Climatic Stress, Structural Change and Farm and Non

ACIAR Report ADP/2015/032

Climatic Stress, Structural Change and Farm and Non-farm Enterprise Uptake by Farmers in India and Bangladesh1

Prepared by Ram Ranjan, Thiagu Ranganathan, Asif Reza Anik, Kanchan Joshi, Deepa Pradhan and Brajesh Jha, with assistance from Anup Tiwari and Bidyut Deb

July 05, 2016

1

Contact details: [email protected]; [email protected]; [email protected];

1


Executive Summary India •

• • • • • • • •

• •

•

•

Even as water scarcity in farming pushes households to seek alternative livelihood avenues, it also makes it harder for them to take up enterprises. This is confirmed through the finding that those who face significant water scarcity in the kharif season are less likely to start an enterprise. High water scarcity can adversely affect the capability of farmers to accumulate significant capital needed for starting enterprises. Similarly, the level of cropping intensity is positively related to enterprise uptake probability (increases uptake by about 6 percent when other variables are kept constant). Higher cropping intensity is also positively associated with higher household income, suggesting that higher farm incomes and household incomes can help households with enterprise uptake decisions. Not all households, however, are able to take up enterprises. Some minimum level of education (secondary or higher secondary level) is needed to take up an enterprise. There is a negative association found between the probability of starting an enterprise and stated drought resilience of farmers. That is, more drought resilient households are less likely to start an enterprise. Households that own more livestock are also less likely to take up enterprises. Water buyers are both less likely to start an enterprise as well as have lower household incomes. Households with SHG membership increase their probability of starting an enterprise as well as have a higher household total income. Households with an enterprise have their incomes higher by 10 percent compared to nonentrepreneurial households. However, when we control for selection bias, the contribution of enterprise to total household income turns out negative. This suggests that households may be better off not taking up such enterprises. Migration contributes far more towards augmenting household incomes as compared to business enterprises. Higher cropping intensity is also positively associated with urban migration decision. It is possible that those who migrate are able to bring back savings that allow them to buffer losses in farming as well invest in farm productivity through higher investments in tools as well as better farming practices. The decision to migrate to urban areas is not influenced by water scarcity. Neither is it found to be a caste specific phenomenon. That is, migration may have become a necessity for a large number of poor households in our study area due to lack of incomes and employment opportunities locally. Not all households may be able to migrate. Having more household members helps with migration decision, whereas, having more dependents within the household becomes a constraint. But for those household, who are able to migrate (to urban areas), their

2


•

•

household income is higher by at least 25 percent as compared to non-migrating households. This finding is robust even after controlling for selection biases. For those households who are not able to migrate, the alternative is to supplement their incomes through rural labor. Contrary to the case of urban migration, the probability of rural labor supply declines by about 13 percent for those with higher cropping intensity. We also do not find any association between rural labor supply and water scarcity. Those who supply rural labor have total household incomes lower by at least 15 percent compared to those households who do not supply rural labor. Despite such a low impact of labor supply on household incomes, we find that almost half the sample in our study area (in India) is engaged in this livelihood activity. This finding indicates high levels of stress currently faced by low and small land holding farming households.

Bangladesh

• • • • • • • • • • •

•

Salinity poses serious challenges to the livelihoods of rural households in Bangladesh, particularly to those who are still into crop farming. Those who have exited crop farming or reduced their reliance upon crop farming as a means of livelihood can in fact benefit from higher salinity levels if they have the requisite resources to invest into fisheries. Saline water fisheries is a lucrative business, however, taking up of saline water fisheries tends to be driven more by community decisions as opposed to individual intentions. Our analysis suggests that households are faced with a few alternative livelihood options in these areas comprising mainly labor supply, enterprise uptake and fisheries. Those who take to enterprises, their household heads tend to have higher secondary or above level of education and also own less land compared to those who do not take up enterprises. For landed farmers, enterprise may not be their first choice. The probability of taking an enterprise is also lower for households who participate in labor force or have salaried employed members or own more livestock. Enterprise uptake is higher in Satkhira and Khulna districts and also higher in more salinity affected villages. In terms of reward from entrepreneurial activity, those with secondary education or higher tend to generate 60 percent more income compared to illiterate households. Despite higher enterprise uptake in Khulna and Satkhira, the profits of these entrepreneurs are lower compared to Patuakhali district, suggesting high level of distressed enterprise uptake in Khulna and Satkhira. Remote village area entrepreneurs tend to earn lower incomes due to lack of demand or other distance related constraining factors. Compared to enterprise uptake, labor force participation seems to be very sensitive to the frequency with which households had to purchase drinking water (due to salinity problem). This signifies the presence of most vulnerable groups of household in the labor force category. A higher level of salinity also increases the chances for such vulnerable households to take to labor supply. This form of livelihood tends to draw from the illiterate group of 3


• • •

• •

• •

• • •

households and the probability of participation for the households with educated household head goes down significantly. Those having more family members are more likely to participate in the labor force for the obvious advantages of more earning hands within the family and also for the obvious challenges of more mouths to feed. Having more land or being in alternate livelihood categories such as enterprise, fisheries or salaried employment reduces the likelihood of labor force participation. Satkhira and Khulna districts draw more households into the labor force, suggesting higher level of salinity stress faced by households in these regions. However, some of the higher labor force participation is also demand driven, given relatively higher level of shrimp farming in these regions. Those with more dependents tend to have lower labor incomes suggesting reduced ability to migrate farther or lower flexibility to take up roles that would take them away from dependents for sustained time periods. Level of education of the household does not seem to play a role in determining the ability to start a fishery based enterprise (unlike other livelihood options such as labor force participation, enterprise uptake or salaried employment). This suggests, ease of access to know-how as well as willingness of households to take to this form of livelihood if they had the required resources. Amongst those who take to fishery based livelihood, those in high salinity affected areas tend to earn less suggesting the presence of adverse effects of high saline water on fishery health. Those taking to enterprise have their incomes higher by 100 percent compared to those who are not into enterprise. Similarly, labor force participation and fishery based enterprises are also lucrative, though their contribution to household incomes is lower at 85 and 60 percent respectively (as compared to those who do not take to these professions). Doing crop farming reduces household incomes by 40 percent, whereas salaried employed tend to do the best with highest contribution to household incomes. Compared to the Indian case study, households in Bangladesh tend to have more livelihood options through which they can generate higher total household incomes. Rural farmers in Bangladesh make more through labor supply and enterprise uptake compared to their Indian counterparts. Fisheries based enterprise is also a significant contributor to the incomes of Bangladesh households.

4


Table of Contents

Chapter 1: Introduction......6 Chapter 2: Basic Characteristics of India and Bangladesh Sample Households……21 Chapter 3: Households’ Financial and Physical Capital Endowments……32 Chapter 4: Climatic Stress faced by households……59 Chapter 5: Livelihood Profiles of households……72 Chapter 6: Sample Characteristics of Enterprise Take-up by Farming Households……95 Chapter 7: Regression Analysis……109 Chapter 8: Experimental Analysis of Risk Aversion of Farmers……186 Chapter 9: A summary of key findings—India and Bangladesh……208 Chapter 10: Policy Recommendations and limitations of study……218 Appendix……226

5


Chapter 1: Introduction This chapter presents the project background, a brief review of the relevant literature, study area description and key research objectives.

