Empowering Women through Microcredit in Bangladesh: An Empirical Study
, Debasish Chandra Acharjee 3, Waqas Umar Latif 1 and Linping Wang 1,*Round 1
Reviewer 1 Report
REVIEW OF “EMPOWERING WOMEN THROUGH MICROCREDIT”
This study uses a sample of 300 “women households” (180 who had borrowed, 120 who had not) in rural Bangladesh to investigate the relationship between microcredit and women’s empowerment. The significance of socio-economic characteristics in determining the likelihood of accessing microcredit is assessed [high-income respondents are less likely to be borrowers; the relationship of education is asserted, but questionable]. A logit model is used to investigate the relationship between accessing microcredit and 14 indicators of “empowerment” – only six of which were significantly positively associated with borrowing (at the 1% level or better), while two were significantly negatively associated. (Of the non-significant coefficients, only two were positively associated, while four were negative.)[Note: it appears that loan amount rather than access (borrower/non-borrower) was used as the explanatory variable, which raises some questions.] Without directly acknowledging the mixed nature of the results, the authors conclude that “Micro credit in Bangladesh [is] mostly helping rural women in establishing their decision making right and increasing legal awareness related issues.” The authors recommend that “policymakers” [why not microfinance institutions?] “redesign loan products pushing more importance on targeting women obviously” without explaining the shortcomings of the present design [which is not explained, beyond mentioning group loans] – although they do suggest that “credit delivery should be combined with non-financial services within the programs such as training .”
The study is potentially publishable as an empirical test of the hypothesis that access to microcredit is associated with empowerment of women. A clearer conceptual framework is needed to explain the hypothesis. And it will be necessarily both to acknowledge that causality may be two-way (i.e., the results using data at one point in time show correlation, not necessarily the direction of causation), and that the results are mixed (i.e., there is evidence of negative as well as positive associations between access and empowerment). The authors need to clarify whether the explanatory variable is access to microcredit (borrowers vs. non-borrowers) or the amount of loan obtained – and this should be consistent with the conceptual framework (or perhaps both variables tested, as an indicator of robustness).
Substantial reworking of the presentation is needed. The research design and methodology should come before, not after the presentation of results. Significant copy-editing is required.
There is some disconnect between the authors’ apparent desire to argue in favor of microcredit as a source of empowerment and their actual results. The results indicate no or negative influence in three of the five categories of empowerment, mixed results in one, and clearly positive only in one:
Line 178 “The results indicate that microcredit does not influence the control over own income by rural women.
L. 192 “ loan amount is negative and significantly influence the frequency of visit at bank/market without husband permission.
L. 206: “the probability of purchasing furniture, children cloth and own cloth is very little and insignificant influence by the microcredit “ (and two other studies show no impact of microcredit on purchases).
L. 213: “micro credit negatively influence the decision to raise livestock On the other hand, micro credit positive and significantly influences four other indicators “
L. 229: “Borrowing from microcredit program has a positive and significant influence on legal awareness”
The article should be revised as a dispassionate academic presentation of the results, negative and positive, rather than as a vehicle for introducing the authors’ views that are not directly substantiated from evidence presented. (See also comment below on Conclusions.)
COMMENTS
Abstract line 26: Are “policymakers” the correct target? Are microfinance programs in Bangladesh determined by government, or by the organizations that operate them? I.e., don’t Grameen Bank, ASA, etc. determine their own methods and policies? So shouldn’t they at least be included, if not the main target for using the findings?
Indeed, empowering women has been an objective of Grameen Bank and others from the beginning (see the “16 decisions”). So it might make more sense to frame the paper as testing the hypothesis that accessing microfinance does indeed facilitate and cause empowerment of women – rather than framing it as making a recommendation to government. Likewise: lines 102-3 state the objective of helping “the government, policy makers and the independent researcher to set priorities to formulate alternative program .” Why are the MFIs themselves omitted? Exactly what government policy is it that will affect the ability of microfinance to empower women?). And why “alternative program”? -- that suggests that MFIS are presently NOT empowering women, and that some alternative (to Grameen Bank, ASA, etc.) is needed. They would likely take issue with the need for “alternative programs”, although they might be receptive to ways they can improve to better empower women.
Research design and methodology
Section 2 should present the research design and methodology. It is not until Section 3 “Materials and Methods” that the necessary information is given – forcing the reader to struggle through the results without a clear understanding of the methodology.
A bit more information is needed on why Jamalpur and Mymensingh districts were selected, and are they considered representative of Bangladesh. We also need to know:
What is the nature of the microfinance institutions were involved (and how many)? Were these NGOs or government programs, and are they representative or typical of microfinance in Bangladesh? Were they targeted toward women, and was empowerment a specific objective? If so, did they include non-financial support (training) – or were these ‘minimalist’ (credit-only) programs?
l. 127: “The primary lending program of microcredit is group-based form.” Is it a Grameen-type group methodology, or other?
The respondents that were selected represent what proportion of the total number of clients on the lists provided by the MFIs?
