 families where income has actually decreased from 2005 to 2009, almost 70% are

unemployed. Thus lower income families have more unemployed women. 20

During the second round of survey information about the size of borrowing of the member from

the group corpus, the loan use pattern and other related information were collected. On the basis

of collected data it has been observed that loans were used for different purpose such as

household purpose including education expenditure of children, medical expenditure,

maintenance and repair of house, marriages, buying of ornaments etc. These types of loans are

clubbed under “Family Consumption and family asset creation purpose or family expenses”.

Secondly loans are taken to meet the working capital expenses of the husband of the loan taker or

for family business purpose. This includes the working capital required for husband’s business as

well as working capital required for farming by the farmer husbands. At the same time we have

observed some group members have taken working capital loan for the purpose (such as poultry,

extracting rice from the paddy) other than their reported occupation category. These two types of

loans are categorized as loan taken for husband’s business and other family business purpose.

Finally the loans are taken for the business purpose of the group member and categorized as loan

for self business. This includes loans which have been used as working capital for making

handicraft commodities, basket binding, sewing, poultry business, small shops are extracting rice

from paddy. Table 3a provides descriptive statistics of different types of loans and table 3b

describes the loan use pattern across different occupational group of 2009 for the borrowers.

Table 3a Descriptive Statistics of Different Types of Loans taken during survey

N Mean Std. Deviation Std. Error

No Loan 44 .00 .00 .00

Loan taken to meet Family Expenses 98 2376.53 2781.93 281.02

Loan Taken for Husband's business or other family Business 63 2690.48 2165.40 272.81

Loan Taken for Self business 63 3471.43 2232.78 281.30

Total 268 2317.54 2508.27 153.22

Data collected during 2009 survey

20 See table A2-2 of appendix 2.

17

Table 3b Distribution of Types of loans among different occupational categories

Loan 09

Occupation 09

Family Expenses Husband and other family

Business

Self Business Total

Housewives 82 35 0 117

Labour 13 19 0 32

Animal Husbandry 0 4 11 15

Self Employed 3 5 52 60

Total 98 63 63 224 Source: As reported during 2009 survey.

The table 3 depicts an interesting feature. The unemployed housewives are also taking loan from

SHGs but they are using it mostly for the family expense purpose whereas the self employed

women are using it as working capital in their own business.

Thus the resurvey brings out the following observations:

A large number of group members have become unemployed even though the average family

income of the resurveyed sample has increased over time.

Group members have taken loans in 2009 but around 44% of such loans are used to meet

different family expenditures. The unemployed group members had used the loan mostly (around

70%) to meet family expenses and rest they had used to meet working capital need of their

husbands.

Section IV: Empirical Analysis

Given the theoretical structure and the empirical observations of the survey we would estimate

the factors determining the employment status of the group member, size of their loan and loan

use pattern of such loans.

Section IV-A: Estimation of the determinants of employment status of the group members:

Empirical model and Estimation Procedure

This subsection presents the empirical analysis on the determinants of the employment status of

group members. In the survey we have observed that many self employed women who are

perusing their activities from home, did not give clear answer on exact allocation of time between

the household activities and income earning activities. This made the estimation of time

allocation model of the SHG members impossible. Hence we have tried to isolate the factors

determining the employment status of the group member using a logistic regression model where

it is assumed that a group member with zero income is unemployed and allocating her time

between household activities and leisure. The dependent variable, employment status takes the

18

value one if the member is employed during 2009 survey or otherwise zero. The dummy

variables21

are as follows: The socio economic demographic variables includes religion, number

of family member, dummy related to education level of the group member, age of the SHG

member, and her employment category in 2005. As already mentioned in 2005 we have found

that all group members were employed and they were under three broad occupational classes:

labour, animal husbandry and self employed. In the regression analysis the first class has been

used as reference category. We have included two more economic variables in the model: income

change dummy and agricultural land holding dummy. Income change dummy assumes a value

equal to one if the family income of the group member in 2009 has increased at constant prices

compared to that of 2005 level. Inclusion of this variable will capture the impact of the economic

status of the family on the continuity of employment status of the group member. Finally we have

included one subjective variable namely status of the group member in the family. This variable

assumes value equal to zero if the group member answered that her status in the family is lower

than her husband, it assumes value equal to one if the group member has equal status in the

family and it is two if she has higher status than her husband. The size of SHG is also included in

the regression analysis to take account of the group related impact on the employment status.