6

ACIAR Report ADP/2015/032 Section 1.1. Background

As economies grow, the contribution of agriculture to their GDPs tends to decline. This has been true for most developing countries in South Asia. From around 30% in 1990s, agriculture now contributes to just 16% and 18% of GDP for Bangladesh2 and India3 respectively. But, the share of population employed in agriculture has remained high in these economies. The agricultural sector provides around 47% employment in both India4 and Bangladesh5. This is still a reduction from a very high share of around 60% in the early 1990s. If economic growth of developed countries is any indicator, we should be expecting further decline in the populations employed within agriculture. However, the rate of decline in the share of agricultural contribution to GDP in India and Bangladesh has been much faster than the rate of decline in the populations employed within agriculture. This phenomenon, termed as structural stagnation, leads to reduced productivity as well as disguised unemployment6. The situation has been further exacerbated by the fact that the average land sizes in these countries are getting smaller. More than 85% of the farms have less than 1 ha in both these countries and the average operated area is 1.15 ha in India 7 and 0.62 ha in Bangladesh8. Given this context, it turns out that farming alone is insufficient to provide sustenance for rural households, and such households are increasingly forced to exploit opportunities in the non-farm sector9.The major challenge faced by these countries is to help a large majority of their rural populations transition or diversify out of farming into more profitable and sustainable non-farm activities or urban employment. This challenge is further exacerbated by climatic changes10 and differences in inter-generational occupational preferences within a rural household (Ranjan, 2014). Yet, the role of climatic stress

2

http://www.mof.gov.bd/en/budget/14_15/ber/en/Ch-02%20(English-2014)_Final_Draft.pdf http://www.pib.nic.in/budget2015/english/EconomicSurvey_Appendix.pdf 4 http://data.worldbank.org/indicator/SL.AGR.EMPL.ZS 5 http://www.mof.gov.bd/en/budget/14_15/ber/en/Ch-03%20(English-2013)_Final_Draft.pdf 6 Binswanger-Mkhize, H. P. (2012): India 1960-2010: Structural Change, The Rural Non-Farm Sector, and the Prospects for Agriculture, Centre for food Security and the Environment, url: https://woods.stanford.edu/sites/default/files/files/India1960-2010.pdf 7 http://agcensus.nic.in/document/agcensus2010/completereport.pdf 8 http://203.112.218.65/userfiles/Image/ArgYearBook11/Chapter-7.pdf 9 http://www.financialexpress.com/article/fe-columnist/from-plate-to-plough-a-wake-up-call-from-the-farms/67093/ 10 http://www.business-standard.com/article/specials/for-india-s-farmers-harder-life-ahead-find-latest-studies-115042900153_1.html 3

7


and natural resource base depletion in this structural transformation process has not received enough attention in the literature thus far. Rural non-farm diversification holds much promise for these countries. Non-farm expansion has indeed become noticeable in India and Bangladesh in the past two decades and almost 50% of farming household incomes are derived through non-farm based sources in India and Bangladesh. However, most of these incomes have been generated through activities involving casual labour based employment and less through nonfarm businesses or regular salaried employment (Jha, 2011; Himanshu, 2013; Ranganathan, 2015a)11. In reality, among farm based households in India, non-farm businesses contributed only 8% of household income in 2012-13 as compared to 11% in 2002-03 (Ranganathan, 2015b). This seems to indicate farming populations may not be trained in skills required to exploit the non-farm business opportunities12. Despite this, the fact is that currently 6 out of 10 jobs in the rural sector are created in the non-farm sector and it has ‘emerged as the largest source of new jobs in the Indian Economy’ (Binswanger 2012). Other studies have also noted an increase in the rate of livelihood diversification into non-farm sectors in India in the recent past as well as its correlation with poverty decline.13 In order to help farmers move out of poverty and reduce their vulnerability to future climate change, there has also been an emphasis on promoting crop diversification as well as enhancing the marketability of the crops through promoting better supply chains and increasing the uptake of farming technologies. However, there are limits to relying upon increasing profitability of agriculture given marginalization of land and the presence of climatic and ground water stresses. In this context, there is a need to enhance the scope of entrepreneurial ventures in rural areas and learn from countries that have been successful at it. China’s experience over the past two decades in this regard sets an example. China’s non-farm employment (through its private rural enterprises as well as through Township Village Enterprise systems) has not only increased rapidly, it has also seen an increase in capital intensity as well as complexity of economic activity (Zhang and Scott Rozelle 2005) 14. In

11

http://www.livemint.com/Opinion/Dk2OPASLtOibQTCI2gK13M/The-livelihood-question.html http://indianexpress.com/article/opinion/columns/untie-the-farmer/99/ Himanshu, P. Lanjouw, A. Mukhopadhyay and R. Murgai (2011): Non-Farm Diversification and Rural Poverty Decline: A Perspective from Indian Sample Survey and Village Survey Data, LSE Asia Research Centre Working Paper, url: http://www.lse.ac.uk/asiaResearchCentre/_files/ARCWP44-HimanshuLanjouwMukhopadhyayMurgai.pdf 14 http://aciar.gov.au/files/node/642/MN116%20Part%202.pdf 12 13

8


comparison to China, property rights systems as well as policy choice sets are significantly different in India, which makes it difficult to directly follow their approach. With respect to the rate and extent of structural transformation in agriculture across India and China, Binswanger (2012) points out that in the case of India, a high population growth rate (of 1.6 percent) and low rural to urban migration stands in stark contrast to China’s zero population growth rate and rural to urban migration of close to 220 million workers (over the last 20 years). The challenges faced by Bangladesh are similar in many ways to India. However, both these countries have moved differently in some aspects in the recent years. Though India has shown a higher economic growth compared to Bangladesh in recent years, Bangladesh has shown a significant improvement in human capital outcomes as compared to India (Dreze and Sen, 2013). This is likely to pose different sets of constraints to these two economies. The similarity of challenges and differences in constraints provides with an opportunity to understand the issue from a broader perspective. For Bangladesh, even as agriculture has lost its relative importance in terms of contribution to GDP, the sector is still a major source of livelihood for its rural populations. During the 6th Five Year Plan period (2011-15), the agricultural growth rate in the country was below the set target. Further challenges await the sector due to declining trend in the availability and quality of natural resources such groundwater and soil. Farmers periodically lose 12–36% of their harvest due to natural hazards such as flooding (Thomas et al., 2012). A World Bank (2010) report predicted that agricultural GDP will be 3.1 percent lower each year as a result of climate change. Non-agricultural sectors, whereas, have much more potential to contribute. The fishing industry, in particular, fueled by increasing shrimp exports, holds great promise. It employed roughly 10 percent of its labor force and accounted for 6 percent of its GDP in 200615. In short, future flooding, droughts, irregular rainfalls, temperature and increasing salinity levels due to sea water intrusion will further challenge the livelihoods of poor rural farmers. Given such challenges, non-farm enterprises in Bangladesh could play a vital role in pulling a significant rural mass out of poverty and making them resilient to natural hazards.

15

Bangladesh Integrated Water Resources Assessment: Final Report, CSIRO Water for a Healthy Country Flagship Report Series ISSN: 1835095X

9


There exists an extensive literature on the challenges and opportunities presented by non-farm enterprises globally. Nagler and Naude (2014)16, using data for six African countries (from 20052012), find that non-farm enterprises there tend to be small in size and mainly serve local economies. These businesses are also often interrupted during the farming season. In the case of Rwanda, Abbott et al. (2012)17observe that enterprising households can be classified as ‘survivalist’, ‘steady employment types’ and ‘entrepreneurial’. While the last category is mainly found in urban areas, the first category mainly in the rural areas (and is afflicted by poverty), the middle category could be found in both regions. Owoo and Naude (2014)18, explore whether productivity in rural entrepreneurship exhibits any kind of spatial autocorrelation. They find that, in Ethiopia, places where farm productivity is lower also have higher non-farm enterprise productivity. Also, individual traits such as education pay a crucial role in enhancing productivity. When the right set of conditions are present, opportunistic and skilled farmers would be the first ones to start a nonfarm business. We may consider such category of farmers as being influenced by the ‘pull’ effect of profitable opportunities outside farming. However, even when lacking in necessary skills and human/financial capital, a significant rural population would be compelled to follow suit into entrepreneurial activity (EA) if the natural resource base that previously sustained their livelihood in agriculture is now threatened. For instance, declining groundwater levels19 in agriculture could trigger mass scale farming exodus. Some of these farmers would be forced to diversify or venture into non-farm related EA. It is normally observed that farming households typically venture into petty trade (small retails shops, petty roadside shops) or transport and motel businesses. These sectors generally provide low profitability, which further declines when many rural households venture into same/similar substitutable activities. This kind of stress induced entrepreneurial shift is generally more vulnerable to failure as well as could lead to significant burden on the local and central governments’ finances and other resources. Despite there being an extensive literature on the topic of rural entrepreneurship, very little is understood over the nature of factors that compel (or attract) different types of farmers to try

