International Review of Business Research Papers Vol. 3 No. 3 August 2007 Pp

Transcription

1 International Review of Business Researh Papers Vol. 3 No. 3 August 2007 Pp Miroredit Programs and Eonomi Indiators: Are the Higher Inome Borrowers Better Off? Evidene from Bangladesh Sayma Rahman* Miroredit program in Bangladesh provides small loans to rural people espeially to the poor with the purpose of eradiating poverty. This study investigates the impat of miroredit on eonomi indiators of the borrowers and ompares if the impat is the same aross borrowers having different inome levels. Household savings, assets and inome are onsidered as ausal fators that may ontribute towards eradiating poverty. To estimate the impat of miroredit on suh indiators we have used simultaneous equations model. Primary data has been olleted from the Grameen Bank and the Bangladesh Rural Advanement Committee (BRAC) borrowers of some seleted villages from three major distrits in Bangladesh. The Two-Stage Least Squares (2SLS) estimation results show that the miroredit programs is effetive in generating higher inome, assets and savings for the borrowers in general. However, that impat is not found to be uniform aross inome levels of the borrowers and higher inome borrowers seem to be better off ompared to the middle and lower inome borrowers. This study further shows that the age and eduation of the and his/her partner in the family are signifiant in bringing about better household impat. Field of Researh: Eonomis 1. Introdution The miroredit program in Bangladesh is a unique innovation of redit delivery tehnique to enhane inome generating ativities. Its uniqueness is refleted in its ollateral-free group-based lending strategy (Hulme and Mosley, 1996; Yunus, 1999; World Bank, 1994). The program extends small loans to poor people, mainly women, for self-employment ativities thus allowing the lients to ahieve a better quality of life (Hussain, 1998; Morduh, 2000; Rahman, 1995). It is the most sensational anti-poverty tool for the poorest, espeially for women (Miro-redit Summit 1997). Miroredit in Bangladesh has drawn attention of researhers throughout the world beause of its distintive redit delivery system, high reovery rate and its speial fous on women. The Grameen Bank - the largest miroredit institution - and the Bangladesh Rural Advanement Committee (BRAC) - the largest nongovernmental organisation (NGO) - are the pioneers of miroredit in Bangladesh for almost three deades. * Sayma Rahman, University of Western Sydney, Australia sa.rahman@uws.edu.au I would like to thank Professor P. N. (Raja) Junankar and Dr. Girasankar Mallik for their omments on earlier drafts of the paper. Any errors are my responsibility.

2 Rahman 310 The impat of miroredit on poverty alleviation is so far found to be ontroversial in the literature. Several studies have found that miroredit program has a positive impat on eradiating poverty (Hossain, 1988; Khandker, 1998; Wahid, 1993; Yaron, 1994) but other studies, for example Morduh (2000), Weiss and Montgomery (2005), report a negative impat. To substantiate suh a ontroversy, it is important to assess the impat of miroredit programs on eonomi and/or household indiators of the borrowers. In doing so, first, it is neessary to identify fators that are essential measures/indiators of poverty seond, to find out whether these fators are ultimately affeted by these miroredit programs and finally, to ompare the impat aross borrowers with different inome-levels. Khandker (2000) onsiders savings as an indiator and finds that this fator has an influene on eradiating poverty. He argues that redit programs do stimulate savings beause miroredit borrowers make mandatory savings every week, whih they are entitled to withdraw at the end of their membership. In addition, he finds miroredit program has a positive impat in generating not only voluntary savings but also additional savings among the borrowers. Apart from savings, it an be argued that there are other fators that may ontribute towards eradiation of suh poverty. For example, inome and aumulation of assets of the household may be onsidered as additional ausal fators. It is likely that with the introdution of miroredit programs, borrowers may have better inome, better savings and more assets. In this bakdrop, it is neessary to analyse how these miroredit programs an influene inome, savings and assets for the borrowers. In this study we have investigated the impat of miroredit on household indiators using the simultaneous equations model suggested by Pitt and Khandker (1996). The Two-Stage Least Squares (2SLS) estimation is used to estimate the model. We also have divided the borrowers into three groups based on their inome level to assess the impat on borrowers from different inome groups. The underlying researh question for this study is as follows: Do the miroredit programs in Bangladesh improve the savings, inome and assets of the borrowers? In order to examine this we have used primary data from the borrowers of the two major institutions, the Grameen Bank and the BRAC. The rest of the paper is strutured as follows. Setion two provides the bakground literature. Setion three desribes the data and its soures. Setion four speifies the model. The results are disussed in Setion five. A onlusion is drawn in the final setion. 2. Bakground Literature 310