How exactly were the non-borrowers selected, and what measures (if any) were taken to make them as comparable as possible to the clients of MFIs? Since it is unlikely that you had a list of all non-borrowers in these areas, you can’t just say you “randomly” selected non-borrowers; the process should be sufficiently clear that another researcher could duplicate it.
The latter point is particularly important if you wish to argue that it was access to microcredit that led to empowerment, rather than empowerment making it easier to access microcredit. I have some concerns that the latter may be the case, because (compared to the borrowers) the non-borrowers were relatively older and less educated. (Line 111 says that a “majority of microcredit holder[s] are illiterate” – but the proportion of illiterate non-borrowers is higher.) On the other hand, they were much more concentrated in the highest income category – which could be interpreted in several very different ways:
· They may not need credit as much [l. 139 says: “high income family were less interested to take microcredit compared with low income family” ]
They would be better placed to access credit if they did need it;
If borrowers tend to be in lower income categories, does that mean that accessing microcredit did not lead to improved incomes? [hopefully, the statistical analysis will help sort that out]
Mainly, we would like to know how representative the sample was – the more representative, the more meaningful the results will be to others (and hence the more publishable they are). If the sampling methodology did not ensure representativeness, it would at least be useful to know how the sample characteristics compare to national statistics (for age, education, occupation, etc.).
The logit model (and why/how used) should also be introduced in this section.
The authors need to clearly define “empowerment,” the basis for the indicators they use, and how they measure it. It is not until section 2.5 that the indicators for empowerment are introduced, with no justification or explanation. This should be laid out clearly in a research design section, with literature cited as justifications for using the indicators (control over financial assets; control over mobility; ability to make purchases independently; participation in decision-making; and legal awareness – all of which seem reasonable).
Measuring empowerment:
The authors need to look further for literature on measuring empowerment and confirm that their methodology is consistent with others. The following statement (line 84) is demonstrably false: “the World Bank or any other major development agency has not developed a rigorous method for measuring tracking changes in levels of empowerment”
For example, Dr.V.Rengarajan says: “World Bank has produced numerous working definitions that make it possible to clarify the meaning of the term empowerment such as it is employed today.”
Soham Banerji, CGAP, January 2018 says: “FII data has a dedicated gender module measuring women’s economic empowerment. The module aims to assess whether women are achieving economic success and have agency over financial decision-making as a result of access to finance.”
https://www.findevgateway.org/blog/2018/jan/women’s-empowerment-indicators
In presenting the results, the authors do cite a number of studies with data relevant to some of the indicators. Use the findevgateway to search for other studies that have attempted to define and measure empowerment, and use these in a methodology section to set up the explanatory variables that are used in this study.
In a country (such as Bangladesh) where women’s social status is problematic and often controlled by men, “empowerment” may often be considered in terms of how much independence women have from their husbands. Indeed, 6 of the indicators used relate to “participation in decision-making.” Women who are heads of households could be considered to be necessarily “empowered” to some extent in that regard. It would be useful to at least briefly note that 80% of borrowers and 73% of non-borrowers are “housewives” (and even those engaged in agriculture and non-agricultural activities may have husbands), so the issue of empowerment is relevant.
Conceptual framework
A particular problem that must be confronted is that of causality:
Does accessing microfinance make women more empowered?
OR: are empowered women more likely to be able to access microcredit?
It appears that both are likely – a self-reinforcing virtuous circle. This makes attribution extremely difficult, unless one uses randomized controlled experiments. Hence one must be very circumspect about attributing causality when all one really has is an association (correlation) between microfinance and empowerment.
Hence we need a conceptual framework that presents the model of how microcredit is thought to lead to empowerment, and that recognizes the possibility of two-way causality. This section should also explicitly provide the hypotheses (and null hypotheses) that are tested using the model.
Although the discussion seems oriented toward access to microcredit being the hypothesized determinant of empowerment, in practice it appears that loan amount is used as the explanatory variable in Sec. 2.5. So is it access vs. non-access that is hypothesized to influence empowerment? Or, is the hypothesis that increasing amounts of borrowing increase empowerment? If the latter, do you mean to imply that highly indebted people must be highly empowered?!? However, since the dependent variable is binary, how can it be used to test whether (steadily) increasing loan size means greater and greater empowerment? If you are using a binary dependent variable, then it would seem that you would then have to have a theory that there is a threshold of some minimum level of credit that triggers “empowerment” (“1” rather than “0”).
It might be useful to state both propositions: (1) just having access to microcredit influences empowerment; and (2) increasing loan size is
Lines 160-1 implicitly assume that income is a causal variable in relation to accessing microcredit: “Negative and significant coefficient of higher income indicates inverse relationship with credit accessibility.” This is used to argue that the microcredit programs are focused on the relatively poor (or that the less poor simply don’t need the credit). But other studies assume the opposite causality when they find that borrowers have higher incomes than non-borrowers, arguing that accessing credit must have resulted in the higher incomes. So in the data presented, could it be argued that those with access to microcredit have lower incomes as a result?? That is, if you are assuming that being a borrower is what causes certain empowerment indicators to be higher, doesn’t that imply that we should likewise assume that it is borrowing that causes the income variable to be lower?
This is just to note that causality is problematic, and you should be explicit about what is assumed.
