Choice constraint and the gender dynamics of labour markets in West Bengal

Lead Research Organisation: London School of Economics and Political Science
Department Name: Gender Institute

Abstract

This research project addresses the first overarching research question in the call specification: the factors which shape pathways into and out of poverty, how people experience these factors and the role of policy in creating sustained routes out of poverty. It is premised on the recognition that well-functioning labour markets are a key institutional channel through which the sustainable reduction of poverty can be achieved. They are also significant from a gender perspective. There is a substantial body of evidence to suggest that gender equality in labour market opportunities has a positive impact on survival, wellbeing and human capital of family members. And where these opportunities relate to regular and reasonably well paid work, they enhance women's agency and bargaining power within family and community.

This research project is motivated by the following puzzle: The Indian economy has experienced extremely strong rates of economic growth in recent years. This has been accompanied by declining rates of fertility and rising levels of education, including female education. These are changes that have been associated with an increase in women's labour force participation rates in much of the world. Yet women's labour force participation has not only remained low in the Indian context, but has declined in recent years. Moreover, women continued to be concentrated in a more limited range of labour market activities than men.

The objective of our research is gain a better understanding of the barriers to women's labour force participation in the Indian context through a detailed case study of the gender dynamics of labour markets in the state of West Bengal. We propose to use an interdisciplinary framework which draws on economic and sociological explanations of labour market behaviour and the careful combination of quantitative and qualitative methods to identify these barriers and what can be done to overcome them. A number of policy case studies will be documented to provide more specific examples of measures that can help to overcome barriers that particularly affect poorer women.

The selection of West Bengal as our case study reflects a number of considerations. First of all, it typifies the trends that make up the puzzle at the national level. West Bengal has one of the highest rates of growth in India, significant declines in fertility and rising levels of female education but continues to report extremely low rates of female labour force participation which have declined further in recent years. It therefore provides a very appropriate context in which to explore the barriers to women's labour force participation and what might help to overcome them.

Secondly, we propose to take advantage of a similar study on the gender dynamics of labour markets on-going in neighbouring Bangladesh to design the West Bengal study as part of a larger comparative project. The two Bengals are not only geographically contiguous, they were once part of the same political entity, they have a shared ethno-linguistic identity and similar cultural norms, including norms regarding women's labour force participation. And while their policy regimes have diverged considerably since their partition in 1947, both have experienced strong growth rates in recent year and declining rates of poverty. Yet while female labour force participation rates in West Bengal have been low and declining, those in Bangladesh have risen steadily since the 1970s to levels now higher than those in West Bengal. While the West Bengal study will constitute a stand-along research project into the barriers to women's labour force participation, we believe that the opportunity to carry out a comparison with Bangladesh will allow us to better understand the interaction between cultural norms and economic motivations in shaping women's labour market behaviour in contexts with similar norms of gender propriety but diverging policy regimes.

Planned Impact

Who will benefit? The aim of the research is to understand the barriers to women's participation in the labour market, particularly those facing women from poorer and socially marginalized groups. These women are our primary beneficiaries. Our secondary beneficiaries are the various stakeholders who are interested in the fairer and more effective functioning of labour markets. They include government officials at local and national levels, development NGOs, women's groups, labour organizations as well as sections of the corporate sector and the wider public. At the international level, they include organizations that work specifically on gender and labour market issues, such as ILO and WIEGO, along with various other organizations , such as the World Bank, UN Women, bilateral donors and private foundations who have put women's economic empowerment high on their agenda.

How will they benefit? We believe that the improved understanding of how labour markets work from the perspective of poorer people, particularly women from socially marginalized groups, together with the practical implications which flow from this understanding, will contribute a) to the better design of current and future policy efforts to address the relevant constraints b) to more informed advocacy on the part of organizations seeking to promote gender equality in labour market opportunities and c) to greater awareness on the part of the public of the realities of poorer women's working lives and the contributions they make. While these impacts are likely to be strongest at the national level, the growing interest in women's economic empowerment on the part of international agencies suggests that they are also likely to benefit from lessons from the South Asia context.