3 Rahman 311 Miroredit program have appeared in the literature as a popular soure of finaning that provides small loans in rural and remote regions of developing ountries (Bornstein, 1998; Johnson and Mosley, 1997; Zeller and Sharma, 2000). As far as developing ountries are onerned, Bangladesh may be onsidered as the pioneer that started this finanial innovation that provides loans to the poor espeially to women engaged in self-employment projets allowing them to generate inome and in many ases, begin to build wealth and eliminate poverty (Hulme and Mosley 1996; Yunus 1983; World Bank 1994). World Bank (Miroredit Summit 1997), lassified the miroredit program in Bangladesh as one of the most effetive anti-poverty tools for the poorest. The program extends small loans to unemployed poor people that are not bankable. These individuals lak ollateral, stable employment and therefore annot meet even the most minimum qualifiations to gain aess to formal redit. Sine loans are provided in a group, aording to Stiglitz, (1993); Varian, (1990); and Ghatak (1999), the members in a group are well plaed to umpire the reditworthiness and to srutinise the ations of their peers, as a onsequene mitigating the problems of adverse seletion and moral hazard. Group lending also gives inentives to the members to avoid exessively risky projets (Stiglitz, 1993). It also provides insurane to other members in the event their projets fail (Coleman, 1999). However, mutual trust among group members reated by their long assoiation with eah other provides strong induement to self-monitoring whih redues the monitoring ost down to zero. Several empirial studies support that redit market involvements improve both onsumption and prodution of the poor via smoothing onsumption and reduing onstraints in prodution (Feder et al., 1988 and Foster, 1995). Even though it is evident from the literature that not all money borrowed is invested by the households, a portion of it is used for onsumption. Therefore, it may be assumed that miroredit may benefit households in terms of inome as well as onsumption. Many studies have attempted to measure the impat of miroredit on poverty, inome, employment, ontraeptive use and fertility (Hossain, 1988; Hulme and Mosley, 1996; Shuler and Hashemi, 1996). One of the limitations of these studies is that they fail to address the ausality effet of program partiipation on the quality of life. Existing literature on the impat of miroredit on poverty in Bangladesh provides a mixed piture. Khandker et al. (1998) have used data from three most important miroredit programs in Bangladesh namely, the Grameen Bank, BRAC and RD-12 projet run by the Bangladesh Rural Development Board (BRDB). The authors have attempted to quantify the village-level impats of these programs using OLS estimates. Their eonometri analysis shows that these programs have positive impats on inome, prodution and employment partiularly in the rural non-farm setor. 311