Results
Sections 2.1 and 2.3 investigate the relationship between the characteristics of the respondents and whether or not they accessed microcredit. That is fine, but there is a problem with the “above secondary” category. Only 3 (out of 180) borrowers and no non-borrowers fall in this category. Therefore it cannot meaningfully be analyzed as a separate category, and should be combined in the analysis with the “up to SSC” category to be: “SSC or above” – i.e., for the analysis in Tables 3 and 4 (fine to leave Table 1 as is). Furthermore, there is a problem with how the level of education was turned into a continuous variable in Table 3 by assigning values to each level (“3” for above SSC). With only 1.7% of observations above “3” and 77% below “1”, the average cannot be 2.88; and since the non-borrowers have zero observations above 2, their average cannot be 2.76. Ideally, years of education would have been more appropriate.
The statistical significance of the apparent relationship between level of education and accessing microfinance is likely to disappear when this adjustment is made. Furthermore, the coefficient for primary education is negative (though not significant) – suggesting that those with less education were less likely to have access to microcredit. [Some explanation of how these were calculated is needed. It appears that this negative coefficient may be relative to those with no education – in which case, having a little (primary) education seems associated (not significantly) with a lower likelihood of borrowing, but more (secondary) education increases the likelihood of borrowing. Which looks like: no clear pattern.] This means that the discussion in section 2.4 needs to be reworked, especially lines 152-5:
the likelihood of demanding a larger amount of credit decreases as the level of farmers’ education increases. Therefore, our result implies that the availability of credit services from the small credit schemes targeted the rural poor and vulnerable women in the rural areas and the majority of them had lower educational levels.
Since you purposively selected women for the study, you cannot conclude from your data that “credit services targeted women” (you would need data on the total population of women and men in the microcredit programs – at least the ones studied, if not all in the area; or you may have information that the programs study in fact did target women as a matter of policy).
Table 1: Why are significance levels shown in the footnotes? Were tests of significance conducted on these variables? If so, it should be stated that there were no significant differences between the borrower and non-borrower groups (which would be positive in terms of confirming that the two groups were not significantly different, and hence comparable); if not, then no need to state the significance levels.
Line 192 says that “the loan amount is negative and significantly influence frequency of visit at bank/market “. Why “loan amount”? According to the note to Table 4: “Dependent variable: borrower = 1, Non-borrower = 0” – not “loan amount.” In Table 5, the dependent variable is the empowerment indicator, so the independent variable is presumably the borrower/non-borrower binary variable – or is it “loan amount” (as stated in sections 2.5.1 and 2.5.2? If loan amount, does that mean that Table 5 is conducted only for those women who had loans? (If loan amount is used, that is a somewhat different question from whether accessing microcredit itself is empowering. Loan amount would be asking whether borrowing more makes one more likely to be empowered.) [Line 229 says “Borrowing from microcredit program has a positive and significant influence on legal awareness “ – which seems to refer to the binary variable, rather than loan amount.
Oh, wait! At the very end of the text – line 289 – it says: “Total loan amount was used as an explanatory variable to capture the effect of micro credit along with some other control variables explained in the earlier Table 8.” This needs more explanation. Are non-borrowers included with “0” as their loan amount? Or excluded? What not try it both ways – using binary loan/no loan as an explanatory variable; and then trying loan amount. (As noted, these represent somewhat different questions.)
This confusion is why the model and variables need to be presented up front and more thoroughly in a methodology section. It is intolerable to have key explanations only at the very end of the paper.
Sec. 2.5.2 The coefficients are negative for both indicators, and in both cases the non-borrowers have a higher probability of mobility than borrowers. This seems to imply a negative influence of accessing microcredit on mobility. But instead of making that conclusion, the authors explain lack of mobility in terms of the “enormous double load for many poor women” due to “entering the public workforce” and taking care of children and housework. This indicates a strong bias of the researchers toward finding the desired conclusion (that microcredit positively influences empowerment), because these “explanations” would apply to both borrowers and non-borrowers, and hence would not explain why borrowers appear to be less mobile (other things presumably being equal). Furthermore, 90% of borrowers are housewives or in agriculture, so obviously not “entering the public workforce,” and the proportion of non-borrowers who are engaged outside agriculture is higher than borrowers (23% vs. 10%) – so why then are they more mobile?
Model
Sec. 3.2.3 does not explain the model adequately, esp. line 281. Reference is made to the “control variables explained in the earlier Table 8”, but Table 8 explains Xi not Hij and Fij. Most importantly, the critical “loan amount” variable is not explained. Is it the actual amount received, or based on the groupings shown in Table 2? How are non-borrowers treated – as 0 loan amount? Why isn’t it a binary variable for loan vs no loan?
Conclusions
The authors say that “training on micro credit management and building up social networks … seem to be absent in the Bangladeshi microcredit programs but are crucial for women’s integration in the social development.” This seems a bit odd, in that Grameen Bank at least is well-known for its social orientation and focus on empowering women (including the “16 decisions”). This shows the need to explain better the nature of the microfinance programs studied (and how they may differ from other models such as Grameen Bank).