16

Non-farm Entrepreneurship in Rural Africa: Patterns and Determinants, IZA Working paper no. 8008. url: http://ftp.iza.org/dp8008.pdf Abbott, P., I. Murenzi and S. Musana (2012): The Role of Non-farm Household Enterprises in Poverty Reduction, Employment Creation and Economic Growth in Rwanda, Rwanda Journal 26, Series B, Social Sciences, url: http://www.ajol.info/index.php/rj/article/viewFile/78925/69243 18 Non-Farm Enterprise Performance and Spatial Auto-correlation in Rural Africa: Evidence from Ethiopia and Nigeria, url: http://www.worldbank.org/content/dam/Worldbank/Feature%20Story/Africa/afr-nkechi-owoo.pdf 19 http://www.livemint.com/Opinion/97fuaF2aQkO9IjPiPAjMyL/Six-charts-that-explain-Indias-water-crisis.html http://www.livemint.com/Opinion/v4nXpXNxSJtxQNlEbvtJFL/Indias-groundwater-crisis.html 17

10


business ventures. It may very well happen that EA becomes a disguising tool to hide a large number of farm-rejects and this sector would suffer from low capital intensity, productivity and high failure rates in the future. In some instances, geographical factors such as proximity to bigger markets may promote uptake and success in EA. Such beneficial effects are likely to be non-existent or weak the farther the villages are from big towns or main markets. However, clustering of businesses can exist even in remote areas, thereby attracting further entrepreneurial ventures. There are several benefits of clustering, such as reduction of costs through sharing of labor pool and other resources, reduction in costs to provide transportation and energy infrastructure, as well as productivity gains through knowledge spillover amongst competing businesses. Examples of such clustering based gains for micro-enterprises have been noted in the case of Handloom sector in Ethiopia20. Another crucial factor that affects EA uptake is the risk faced by farmers both within farming as well as in the particular EA. Objective as well as subjective perceptions of risks influence investment into non-farm enterprises. The literature holds conflicting theories on their roles, however. One line of thought argues that when farmers face risk of low productivity in farming (say due to reduced rainfall), they would diversify their portfolio by starting non-farm enterprises21. However, there also exists evidence that risk in agriculture can deter investment into non-farmer enterprises when productivity shocks co-vary with profits in non-farm enterprises. This could occur when lower agricultural output reduces incomes of the population and dampens demand for consumption goods which are produced by non-farm enterprises (Rijkers 2013)22. Finally, even if some households would like to diversify their livelihood portfolio when one particular source of income became risky (such as farming), they may not be able to do so if there exist constraints to their abilities to diversify into more productive and less risky avenues. This point has been highlighted by Dercon and Krishnan (1996), using the case of rural households in Ethiopia and Tanzania. They demonstrate that while ownership of livestock was considered

20

Ali, M. and J. Peerlings (2001): Value Added of Cluster Membership for Micro Enterprises of the Handloom Sector in Ethiopia, World Development, 39 (3), pp. 363-374. 21 Rosenzwieg, R. and H. Binswanger (1993): Wealth, Weather Risk and the Composition and Profitability of Agricultural Investments, The Economic Journal, 103(41), pp. 56-78 22 The effects of risks and shocks on non-farm enterprise development in rural Ethiopia, World Development, 45, pp. 119-136, url:http://www.sciencedirect.com/science/article/pii/S0305750X12002574

11


desirable by the households, due to the safety it offers in lean farming periods, not all households are able to own livestock due to low wealth or problems with accessing common pool resources23. Risk also affects farmers’ decisions related to diversification into other livelihood options, such as through wage based labor income or migration income. The literature is largely unanimous in the observation that an increase in income shocks or risks (either ex-ante or ex-post) leads to households increasing their labor supply (Bardhan 1983, Saha 1994 and Rose 2001)24. However, there could be factors other than risk which could also influence households’ labor supply decisions. These may include, poverty, lack of productivity in farming, landlessness, etc. A study by DFID25 (2006) conducted on the ‘role of migration and remittances in promoting rural livelihoods’ finds that migration in Bihar has substantially increased, especially amongst lower caste communities, and despite its risks and hardships, offers upward mobility to migrating households. Migrants bring back valuable skills as well as savings to finance agricultural assets. Additionally, they also acquire new skills in the urban areas that help them diversify into more secure and better paying jobs (DFID 2006). Similarly, for those unable to migrate, wage based labor provides supplementary source of income. Very often such earnings could exceed the expected earnings if they were to start an enterprise according to their capabilities and assets. In summary, it is not clear whether (and which category of) farmers would take to different livelihood venues based upon opportunistic reasons or through sheer compulsion. Nevertheless, understanding the key pull and push related determinants of rural enterprises would help with targeting policy measures that can mitigate resource scarcity related mass scale exodus and crowding into already saturated entrepreneurial options. By quickly identifying what are the key triggers for farmers’ EA, and linking them with available opportunities, it would become possible to ensure a rapid and successful transition for most.

23

Dercon, S. and P. Krishnan (1996): Income Portfolios in Rural Ethiopia and Tanzania: Choices and Constraints, Journal of Development Studies, pp. 850-875, url:http://www.tandfonline.com/doi/pdf/10.1080/00220389608422443 24 Rose, E. (2001): Ex ante and ex Post Labor Supply Response to Risk in Low-Income Area, Journal of Development Economics, 64, pp. 371-388. Sha, A. (1994): A Two-Season Agricultural Household Model of Output and Price Uncertainty: Journal of Development Economics, 45, pp. 245269. Bardhan, P. (1983): Labor Tying in a Poor Agrarian Economy: A Theoretical and Empirical Analysis, Quarterly Journal of Economics, 98(3), pp. 501-514. 25 Deshingkar, P. S. Kumar, H.K. Chobey and D. Kumar (2006): The Role of Migration and Remittances in Promoting Livelihoods in Bihar, url: https://www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/2354.pdf

12


Section 1.2. Project Objectives The overall aim of this project is to explore the key factors that affect farmers’ decisions to take up entrepreneurial ventures and to understand how water scarcity and prolonged droughts influence their overall livelihoods choices. Specifically, we intend to identify interactions between households’ natural, physical, human, financial and institutional capital bases and how these could influence the uptake, competitiveness and profitability of their EA. It is important to keep in mind that apart from farm and non-farm entrepreneurship, migration and wage based incomes also serve as significant sources of livelihood incomes for the farming households. Therefore, the propensity to rely upon entrepreneurial incomes and its prospects cannot be studied in isolation of these alternate livelihood options. It may very well be true that migration and wage based non-farm incomes are still heavily relied upon and preferred by a majority of rural households. Therefore, the objective of this project is to compare and understand the opportunities offered by non-farm business enterprises in the context of the alternate livelihood options such as migration and wage based earnings. Further, from a policy perspective, our goal is understand what household characteristics (along with community, geographical ones) determine that farmers would make a certain type of livelihood choice and would be successful in it. This is an important policy concern as these decisions determine their future resilience and vulnerability to climate and structural shocks. Therefore, the context of water scarcity (and salinity in the case of Bangladesh) is given special consideration in our research objective. Below we identify specific questions and hypotheses that are tested in this study: 1. To what extent is non-farm or high-value farm enterprise take-up affected by water scarcity or salinity challenges? 2. To what extent is migration take-up by farming households water stress or salinity driven? 3. Are migrating households better off compared to those who prefer to take up enterprise or those who just concentrate on farming? 4. Are wage based migrants less likely to take up an enterprise?

13


5. Are farmer-turned-entrepreneurs and farmers who supplement their livelihoods through migration based incomes mutually exclusive categories? What are their distinguishing traits? 6. How do household, community, regional and geographical characteristics determine the likelihood of occupational choices for water starved farmers? 7. Which livelihood option makes farming households more drought resilient? Which livelihoods options are preferred by drought resilient farmers? 8. How do entrepreneurial opportunities in rural areas compare with those in nearby urban centres? 9. What are the challenges and determinants for entrepreneurial success for those who start businesses in nearby towns? What lessons could be derived from their experience for the rural entrepreneurs? 10. Is there a significant difference in the risk aversion of the water stressed farmers compared to those with assured water supply? 11. How do higher measures of risk aversion, such as prudence and temperance, vary across farming households? 12. Do water stressed farmer exhibit more prudence and temperance when faced with risky situations?

Section 1.3. Study Area Description Section 1.3.1. India

Two states within India, Bihar and West Bengal, were selected for our study. Three districts were selected from within these two states. Birbhum and Purulia districts from West Bengal and Nalanda district in Bihar are crucial districts facing varying levels of water scarcity. (Tables A1.1 – A1.5 in Appendix A provide relevant indicators related to these districts). Nalanda has a population of roughly 2.8 million and an area of about 2400 square kms. It has a forest area of 47 sq. km, a cultivable area of 2000 sq. km. and a net sown area of 1600 sq. km.