4 Rahman 312 Pitt and Khandker (1998), in a separate study, have estimated the impat of partiipation by gender in eah of the three group-based redit programs on women and men s labour supply, boys and girls shooling, expenditure and assets using weighted exogenous sampling maximum likely-hood information maximum likelihood-fixed effets (WESML-LIML-FE). They have found that miroredit is a signifiant determinant of many outomes suh as household expenditure, non-land assets held by women, male and female labour supply, and boys and girls shooling. Furthermore, redit provided to women is more likely to influene these behaviours than redit provided to men. Apart from the studies that suggest aess to redit has the potential to signifiantly redue poverty (as mentioned above) others argue that miroredit has a minimal impat on poverty redution (Morduh, 1999 and 2000; Wiess and Montgomery 2005). Morduh (2000) using the same data set but obtains very different result. Morduh s result indiates that program partiipants do not benefit in terms of greater onsumption level, but they partiipate beause they benefit form risk redution. To substantiate suh a ontroversy, it is important to assess the impat of these redit programs on some eonomi indiators. In this study we used inome, asset aumulation and savings patterns of the miroredit borrowers as eonomi indiators. 3. Data In this study we used data from the borrowers of two major miroredit institutions in Bangladesh suh as the Grameen Bank and the BRAC. These two large institutions have overage all over the ountry. Based on different agro-limati and soio-eonomi onditions we have seleted three distrits of Bangladesh to ollet data. The distrits are Gazipur, Dinajpur and Chokoria. Gazipur is lose to Dhaka, the apital ity and Dinajpur and Chokoria are hosen from far north and south part of the ountry to have different soio-eonomi effet. From eah distrits five villages are hosen at random. The borrowers are seleted in a luster from eah village. The samples of borrowers are randomly seleted without replaement from the list of households available from the programs loal offie of eah village. Randomness and soio-eonomi and ultural bakground are used as ontrol variables to failitate better omparison. From all three distrits 387 borrowers were interviewed through a strutured questionnaire. 4. Model Speifiation The primary fous of this study is to estimate the impat of miroredit on various household outomes suh as inome, assets and savings from two major miroredit institutions in Bangladesh, viz., the Grameen Bank and BRAC. 312

5 Rahman 313 To analyse the impat of miroredit using data from borrowers of the above mentioned institutions, we have adopted the model suggested by Pitt and Khandker (1996). Pitt and Khandker (1996) onsider the redit C (endogenous variable) depends on some household harateristis, some villages speifi harateristis and on some other variables. So the model used by them is: C = α + β X + γ V + πz + ε C ( 1) Where X is the vetor of exogenous household harateristis (e.g., some demographi fators of the household), V is the vetor of village harateristis (ommunity infrastruture), Z is also a vetor of a set of household or village harateristis whih are different from the X s and V s in that they affet C. β, γ and π are unknown parameters, and ε is a random error omposed of three omponents ε = η + µ + e ( 2) j j Where η j and µ j are unobservable household-speifi and village-speifi e effet respetively, and is a non-systemati error unorrelated with the other error omponents or the regressors. Aording to Pitt and Khandker (1996), the household outome Y (we have used inome, savings and assets whereas Pitt and Khandker have used savings alone) depends on amount of C as well, whih may be explained as: Y = β X + γ V + δc + ε y y j y...(3) Where β y, γ y and δ are unknown parameters and y ε is the error term. Models ontaining simultaneous equations are estimated through instrumental variables in general. However, maximum likelihood estimates ould also be a possible alternative, whih provides effiient results. In this study we have used 2SLS to estimate the above mentioned simultaneous equations model. Speifiation of the Instruments and the Variables In the above-mentioned model, the exogenous regressors Z are the instrumental variables. General norms of hoosing instruments are to use variables, whih are unorrelated with the error term but orrelated with the explanatory variables (independent variables). Vetors of the dependent and 313

6 Rahman 314 independent variables, as well as the instrumental variables that are used in the model are speified below. Desription of the Variables In this study we onsider that the amount of borrowing depends on demographi variables suh as the age of the male and the age of the partner of the. It may also depend on the eduation of both female and male. Unlike Pitt and Khandker (1996), we onsider the number of adult male/female as a perentage of family size (MAPFS/FAPFS) instead of the number of adult male/female (ADMALE/ADFEM). Sine family size is likely to vary, we onsider MAPFS/FAPFS as an appropriate measure ompared to ADMALE/ADFEM. We also have onsidered gender of the as a determining fator of the amount of borrowing. In relation to village speifi variables, we onsider the amount of borrowing depends on infrastrutural failities of the village suh as eletriity, paved roads and lean drinking water. Dummy variables are used to onstrut these variables. We have used the empowerment index 1, number of earners and the types of houses 2 as instrumental variables. Speifiation of the Variables X = vetor of demographi harateristis are: Number of adult females in the household (ADFEM). Number of adult males in the household (ADMALE). Number of male as perentage of family size (MAPFS). Number of female as a perentage of family size (FAPFS). Age of the female (AFHH). Age of the male (AHH). Eduation of the female (EFHH). Eduation of the male (EHH). Gender of the household (GHH). Land size of the household (LANDSIZE). V = vetor of village level harateristis (ommunity infrastruture) are: Village has eletriity (V1 =1, if there is eletriity and zero otherwise). Village has paved road (V2 =1, if there is pave roads and zero otherwise). Villagers have aess to lean water, arseni free tube well (V3=1, if yes and zero otherwise). Y = dependent variable are the sets of outomes like: Inome of the households. 314