EDITORIAL
Be careful not to overstate causality when what one has is simply an association.
There are substantial problems with subject-verb agreement (plural vs. singular); I’ve highlighted a few cases in the abstract. Although reasonably clearly written, the paper as a whole needs substantial copy-editing for these and many other grammatical errors, as well as incorrect usages. Lines that especially need rephrasing include 46, 49-50….and more.
Below is just a sample of the many edits that would be needed:
Abstract
Line 13: “of” instead of “persuading”
l. 19: “were inversely related with” instead of “inversely influenced”
lines 20-21: was positively and significantly associated with” instead of “ has positive and significant effect on”
line 21: “indicate” instead of “indicates”
lines 22 & 23: “borrowers” (plural)
l. 24: “influence” instead of “influences”
Text
l. 39: “rose” instead of “growths”
l. 44: “provides” instead of “plays”
lines 73-74: “more women lost control over their loans” instead of “the more women lose their control over loan”
l. 88: “faith privileges” ???
l. 160: “tendency” (or “likelihood of”) instead of “ability”
l. 177: “constant” instead of “persistence”
Author Response
Author’s Response to Reviewer-1 Comments
Manuscript ID：ijfs-510923
Manuscript Title: Empowering Women through Microcredit in Bangladesh: An Empirical Study
The authors would like to thank the Editor and Reviewer-1 for their deep review of our manuscript with very constructive comments and suggestions. The general and specific questions or concerns and recommendations are highly appreciated, and corrections have been made accordingly in the revised version of the manuscript to improve its quality. The point-by-point responses to the reviewer’s comments and suggestions are listed as follows:
Reviewer #1:
Comment 1
Abstract line 26: Are “policymakers” the correct target? Are microfinance programs in Bangladesh determined by government, or by the organizations that operate them? I.e., don’t Grameen Bank, ASA, etc. determine their own methods and policies? So shouldn’t they at least be included, if not the main target for using the findings?
Indeed, empowering women has been an objective of Grameen Bank and others from the beginning (see the “16 decisions”). So it might make more sense to frame the paper as testing the hypothesis that accessing microfinance does indeed facilitate and cause empowerment of women – rather than framing it as making a recommendation to government. Likewise: lines 102-3 state the objective of helping “the government, policy makers and the independent researcher to set priorities to formulate alternative program .” Why are the MFIs themselves omitted? Exactly what government policy is it that will affect the ability of microfinance to empower women?). And why “alternative program”? -- that suggests that MFIS are presently NOT empowering women, and that some alternative (to Grameen Bank, ASA, etc.) is needed. They would likely take issue with the need for “alternative programs”, although they might be receptive to ways they can improve to better empower women.
Response 1
Thank you for your kind thoughtfulness. We have corrected the text as suggested by the reviewer. We agree that microcredit providers develop their own program. Therefore, we rewritten the term ‘policymakers’ as ‘microcredit providers’ for better understanding.
Comment 2: Research design and methodology
Section 2 should present the research design and methodology. It is not until Section 3 “Materials and Methods” that the necessary information is given – forcing the reader to struggle through the results without a clear understanding of the methodology.
A bit more information is needed on why Jamalpur and Mymensingh districts were selected, and are they considered representative of Bangladesh. We also need to know:
• What is the nature of the microfinance institutions were involved (and how many)? Were these NGOs or government programs, and are they representative or typical of microfinance in Bangladesh? Were they targeted toward women, and was empowerment a specific objective? If so, did they include non-financial support (training) – or were these ‘minimalist’ (credit-only) programs?
o l. 127: “The primary lending program of microcredit is group-based form.” Is it a Grameen-type group methodology, or other?
• The respondents that were selected represent what proportion of the total number of clients on the lists provided by the MFIs?
• How exactly were the non-borrowers selected, and what measures (if any) were taken to make them as comparable as possible to the clients of MFIs? Since it is unlikely that you had a list of all non-borrowers in these areas, you can’t just say you “randomly” selected non-borrowers; the process should be sufficiently clear that another researcher could duplicate it.
The latter point is particularly important if you wish to argue that it was access to microcredit that led to empowerment, rather than empowerment making it easier to access microcredit. I have some concerns that the latter may be the case, because (compared to the borrowers) the non-borrowers were relatively older and less educated. (Line 111 says that a “majority of microcredit holder[s] are illiterate” – but the proportion of illiterate non-borrowers is higher.) On the other hand, they were much more concentrated in the highest income category – which could be interpreted in several very different ways:
• They may not need credit as much [l. 139 says: “high income family were less interested to take microcredit compared with low income family” ]
• They would be better placed to access credit if they did need it;
• If borrowers tend to be in lower income categories, does that mean that accessing microcredit did not lead to improved incomes? [hopefully, the statistical analysis will help sort that out]
Mainly, we would like to know how representative the sample was – the more representative, the more meaningful the results will be to others (and hence the more publishable they are). If the sampling methodology did not ensure representativeness, it would at least be useful to know how the sample characteristics compare to national statistics (for age, education, occupation, etc.).