14


Major soil types are clay loam, fine loam, coarse loam and loam26. There are 114,000 tubewells, 4,000 tanks and 16,000 canals in the district. Groundwater is the main source of irrigation, however, the average groundwater utilization is only 65% indicating much scope for further exploitation.27 The average rainfall is about 1000 mm. Main crops grown in Nalanda are paddy, potatoes and onion. The region has seen farm productivity increases in the recent past owing to adoption of better technology and practices. Some crops such as bitter gourd and onions are being exported to Bangladesh and Nepal28. Birbhum district in West Bengal has a total area of 4,500 sq. km and a population of roughly 3.5 million. It has a forest area of 160 sq. km. and about 75 percent of population is farming dependent. It has a number of cottage industries including textile. Average annual rainfall is roughly 1400mm. Paddy, oilseeds, wheat and pulses are major crops grown in the district. Nearly 75 percent of the land is owned by small and marginal farmers, with average landholding being 1 ha. 29 In contrast to Nalanda, groundwater is over-exploited and groundwater levels are declining rapidly. Purulia district in West Bengal has a total area of about 6,500 sq. km with a population of about 3 million. It is the most drought prone and driest of the three districts chosen for our study. Average rainfall received is 1300mm. Paddy is the main crop grown in the district. It has an undulating topography which is not conducive for groundwater recharge. The three districts vary significantly in soil type, topography and natural resource endowments. While Nalanda predominantly has alluvial soils, Purulia is less fertile and mainly has red and lateritic soil. Birbhum has both alluvial and lateritic types of soil. The district of Purulia is largely undulated with sharp ups and downs; such undulations are however less frequent in Birbhum and Nalanda districts. The above diversity is also apparent in socio-economic and geographical endowments of the regions. Birbhum has couple of large dams for irrigation while Purulia has smaller check dams. Farmers in Nalanda depend largely on private sources of irrigation. Birbhum has a few small towns such as Siuri, Bolpur, and is endowed with limited network of roads; while Purulia has a bigger town and is located in close proximity from a number of large industrial towns such as Jamshedpur, 26

http://www.cgwb.gov.in/District_Profile/Bihar/Nalanda.pdf http://www.cgwb.gov.in/District_Profile/Bihar/Nalanda.pdf 28 http://timesofindia.indiatimes.com/city/patna/Nalanda-emerges-as-export-hub-of-agriculture-products/articleshow/13281568.cms 29 http://www.birbhum.gov.in/DDAgri/PAO.htm 27

15


Bokaro, and Asansol. Nalanda is close to Patna (which is the capital of Bihar) and also to a number of international tourist destinations such as Rajgir and Gaya. Blocks in these districts were selected on the basis of varying levels of cropping intensities30. Cropping intensity can be used as a proxy for water scarcity, as it reflects the status of both types of water sources (ground and surface). Cropping intensity is also a reflection of temperature and rainfall in the area, as both the nature and intensity of cultivation is governed by these climatic factors. The cropping intensity in these regions has therefore been used as main criteria to select blocks based upon low or high water scarcity. As indicated in Table A1.3, cropping intensity in Birbhum varies between 108 and 222. In some blocks of Birbhum, water, specifically surface water, is available for cultivation of crops in both the seasons (rainy and winter); the cropping intensity in such blocks is 200 and above (Nalhati-II). Contrary to this, Rajnagar is a block with cropping intensity of just 104. These blocks were selected as they represent significant variations in cropping intensity in Birbhum. As compared to Birbhum, Purulia is more water stressed and almost all blocks suffer from a uniform level of water scarcity. Cropping intensity in different blocks of Purulia ranges from 103 to 121 percent, indicating that there are very few farms which are cultivated more than once in a year. In terms of cropping intensity, blocks such as Balrampur, Manbazar are better off, but these blocks were excluded for survey purposes as these blocks are affected by naxal31 problems. Taking note of the above practical challenge, Raghunathpur-I and Kashipur were chosen as the blocks with high and low cropping intensities respectively in Purulia district (TableA1.4). Similarly, in Nalanda on the basis of cropping intensity, Silao and Islampur blocks were chosen to present variations in water stress (Table A1.5). List of Selected villages and districts is presented in table 1.3.1. Further details on sampling strategy including study area maps are provided in the Appendix A3.

30

Cropping intensity is calculated as the ratio of gross sown area to net sown area multiplied by 100. A Naxal or Naxalite is a member of any of the Communist guerrilla groups in India, mostly associated with the Communist Party of India (Maoist). The term Naxal derives from the name of the village Naxalbari in West Bengal, where the movement had its origin. 31

16

ACIAR Report ADP/2015/032 Table 1.3.1. A list of selected districts, blocks, villages and Village clusters (VC) S.N.

1

District

Birbhum

Blocks

Rajnagar

Village Cluster I (near block

Village Cluster II (away

office)

from block office)

Rajnagar gram panchayat

Tantiapara GP

(GP)

2

Purulia

Nalhati –II

Bhadrapur –II GP

Sheetal gram GP

Kashipur

Rambani village cluster (VC)

Ranjandih VC

comprising Bauripara,

village

like

Bathandih,

Mahtodih, Puratanbasti Raghunathpur-II

Lachua

VC

under

Bero

panchayat

Thekra

VC

Bansgram,

comprising Beldanga,

Kulgram Ocha, Korapara 3

Nalanda

Silao

Bargaon, Sarinchak, Nirmal

Pakrisarai,

Lodipur,

bigha (Surajpur panchayat)

Govindpur

(Mirpur

Panchayat) Islampur

Panhar VC

Mehmuda VC

Section 1.3.2. Bangladesh Study Area Description The Bangladesh study area selection is based around climate change induced salinity problem in agriculture. In the following sections, we provide, a history of the salinity problem, identification of salinity affected districts and the socio-economic and demographic characteristics of the area. Section 1.3.2.1. Salinity Problem in the Coastal Areas Table 1.3.2.1 presents salinity status and changes in salinity levels in the 18 salinity affected coastal districts of Bangladesh. In a span of past four decades, the area affected by salinity has increased from 833.45 thousand ha to 1,056.26 thousand ha. Except for three districts, salinity levels have increased in the remaining fifteen districts. Increase in salinity is more severe in the western part as compared to the eastern part.

17

ACIAR Report ADP/2015/032 Table 1.3.2.1. Changes in salinity status in different districts of Bangladesh in terms of hectares affected District

Salinity affected areas (000 ha)

Salinity change between 1973-2009

1973

2000

2009

(000 ha)

Khulna

120.04

145.25

147.96

27.92

Bagerhat

107.98

125.13

131.12

23.14

Satkhira

146.35

147.08

153.11

6.76

Jessore

0

10.86

14.99

14.99

South-Western Districts

Narail Pirojpur Jhalakathi Barishal

0

16.05

18.71

18.71

21.30

28.64

35.83

14.53

0

3.52

4.69

4.69

0

10.82

13.96

13.96

Bhola

40.33

93.64

94.57

54.24

Patuakhali

115.40

139.35

155.18

39.78

Borguna

103.55

104.22

95.62

-7.93

Gopalganj

0

10.20

6.27

6.27

Madaripur

0

1.19

0.72

0.72

Laxmipur

19.30

17.50

18.43

-0.87

Feni

9.50

7.30

5.75

-3.75

Noakhali

49.60

53.54

52.52

2.92

Chittagong

45.70

46.50

51.48

5.78

Cox’s Bazar

54.40

59.96

55.35

0.95

Grand Total

833.45

1020.75

1056.26

222.81

South-Eastern Districts

Source: SRDI 2010

Most of the coastal parts in Bangladesh are affected by salinity. Compared to the south-western coastal region, salinity is less severe in the south-eastern coastal region. Most of south-western coastal districts are low-lying and surrounded by hundreds of rivers and canals carrying salty water to the land from the adjacent sea. Salinity in the area is closely related with season and polder (dike) management. Saline water intrusion exacerbates during the dry season (winter season) when water flow is low from the upstream. The low flows allow saline water to travel up the distributaries of the western part of the delta in southwest Bangladesh and ultimately thousands of hectares of lands are rendered unproductive during the season (Figure B4.1, Appendix B4). These coastal polders, with sluice gates for controlling river flows and to protect low-lying agricultural lands from tidal inundation and saline water intrusion, were built during 1960s. The gates are supposed to be closed during mid-November and again opened in May when the monsoon begins. In the monsoon, accelerated water flow in upstream rivers and heavy rainfall makes water