Finally to capture the impact of political change on the employment status of the group member

we have included one dummy variable “Village Panchayat Member Change Dummy”. The

village where the group member lives takes value one if the political party of the village

panchayat member corresponding to that particular village has changed during 2008 panchayat

election. The value zero of the dummy indicates that the political party at the village level has not

changed between two subsequent surveys. The change in political party for the panchayat village

member from the village indicates political instability of the corresponding village. This variable

will help us to understand the impact of change in village political scenario on the employment

status of the group member. Section III also explains the importance of the village level politics

on the activities of the group member. In this context we must mention that we have not included

the training dummy in the analysis as it is significantly and negatively correlated with the village

panchayat change dummy. Most interestingly it has been observed that in the villages where the

village panchayat member has changed during 2008 panchayat election, most of the group

members have not received training since inception of the group.22

The members have been

deprived of training in the politically unstable villages.

Table 4 gives the results of regression analysis:

21 Appendix , A3 gives the description of dummy variables. 22 See Appendix 2 table A2-3.

19

Table 4: Results of Logistic Regression

Dependent Variable: Employment status= 1 of SHG member is employed

0 otherwise

Number of observation 268

Explanatory Variables Co-efficient Significance Marginal_Effects

Group Size 0.110851 0.276 272252

Religion_dummy 0.175123 0.653 431663

Income Change Dummy 0.850402* 0.007 1991839

Education_Dummy 0.244266 0.477 602411

Village Panchayat Member Change Dummy -1.75093* 0 4112524

Age -0.02698 0.122 66256

Number of family Members -0.22097 0.138 542716

Self_Employed_05 1.68993* 0 3975084

Animal_Husbandry_05 1.129714** 0.023 2748515

Land_dummy 0.212271 0.57 524318

Status_Dummy_1 -0.09397 0.868 231488

Status_Dummy_2 0.171746 0.789 0.042462

Constant -0.45349 0.78

Log likelihood -150.6489

Pseudo Rsquare 0.182

*Significant at 1% level, ** significant at 5 % level.

Note: The estimates are robust in nature.

Table 4 represents that income change dummy has positive and significant effect on employment

status of group member. This implies that for those members whose absolute family income has

increased over the period 2005-2009 are more likely to remain employed.

The negative and significant village panchayat member change dummy infers that members

belonging to the villages where political affiliations of panchayat members have changed during

the last panchayat election are getting unemployed. Table A2-3 of the appendix also shows that

the political party in power in villages has changed mostly where group members did not receive

any training. These results are in line with the explanation regarding political clientilism

discussed in section 3. As already discussed distribution of different benefits to group members

depends on the connection of group members with village panchayat member and political

identification of the group members become crucial in this respect due to electoral clientelism.

The panchayat member of the villages has not changed where the members have received training

and other benefits from the respective panchayat members of their villages. So in the politically

stable villages members are found be more employed than other villages. In villages where

people did not receive any training after SHG joining, the gram panchayat representatives of

those villages have been mostly replaced by new members from opponent parties. In villages

where village panchayat representative has changed after election of 2008, only 30% of group

20

members are employed. The change in village panchayat representative may be causing lack of

coordination between SHGs and gram panchayats. Thus reducing the profitability of projects of

SHG members, and resulting in unemployment.23

In this respect, we have another interesting

observation. Among those employed group members from villages where panchayat

representative has changed, for 77.5% group members the village panchayat representative

belongs to the same party as that gram panchayat.24

As we have already mentioned village

panchayat representative coordinates between the gram panchayat and group members. If same

party rules at both levels, the group members would be benefited. Study by Swain and Adel

Varghese(2009 ) stated the importance of training on asset formation of group members, but the

impact of local village politics was not taken into account. Thus our study brings out the impact

of local political dynamics on the employment status of the group members which has not been

covered in the existing literature.

Finally the occupation dummies of 2005 are positive and significant which means that the group

members who were self employed and in animal husbandry in 2005 are more likely to remain

employed compared to members who were in the labour class in 2005.