Strategy for impact: Our strategy will be to work those who have an interest in seeing change happen, to involve them as early as possible in the research process and to use our collective resources to influence those who have the power to make change happen. The National Reference Group that will be set up for this project are part of this strategy. They have been selected for their knowledge of the research context and their experience in the policy world. We will work with the Group throughout the project to engage the wider constituency of national stakeholder outlined above and to help us organize two consultation workshops, one held in the inception phase of the project with a view to sharing our research objectives and seeking advice on key research questions and a second towards the end of the project to share our findings and explore its practical implications for different groups of stakeholders. We will seek to maximize our impact with different stakeholders by disseminating our findings and recommendations through a variety of means: policy briefs, blogs, seminars, workshops, informal meetings and so on.

We will also seek to expand and strengthen the public constituency for women's economic empowerment in India by using the 'women in media' network in India (we have been in contact with its West Bengal members) to engage the wider public to discuss cultural preconceptions about women's work in India specifically, and South Asia more generally, and whether our findings support the widespread belief that their absence from the labour market is largely a matter of choice.

Internationally, we would seek to engage with organisations that are interested in, or would benefit from, the policy implications of our findings through seminars, policy briefs and our participation in various conferences and networks. An International Reference Group, drawn from the World Bank, ILO and WIEGO, has already been constituted for the Bangladesh component of the study. We will draw on this group to provide advice and assist us in our efforts to disseminate relevant findings and recommendations from the West Bengal study as well.

Publications

10 25 50

 
Description The key findings reported here are based on preliminary analysis of two surveys and qualitative interviews. The surveys were of 3599 women in 7 districts (we could not cover the planned 8th district due to political problems at the time of our survey) and 1642 men in 5 of these 7 districts. Qualitative interviews were carried out with 68 women and 40 men from the different districts covered by the surveys. Interviews were also carried out with 20 employers of various kinds from these districts.
Patterns of women's labour force participation
The overarching objective of our research was to examine why female labour force participation in West Bengal, and India more generally, was so low and to analyse the role of structural inequalities in the market-place as a factor in reproducing poverty. Our decision to collect data on male labour market behaviour, although there has been quite a lot of research on this topic, was because we believed that analysing male behaviour through the same conceptual lens as we brought to our analysis of female behaviour would give us a deeper understanding of the gender dynamics of labour markets in the West Bengal context.
The analysis of our West Bengal data was framed by a number of stylized facts about gender in Indian labour market. First, as we mentioned, there has been steady decline in female (but not male) labour force participation rates over the last few decades. Much of this decline appears to have happened among married women in rural areas. Second, there is a wide body of literature attesting to the existence of a U-shaped relationship between female education and their labour force participation: women with little or no education have much higher than average rates of labour force participation, their rates decline as education levels increase and then rise again once their education levels exceed the secondary levels. And finally, the less well known of these stylized facts comes from research carried out by Professor Ashwini Deshpande which suggests that variations in the size of the gender wage gaps across the wage distribution for regular workers points to the existence of a 'sticky floor' in the India wage distribution: in other words, gender wage gaps are larger for lower wage earners and smaller for those at the higher earning end. Professor Deshpande has published work on this using India-wide data but is in the process of writing up a paper that finds the same phenomenon for the wage distribution in West Bengal. While the data used excludes casual and irregular wage earners, there is robust evidence to suggest that the gender wage gap is even higher in informal waged work where wages are lowest.
For our surveys, we drew on definitions of economic activity used by the National Sample Survey of India but greater care in the interviewing process meant that our estimate of women's economic activity rate in West Bengal was 52%, considerably higher than the 17% estimated by the NSS for West Bengal. Nevertheless, only 5% were in formal salaried employment, 13% were in informal wage work, 2% were in outside self-employment, 6% were in home-based self-employment, 2% were in unpaid family labour while the bulk of working women were in expenditure saving activities (25%) while 48% were outside the labour force.
Thus around 30% of women (in expenditure-saving work and unpaid family labour) in our sample were not earning any income. The overwhelming majority of these two categories described household work as their primary activity and did not report any other activity. These are women who would have been classified as inactive by conventional labour force surveys but were classified as active by our survey because of additional probing questions.
Our cross sectional survey data clearly cannot tell us very much about the decline in the rural female labour force participation recorded in West Bengal, but the difference between NSS estimates and our own would suggest strongly that many of the women deemed to have dropped out of the labour force by NSS may, in fact, have entered these expenditure-saving/unpaid labour activities. These women do not differ from the rest of the population in our survey in terms of caste or religion.
However, they were more likely to be married than the rest of the sample and were less likely to have young children nor was there any systematic differences in their responsibility for child care or care of the elderly. The marriage effect on women's labour force participation is generally strong in South Asia , reflecting a combination of status considerations and the particular responsibilities of married women. The decline in female labour force participation in India more generally, and in West Bengal in particular, suggests that there seems to have been a strengthening of the marriage effect over time. We will investigate why this might be the case in future research.
Bearing in mind the likelihood that those in expenditure saving activities, unpaid family labour and possible home-based self-employment are likely to have been missed by conventional labour force surveys and classified as inactive, our cross tabulations also suggest an explanation for the U-shaped relationship between female education and labour force participation rates. It suggests that women at the two ends of the educational spectrum are most likely to be in wage labour outside the home and hence most likely to be recognized as economically active by conventional labour force surveys.
50% of women in informal wage labour in our survey are illiterate (compared to 25% of the overall sample). These are likely to be the women who account for the high rates of labour force participation recorded among those with the least education. Concentrated at the other end of the educational spectrum are women in formal salaried employment, of whom nearly half have post-secondary education (compared to an average 18% of the overall sample): these women are most likely to account for the high rates of labour force participation recorded for those with the highest levels of education. The women in the declining arm of the U-curve and at its bottom are mainly in expenditure saving/unpaid/home-based self-employment; they are more likely to be literate than women in formal wage labour (between 21-30%) and less likely to have post-secondary than those in formal waged labour. Their education levels and those of the economically inactive are similar with over 50% with either primary or secondary education.