18


in surrounding canals fresh; whereas in winter, as rainfall and upstream water flow reduces, sea water enters the locality. Salinity in groundwater is a major concern in the coastal region, particularly in the dry season. Figure B4.2 (Appendix B4) depicts groundwater salinity during the dry season (November to May) of 2011-12 and the wet season (June to October) for 2012 up to an aquifer depth of 350 m, based on data collected by BWDB, DPHE and IWM. Groundwater salinity is relatively higher near the coast and lower away from the coast. It is generally lower at shallow depths and higher at greater depths. Not all of the coastal area is suitable for farming. Except for the green colored areas (see figure B4.3, Appendix B4), farming in other areas is difficult. People in highly saline areas are largely dependent upon nature for their livelihood sustenance. Capturing fish and collecting wood, honey, leafs, etc. from the Sundarbans are the most important livelihood options for them. Some of them migrate to other areas to work as day laborers. In the green colored areas, farming is possible only during the monsoon season. Aman paddy grown during monsoon is the major crop here. This is the only crop grown in some of the areas. But due to increasing labour costs, farming is becoming less and less profitable. Unlike other parts of the country, farmers here cannot grow paddy in the winter season due to salinity and lack of irrigation water. As winter is the drier season, with less rainfall, farmers have to rely on ground water irrigation. But groundwater irrigation is costly in the saline prone areas as fresh water level is very low. Inability to cultivate paddy in the winter season, forces farmers to look for alternative crops, particularly crops with lower water requirements. Farmers cultivate different types of pulses and nuts. Of these, some common ones are mung, sesame, groundnut, cowpea, sunflower, etc. Mainly, the traditional varieties are cultivated. They do practice modern varieties, but only when input packages (or subsidies) are provided by the agricultural department or NGOs. They are not very much familiar with the modern varieties, neither are these varieties widely available in the local markets. Even those farmers who have knowledge of modern varieties, do not prefer them over the traditional varieties, as the latter require less capital and labour. For some crops, the traditional ones have a higher market price, though provide lower yield. But, in areas where shrimp and other saline water fisheries are possible, farmers are unwilling to go for crop farming. 19


Basically, farmers do farming in those areas where shrimp farming is not possible. In Bangladesh, vegetables are mostly grown during the winter season. Most of the vegetables are saline intolerant and hence cannot be grown in saline areas during the winter season. In summer and in monsoon, options are relatively fewer and limited to local varieties. Moreover, as paddy is grown only once in a year in the saline areas, farmers prioritize paddy over vegetables. Homestead gardening is practiced in the high lands, though to a very limited extent. In the low lying areas, due to water logging, homestead gardening in monsoon is not possible. Homestead gardening is gaining popularity in these areas following interventions from NGOs and the extension department. Sack gardening could be another appropriate technology in the water logged areas, however, it is yet to be widely tested.

20


Chapter 2: Basic Characteristics of India and Bangladesh Sample Households This chapter discusses basic sample characteristics of the households for the study areas spanning districts of Nalanda (Bihar), Birbhum and Purulia (West Bengal) and the districts of Khulna, Satkhira and Patuakhali in Southwest Bangladesh.

21


Section 2.1. Sample Characteristics (India) In all, 1,604 households were surveyed across the three districts in India for the present study. Out of these, 47 inconsistent data were omitted. The remaining 1,557 households were used for the analysis. Table 2.1 provides the characteristics of the overall sample surveyed in our study. Table 2.1 Characteristics of the Sampled Households (Overall) Religion Hindu

87.6%

Muslim

12.3%

Christian

0.1%

SC

24.2%

ST

6.4%

OBC

34.7%

Forward

34.6%

Caste

Household Size (number of members)

5.9

Dependency Ratio (age)—defined according to working age Dependency Ratio (work)—defined on whether working Age of the Household Head Gender of the Household Head

0.27 0.54 50.6 Male

95.2%

Female

4.8%

Illiterate

15.1%

Informally Literate

3.7%

Primary

20.6%

Secondary

44.3%

Higher Secondary

9.6%

Graduate

6.8%

Education of the Household Head

N (number of observations)

1557

In the sample, around 88% households were Hindus and 12% were Muslims. In terms of caste wise distribution, around one-fourth of the households belonged to Scheduled Castes (SCs), 6% belonged to scheduled tribes (STs), around 35% belonged to Other Backward Castes (OBCs) and 35% belonged to forward castes. The average household size in the sample was 5.9. The average dependency ratio, defined as ratio of number of working age (15-59 years) members in the household to the total number of members 22


in the households, was 0.27. So, roughly one in four members of a household was dependent on the other three working age members of the households. We also calculated an alternative dependency ratio, defined as the ratio of number of members who were not working to the total number of members in the households. The average of this ratio was double that of dependency ratio based on working age (0.54). We find that slightly more than one in two persons in the household are dependent on working members of the household. The average age of the household head was 50 years. More than 95% of the households had a male household head. Around two-thirds of the household heads were educated at primary and secondary levels. 15% of the household heads were illiterate. Around 16% of the household heads were educated at a level of higher secondary and above. There is heterogeneity in these characteristics across different districts of the study. The following sub-sections explore this heterogeneity in detail by looking at the sample characteristics across the three districts individually.

2.2. Sample characteristics for Nalanda (Bihar) In Nalanda district of Bihar, 604 households were surveyed in the two blocks of Silao and Islampur. Of these, data from 568 households were used in the analysis. The sample characteristics of these households are presented in table 2.2. Table 2.2 Characteristics of the Sample Households in Nalanda Silao

Islampur

Combined

Religion Hindu

100.0%

93.9%

97.0%

Muslim

0.0%

5.8%

2.8%

Christian

0.0%

0.4%

0.2%

SC

9.5%

11.9%

10.7%

ST

1.1%

1.4%

1.2%

OBC

59.3%

76.3%

67.7%

Forward

30.2%

10.1%

20.2%

Household Size

6.4

7.1

6.8

Dependency Ratio (age)

0.30

0.36

0.33

Caste

23


Dependency Ratio (work)

0.64

0.63

0.63

Age of the Household Head

51.5

48.3

49.9

Male

93.7%

92.9%

93.3%

Female

6.3%

7.1%

6.7%

Illiterate

12.7%

17.1%

14.9%

Informally Literate

1.4%

4.3%

2.8%

Primary

9.2%

16.7%

12.9%

Secondary

48.9%

49.1%

49.0%

Higher Secondary

14.8%

6.0%

10.4%

Graduate

13.0%

6.8%

9.9%

287

281

568

Gender of the Household Head


N

From table 2.2, we observe that all the sampled households in Silao block were Hindus while 6% of households in Islampur were Muslims and 1 household was Christian. Around 10% of sample households in Silao and Islampur were from SC category and around 1% households in both these blocks came from ST category. While Silao had a high percentage of General (30%) and OBC (59%) households, Islampur had a very high percentage of OBC households (76%) and only a low percentage (10%) of households belonging to the General caste group. The average household size in the sample in Nalanda was higher than that in the two districts of West Bengal. It was higher in Islampur (7.1) compared to Silao (6.4). Higher household size also meant a higher dependency ratio when calculated using the working age criteria. The average dependency ratio based on working age criteria was 30% in Silao and 36% in Islampur. The dependency ratio based on working member criteria was similar across Silao (64%) and Islampur (63%). It was also higher in the Nalanda overall sample as compared to the sampled districts in West Bengal. The average age of the household head was slightly higher in Silao (51.5) compared to Islampur (48.3). Compared to the whole sample, there were slightly higher percentage of female headed households in Silao (6.3%) and Islampur (7.1%). This could be indicative of higher migration to 24


urban areas by the men in these districts. In terms of education of the household head, Nalanda had a much lesser percentage of household heads who were illiterate, informally literate or educated until primary education compared to the whole sample and had a higher percentage of household heads who were educated in secondary, higher secondary or graduate level. But, there were differences between Silao and Islampur. In Islampur, there were lesser number of household heads who were educated at higher secondary level or above and more household heads who were illiterate or educated up to primary education.