Section IV-B: Estimation of the determinants of loan size of the group members:

Empirical model and Estimation Procedure

In this subsection we will try to estimate the impact of different variables on the size of loans

taken by the group members from group corpus during 2009 survey. All the 268 members have

not borrowed so the loan amount of these non-borrowers is zero. Our sample is a classic

example of censored data and OLS method of estimation or any of its variants cannot be

used, as they will give biased and asymptotically inconsistent estimates of the parameters.

So our obvious choice of methodology is Tobit regression analysis. We have the sizes of

formal loans taken during 2009 survey as the dependent variable, which take either a

positive value or a value of zero. Explanatory variables are same as that of logistic regression.

Table 5 presents the results of tobit regression analysis:

Table 5: Tobit _Model

23 It has been observed that among the training recipients, village panchayat member has changed for 12 group members only and all of them had become unemployed. 24 It has been observed that village panchayat representative has changed for 189 group members. Among

them ordinal correlation coefficient between employment status and dummy variable corresponding to

same party in village and gram panchayat level shows positive (.287) and significant (pvalue=0.00) ordinal

correlation.

21

Dependent Variable is the amount of loan by a group member

Coefficient Significance

Group Size -563.721* 0.002

Religion_dummy 284.371 0.499

Income Change Dummy -155.127 0.68

Education_Dummy 19.396 0.962

Village Panchayat Member

Change Dummy -1468.063* 0

Age -17.609 0.294

Number of family Members 115.892 0.528

Self_Employed_05 1107.721* 0.005

Animal_Husbandry_05 -59.434 0.89

Land_dummy 1196.358** 0.012

Status_Dummy_1 1176.886** 0.028

Status_Dummy_2 424.641 0.447

Constant 7669.999 0

Number of observation 268

Pseudo Rsquare

0.0147

Correlation between predicted loan amount and actual loan

amount

.5

*Significant at 1% level,** Significant at 5% level.

Note: Estimates are robust in nature

From the tobit model it is observed that the size of a SHG negatively influences the loan

amount. Higher the number of the group members lower will be the amount of the loan

per member.

The village panchayat representative change dummy is negative which states that as the

panchayat member of the village changes the loan amount declines. Table 4 brings out

the fact that change in the political regime in the villages after panchayat election of

2008, has increased incidence of unemployment among group members. Thus Tobit

result is emphasizing the fact that these unemployed group members are taking lower

amount of loans also25

. This is in direct confirmation with the results obtained by

Bardhan et al (2007). They have stated that politics at grass root level promotes

clientelism by provision of small repetitive benefits. So when the panchayat member

from a particular village changes the probability of getting subsidized benefits like loan

from self help groups declines for the existing self help group members as the groups had

been formed and nourished previously by the opponent political party. So at present with

25 Table A25 in the appendix 2 shows that average loan size for employed members is significantly higher

than that of unemployed group members.

22

the change in the political reign of one party the existing members are in disadvantageous

position as they have enjoyed the benefits previously from the other political party. This

is increasing the incidence of unemployment and unemployed members are taking small

sized loans because the loans will now have to be repaid with the income of the other

earning members of the family.

Further the occupation dummy, self employed in 2005 is positive and significant. This

infers that the people who were in self employment are taking higher amount of loan

than the people who were labourers. Table 1 shows that around 68% of people who were

labourers are in 2005 have become unemployed housewives whereas 59% of people who

were self employed in 2005 are in employment in 2009 also. Thus again this result

signifies the fact that unemployed housewives are taking lower amount of loans from

SHGs than the employed group members.

Moreover the land ownership increases amount of loan taken from the SHGs. It has been

observed many group members from land owing families are taking loan to meet their

husband’s working capital need in agriculture. Secondly family land holding may be

acting as pseudo collateral to get higher amount of loan.

The status dummy _1 is significant which implies if the status of the woman member is

same as that of her husband within the family then the loan amount increases compared

to the group of members whose status in the family is lower than that of her husband.

Next section will try to find out the determinants of loan use pattern by the group

members.

Section IV-C: Estimation of the determinants of loan use pattern of the group members:

Empirical model and Estimation Procedure

Table 3a and 3b show that the loans are utilized mostly for three purposes: to meet

different types of family expenses, to meet the working capital need of the husband and

other family business which is not dominant occupation of group member and finally to

meet the working capital need of the group member herself. In this case we use a

23

multinomial logistic regression26

model to identify the determinants of loan use pattern.

The explanatory variables are same as the earlier cases.