Structural inequalities and labour market segmentation
Given the broad theme of the ESRC/DFID call, we were particularly interested in the role of labour markets in reproducing structural inequalities. Analysis of occupational preferences among the survey population suggested that women's preferred occupations reflected a concern with cultural constraints and status considerations: in order to preferences, they were semi-skilled home-based self-employment (crafts, tailoring etc); household work (domestic labour); and formal salaried work. Their least preferred were daily wage labour, within and outside agriculture, and animal husbandry. Given that the most educated women in our sample overwhelmingly cited formal salaried employment, preferably for government, as their most preferred form of employment (and it was also cited as most preferred by men in our survey), we will take this occupation as a marker of privilege, the form of work that women would take up if they could. At the lower end of the preferred hierarchy of occupations was informal daily wage labour: we take this as a marker of dis-privilege, the form of work women would avoid if they could.
Regression analysis of our female survey data used the Heckman selection model to control for the effects of various individual and household characteristics on the pattern of labour force participation, condition on their participation in the labour force. This analysis helped to sketch out the outlines of a highly segmented labour market in which gender intersected with other forms of spatial and social inequalities to create labour market sub-segments among the female labour force.
First there was a clear geography to the pattern of women's work activities. Urban areas reported higher levels of female labour force participation than rural and also reported significantly higher levels of formal employment than rural areas (regression analysis of male labour force activities suggested that there was less of an urban rural divide in men' access to formal employment). Urban areas were also less likely to report home-based self-employment and unpaid family labour Rural districts, on the other hand, all reported higher levels of informal wage labour than urban Expenditure saving work did not seem to vary a great deal by district with the except of 24 South Parganas which reported higher levels.
Secondly, if we use education and assets as a proxy for the economic position of households, there is also a class dimension to labour market segmentation for women. Education operates more clearly than assets to distinguish between households at different ends of the economic spectrum. According to our survey data, post-secondary education held the key to accessing formal salaried employment, while productive assets allowed women in engage in home-based self employment, the preferred activity. By contrast, the absence of education condemned women to the least preferred activity mentioned by both men and women: informal wage labour.
There is also a social dimension to stratification based on identity and adherence to cultural norms. While social identity is likely to be correlated with economic position, the regression suggests that, controlling for education and assets, social identity had an independent effect on labour market activities. Religion and caste did not prove significant in distinguishing between women within and outside the conventional labour force: Muslim women were as likely to be in the labour force as Hindu women while women who used some form of veiling were as likely to work as women who did not veil.
However, both caste and religion proved important in channelling women into different labour market activities. First, while there was little difference in activities among lower caste Hindus (Scheduled castes, Scheduled Tribes and Other Backward Castes), higher castes, both Muslim and Hindu, were less likely than these groups to engage in informal waged labour. Secondly, religion also differentiated labour force activities. Muslims from both higher and lower castes were more likely than other groups to engage in home-based self-employment. And thirdly women who practiced some form of veiling , both Muslims and Hindus, were less likely to be in work outside the home.
Finally, while women's unpaid domestic responsibilities did not have much impact on the likelihood of participating in the labour force, it emerged as extremely significant in shaping the work patterns of those who did participate. Those who had primary responsibility for child-care were significantly less likely to engage in work outside the home than those who either did not have such responsibilities or hd others to help them. Primary responsibility for domestic chores increased the likelihood of expenditure saving work. However , the impact of primary responsibility for the elderly appeared counter-intuitive: women who reported such responsibility were more likely to engage in informal wage work and outside self-employment. One plausible interpretation for this reflects the fact that looking after older people, unless they are very old, is a different order of responsibility compared to looking after children. Since the main impact of looking after the elderly related to forms of work carried out by poorer women, it is possible that having an older person in the house reduces women's unpaid domestic work burdens (including looking after children) and allows women who cannot afford to work within the home to seek an income outside the household.