Section 2.3. Sample characteristics for Birbhum (West Bengal) In Birbhum district of West Bengal, 500 households were surveyed, of which 489 were considered for the final analysis. The survey was conducted in the blocks of Rajnagar and Nalhati-2. The characteristics of these households are presented in the Table 2.3. Table 2.3. Characteristics of the Sample Households in Birbhum Rajnagar

Nalhati-2

Combined

Religion Hindu

86.5%

53.5%

69.9%

Muslim

13.5%

46.5%

30.1%

SC

24.6%

28.6%

26.6%

ST

9.4%

2.0%

5.7%

OBC

13.5%

27.8%

20.7%

Forward

Caste

50.8%

41.6%

46.2%

Household Size

4.8

5.2

5.0


0.23

0.24

0.23


0.54

0.47

0.50


49.9

49.5

49.7

Male

95.0%

99.2%

97.1%

Female

4.5%

0.8%

2.7%



25


Illiterate

13.6%

9.4%

11.5%

Informally Literate

7.4%

1.2%

4.3%

Primary

26.9%

16.8%

21.8%

Secondary

40.1%

50.8%

45.5%

Higher Secondary

7.0%

13.1%

10.1%

Graduate

5.0%

8.6%

6.8%

The Birbhum sample had the highest percentage (30%) of Muslim households in the survey compared to the three districts. In particular, 46.5% of Nalhati-2 block sample comprised Muslim households. Rajnagar sample also had 13.5% Muslim households. Birbhum had a higher percentage of SC households compared to Nalanda. Roughly, one in four households in Birbhum was an SC household. Rajnagar had 10% ST households while Nalhati-2 sample had only 2% ST households. There were 15% OBC households and 50% forward caste households in Rajnagar and 28% OBC households and 42% forward caste households in Nalhati-2. The average household size in both the blocks was around 5. The dependency ratio based on working age criteria was 23% and 24% in Rajnagar and Nalhati-2 respectively while that based on working member criteria was 54% and 47% in the two blocks. The average age of household head was about 50 years in both the blocks. There were very few female headed households in Birbhum. Only 2 households were female headed in Nalhati-2 block. Education of household heads was generally lower in Rajnagar compared to Nalhati-2.

Section 2.4. Sample characteristics Purulia (West Bengal) In Purulia district of West Bengal, 500 households were surveyed in the blocks of Kashipur and Raghunathpur-I. The characteristics of these households are presented in the Table 2.4. Table 2.4. Characteristics of the Sample Households in Purulia Kashipur

Raghunathpur-I

Combined

Hindu

99.6%

89.2%

94.4%

Muslim

0.4%

10.8%

5.6%

21.2%

53.2%

37.2%

Religion

Caste SC

26


ST

24.8%

1.2%

13.0%

OBC

12.3%

5.6%

11.4%

General

34.4%

40.0%

37.2%

Household Size

5.7

5.9

5.8


0.22

0.24

0.23


0.49

0.44

0.47


51.7

53.1

52.4

Male

94.3%

96.4%

97.3%

Female

5.7%

3.6%

4.6%

Illiterate

15.9%

21.6%

18.8%

Informally Literate

6.1%

2.0%

4.0%

Primary

26.4%

29.6%

28.0%

Secondary

39.0%

36.8%

39.9%

Higher Secondary

8.9%

7.2%

8.1%

Graduate

3.7%

2.8%

3.2%



In Kashipur, there was only one Muslim household. In Raghunathpur-I block, there were 90% Hindu households and 10% Muslim households. Kashipur had 21% SC households while Raghunathpur-I had around 53% SC households. Kashipur had 24% ST households while Raghunathpur-I had only 1% ST households. There were 12% OBC households and 34% forward caste households in Kashipur while 6% OBC households and 40% forward caste households in Raghunathpur-II. The average household size in both the blocks of Purulia was around 6, more than that in Birbhum and less than that in Nalanda. The dependency ratio in Purulia, both in terms of working age and working members, was the least among the three blocks. The dependency ratio in terms of working members was the least in Raghunathpur-I at 44%, which meant that on average 56% of household members were working in some job. The average age of the household head was 52 years and 53 years in Kashipur and Raghunathpur-I blocks respectively. Most of the households surveyed were male headed. In terms of educational outcomes of household head, Purulia fared the worst among 27


the three surveyed districts. Only 3.7% and 2.8% household heads in Kashipur and RaghunathpurI had obtained a degree or a diploma.

Section 2.5. Sample characteristics for Bangladesh Section 2.5.1. Socio-economic and Demographic Profile of the Surveyed Households Table 2.5.1 presents socio-economic and demographic characteristics of the households. On an average a household has 4.47 members. Among the household members, 51% are male. Average family size and gender composition of the family members are almost identical across districts. The average age of the household members is 30.5 years. Patuakhali has the relatively lowest average age (27.7 years) and Khulna has the highest average age (33.8 years) among the three districts. Dependency ratios have been estimated based on household members’ age and occupation. The estimated dependency ratio using age implies that 30% of the total household members belong to the dependent age group (age group of below 15 years and above 65 years). The other estimated dependency ratio (using work) reveals that 68.2% of the members are not involved in any income generating activities, i.e., they are mostly students, or retired professionals, or unemployed and involved in household activities which do not yield any cash returns. Almost all the households have male heads (97.3%).The head’s average age is around 45 years. Compared to the national level statistics, the heads have better literacy rate. In Bangladesh 57.91% of the population at 7 years and above have literacy (HIES, 2010)32, whereas among the heads 74.7% are literate. On an average, a household head studied up to the primary level. Around 36% of the educated household heads studied up to the primary level, whereas another 32.4% have formal education up to the secondary level. Hardly any of them have studied beyond the secondary level. Among the three districts, heads in Khulna have better educational status. In Khulna, relatively lower proportion of the heads are illiterate and higher proportion of the heads have studied above the primary level. The average schooling level for the heads is primary.

32HIES

(2010). Household Income and Expenditure Survey 2010. Bangladesh Bureau of Statistics. Dhaka.

28


Casual labour supply in agriculture and/or outside agriculture is the main occupation for almost half (47.8%) of the household heads. Compared to the other two districts, in Khulna relatively lower proportion of the heads (39.7%) indicated casual labour participation as their main occupation. In other two districts, at least half of the heads indicated casual labour participation. Involvement in business or entrepreneurial activities as main income source is mentioned by 14% households. Satkhira has relatively higher proportion of heads (17.19%) into business, whereas in other two districts the proportions are almost similar. Almost one out of every four heads (24.3%) reported agriculture as their main income source. Crop farming is the main occupation for 12.8% of the heads, whereas another 11.5% reported non-crop farming.

Table 2.5.1. Socio-economic and Demographic Profile of the Surveyed Households Khulna

Patuakhali

Satkhira

All

Household Characteristics

Gender Composition of the household Members (%) Male

0.52

0.50

0.52

0.51

Female

0.48

0.50

0.48

0.49

Average Household Size (no.)

4.18

4.65

4.58

4.47

Average Age of the Household Members

33.8

27.7

29.9

30.5


25.4%

34.8%

29.9%

30.0%


66.2%

67.5%

70.8%

68.2%

46.2

42.4

45.7

44.8

Male

96.6%

96.6%

98.8%

97.3%

Female

3.4%

3.4%

1.3%

2.7%

Illiterate

19.1%

27.5%

29.4%

25.3%

Primary

30.3%

46.6%

30.9%

35.9%

Secondary

40.0%

24.7%

32.5%

32.4%

Higher Secondary & above

10.6%

1.3%

7.2%

6.4%

Average Year of Schooling

5.62

3.92

4.58

4.70

Crop Farming

15.3%

20.3%

2.81%

12.8%

Non-crop Farming (Livestock, Poultry & Fisheries)

18.4%

2.8%

13.13%

11.5%

Business

12.2%

12.5%

17.19%

14.0%

Head’s Characteristics Age of the Household Head Gender of the Household Head


Main Occupation of the Head

29


Service

5.3%

5.0%

5.31%

5.2%

Day Labour

39.7%

50.0%

53.75%

47.8%

Others

9.1%

9.4%

7.81%

8.8%

Table 2.5.2 presents information regarding household’s access to pure drinking water. Except West Lunda, in all the villages of Patuakhali, all the interviewed households have to purchase drinking water. They have to purchase water for around five months in a year. In Bethbunia, Shymnagar and Guhalbathan villages in Khulna, almost all households had to purchase drinking water. Drinking water scarcity in these three villages in Khulna is much higher than the villages in other two districts, as the villagers purchase water for more than seven months in a year. Dumuria and Parshemari are the two severely water scarce villages in Satkhira district. Mainuddin et al. (2014a)33 have reported that the south west districts are affected by low water availability in the dry season. Salinity problems in surface and groundwater, along with groundwater arsenic contamination, has resulted in limited irrigation development in the region.