Table 6 summerises the result:

Table 6: Multinomial Model

Multinomial Logistic Regression

Loan for family Expenditure is the reference category.

Coefficients Significance

Loan used for husband’s Business

and Family Business Purpose

Group Size -0.108 0.414

Religion_dummy -0.441 0.315

Income Change Dummy 0.280 0.484

Education_Dummy 0.599 0.119

Village Panchayat Member

Change Dummy -1.205** 0.018

Age -0.020 0.396

Number of family Members 0.317 0.101

Self_Employed_05 -0.579 0.267

Animal_Husbandry_05 -0.424 0.4

Land_dummy 1.324* 0.002

Status_Dummy_1 -0.159 0.808

Status_Dummy_2 -0.979 0.263

Constant 0.968 0.652

Loan for self Business

26 Multinomial logistic regression analysis requires that the dependent variable be non-metric.

Dichotomous, nominal, and ordinal variables satisfy the level of measurement requirement. Multinomial

logistic regression is used to analyze relationships between a non-metric dependent variable and metric or

dichotomous independent variables. Here in this case the dependent variable is qualitative in nature that is

pattern of loan use. There are three categories of loan use. The loan use for the family expenses has been

taken as the base category. Multinomial logistic regression provides a set of coefficients for each of the

two comparisons. The coefficients for the reference group are all zeros, similar to the coefficients for the

reference group for a dummy-coded variable. Thus, there are three equations, one for each of the groups defined by the dependent variable. The three equations can be used to compute the probability that a

subject is a member of each of the three groups. A case is predicted to belong to the group associated with

the highest probability.

24

Group Size 0.109 0.398

Religion_dummy -0.441 0.497

Income Change Dummy 1.298* 0.003

Education_Dummy -0.373 0.469

Village Panchayat Member

Change Dummy -2.765* 0

Age 0.002 0.93

Number of family Members -0.205 0.489

Self_Employed_05 3.647* 0

Animal_Husbandry_05 2.704* 0.007

Land_dummy 1.112 0.05

Status_Dummy_1 0.467 0.465

Status_Dummy_2 1.019 0.16

Constant -3.723 0.097

Number of observation 224

Loglikelihood -173.712

Pseudo Rsq 0.2787

*Significant at 1% level ,** Significant at 5% level.

Note: Estimates are robust in nature

The base category is the loan used for family expenses. The variables significant for

loans used for husband’s business or family business purpose are village panchayat

member change dummy and land dummy.

Land dummy positively influences the loan use for husband’s business and family

business purpose compared to family expenses. It has been observed in the survey that

many group members are transferring their loan to meet the agricultural working capital

need of their husband who is a small farmer. This result is in line with the result we

obtained in the tobit regression analysis which shows that the amount of loan also

increases with the holding of agricultural land. Thus the families which hold agricultural

land are more likely to use the loan taken by the member for husband’s business( here

loan use by farmer husbands for farming purposes are also included) and family

business purposes than to use it for family expenses.

The village panchayat member change dummy is negative and significant. This implies

that as the panchayat member from the village changes the probability of using the loan

for husband’s business and family business purpose decreases compared to that of using

the loan for family expenses. These results are in confirmation with the results obtained

from last two subsections where we have observed that as village panchayat member

changes, group members have become unemployed and they have borrowed small sized

25

loans. Now the multinomial regression results have shown that those loans were mostly

used to meet different family expenses.

The variables significant for self business loans are village panchayat member

change dummy, income change dummy, occupation dummies and the land holding

dummy.

The village panchayat member change dummy is negative and significant. This is in

confirmation with previous results. This states that if the panchayat member from the

village changes the existing self help group members stand to loose. The literature on

political dimension and the public service delivery has shown that the party members

always shower benefits on the followers of the parties. At the village level the panchayat

member representing a village has the power to provide benefits to the villagers. The self

help group members may get benefits from the panchayat member from the village. But

as the party changes the existing group members may not get benefits any more as they

have already received benefits from the opponent party. So the social benefits get

diminished and so the probability of using the loan for self business declines compared to

that of family expenses.

The income change dummy is positive and significant. This infers that the positive

change of family income over a period of time has positive influence on use of loan for

self business purpose.

The land holding also positively influences the probability of loan use for self business.