Qualitative findings
The variable-based outline of labour markets sketched out by our survey analysis was fleshed out in greater detail by our qualitative interviews. Its confirmed that labour markets are indeed geographically segmented, as suggested by the survey but are characterised by a more fine grained segmentation than suggested by district level data or by the urban-rural divide , although these provide a first level of analysis.
Women in urban areas were more likely to be found in formal wage employment. As we suggested, these women are most likely to be located in the upward sloping arm of the U-relationship between education and women's labour force participation. Our qualitative interviews suggested that these women worked in both public and private sector jobs, exercised considerable agency in their decision-making, many spoke of flexible hours of work and most could afford to hire domestic help. The greater formality of their work places, including regulation by labour legislation as well as their ability to hold out for the kind of jobs that made it worth their while to take up work (since most had an earning male adult) may explain why the gender gap in earnings in jobs in which women with post-secondary levels of education were most likely to be found tended to be smaller than in jobs lower down the wage hierarchy.
The question of choice was far less in evidence for most of the women in our rural locations. In many of these locations, the women interviewed reported that women ('like themselves') were involved in just one or maybe two income-earning activities - most often home-based piece work provided by a local industry, a trader or a middle man. These might be rolling bidis (local cigarettes), sewing buttons onto garments and so on. While such activities were classified as home-based self-employment, they are more accurately classified as a disguised form of wage labour, work that is subcontracted out to women directly by a factory or via an intermediary to be carried out on a piece rate basis within the home. It is therefore likely to account for a great deal of what is described as home-based self-employment in our survey . Its advantage over actual home-based self-employment was that it did not require women to have to find markets for their products. Markets were assured, at least as long as orders came in. However, our evidence suggests that intense competition in these markets had led to static or declining rates of payment.
The qualitative research also picked up on market segmentation by household economic position but highlighted more clearly how closely bound up economic position was with social identity. First of all, social identity, particularly along caste lines, was itself correlated with economic position - particularly distinguishing lower castes (Hindu or Muslim) from higher castes. What is picked up by our descriptive statistics, but not by the regression analysis, is that while upper caste Hindus and Muslims are both less likely than Scheduled caste Hindus to engage in informal wage labour, women from Brahmin and Forward caste Hindu households are over-represented (the percentage of Brahmins is three times their share in our sample population) in formal salaried employment while Muslims, Scheduled Tribes and Other Backward Castes are seriously under-represented (half their share in the overall population).
Clearly there are cultural as well as economic aspects to questions of identity that combine in demarcating identity-based segments to the labour market. While concerns with status and respectability should be added to economic motivations in explaining women's labour market choices, these constraints operate in very different ways across the socio-economic hierarchy. In families who enjoy social status or aspire to social status, women are constrained them to 'respectable occupations' - in the absence of which they are required to stay at home. These constraints may be internalized as part of growing up in these families or imposed by husbands or in-laws. As one of the women interviewed in Kolkata told us, she herself was very anxious to take up paid work, she was bored at home but she was not a graduate and was forbidden to look for a job by her husband and in laws, as with her level of education, she was unlikely to get the kind of job a family like theirs considered appropriate. So it is the dearth of respectable, socially acceptable jobs combined with the constraints imposed by families or internalized by women themselves with regard to less respectable jobs , that contribute to the high levels of inactivity or invisible economic activity, among women with primary and secondary education.