7 Rahman 315 Assets of the households without land. Savings of the households. Z = the instruments are the sets of variable other than X s and V s. Empowerment Index (for empowered EAI=1, and zero otherwise). Number of earners in the household. Types of houses (for household with tin/brik roofs and/or walls =1, zero otherwise). The following table shows the orrelation matrix of the variables. It is observed that most of the variables are not orrelated to eah other exept for the the eduation and age of the male and female. These two variables are found to be highly o-related. To avoid the possible multiollinearity we have not used these two variables at the same time in any estimation proess. Table 1 Correlation Matrix of the Variables Credit Inome Asset Saving AFHH EFHH EHH AHH GHH Land size Credit 1 Inome Asset Saving AFHH EFHH EHH AHH GHH Land Size Results and Disussion The desriptive statistis of the dependent variables and the independent variables are provided in Table 2. Table 3 shows the impat of miroredit on household outomes suh as inome, assets 3, and savings 4. Borrowers are divided into three equal groups based on their inome level. Eah group ontains perent of the sample and are lassified as high, medium and low inome groups. Table 4 provides estimation results based on different 315

8 Rahman 316 inome group borrowers using 2SLS. White heterosedastiity orreted standard errors are reported. STATA 8.0 is used for the estimation purpose. Desriptive statistis of Dependent and Independent Variables Table 2 provides the summary statistis of the variables. The average age of the female in the study is 37 years with the highest age of 75 years and the lowest of 16 years. The average eduation of the female is 4 years of shooling with the highest eduation to the level of a bahelor degree (14 years of shooling). Average age of the male is 43 years with highest of 80 and lowest of 20 years. Average eduation of the male is approximately 5 years of shooling with the maximum of a masters degree (16 years of shooling). 91 per ent of the household is headed by a male while the rest by female. Average number of adult male in the household is 1.53, while average number of adult female in the household is found to be Average number of adult males as a perentage of family size is 29 while for female it is 27. Average number of earners in a family is 1.73 and the average family size is 5.4. Out of total sample of observations 56 perent of female are found to be empowered. Average amount of borrowing is eight thousand, seventeen taka 5 and fifty-one paisa 6 with the highest of one hundred thousand Taka. Average land size of the families is deimals 7 with the highest of 950 deimals. Average monthly inome of the household is found to be 6,743 Taka with the highest of 27,000 Taka. Average aumulated savings (amount of money in their bank aount or saved at home till date) of the households is 7,126 Taka with the lowest of zero and the highest of 16,25,000 Taka. 52 per ent of the houses are found to have tin or brik roofs and/or walls and the rest are built with other endogenous housing materials suh as straws, mud et. In this study we have estimated the simultaneous equations model to see the impat of borrowing on different household outomes suh as inome, savings and asset aumulation. The equation 3 of Setion 4 is estimated separately for inome, savings and assets as the dependent variables. Variables (amount of redit, inome, savings and assets) are log transformed. Table 3 show the impat of redit household savings, inome and assets without differentiating the inome level of the borrowers. From Table 3 we see that the impat of miroredit on inome is signifiant and positive. It implies that as the amount of redit inreases, inome of the borrowers also inreases. Age of the male is found to be negative and signifiant. This implies that as age inreases, inome of the household dereases. Eduation of the male is found to be positive and signifiant, whih means that the inome of the household inreases as eduation of the male inreases. Number of male as a perentage of family size is also found to be positive and signifiant. Table 3 further shows that the miroredit program has a positive influene in inreasing the assets of a household. Age of the male is signifiant and negative on assets that has no plausible explanation for this 316