Basically employed group members who are taking loan for self business purpose have

significantly higher average amount of loan than the unemployed group members.27

In

this case family land holding may be acting as pseudo collateral to get large amount of

loan hence positively influencing the probability of taking a self business loan against

family expense loan.

The probability of the loan use for the self business purpose increases relative to that of

the family expenses if the members are self-employed or in animal husbandry during

2005. As already mentioned this type of loan is taken by the group members who are

presently in self-employment or in animal husbandry. Table 1 shows that people who

27 Analysis shows that average size of loan taken for self business purpose is significantly higher than that

of other types of loans.

26

were in these two occupational categories are mostly employed in 2009. Most of the

people who were in self-employment had remained so and some people from the animal

husbandry category had shifted to self employment in 2009 also. Hence follows the

results.

Section V: Conclusion

This paper has tried to identify the factors responsible for continuity of employment status of the

woman members of the self help groups considering maximization of the intra household utility

function. The decision of the woman to work depends on probability of success of her

employment activity. The probability of success depends on various socio economic demographic

factors of the household as well as on social benefits. The social benefits depend on local politics.

The most striking result that has been obtained here is that the local politics plays a very crucial

role in determining the employment status of the members. The political stability of the village

ensures employment of the members whereas, as the village representative changes from a

particular village the employment of the self help group members gets affected. Moreover local

politics also affects use of loan. The political stability in terms of no change of village

representative from a particular village during panchayat election 2008 positively affects the size

of the loan amount as well as positively influences the probability of loan use for self business or

for husband’s business or for family business compared to that of the loan use for family

expenses.

The other factor that influences the employment status of the member is the past

occupation. Those women who were working as labourers during 2005 survey have mostly

become unemployed compared to other categories of profession.

The self help groups which were formed under SGSY programme with the objective of

providing financial assistance to the poor woman to engage them into economic activities have

failed in many cases. In order to fulfill this objective the operation and functioning of the self help

groups have to be devoid of any local politics.

27

Appendix I

First Order Conditions of maximization:

The Lagrangean is given as follows:

A1. kk1k1k1cBkk1cLr1

Bi1ytsfKpmKFffdfdmdffccU

2142413T212p

hlFlM1lwFlFlMlwTp

)()()())(())(

)(),,,(),(()()()(),,(

−−λ+−λ+−λ+−−−λ+−+−

+−+λ+−β+α−−−=ψ

The first order conditions are as follows:

1

pp c

U 0

cλ=

δ

δ⇒=

δ

δψ A2

0c

c c

U0

c T

T 2

TT

=δ

δψλ≤

δ

δ⇒≤

δ

δψ A3

F

h

FhlFf

pl

df

UdytsfKp

c

U0

f l''),,,( β−

δ

δ≤α−

δ

δ⇒≤

δ

δψ 0

ff

l

l =δ

δψ A4.

0L

L r1K

F0

L M

1 =δ

δψ+≤

δ

δλ⇒≤

δ

δψ)( A5.

0k1

≤δ

δψ 0

kk 0Byp

1

1532hK1 F=

δ

δψ≤λ−λ−λ−λ A6.

28

0k 2

≤δ

δψ 0

kk 0B

K

F

2

2542

M

1 =δ

δψ≤λ−λ−λ−

δ

δλ A7

01

=δλ

δψ 0cLr1Bi1ytsfKpmKF phlFlM =−+−+−+ ))()(),,,(),(( A8

02

=δλ

δψ 0cBkk1

T21=−−− ))(( A9.

0)k1( 0)k-(1 0 131

3

=−λ≥≥δλ

δψ A10.

0)k1( 0)k-(1 0 242

4

=−λ≥≥δλ

δψ A11.

0)kk1( 0)kk-(1 0 21521

5

=−−λ≥−≥δλ

δψ A12.

Proof of the Proposition 2

From First order Conditions if f1=0, from the properties of the production function we must have

k1=0

For f1 >0 we have k1>0.