These findings bring into view a further aspect of the U-shaped relationship we have been discussing. A great deal of the literature on this has focused on supply side explanations to explain why educated women are less likely to be in the labour force - either in terms of an income effect since they are likely to belong to better off households and don't need to work or a status effect to signal the ability of their households to dispense with women's labour. However, there is increasing attention to the demand side and the extent to which jobs appropriate to higher levels of education are available for women as their education levels rise. Our survey tells us that the wider literature has found: men are much more likely to dominate in the kinds of formal salaried employment sought by educated men and women.
So our findings emphasize the need for a greater focus on the demand side for educated female labour - both the overall availability of jobs but also the extent to which these jobs are open equally to educated men and women. Additional regression analysis suggests that many of the women in our survey who are currently outside the labour force would be willing to take up paid work if it was available, with most of them favouring part time work. Our qualitative data supported both the finding that women in many of our study locations believe that there is simply not enough work available for them although a number of them believe that suitable training and skills might help them expand their work options.
Our qualitative data also pointed to a further explanation for the decline in female labour force participation recorded in the official data. Women in our interviews pointed to various changes in their local economy that had reduced the work available to both men and women. One important change was mechanization of agriculture: tractors, threshing machines, rice husking mills - all these had reduced the demand for wage labour. The other was the emergence of synthetic substitutes for products that used to be made within the home or in local industry - plastic mats, paper bags and so on. Geographical constraints meant that women were hit harder than men. Many of those we interviewed spoke of men from their villages migrating to Kolkata and other towns in West Bengal but also further afield to Kerala, Rajasthan and other states in India where they found jobs as well as higher incomes. Women, on the other hand, even those from SC and ST households who were accustomed to working outside the home, were restricted - by male family members as well as their domestic responsibilities - from seeking jobs in more distant places. As a result, not only were earnings declining among the categories of women who did this kind of work (widening local gender gaps in earnings) but also leading them to withdraw into expenditure saving work or economic inactivity.
The qualitative interviews highlighted a number of policy initiatives that held out the promise of expanding women's labour market options. Many spoke of having benefited from the National Employment Guarantee Scheme but reported that there had been a major decline in the availability of these jobs in recent years. One woman who worked as a domestic in Kolkata believed that working for an agency (a Centre) had given her greater security and protection in a form of work generally avoided by those who could because of low earnings and social stigma. There was widespread evidence of self-help groups and access to loans through membership but these seemed to be accessed as a form of safety net in times of crisis or to finance lumpy consumption expenditures more often than productive investments. We also revisited an asset transfer programme for women in extreme poverty (part of the BRAC graduation programme) which Prof. Kabeer had evaluated qualitatively in 2011. This component of the research allowed us to make some conclusions about the longer term effectiveness of the programme: that the two year pilot has operated as a particularly effective form of hands-on training that combined livelihood skills with important life skills that allowed many of the participants to take advantage of new opportunities as they emerged.
Findings from the component on male workers.
This is still in a preliminary stage. Our survey tells us that 35% of male workers were in informal wage employment, 29% in outside self-employment, 15% in formal salaried employment and 9% in home based self-employment. Only 2% said they worked as unpaid family labour, 2% worked in expenditure saving activities and 7% were inactive. Men were thus spread across a wider range of outside work, signalling their greater mobility, and many more had access to formal employment. We will be exploring the interaction between social identity and economic constraint in greater detail for men in the coming months.
Exploitation Route My co-investigator and I would like to work with NSS officials in West Bengal to persuade them to experiment with 'tweaking' their questions so as to obtain a better estimate of women's economic activity. The findings about women's demand for work that is compatible with their domestic chores as well as the effect of labor-saving technologies on their domestic work load are also worth taking forward. The documentary and working paper on the short and long term assessment of the asset transfer pilot to women in extreme poverty has been produced for teaching and training purposes. It is being used by Open University, Birkbeck College and SOAS. I also used it to discuss RCTs and qualitative assessments with the evaluation unit at the Asian Development Bank
Sectors Communities and Social Services/Policy