9 Rahman 317 result. A positive and signifiant oeffiient of eduation of the male household head on assets implies as eduation inreases asset aumulation also inreases. Further it shows that the miroredit program is effetive in aumulating savings for the borrowers. Table 2 Desriptive Statistis of Dependent and Independent Variables (Figures are in years; number of shooling; Taka and in deimals in respetive areas) Variables Mean Standard Minimum Maximum Deviation Age of the female Eduation of the female Age of the male Eduation of the male household head Gender of the Number of adult males in the household Number of adult females in the household Number of male as a perentage of family size Number of female as a perentage of family size Number of earners in the household Family size Empowerment Index Amount of ,000 borrowing Land size Inome Savings ,25,000 Type of houses It may be onluded from Table 3 that the redit program is suessful in inreasing inome, assets and saving of the borrowers. Age and eduation of the male has a signifiant positive impat on inome and assets of the household. 317

10 Rahman 318 Table.3 Two-Stage Least Square Estimation of Amount of Borrowing on Household Outome (Log of Inome, Log of Assets and Log of Savings) (Figures in the parentheses show the t-value on white orreted standard error) Explanatory Log of Inome Log of Asset Log of Savings Variables Constant 0.14* (0.07) -7.37* (-1.68) (-1.43) Log of amount of redit borrowed 0.95*** (3.74) 1.94*** (3.66) 1.52*** (3.13) Age of the male ** -0.02** (-2.05) (-2.00) (-0.39) Eduation of the 0.022** 0.03** male (2.28) (2.00) (1.01) Gender of the (1.42) (1.13) (0.03) Number of male as a perentage of family 0.008*** (2.89) (1.41) (1.09) size Village has eletriity (-1.49) 0.31 (1.09) (-0.73) Village has paved road (0.28 (-0.19) (0.64) R-squared F-statistis (7, 377) Number of observations (***) Signifiant at 1 perent level, (**) Signifiant in 5 perent level, (*) Signifiant in 10 perent level and the t-statistis are shown in the parentheses. Impat of miroredit on different inome group borrowers Table 4 shows the 2SLS estimation of the effets of miroredit programs on various household outomes based on different inome levels of the borrowers. From the table we see that miroredit affets inome positively. This an be seen from positive and signifiant o-effiient of redit program on inome. The table also shows that the age of the male and gender of the is signifiant. It further shows that there is a differene in terms of impat of redit on different inome groups. Table 4 shows the high-inome group borrowers are better off ompared to middle and low-inome group borrowers. The table further shows that as the amount of redit inreases, inome of the low and middle inome group borrowers dereases relative to the high-inome group borrowers. This is an interesting finding that onfirms that miroredit is helping only the wealthier borrowers and not effetive in generating higher inome for the low and 318

11 Rahman 319 middle inome borrowers. It is also evident that having eletriity in the village is signifiantly enhaning the inome of the borrowers. From table 4 it an further be inferred from positive and signifiant o-effiient of the amount of redit that it has a signifiant role in inreasing household assets. Age and eduation of the male are signifiant. As eduation of the inreases, amount of assets of the household also inreases. The table further shows that redit program has a differing impat in terms of asset aumulation depending on the inome level borrowers. High-inome group borrowers are better off in terms of asset aumulation relative to low and middle inome group borrowers. For the low and middle inome group borrowers, as the amount of redit inreases, asset aumulation dereases relative to high-inome group borrowers. Table 4 also shows the impat of redit on household savings. Credit program has signifiant impat on savings aumulation of the borrowers. However, by looking at different inome level of borrowers it is observed that the impat of redit on savings does not vary between different inome groups. Table 4 Two-Stage Least Square Estimation of Amount of Borrowing on Different Inome Level Borrowers (Figures in the parentheses show the t-value on white orreted standard error) Explanatory Log of Inome Log of Asset Log of Savings Variables Constant 5.00*** (3.69) (0.96) (0.74) Log of amount of redit borrowed 0.455*** (2.86) 1.69*** (2.85) 1.32** (2.22) Age of the male * -0.02* (1.70) -(1.73) -(0.13) Eduation of the * 0.01 male (1.18) (1.67) (0.55) Gender of the 0.18*** (3.17) (1.25) (0.12) Number of male as a perentage of family (1.53) (0.95) (0.61) size Low inome dummy -0.99*** -(13.56) -0.56* -(2.02) (1.40) Middle inome -0.51*** dummy -(8.29) -(1.01) -(1.15) Village has eletriity -0.12* -(1.68) 0.34 (1.19) (0.59) Village has paved road (0.40) -(0.26) (0.74) R-squared F-statistis(9, 375) Number of observations (***) Signifiant at 1 perent level, (**) Signifiant in 5 perent level, (*) Signifiant in 10 perent level and the t-statistis are shown in the parentheses. 319