Under this situation there can be two cases:

1kor 1k11

<=

If 1k1

= , the entire loan taken by the female group member is applied for his own production

purpose, thus k2=0 , cT=0, 0 043

=λ>λ and 05

>λ

Thus k1=1 and k2=0 implies that

c

U 2

T

λ<δ

δ

0Byp 532hK1 F=λ−λ−λ−λ

0BK

F52

M

1 <λ−λ−δ

δλ

This further implies that

BypBK

F3hK152

M

1 Fλ−λ=λ+λ<

δ

δλ

Or ypK

FhK

M

F<

δ

δ

The marginal product of capital of male output is less than the marginal product of female output

29

And Bypc

U

c

UhK

pT

Fδ

δ<

δ

δ

That is marginal utility of consumption of cT is less than marginal utility of consumption of cp

when one unit of borrowed capital is used for female production and that raises female output and

family income and hence cp by the amount Byp hK F

Using the same logic, in equilibrium when 1kk ,0k ,0k 2121 <+>>

K

F

c

U Byp

c

U

c

U

Mp

hK

pTF δ

δ

δ

δ=

δ

δ=

δ

δ A13..

In this case the loan is allocated between family expenditures, a part is used as capital of male

member and the rest is used as working capital of the female member.

Alternatively, the entire loan is used for consumption purpose (i.e. k1=0 and k2=0) if the

following conditions are satisfied.

K

F

c

U

c

U

Bypc

U

c

U

MpT

hK

pTF

δ

δ

δ

δ>

δ

δ

δ

δ>

δ

δ

A14.

Finally, entire loan is used as the working capital of the male member of the family (i.e.k2=1 ) if

TMp

hK

pMp

c

U

K

F

c

U

Bypc

U

K

F

c

UF

δ

δ>

δ

δ

δ

δ

δ

δ>

δ

δ

δ

δ

A15.

Last part of the Proposition follows from A5.

Appendix 2

Table A21: Paired sample t test for change in mean Adjusted family income level

for 2005 and 2009

a)Paired Samples Statistics

Mean N Std.

Deviation

Std. Error

Mean

Pair 1 Income 09 6802.702 268 3327.97 203.29

Income 05 4851.70080 268 1653.09 100.97857

b) Paired Samples Correlations

N Correlation Sig.

Pair 1 Income 09-

Income 05

268 .061 .316

30

c) Paired Samples Test

Paired

Difference

s

t df Sig. (2-

tailed)

Mean Std.

Deviation

Std. Error

Mean

95%

Confidence Interval of

the

Difference

Lower Upper

Pair 1 Income 09-

Income 05

1951.0007

3

3623.797 221.359 1515.170 2386.831 8.814 267 .000

Table A2-2

Cross tabulation and correlation between employment status dummy and income change dummy

Values of

the dummy

variable

EMPLOY_

S

Total

0 1

Dummy related to

income change

0 51 22 73

1 97 98 195

Total 148 120 268

b) Correlations Value Asymp.

Std. Error

Approx. T Approx.

Sig.

Ordinal by Ordinal Kendall's

tau-b

.180 .058 3.060 .002

N of Valid Cases 268

a Not assuming the null hypothesis. b Using the asymptotic standard error assuming the null hypothesis.

Table A2-3

Cross tabulation and correlation between training dummy and village panchayat member change dummy Count

TRAINING Total

0 1

Village member

change dummy

0 38 67 105

1 151 12 163

Total 189 79 268

Value Asymp.

Std. Error

Approx. T Approx.

Sig.

Ordinal by Ordinal

Kendall's tau-b

-.604 .049 -10.571 .000

N of Valid 268

31

a Not assuming the null hypothesis.

b Using the asymptotic standard error assuming the null hypothesis.

Table A2-4

Cross tabulation and correlation between training dummy and employment Status Dummy TRAINING Total

0 1

Employment

Status

0 125 23 148

1 64 56 120

Total 189 79 268

Symmetric Measures Value Asymp.

Std. Error

Approx. T Approx.

Sig.

Ordinal by Ordinal

Kendall's tau-b

.339 .057 5.705 .000

N of Valid

Cases

268

a Not assuming the null hypothesis.

b Using the asymptotic standard error assuming the null hypothesis.

Table A2-5: Comparison of loan size for employed group members and unemployed

group members

Group Statistics EMPLOY_S N Mean Std.

Deviation

Std. Error

Mean

LOAN 1 120 2853.33 2595.73 236.96

0

148 1883.11 2355.52 193.62

Cases

Levene's

Test for

Equality of

Variances

t-test for

Equality of

Means

F Sig. t df Sig. (2-

tailed)

Mean

Difference

Std. Error

Difference

95%

Confidence

Interval of

the Difference

Lower Upper

LOAN Equal 5.407 .021 3.203 266 .002 970.23 302.91 373.81 1566.64

32

Independent Samples Test

Appendix 3

Religion Dummy- It takes value one if the woman is Muslim, 0 other wise.