Response 2
Thank you for your kind deep observation. We have done a significant copy editing as suggested by the reviewer. In this revised version Materials and methods section is placed before results and discussion. Justification of selecting locations and non-borrowers are presented in the text (see line 114-126).
Comment 3
The logit model (and why/how used) should also be introduced in this section.
Response 3
Thank you for your kind concern. Logit model was introduced in this section (see line 129-144).
Comment 4
The authors need to clearly define “empowerment,” the basis for the indicators they use, and how they measure it. It is not until section 2.5 that the indicators for empowerment are introduced, with no justification or explanation. This should be laid out clearly in a research design section, with literature cited as justifications for using the indicators (control over financial assets; control over mobility; ability to make purchases independently; participation in decision-making; and legal awareness – all of which seem reasonable).
Response 4
Explanations are provided in the text (see line 145-161)
Comment 5: Measuring empowerment
The authors need to look further for literature on measuring empowerment and confirm that their methodology is consistent with others. The following statement (line 84) is demonstrably false: “the World Bank or any other major development agency has not developed a rigorous method for measuring tracking changes in levels of empowerment”
For example, Dr.V.Rengarajan says: “World Bank has produced numerous working definitions that make it possible to clarify the meaning of the term empowerment such as it is employed today.”
https://www.findevgateway.org/sites/default/files/mfg-en-paper-micro-finance-syndrome-is-it-for-empowerment-or-disempowerment-of-women-jan-2010.pdf
Soham Banerji, CGAP, January 2018 says: “FII data has a dedicated gender module measuring women’s economic empowerment. The module aims to assess whether women are achieving economic success and have agency over financial decision-making as a result of access to finance.”
https://www.findevgateway.org/blog/2018/jan/women’s-empowerment-indicators
In presenting the results, the authors do cite a number of studies with data relevant to some of the indicators. Use the findevgateway to search for other studies that have attempted to define and measure empowerment, and use these in a methodology section to set up the explanatory variables that are used in this study.
In a country (such as Bangladesh) where women’s social status is problematic and often controlled by men, “empowerment” may often be considered in terms of how much independence women have from their husbands. Indeed, 6 of the indicators used relate to “participation in decision-making.” Women who are heads of households could be considered to be necessarily “empowered” to some extent in that regard. It would be useful to at least briefly note that 80% of borrowers and 73% of non-borrowers are “housewives” (and even those engaged in agriculture and non-agricultural activities may have husbands), so the issue of empowerment is relevant.
Response 5
Thank you for your kind deep review. We have removed the line 84 ‘world bank and any other development agency………’ as suggested by reviewer and tried to provide explanation of measuring empowerment in material and methods section with appropriate references.
Comment 6: Conceptual framework
A particular problem that must be confronted is that of causality:
• Does accessing microfinance make women more empowered?
• OR: are empowered women more likely to be able to access microcredit?
It appears that both are likely – a self-reinforcing virtuous circle. This makes attribution extremely difficult, unless one uses randomized controlled experiments. Hence one must be very circumspect about attributing causality when all one really has is an association (correlation) between microfinance and empowerment.
Hence we need a conceptual framework that presents the model of how microcredit is thought to lead to empowerment, and that recognizes the possibility of two-way causality. This section should also explicitly provide the hypotheses (and null hypotheses) that are tested using the model.
Although the discussion seems oriented toward access to microcredit being the hypothesized determinant of empowerment, in practice it appears that loan amount is used as the explanatory variable in Sec. 2.5. So is it access vs. non-access that is hypothesized to influence empowerment? Or, is the hypothesis that increasing amounts of borrowing increase empowerment? If the latter, do you mean to imply that highly indebted people must be highly empowered?!? However, since the dependent variable is binary, how can it be used to test whether (steadily) increasing loan size means greater and greater empowerment? If you are using a binary dependent variable, then it would seem that you would then have to have a theory that there is a threshold of some minimum level of credit that triggers “empowerment” (“1” rather than “0”).
It might be useful to state both propositions: (1) just having access to microcredit influences empowerment; and (2) increasing loan size is
Lines 160-1 implicitly assume that income is a causal variable in relation to accessing microcredit: “Negative and significant coefficient of higher income indicates inverse relationship with credit accessibility.” This is used to argue that the microcredit programs are focused on the relatively poor (or that the less poor simply don’t need the credit). But other studies assume the opposite causality when they find that borrowers have higher incomes than non-borrowers, arguing that accessing credit must have resulted in the higher incomes. So in the data presented, could it be argued that those with access to microcredit have lower incomes as a result?? That is, if you are assuming that being a borrower is what causes certain empowerment indicators to be higher, doesn’t that imply that we should likewise assume that it is borrowing that causes the income variable to be lower?
This is just to note that causality is problematic, and you should be explicit about what is assumed.
Response 6
We agreed that measuring causal impact when baseline is not available is always difficult. In this study, we assume that access to microcredit helps empowering women based on the previous literatures and tried to test the hypothesis access to microcredit empowered rural women. In this revised version interpretations are provided according to the hypothesis.