12 Rahman Conlusion In this study we analyse the impat of miroredit on various household outomes suh as inome, assets and savings. We have modified the simultaneous equations model suggested by Pitt and Khandker (1996) in this study by assessing the impat of miroredit separately on household, savings, inome and assets. We have further differentiated the impat of redit on different inome level borrowers. In addition to that we also have used some refined variables to estimate the model ompared to previous studies. From the estimation it may be onluded that the miroredit program is effetive in generating higher inome, assets and savings for the borrowers. Our estimation also suggests that the age of the male has a positive impat on inome and assets but not on the household savings. Eduation of the male is also an important fator in effeting inome and assets positively. The household outomes due to redit program have been different aross inome groups. Miroredit program helps bringing better outomes for the high-inome group borrowers ompared to medium and low-inome group borrowers. Inome and assets derease as amount of redit inreases for the middle and low-inome group borrowers relative to high-inome group borrowers. This is a very important finding in terms of impat of miroredit as it shows that the miroredit program is only effetive for the high inome group borrowers. Our findings suggest, therefore, the low and middle inome group borrowers may have to make further borrowing to pay-off their outstanding loans. Infrastrutural failities suh as eletriity and paved roads in the villages are important fators for redit program to be more effetive to bring about better outomes. Also the age and eduation of the female is signifiant in having better household outomes. The number of earners in a family also has a signifiant positive impat on household outomes. Consistent with Pitt and Khandker (1996) our study also suggest that redit program enhanes savings. In general in may be onluded that miroredit has a pronouned positive effet on inome, assets and savings of the borrowers. However, miroredit seems to be more effetive for the higher-inome group borrowers ompared to the middle and low-inome group borrowers. Miroredit is more effetive for younger and eduated borrowers ompared to older and less eduated borrowers. Infrastrutural failities failitate household outomes of redit program. It is also found that families having more earning members enjoy better household outomes. Over all this study suggest that even though miroredit is an attrative tool to produe better outomes in terms of inome, assets and savings but it is more effetive for relatively wealthier borrowers ompared to non-wealthy borrowers. It would be a good researh 320

13 Rahman 321 question to peruse in future to unover why this is the ase. At the same time, our results suggest that there should be some adjustment to the existing miroredit programs to ahieve the intended outome, that is, to serve the purpose of those who are at the bottom of the soiety. End-Notes 1 Empowerment index has been derived but not reported in the study and will be available from the author. 2 The variable types of houses is expressed in terms of a dummy, where the dummy is one for brik and/or tin (roofs and walls) houses and zero for non-brik and non-tin houses. 3 Asset is defined here as values of furniture, television, radio and other household items other than land and houses. 4 Savings is defined here as money saved by the households in banks or ash saved at home. 5 Taka is the urreny of Bangladesh. 6 Paisa is the lowest denominator. 7 One deimal equals 1/100 are. Referenes Bornstein, D 1998, The Miroredit Movement is Revolutionising International Development, Coleman, B.E 1999, The Impat of Group Lending in Northeast Thailand, Journal of Development Eonomis, vol. 60, pp Feder, G., Onhan, T., Chalamwong, Y., and Hongladarom, C 1988, Land Poliies and Farm Produtivity in Thailand, the Johns Hopkins University Press, Baltimore. Foster, A. D. 1995, Pries, Credit Markets and Child Growth in Low Inome Rural Areas, The Journal of the Royal Eonomi Soiety, vol. 105, pp Ghatak, M. 1999, Group Lending, Loal Information and Peer Seletion, Journal of Development Eonomis, vol. 60, pp Hashemi, S., M. Shuler, S. R, and Riley, A.P. 1996, Rural Credit Programs and Women's Empowerment in Bangladesh, World Development, vol. 24, no. 4, pp Hossain, M. 1988, Credit for Alleviation of Rural Poverty: The Grameen Bank in Bangladesh, International Food Poliy Researh Institute, Washington, D.C. Hulme, D. and Mosley P. 1996, Finane against poverty, London, Routledge Int. Publishing Company. 321