Out of 268 samples 82 women belongs to Muslim community and rest belong to Hindu families.

Income Change Dummy- The difference of the absolute value of family income for period two

years 2009 and 2005 was obtained. If the difference is positive then the value takes 1 and 0

otherwise. There are 195 members for whom the absolute income change is positive.

Village Dummy- This dummy takes the value of one if the representative member of the village

changed during 2008 panchayat election. There are 163 observations for whom the village change

dummy is positive.

Training Dummy= 1 if the members got training during the period of membership from 2005

to 2009.

=0 , otherwise.

There are 79 members who have got training.

Occupational Dummy:

The interviewed persons are classified in four broad occupational categories namely Self

Employed.) Farmer, Laborer (the wage earners) and Business (small business establishments like

shop owners, sellers of different goods, contractors etc.) Occupational dummies are defined as

follows:

Labourer_05=1 if the interviewed person is non agricultural wage or agricultural labourers during

the survey 2005.

=0 otherwise.

In the survey it is found that 68 individuals are in this category.

Self Employed-1 if the members are self employed (covering occupations like bidi binding, van

driving, basket binding, stitching and embroidering, etc). There are in total 113 women in this

category.

Animal husbandry =1 if the members are engaged in animal husbandry.

There are 87 members under this category.

Education above =1 if the interviewed person has above primary education level

variances

assumed

Equal

variances

not

assumed

3.171 243.194 .002 970.23 306.00 367.47 1572.98

33

=0 otherwise.

86 individuals who have education above primary level.

Land holding=1 if the family of self help group member holds some agricultural land.

=0 otherwise

In the survey it is found that 64 families who hold some agricultural land among borrowers.

Status_Dummy_1=1, if the woman has same status in the family than husband .

=0 , otherwise,

Status_Dummy_2=1 , If the woman has higher status as that of her husband in the family

=0, otherwise.

34

Appendix 4

Correlations: Kendall's tau_b group size RELIGION DUMINCHA EDU_DUMM VILLAGE AGE_OF_S FAMMEM SE05 AGRI05 LAB05 LAND_DUM Status in

Family

group size Correlation

Coefficient

1.000 .290 .158 -.100 -.367 -.030 .189 -.010 -.332 .317 -.143 -.046

Sig. (2-tailed) . .000 .004 .066 .000 .509 .000 .851 .000 .000 .009 .386

RELIGION Correlation

Coefficient

.290 1.000 .133 .081 -.214 -.152 .216 -.272 -.020 .301 -.049 -.214

Sig. (2-tailed) .000 . .029 .184 .000 .003 .000 .000 .742 .000 .423 .000

DUMINCHA Correlation

Coefficient

.158 .133 1.000 .097 -.182 -.003 .207 -.055 -.090 .126 .067 -.025

Sig. (2-tailed) .004 .029 . .111 .003 .952 .000 .372 .141 .040 .270 .670

EDU_DUMM Correlation

Coefficient

-.100 .081 .097 1.000 .191 -.176 -.047 -.020 .087 -.015 .234 .128

Sig. (2-tailed) .066 .184 .111 . .002 .001 .405 .739 .156 .805 .000 .031

VILLAGE Correlation

Coefficient

-.367 -.214 -.182 .191 1.000 -.058 -.154 -.151 .314 -.147 .234 .142

Sig. (2-tailed) .000 .000 .003 .002 . .258 .006 .014 .000 .016 .000 .017

AGE_OF_S Correlation

Coefficient

-.030 -.152 -.003 -.176 -.058 1.000 -.080 -.007 .080 -.059 -.009 -.056

Sig. (2-tailed) .509 .003 .952 .001 .258 . .092 .896 .121 .249 .865 .258

FAMMEM Correlation

Coefficient

.189 .216 .207 -.047 -.154 -.080 1.000 .016 -.136 .133 .030 -.097

Sig. (2-tailed) .000 .000 .000 .405 .006 .092 . .773 .016 .018 .594 .078

SE05 Correlation

Coefficient

-.010 -.272 -.055 -.020 -.151 -.007 .016 1.000 -.669 -.498 -.124 -.009

Sig. (2-tailed) .851 .000 .372 .739 .014 .896 .773 . .000 .000 .043 .883

AGRI05 Correlation -.332 -.020 -.090 .087 .314 .080 -.136 -.669 1.000 -.192 .156 -.028