Education

 
Description The first published output from the research came out as a co-authored working paper for the Department of Economics at Ashoka University, India in 2019. The second version of this paper was published as a working paper for UNU-WIDER, Helsinki (2021) and then a revised version came out as an article in World Development in 2024. These versions have been widely cited. The 2019 version was cited 53 times as of March 2024 while the 2024 version has already been cited 27 times (Google Scholar). This paper feeds into an ongoing debate about the role of gender norms in restricting women's life choices and life chances in various domains of life, including the labour market. Those that argue in favour of a norms-based explanation tend to focus on restrictions on women's labour supply. So for instance the 2021 paper has been cited in a paper by World Bank economists in the World Bank Research Observer, an influential journal in the policy domain , to support their argument that gender norms have blocked progress on gender equality, including in labour markets, in South Asia (1). This however is not the main argument of our paper. We argue that the jobless pattern of economic growth in India has been a major factor in explaining why female labour force participation rates in India is lower than most other countries in south Asia. In our next paper, we will be using detailed micro-level analysis from West Bengal to support this point. Ashwini Deshpande has included citations to the paper in several opinion pieces, such as the recent one in Project Syndicate (2), which got reproduced widely across platforms and also got included in the monthly selection of the CGD blog (3). Additionally, her column in VoxEU (4)got republished in VoxDev (5), which summarises the results of the paper. These blog posts are further amplifying the outreach of the paper. References: (1)Bussolo, M.; J. Amarachi Ezebuihe, AM Muñoz Boudet, S. Poupakis; T. Rahman and N. Sarma (2023) 'Social Norms and Gender Disparities with a Focus on Female Labor Force Participation in South Asia' World Bank Research Observer pp. 1-35, https://doi.org/10.1093/wbro/lkad010 0:1-35 (2)Are Social Norms Really the Main Cause of Low Female Employment? by Ashwini Deshpande - Project Syndicate https://www.project-syndicate.org/commentary/social-norms-not-only-cause-india-low-women-employment-rate-by-ashwini-deshpande-2024-01 (3) Economics & Marginalia: January 19, 2024 | Center For Global Development. https://www.cgdev.org/blog/economics-marginalia-january-19-2024 (4) Indian women are not dropping out of paid work voluntarily | CEPR https://cepr.org/voxeu/columns/indian-women-are-not-dropping-out-paid-work-voluntarily (5) Indian women are not dropping out of paid work voluntarily | VoxDev https://voxdev.org/topic/labour-markets/indian-women-are-not-dropping-out-paid-work-voluntarily
First Year Of Impact 2020
Sector Education
Impact Types Societal

Policy & public services

 
Title Survey questions relating to women's economic activity 
Description India has long reported very low rates of female labor force participation. While this reflects a genuine phenomenon, it also reflects the way that labor force participation is defined. We modified the National Sample Survey (NSS) instrument to better reflect women's economic activity rates, using ILO definitions of labor force participation as our standard. We measured their participation in economic activity through a series of questions which we asked sequentially. First we asked if women were involved in any economic activity in the past 12 months, where economic activity was defined as productive work that either earned an income or saved on household expenditure. These would be defined as economically active according to conventional ILO definition but the latter element is excluded from the NSS definition. Secondly, of those women who answered 'no' to this question, we asked a series of explicit questions about economic activities that tend to be treated as 'housework': working on kitchen gardens/orchards, rearing poultry, free collection of fish, small game, wild fruit, vegetables for household consumption, husking paddy, preparing jaggery (gur), preservation of meat or fish, weaving baskets/mats, making cowdung cakes for fuel, tailoring/weaving and tutoring of own or other children free of charge. We asked whether they were involved in any of these activities and whether they did it to help to generate income rather than just home consumption. Those who undertook these activities to generate income were counted as economically active according to the extended definition. These are all activities that would fall within the SNA boundaries. Thirdly we classified all those engaged in expenditure saving activities but not doing paid work because of non-availability of such work as involuntarily unemployed. The rest were classified as out of the labour force ie. economically inactive. Using the conventional definition, our estimate of FLFPR is 27.85 percent. This estimate is significantly higher than the 17 percent for the entire state of West Bengal from the 68th round of NSS Employment-Unemployment Survey of 2011-12, which is the closest survey for which data are publicly available. Based on the expanded definition of economic activity, we get a female labour force participation rate of 52 percent. We plan to publish this paper but also to engage with NSS officials in West Bengal to 'tweak' their survey instrument to better capture women's activity. 
Type Of Material Improvements to research infrastructure 
Year Produced 2018 
Provided To Others? No  
Impact Not public yet but widely disseminated among policy-oriented researchers 
 