33

Mainuddin M, Kirby M, Chowdhury RAR, Sanjida L, Sarker MH, Shah-Newaz SM (2014a) Bangladesh integrated water resources assessment:

Supplementary report on land use, crop production and irrigation demand. CSIRO: Water for a Healthy Country Flagship.

30

ACIAR Report ADP/2015/032 Table 2.5.2. Household’s Access to Drinking Water District Patuakhali

Khulna

Satkhira

Village

% of HHs purchasing drinking water

Average no. of months HHs purchase drinking water

BoroKajal Boroshiba North Panpatty South Panpatty West Lunda Nijampur Dhankhali Lunda Shibnagar Magurkhali Tulna Boruna Village Paikgacha Bethbunia Shymnagar Guhalbathan Kachrahati Nandigram

100.00 100.00 100.00 100.00 00 100.00 100.00 100.00 2.75 00 00 00 00 98.67 96.90 100.00 00

4.70 4.79 5.22 4.26 00 4.81 4.65 4.37 0.05 00 00 00 00 7.67 7.43 7.00 00

Anantapur Dumuria Parshemari

00 94.29 100.00

00 6.99 6.15

Godara Bodortola

00 00

00 00

Kheyargati East & South Baintola

00 00

00 00

31


Chapter 3: Households’ Financial and Physical Capital Endowments In this chapter, the physical and financial capital endowments of the households in India and Bangladesh study regions are detailed and discussed. Where applicable, distributions of assets and incomes are presented for each quintile and Gini coefficients provide further indication of inequality in asset ownership.

32


Section 3.1. Physical and Financial Endowments of Surveyed Households in India This section describes the financial and Physical assets of the households in the survey. Specifically, we describe the asset holdings including land, livestock, housing, agricultural and nonagricultural assets, and debt portfolios of the households. Table 3.1.1 provides the details of landholdings for the entire sample as well as for the three surveyed districts individually. Table 3.1.1 Landholdings of the entire Study Sample (India) Nalanda Silao

Islampur

Birbhum All

Rajnagar

Nalhati-2

Purulia All

Kashipur

Raghunathpur-I

All Sample All

Landless

2.1%

1.8%

1.9%

0.0%

0.8%

0.4%

0.8%

0.0%

0.4%

1.0%

Marginal

74.2%

72.2%

73.2%

72.5%

75.1%

73.8%

74.8%

73.6%

74.2%

73.7%

Small

15.3%

17.8%

16.5%

23.4%

18.8%

21.1%

20.8%

18.4%

19.6%

19.6%

Medium and larger Average Land Size (in Acres) All households

8.4%

8.2%

8.3%

4.1%

5.3%

4.7%

3.6%

4.0%

3.8%

5.7%

2.31

2.60

2.45

2.04

1.76

1.90

1.92

1.84

1.88

2.10

SC households

1.82

1.36

1.57

2.11

1.54

1.80

1.64

1.95

1.86

1.80

ST households

0.98

1.80

1.45

1.58

2.03

1.66

1.95

1.44

1.92

1.82

OBC households

1.87

2.43

2.18

2.23

2.25

2.24

1.85

2.26

1.95

2.17

Forward caste households Gini Land

3.39

5.66

3.95

2.04

1.57

1.83

2.12

1.66

1.87

2.30

51.0%

48.8%

49.9%

36.0%

45.2%

40.8%

41.7%

40.4%

41.1%

45.0%

287

281

568

244

245

489

250

250

500

1557

N

We classified the farm households into different categories (small, marginal and medium, etc.) based on the land cultivated in Kharif. This was different from land owned by the households as it included the land rented in or sharecropped by the household and excluded the land rented out by the farmer. From table 3.1, we observe that a majority of farmers in the sample across the three districts are marginal with landholdings of less than 1 hectare. Across all the six blocks, there were 72% to 75% farmers with marginal landholding. In all, 75% of the sampled households had a

33


landholding of less than 2.5 acres. Around 20% of the farmers have land holdings in the range of 2.5 to 5 acres and 6% farmers had more than 5 acres. Among the sampled farmers in Nalanda, there was a slightly higher share of farmers (8%) with more than 5 acres compared to that in Birbhum (5%) and Purulia (4%). The inequality in land ownership was highest (Gini: 51%) in the Islampur block of Nalanda and lowest (Gini: 36%) in Rajnagar block of Birbhum. There were also caste-based inequalities in land-holdings when looked at the sample as a whole. The average land size of SC households was 1.8 acres and that of ST households was 1.82 acres while the same for OBC households was 2.17 acres and for forward caste households it was 2.3 acres. These inequalities were more prominent in the two blocks in Bihar and less so in the four blocks in West Bengal. For instance, the average land size of SC households in Islampur was 1.36 acres while that of ST households was 1.8 acres. The same for OBC households in Islampur was 2.43 acres and that of forward caste households was 5.66 acres. In the blocks of West Bengal though, except for Kashipur, the average land holdings of SC households was similar and in some cases larger than that of forward caste households. Such distribution could be attributed partly to the land reforms that happened in West Bengal and which did not happen to a similar extent in Bihar (Thimmaiah, 2001; Hanstad, 2005). Along with land, livestock is an integral asset for farming as well as a risk hedging tool for marginal farmers. It is important to understand the livestock possessions of households in the sample. Table 3.1.2 provides the details of livestock possessed by the study sample. Table 3.1.2. Livestock assets possessed by the Households in the Study Sample (India) Nalanda Silao

Islampur

Birbhum All

Rajnagar

Nalhati-2

Purulia All

Kashipur

Raghunathpur-I

All Sample All

Poultry

3.1%

4.3%

3.7%

42.2%

49.4%

45.8%

57.6%

36.8%

47.2%

30.9%

Goat

4.5%

13.5%

9.0%

36.9%

28.2%

32.5%

38.4%

18.0%

28.2%

22.5%

Cows

51.7%

33.5%

42.6%

61.5%

55.1%

58.3%

47.2%

41.6%

44.4%

48.1%

Buffaloes

29.3%

43.8%

36.4%

4.9%

2.4%

3.7%

6.0%

4.0%

5.0%

16.1%

Bullock

9.1%

13.2%

11.1%

45.9%

16.3%

31.1%

43.6%

33.6%

38.6%

26.2%

Other

0.3%

1.8%

1.1%

8.2%

1.2%

4.7%

7.2%

5.2%

6.2%

3.9%

Any Livestock

74.6%

77.6%

76.1%

89.8%

83.3%

86.5%

84.4%

74.4%

79.4%

80.4%

Value of Livestock (INR)

26928

26336

26635

24157

14122

19129

25467

16788

21128

22509

34

ACIAR Report ADP/2015/032 Livestock Gini N

35.3%

45.2%

40.5%

44.1%

48.3%

47.3%

46.6%

44.9%

46.3%

45.2%

287

281

568

244

245

489

250

250

500

1557

In the study sample, around 80% households owned some form of livestock. 31% households had poultry, 23% had goats, 48% had cows, 16% had buffaloes, 26% had bullocks and 4% had other livestock including sheep, horses, mules, donkeys, oxen and pigeons. In Nalanda, 76% households owned livestock, most of them owning cows, buffaloes and bullocks. In Silao block, more than 50% of households owned cows while in Islampur, 44% households owned buffaloes. 29% households in Silao owned buffaloes, while 34% households in Islampur owned cows. Across all districts, Nalanda had a high proportion of households owning buffaloes compared to other two blocks, while it had a much lower proportion with goats. This could also be indicative of water scarcity as goats can thrive under rainfed conditions while it would be difficult for buffaloes to do so (Shankaranarayan, Bohra and Ghosh, 1985; Kumar and Singh, 2008). Around 11% and 9% households in Nalanda owned bullocks and goats respectively. Very few households owned poultry (4%) and other livestock (1%) in Nalanda. In Birbhum, 87% households owned livestock, including cows, poultry, goats and bullocks. 62% households in Rajnagar and 55% households in Nalhati-2 possessed cows. 42% households in Rajnagar and 49% in Nalhati-2 owned poultry. 46% households in Rajnagar and 16% households in Nalhati-2 owned bullocks. 37% households in Rajnagar and 28% in Nalhati-2 block owned goats. Very few households in Birbhum owned buffaloes (4%) and other livestock (5%). In Purulia, 79% households owned livestock, most of them owning poultry, cows, bullocks and goats. 58% in Kashipur and 37% in Raghunathpur-I owned poultry. 47% in Kashipur and 42% in Raghunathpur-I owned cows. 44% and 34% households in Kashipur and Raghunathpur-I respectively owned bullocks while 38% and 18% households in the blocks owned goats. Few households owned buffaloes (5%) and other livestock (6%). The average value of the livestock possessed by the households for the entire India sample households was INR 22,509. This value was the lowest in the Nalhati-2 block (INR 14,122) of Birbhum and highest in the Silao block (INR 26,928). The inequality in livestock values across households had a Gini that was similar to Gini of landholding for the whole sample (45.2%). This 35