Comment 7: Results
Sections 2.1 and 2.3 investigate the relationship between the characteristics of the respondents and whether or not they accessed microcredit. That is fine, but there is a problem with the “above secondary” category. Only 3 (out of 180) borrowers and no non-borrowers fall in this category. Therefore it cannot meaningfully be analyzed as a separate category, and should be combined in the analysis with the “up to SSC” category to be: “SSC or above” – i.e., for the analysis in Tables 3 and 4 (fine to leave Table 1 as is). Furthermore, there is a problem with how the level of education was turned into a continuous variable in Table 3 by assigning values to each level (“3” for above SSC). With only 1.7% of observations above “3” and 77% below “1”, the average cannot be 2.88; and since the non-borrowers have zero observations above 2, their average cannot be 2.76. Ideally, years of education would have been more appropriate.
The statistical significance of the apparent relationship between level of education and accessing microfinance is likely to disappear when this adjustment is made. Furthermore, the coefficient for primary education is negative (though not significant) – suggesting that those with less education were less likely to have access to microcredit. [Some explanation of how these were calculated is needed. It appears that this negative coefficient may be relative to those with no education – in which case, having a little (primary) education seems associated (not significantly) with a lower likelihood of borrowing, but more (secondary) education increases the likelihood of borrowing. Which looks like: no clear pattern.] This means that the discussion in section 2.4 needs to be reworked, especially lines 152-5:
the likelihood of demanding a larger amount of credit decreases as the level of farmers’ education increases. Therefore, our result implies that the availability of credit services from the small credit schemes targeted the rural poor and vulnerable women in the rural areas and the majority of them had lower educational levels.
Response 7
Thank you so much for your kind suggestions. We have rearranged the data and removed the educational category of above secondary. As a result, findings are changed to some extent (table 5). We also changed the interpretation accordingly.
Comment 8
Since you purposively selected women for the study, you cannot conclude from your data that “credit services targeted women” (you would need data on the total population of women and men in the microcredit programs – at least the ones studied, if not all in the area; or you may have information that the programs study in fact did target women as a matter of policy).
Response 8
The present study purposively selected women. So we have changed the term ‘credit services targeted women’.
Comment 9
Table 1: Why are significance levels shown in the footnotes? Were tests of significance conducted on these variables? If so, it should be stated that there were no significant differences between the borrower and non-borrower groups (which would be positive in terms of confirming that the two groups were not significantly different, and hence comparable); if not, then no need to state the significance levels.
Response 9
In table one we did not conduct any test. So, the footnotes of significance have been removed.
Comment 10
Line 192 says that “the loan amount is negative and significantly influence frequency of visit at bank/market “. Why “loan amount”? According to the note to Table 4: “Dependent variable: borrower = 1, Non-borrower = 0” – not “loan amount.” In Table 5, the dependent variable is the empowerment indicator, so the independent variable is presumably the borrower/non-borrower binary variable – or is it “loan amount” (as stated in sections 2.5.1 and 2.5.2? If loan amount, does that mean that Table 5 is conducted only for those women who had loans? (If loan amount is used, that is a somewhat different question from whether accessing microcredit itself is empowering. Loan amount would be asking whether borrowing more makes one more likely to be empowered.) [Line 229 says “Borrowing from microcredit program has a positive and significant influence on legal awareness “ – which seems to refer to the binary variable, rather than loan amount.
Oh, wait! At the very end of the text – line 289 – it says: “Total loan amount was used as an explanatory variable to capture the effect of micro credit along with some other control variables explained in the earlier Table 8.” This needs more explanation. Are non-borrowers included with “0” as their loan amount? Or excluded? What not try it both ways – using binary loan/no loan as an explanatory variable; and then trying loan amount. (As noted, these represent somewhat different questions.)
This confusion is why the model and variables need to be presented up front and more thoroughly in a methodology section. It is intolerable to have key explanations only at the very end of the paper.
Response 10
We have corrected the interpretation and replaced the ‘total loan amount’ with ‘access to microcredit’ which is a binary independent variable. Explanations are provided in the materials and methods section.
Comment 11
Sec. 2.5.2 The coefficients are negative for both indicators, and in both cases the non-borrowers have a higher probability of mobility than borrowers. This seems to imply a negative influence of accessing microcredit on mobility. But instead of making that conclusion, the authors explain lack of mobility in terms of the “enormous double load for many poor women” due to “entering the public workforce” and taking care of children and housework. This indicates a strong bias of the researchers toward finding the desired conclusion (that microcredit positively influences empowerment), because these “explanations” would apply to both borrowers and non-borrowers, and hence would not explain why borrowers appear to be less mobile (other things presumably being equal). Furthermore, 90% of borrowers are housewives or in agriculture, so obviously not “entering the public workforce,” and the proportion of non-borrowers who are engaged outside agriculture is higher than borrowers (23% vs. 10%) – so why then are they more mobile?
Response 11
Interpretations are rearranged based on results (see line 286-297).
Comment 12: Model
Sec. 3.2.3 does not explain the model adequately, esp. line 281. Reference is made to the “control variables explained in the earlier Table 8”, but Table 8 explains Xi not Hij and Fij. Most importantly, the critical “loan amount” variable is not explained. Is it the actual amount received, or based on the groupings shown in Table 2? How are non-borrowers treated – as 0 loan amount? Why isn’t it a binary variable for loan vs no loan?