Title Used methods and findings of RCT and qualitative assessments of two very similar pilot programs to suggest how each could strengthen the contributions of the other 
Description I have used this component of the research to both carry out a qualitative impact assessment of an asset transfer programme for women and families in extreme poverty, using the current research to revisit in 2017 participants of the project who had been evaluated in 2011 in order to compare how they had fared over the longer term. Along with the actual evaluation, I have produced a number of outputs intended to feed into teaching and training. These include a journal article comparing RCT and qualitative approaches in West Bengal and Sindh, a reflection on our previous research. I also produced a blog and pod cast on the Oxfam blogsite (From poverty to power). develop training and teaching materials My working paper on this topic provided a detailed account of the methodology and findings of an RCT and a qualitative assessments - of both short and long term - impacts of an asset transfer programmes for women in extreme poverty in west bengal. Both were carried out around the same time and within close proximity to each other. My working paper shows how the two approaches differed, and what each could contribute, to our understanding of a) distribution of impacts b) interpretation of impacts and c) causal mechanisms of impact 
Type Of Material Model of mechanisms or symptoms - human 
Year Produced 2018 
Provided To Others? Yes  
Impact The blog has had over 5000 reads and the pod cast over 900 listens. The Open University will be included the podcast as part of its teaching materials. I have also produced a working paper and documentary on our west Bengal evaluation - it is already being used in SOAS and Birkbeck for teaching and I am hoping to work with Impact Initiative to make it more accessible. I also presented our findings on West Bengal to senior analysts at the Asian Development Bank in Manila and to the BRAC Targeting the Ultra poor team in Bangladesh. 
URL http://www.lse.ac.uk/international-development/Assets/Documents/PDFs/Working-Papers/WP199.pdf
 
Title Primary survey data 
Description We have collected labor force data on a randomly selected sample of 3701 women along with 1817 men randomly selected from households covered by the women's survey. The data was collected from 7 districts in West Bengal, 57% from rural areas and 43% from urban. We also collected qualitative interviews with 60 women and 40 men from our surveys which is being entered into NVIVO 
Type Of Material Database/Collection of data 
Year Produced 2018 
Provided To Others? No  
Impact None yet 
 
Description A blog summarizing findings of our workshop report on constraints, opportunities and poiicy challenges relating to women and labour markets in South Asia and MENA region 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact This blog was intended to draw attention to the report of a workshop we organized (LSE International Development and Middle East Centre; Ashoka University, India and Economic Research Forum) to bring in scholars from the MENA and South Asia regions to carry out an informed discussion on what could be done to overcome barriers to women's market opportunties in two regions with historically low rates of female labour force participation
Year(s) Of Engagement Activity 2020
URL http://www.theimpactinitiative.net/blog/blog-gender-dynamics-labour-markets-south-asia-and-middle-ea...
 
Description Blog on World Bank website 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact The blog, titled Measuring women's work: more vexing than you think, was authored by myself and my co-applicant from Bangladesh and was posted in May 2016 on the World Bank website, Lets talk development, hosted by the Bank's Chief Economist, and dealt with measurement issues raised by our research in West Bengal and Bangladesh (http://blogs.worldbank.org/developmenttalk/measuring-women-s-work-more-vexing-you-might-think). It was widely downloaded. In addition, versions of the blog were reproduced on the LSE Engendering website (http://blogs.lse.ac.uk/gender/2016/06/13/measuring-womens-work-more-vexing-than-you-might-think/) and on the website of the International Association of Feminist Economics (http://feministeconomicsposts.iaffe.org/2016/06/02/the-persisting-challenges-of-measuring-womens-work-a-bangladesh-story/) while the Bangladesh section of the blog was reproduced on the Open Society Foundations website (https://www.opensocietyfoundations.org/voices/are-we-undercounting-work-done-women). I was also interviewed about my research on Bangladesh and West Bengal in a video posted on the LSE's South Asia Centre website
Year(s) Of Engagement Activity 2016
URL http://blogs.worldbank.org/developmenttalk/measuring-women-s-work-more-vexing-you-might-think
 