14 Rahman 322 Hussain, A. M. 1998, Poverty Alleviation and Empowerment: The Seond Impat Assessment Study of BRAC s Rural Development Programme, BRAC, Dhaka. Johnson, S. and Mosley P. 1997, Mirofinane and Poverty Redution, London: Oxfam. Khandker, S.R Fighting Poverty with Miroredit, Oxford University Press, Washington, D.C. Khandker, S. R. 2000, Savings, Informal Borrowing, and Mirofinane, the Bangladesh Development Studies, vol. XXVI, no. 2 &3, pp Khandker, S. R. Samad H. A. et al. 1998, Inome and Employment Effets of Miroredit Programs; Village-Level Evidene from Bangladesh, Journal of Development Studies, vol. 35, no. 2, pp Khandker, S. R. 2003, Miro-Finane and Poverty: Evidene Using Panel Data from Bangladesh, World Bank Poliy Researh Working Paper Morduh, J. 1999, The Mirofinane Promise, Journal of Eonomi Literature, vol. 37, no. 4, pp Morduh, J. 2000, The Mirofinane Shism, World Development, vol. 28 no. 4, pp Pitt, M. and. Khandker, S.R. 1996, Household and Intra-Household Impats of the Grameen Bank and Similar Targeted Credit Programs in Bangladesh, World Bank Disussion Papers 320, Washington, DC. Pitt, M. and Khandker, S. 1998, The Impat of Group-Based Credit Programmes on Poor Households in Bangladesh: Does Gender of the partiipant Matter? Journal of Politial Eonomy vol. 106, pp Rahman, H. 1995, Mora Kartik: Seasonal Defiits and the Vulnerability of the Rural Poor, in Rahman H. &. Hossain, M. Rethinking rural poverty Bangladesh as a ase study, Dhaka: University Press Limited, (UPL). Stiglitz, K. 1993, Peer Monitoring and Credit Markets: Oxford University Press. The Miro-redit Summit, 1997, The Miro-redit Summit: Delaration and plan of ation, Washington, D.C. Varian, H. R. 1990, Monitoring Agents with Other Agents, Journal of Institutional and Theoretial Eonomis, vol. 146, pp Wahid, N.W. 1993, The Grameen Bank Poverty Relief in Bangladesh, West View Press. 322

15 Rahman 323 Weiss, J. and Montgomery, H. 2005, Great Expetations: Mirofinane and Poverty Redution in Asia and Latin Ameria, Oxford Development Studies, vol. 33, no. 3 and 4, pp World Bank, 1994, The World Bank's Strategy for Reduing Poverty and Hunger, Environmentally Sustainable Development Studies, Monograph series, vol. 4, no. 3. Yaron, J. 1994, What Makes Rural Finane Institute Suessful? The World Bank Researh Observer, vol. 9, no 1, pp Yunus, M. 1999, Banker to the Poor, The Autobiography of Muhammad Yunus, Founder of the Grameen Bank, Aurun Press Ltd. London. Yunus, M. 1983, Group -based Savings and Credit for the Rural Poor: Grameen Bank in Bangladesh. Group-based Savings and Credit for the Rural Poor, Papers and Proeedings of a workshop, Bogra (Bangladesh), Geneva: ILO. Zeller, M., and Sharma, M. 2000, Many Borrow, More Save, and All Insure: Impliations for Food and Mirofinane Poliy, Food Poliy, vol. 25, pp