Response 12
Empirical model has been corrected and proper explanations are provided in the text (see line 164-176).
Comment 13: Conclusions
The authors say that “training on micro credit management and building up social networks … seem to be absent in the Bangladeshi microcredit programs but are crucial for women’s integration in the social development.” This seems a bit odd, in that Grameen Bank at least is well-known for its social orientation and focus on empowering women (including the “16 decisions”). This shows the need to explain better the nature of the microfinance programs studied (and how they may differ from other models such as Grameen Bank).
Response 13
The conclusion has been rearranged (see line 347-363).
Comment 14: EDITORIAL
Be careful not to overstate causality when what one has is simply an association.
There are substantial problems with subject-verb agreement (plural vs. singular); I’ve highlighted a few cases in the abstract. Although reasonably clearly written, the paper as a whole needs substantial copy-editing for these and many other grammatical errors, as well as incorrect usages. Lines that especially need rephrasing include 46, 49-50….and more.
Below is just a sample of the many edits that would be needed:
Abstract
Line 13: “of” instead of “persuading”
l. 19: “were inversely related with” instead of “inversely influenced”
lines 20-21: was positively and significantly associated with” instead of “ has positive and significant effect on”
line 21: “indicate” instead of “indicates”
lines 22 & 23: “borrowers” (plural)
l. 24: “influence” instead of “influences”
Text
l. 39: “rose” instead of “growths”
l. 44: “provides” instead of “plays”
lines 73-74: “more women lost control over their loans” instead of “the more women lose their control over loan”
l. 88: “faith privileges” ???
l. 160: “tendency” (or “likelihood of”) instead of “ability”
l. 177: “constant” instead of “persistence”
Response 14
We have made a substantial editing to improve the sentence structure of the paper. All the mistakes mentioned above by the reviewer were corrected along with other mistakes which were identified by the authors during revision.
We are really grateful for your kind deep review, improvements and suggestions. Hopefully, this version as attached will be acceptable for your kind consideration.
With Thanks
Professor Linping Wang
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Unfortunately the manuscript does not fulfill the criteria for an academic paper. The methodology, i.e.rationale and design of the survey, sampling techniques etc. are only superficially explained and justified. It is therefore next to impossible to make sense of the results. Is the lower education and income of credit receivers the result of proper targeting, or of dodging the procedures? How can the negative effect on mobility be explained. The sketchy interpretations offered are not very helpful and occasionally misleading, for instance: 'Positive and significant value of family size (P ＜0.05) indicates that having more members in the family is assumed to be a credit constraint.' Eventually, the conclusions could have been written without the research.
Author Response
Author’s Response to Reviewer-2 Comments
Manuscript ID：ijfs-510923
Manuscript Title: Empowering Women through Microcredit in Bangladesh: An Empirical Study
The authors would like to thank the Editor and Reviewer-2 for their deep review of our manuscript with very constructive comments and suggestions. The general and specific questions or concerns and recommendations are highly appreciated, and corrections have been made accordingly in the revised version of the manuscript to improve its quality. The point-by-point responses to the reviewer’s comments and suggestions are listed as follows:
Reviewer #2:
Unfortunately the manuscript does not fulfill the criteria for an academic paper. The methodology, i.e.rationale and design of the survey, sampling techniques etc. are only superficially explained and justified. It is therefore next to impossible to make sense of the results. Is the lower education and income of credit receivers the result of proper targeting, or of dodging the procedures? How can the negative effect on mobility be explained. The sketchy interpretations offered are not very helpful and occasionally misleading, for instance: 'Positive and significant value of family size (P ＜0.05) indicates that having more members in the family is assumed to be a credit constraint.' Eventually, the conclusions could have been written without the research.
Response
We have done a significant copy editing as suggested by the reviewer. In this revised version, interpretations were rearranged based on results. The conclusion has been rearranged. Hopefully, this version as attached will be acceptable for your kind consideration.
With Thanks
Professor Linping Wang
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The previous comments have been well received by the authors and the appropriate changes made. It is now a clearly-presented, compact discussion of evidence on the relationship between microcredit and empowerment in a particular case. A useful contribution is the detailed breakdown and discussion of the empowerment indicators.
Publication is recommended, subject to further copy-editing to clean up numerous, albeit minor language errors that are mainly the fault of the irrational complexities of English language grammar (only some of which are indicated below).
EDITORIAL
Lines 178-9: I suggest replacing the last sentence with: “Higher-income women may demand less microcredit because they are less dependent upon credit for their income-generating activities or because they have better access to other sources of finance.”