Description Contribution to ESRC-DFID Power and Partnerships conference on findings from West Bengal 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Study participants or study members
Results and Impact Professor Deshpande presented preliminary findings from our research on choice, constraints and women's labor force participation in West Bengal at the ESRC-DFID conference on The Power of Partnerships in Dehli (Dec. 3-5th 2018). This raised questions which led us to revisit some of our estimates but considerable interest was expressed with regard to our efforts to offer a more refined set of questions to capture women's labor force participation in West Bengal
Year(s) Of Engagement Activity 2018
 
Description Documentary film following up women participants of an asset transfer program for the ultra-poor first interviewed in 2011 and re-interviewed in 2018 to establish longer term impacts 
Form Of Engagement Activity A broadcast e.g. TV/radio/film/podcast (other than news/press)
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact I was granted £12,000 by LSE Knowledge and Research Exchange to work with a documentary film maker in West Bengal who has been involved from the outset with an asset transfer for women in extreme poverty (the Graduation programme) which was piloted in 2007-2010 and which I evaluated in 2011 using in-depth qualitative interviews with 20 participants over a year. Our evaluation was very favourable. I have now returned to re-interview these 20 women in 2018 as part of this project in order to see if the positive impacts lasted. The documentary will follow 6 of these women from when they were first filmed in 2007 to our re-interviews in 2018. The documentary will be completed in April 2019 but we entered a 3 minute version of the film for the LSE Research Competition and won 'highly commended' from the judges. Over 100 students and members of the public attended the exhibition. Both BRAC International and World Bank Social protection platform have viewed the short film and uploaded it on their websites so it has reached many more. I have used the film to present findings about impact assessment to senior staff at the Asian Development Bank. It is also being used in universities for teaching purposes.
Year(s) Of Engagement Activity 2019
URL http://media.rawvoice.com/lse_internationaldevelopment/richmedia.lse.ac.uk/internationaldevelopment/...
 
Description Electronic journal Research Outreach 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact I published a summary of main lessons on randomized control trials learnt from carrying out a qualitative evaluation of an asset transfer programme first carried out in 2011 and revisited 7 years later as part of this project. Research Outreach collaborated to help me write up lessons in professional but easily understood text.
Year(s) Of Engagement Activity 2022
URL https://researchoutreach.org/wp-content/uploads/2021/11/Naila-Kabeer.pdf
 
Description Joint presentation on research findings to annual conference of the International Association of Feminist Economists in New York 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other audiences
Results and Impact 'Gender dynamics of labor force participation in West Bengal' Presentation by Professors Ashwini Deshpande and Naila Kabeer at the International Association of Feminist Economists at University of New Paltz, New York
Year(s) Of Engagement Activity 2018
 
Description Stake holderworkshop on Choice, constraints and gender dynamics of the labour market in West Bengal 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact The workshop was organised by our partners in the Centre for Development Economics at the Delhi School of Economics in collaboration with the Centre for Social Science Research in West Bengal on the 7trh November. We presented our research proposal and methodology and was able to get feedback and suggestions from an audience made up of policy makers, professional practitioners and graduate students. We were also put in touch with an excellent survey firm that had been used by others in the city and have now contracted them to conduct our surveys. The policy turnout was disappointing but we hope that as the research gathers momentum we will be able to invoke greater interest.
Year(s) Of Engagement Activity 2016
 
Description The importance of human agency in impact assessment activities (blog and accompanying podcast) 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact This blog and accompanying podcast used our revisit in 2018 of the Targeting the Ultra Poor pilot which we had first studied through qualitative methods in 2011 to reflect on differences in our findings from those reported by an RCT on a similar pilot in West Bengal (and in Sindh ) and discuss problems arising from failure to consider how human agency can confound the best designed RCT. Both were downloaded numerous times.
Year(s) Of Engagement Activity 2019
URL https://oxfamblogs.org/fp2p/naila-kabeer-on-why-randomized-controlled-trials-need-to-include-human-a...