35

Coefficient

Sig. (2-tailed) .000 .742 .141 .156 .000 .121 .016 .000 . .002 .011 .639

LAB05 Correlation

Coefficient

.317 .301 .126 -.015 -.147 -.059 .133 -.498 -.192 1.000 .015 .027

Sig. (2-tailed) .000 .000 .040 .805 .016 .249 .018 .000 .002 . .802 .651

LAND_DUM Correlation

Coefficient

-.143 -.049 .067 .234 .234 -.009 .030 -.124 .156 .015 1.000 .015

Sig. (2-tailed) .009 .423 .270 .000 .000 .865 .594 .043 .011 .802 . .798

Status in

Family

Correlation

Coefficient

-.046 -.214 -.025 .128 .142 -.056 -.097 -.009 -.028 .027 .015 1.000

Sig. (2-tailed) .386 .000 .670 .031 .017 .258 .078 .883 .639 .651 .798 .

N 268 268 268 268 268 268 268 268 268 268 268 268

36

References

Banerjee (Chatterjee), T (2009), “Economic Impact of Self Help Groups—A Case

Study”, Journal of Rural Development, Vol 20, No. 4, pp.451-467

Bardhan, P, S. Mitra, D. Mookherjee and A. Sarkar (2008), “Political Participation,

Clientelism And Targeting of Local Government Programs: Analysis of Survey Results from

Rural West Bengal”, India, Discussion Paper, ERU/2008-03 April 15, 2008,

http://www.isical.ac.in/~eru/erudp/2008-03.pdf

Garikipati, S (2009), “My Money But Not My Time: Microcredit and Woman’s Time-

Use”,wp2009/41,http://www.liv.ac.uk/managementschool/research/working%20papers/working_

papers.htm

Garikipati, Supriya(2008), “The Impact of Lending to Women on Household

Vulnerability and Women’s Empowerment: Evidence from India” , World Development ,Vol. 36,

No. 12, pp. 2620–2642.

Imai, S Katsushi , Thankom Arun and Samuel Kobina Annim(2010), “Microfinance and

Household Poverty Reduction: New Evidence from India”, doi:10.1016/j., World

Development,.2010.04.006.

Khandker, S (2003), “Micro-Finance and Poverty Evidence Using Panel Data from

Bangladesh”, Policy Research Working Paper 2945, World Bank, Research Development Group.

Khandker, S. (2005), “Microfinance and poverty: Evidence using panel data from

Bangladesh”, World Bank Economic Review, Vol- 19, No. 2, pp.263–286.

McKernan Signe-Mary (2002), “The Impact of Microcredit Programs on Self-

Employment Profits: Do Noncredit Program Aspects Matter?”, The Review of Economics and

Statistics, Vol. 84, No. 1, pp. 93-115

Parida. C Purna and Anusree Sinha (2010) , “Performance and Sustainability of Self-

Help Groups in India: A Gender Perspective”, Asian Development Review, Vol. 27, No. 1, pp.

80–103

Srinivasan, N (2009), Micro Finance India State of the Sector Report 2009, Sage.

Swain Bali Ranjula and Adel Varghese(2009) , “Does Self Help Group Participation

Lead to Asset Creation?” , World Development ,Vol. 37, No. 10, pp. 1674–1682.

Swain Bali Ranjula and FanYang Wallentin(2009), “Does Microfinance Empower

Women?”, International Review of Applied Economics, Vol- 23, No.5,pp. 541-556.

Swain Bali Ranjula and Maria Floro (2007), “Effect of Microfinance on Vulnerability,

Poverty and Risk in Low Income Households”, Working Paper 2007:31, Department of

Economics, Uppsala University, 2007.

37

Tassel, Eric Van (2004),"Household bargaining and microfinance", Journal of

Development Economics, Vol- 74, No. 2, pp. 449-468.

Tedeschi, G. (2008), “Overcoming selection bias in microcredit impact assessments: A

case study in Peru”, Journal of Development Studies, Vol-44, No. 4, pp.504–518.

Yunus, Muhammad(1998), “Banker to the Poor”, Pengiun Books