[ that’s the demand side; l. 176 gives a supply-side explanation]
l. 210 ‘did not significantly affect’ ?? The evidence from Table 8 shows a higher likelihood for borrowers to purchase independently, but Table 7 indicates that the differences are not statistically significant.
l. 20: borrowers
l. 21: influences [singular]
l. 33: achievements
l. 37: provides an influential
l. 40: important instrument for the poor
l. 42: microcredit instituteschange ….. institutions, which has
l. 47: programs
l. 52: compared
l. 57: delete “in”
l. 62: women tended to lose instead of “the more women lost
l. 64: amounts [plural in both cases]
l. 65: delete “of”
l. 70: delete “the” ‘observe’ instead of ‘observed’
l. 72: successful
Despite improvements, the text still needs substantial copy-editing by a native English speaker. The meaning is usually clear, but minor changes such as the above are needed for full fluency. I cannot continue making all corrections needed; going forward, I am only noting major changes.
l. 184 model [singular]
l. 189: ‘constant’ instead of ‘persistence’
l. 204: to the city
l. 208: ‘have less time” instead of “loss of control”
[also attached]
Comments for author File: 
Comments.docx
Author Response
Author’s Response to Reviewer-1 Comments
Manuscript ID：ijfs-510923
Manuscript Title: Empowering Women through Microcredit in Bangladesh: An Empirical Study
The authors would like to thank the Editor and Reviewer-1 for their kind further deep review of our manuscript with very constructive comments and suggestions. The general and specific questions or concerns and recommendations are highly appreciated, and corrections have been made accordingly in the revised version of the manuscript to improve its quality. The point-by-point responses to the reviewer’s comments and suggestions are listed as follows:
COMMENTS ON REVISED “EMPOWERING WOMEN THROUGH MICROCREDIT”
The previous comments have been well received by the authors and the appropriate changes made. It is now a clearly-presented, compact discussion of evidence on the relationship between microcredit and empowerment in a particular case. A useful contribution is the detailed breakdown and discussion of the empowerment indicators.
Publication is recommended, subject to further copy-editing to clean up numerous, albeit minor language errors that are mainly the fault of the irrational complexities of English language grammar (only some of which are indicated below).
Response: Thank you so much for your kind consideration.
EDITORIAL
Lines 178-9: I suggest replacing the last sentence with: “Higher-income women may demand less microcredit because they are less dependent upon credit for their income-generating activities or because they have better access to other sources of finance.”
[ that’s the demand side; l. 176 gives a supply-side explanation]
Response: Replaced according to your kind improvement. Please see the line 174-176.
l. 210 ‘did not significantly affect’ ?? The evidence from Table 8 shows a higher likelihood for borrowers to purchase independently, but Table 7 indicates that the differences are not statistically significant.
Response: Modified as suggested. Please see the line 206-208.
l. 20: borrowers
Response: Corrected. Please see the line 19.
l. 21: influences [singular]
Response: Changes the word by “impact” during editing. Please see the line 21.
l. 33: achievements
Response: Corrected. Please see the line 33.
l. 37: provides an influential
Response: Corrected. Please see the line 37.
l. 40: important instrument for the poor
Response: Corrected. Please see the line 40.
l. 42: microcredit institutes change ….. institutions, which has
Response: Corrected as suggested. Please see the line 42-43.
l. 47: programs
Response: Corrected. Please see the line 47.
l. 52: compared
Response: Corrected. Please see the line 52.
l. 57: delete “in”
Response: Deleted. Please see the line 57.
l. 62: women tended to lose instead of “the more women lost
Response: Replaced according to your kind improvement. Please see the line 62.
l. 64: amounts [plural in both cases]
Response: Corrected. Please see the line 64.
l. 65: delete “of”
Response: Deleted as suggested. Please see the line 65.
l. 70: delete “the” ‘observe’ instead of ‘observed’
Response: Corrected. Please see the line 70.
l. 72: successful
Response: Corrected as suggested. Please see the line 72.
l. 184 model [singular]
Response: Corrected. Please see the line 183
l. 189: ‘constant’ instead of ‘persistence’
Response: Replaced. Please see the line 186.
l. 204: to the city
Response: Corrected. Please see the line 200.
l. 208: ‘have less time” instead of “loss of control”
Response: Replaced according to your kind suggestion. Please see the line 204.
We are really grateful for your kind deep review, improvements and suggestions. Hopefully, this version as attached will be acceptable for your kind consideration.
With Thanks
Professor Linping Wang
Author Response File: Author Response.pdf
Reviewer 2 Report
The paper has been significantly improved and is now publishable after thorough language editing.
Author Response
Author’s Response to Reviewer-2 Comments
Manuscript ID：ijfs-510923
Manuscript Title: Empowering Women through Microcredit in Bangladesh: An Empirical Study
The authors would like to thank the Editor and Reviewer-2 for their kind further deep review of our manuscript with very constructive comments and suggestions. The general and specific questions or concerns and recommendations are highly appreciated, and corrections have been made accordingly in the revised version of the manuscript to improve its quality. The point-by-point responses to the reviewer’s comments and suggestions are listed as follows:
Comments:
The paper has been significantly improved and is now publishable after thorough language editing.
Response: We have made a substantial editing to improve the sentence structure of the paper. All the mistakes mentioned by the reviewer were corrected along with other mistakes which were identified by the authors during revision. Hopefully, this version as attached will be acceptable for your kind consideration.
We are thankful the Editor and Reviewers for your favorable attention and consideration. Please feel free to contact, if you have any question regarding this submission.
With Thanks
Professor Linping Wang
Author Response File: Author Response.pdf