inequality was least in Silao (Gini: 35.3%) block in Nalanda and highest in Nalhat-2 (Gini: 48.3%) block in Birbhum district. Along with land and livestock, households also possess various other agricultural and nonagricultural assets. Table 3.1.3 provides details over the same for the households in the study sample.

36

ACIAR Report ADP/2015/032 Table 3.1.3. Ag and Non-Ag Asset Holdings of the Households in the Study Sample (India) Nalanda Silao Agricultural Assets Access to Openwells

7.0%

Islampur 3.9%

Birbhum All

Rajnagar

Nalhati-2

Purulia All

Kashipur

Raghunathpur

All Sample All

5.5%

28.3%

0.0%

14.1%

24.8%

5.2%

15.0%

11.2%

On-farm Ponds

1.7%

0.0%

0.9%

18.4%

12.7%

15.5%

24.0%

2.4%

13.2%

9.4%

Borewells

24.7%

32.0%

28.3%

9.4%

0.8%

5.1%

7.2%

3.6%

5.4%

13.7%

Tubewells

50.2%

54.4%

52.3%

7.4%

29.4%

18.4%

4.4%

6.4%

5.4%

26.6%

Diesel Engine Electric Motor

27.9% 22.0%

66.5% 8.2%

47.0% 15.1%

10.7% 2.0%

6.5% 2.0%

8.6% 2.0%

14.0% 0.8%

1.6% 0.8%

7.8% 0.8%

22.4% 6.4%

Submersible

11.8%

0.4%

6.2%

0.4%

6.1%

3.3%

2.4%

0.8%

1.6%

3.8%

Tractors

5.9%

4.3%

5.1%

1.2%

2.0%

1.6%

2.0%

0.4%

1.2%

2.8%

Threshers

9.1%

12.1%

10.6%

8.6%

21.2%

14.9%

34.0%

13.6%

23.8%

16.2%

Sprayers

10.1%

10.7%

10.4%

17.6%

26.1%

21.9%

10.0%

3.2%

6.6%

12.8%

Value of Agri Assets-INR (Including tractors, sprayers, threshers, motors, ploughs, small implements, etc.,)

53384

41467

47488

8272

9155

8715

7297

2977

5137

21711

Agri Assets Gini

77.4%

69.3%

74.1%

81.1%

77.3%

79.6%

82.5%

81.9%

83.6%

83.0%

Mobile

96.5%

97.9%

94.2%

88.5%

92.7%

90.6%

88.4%

92.0%

90.2%

92.9%

Cycle

56.4%

74.4%

65.3%

88.1%

91.8%

90.0%

88.8%

92.8%

90.8%

81.2%

Bike/Scooter

19.2%

11.0%

15.1%

11.1%

12.7%

11.9%

14.0%

9.6%

11.8%

13.0%

Television

45.3%

17.1%

31.3%

51.2%

51.8%

51.5%

41.2%

40.0%

40.6%

40.7%

Radio

6.6%

37.4%

21.8%

3.3%

2.0%

2.7%

0.8%

1.2%

1.0%

9.1%

Cattle Shed

42.9%

41.3%

42.1%

77.5%

58.4%

67.9%

69.6%

52.0%

60.8%

56.2%

Nonfarm Building

6.6%

13.5%

10.0%

10.7%

9.4%

10.0%

2.4%

2.8%

2.6%

7.6%

Value of Non-Ag Assets-INR

59695

46176

53007

25265

28536

26904

28516

16509

22513

35016

Non Agri Assets Gini

78.7%

76.8%

78.0%

58.6%

65.8%

62.5%

63.2%

53.5%

60.3%

71.6%

Value of Total Assets

113079

87642

100495

33537

37691

35619

35813

19487

27650

56727

Total Assets Gini

73.6%

66.9%

70.9%

59.3%

62.1%

61.0%

64.9%

54.6%

62.0%

70.6%

287

281

568

244

245

489

250

250

500

1557

Non Agricultural Assets

N

37


Ownership or access to various sources of irrigation is an important asset in agriculture. For the entire India sample households, 27% households owned tubewells, whereas only 14% owned bore wells. 11% had access to open wells, while 9% had access to on-farm ponds. In Nalanda, more than half of the households (52%) had access to tubewells and more than onefourth of households (28%) had access to borewells. Very few households in Nalanda had access to open-wells (6%) and on-farm ponds (1%). Access to the same irrigation sources is much lower in Birbhum with only 18% households having tubewells. 14% households had access to open wells and 15% had access to on-farm ponds. None of the households in Nalhati-2 had access to open wells though and higher share of households in Nalhati-2(29%) owned tubewells. In Purulia, which is the most water scarce district in our study area, there were very few households owning tubewells (5%) or borewells (5%). This was due to lack of aquifers and very limited access to groundwater. In Kashipur, around one-fourth of the households had access to open wells and on-farm ponds while in Raghunathpur, only 5% and 2% households had access to open wells and on-farm ponds respectively. Amongst mechanized assets, 22% of the households possessed diesel engines (for the entire sample). Diesel engines ownership was higher in the blocks of Nalanda (given its relatively better groundwater situation). 28% households in Silao and 66% in Islampur owned diesel engines. A higher presence of diesel engines could also be due to lower electricity availability. Even after considering recent improvements, Bihar lags behind West Bengal in access to electricity (Jainet al., 2015). But, since groundwater is accessible at lower levels34, it is extracted using diesel engines in Islampur and using electric motors and submersibles in Silao. This could be due to better electricity access in Silao. There is very little use of electric motors and diesel engines in both Birbhum and Purulia districts. The extraction of groundwater is not economical in Purulia. In particular, there seems to be no use of motors or engines in Raghunathpur-I which would imply a complete lack of usage of groundwater. There were only 3% of households for the entire sample that owned a tractor. 5.9%

34

While the depth to water level in Nalanda was generally in the range of 2-5 meter below ground level (mbgl), the same in Purulia and Birbhum was either in the 5-10 mbgl range or in the 10-20 mbgl range (Source: Central Ground Water Board, India).

38


households in Silao owned a tractor while only 0.4% households in Raghunathpur-I had a tractor. 16% households in the entire sample owned a thresher. Thresher ownership was relatively higher in Purulia district where 34% households in Kashipur owned a thresher while only 9% households in Rajnagar owned a thresher. 13% households (in the entire sample) owned sprayers. 26% households in Nalhati-2 block owned sprayers while only 3% households in Raghunathpur-I owned sprayers. The average value of agricultural assets owned by households for the entire sample was INR 21,711. The average value of agricultural assets was the highest in Silao block in Nalanda (INR 53,384) and lowest in Raghunathpur-I block in Purulia (INR 2,977). The value of agricultural assets was more unequally distributed than land and livestock with a high Gini of 83%. The agricultural assets were most unequally distributed in Raghunathpur-I block (Gini 83.6%) and least unequally distributed in Islampur (Gini: 69.3%). Amongst non-agricultural asset ownership of the households, a very high percentage (93%) of households owned mobile phones. Around 81% households owned bicycles. In Nalanda, only 65% owned a bicycle, whereas in Birbhum and Purulia, 90% owned bicycles. Only 13% of all sampled households owned a bike/scooter. In Silao, 20% households owned a bike while 10% households owned a bike/scooter in Raghunathpur-I. Around 40% of the overall sampled households owned a television. 52% of households in Nalhati-2 block owned a television while only 17% households in Islampur owned a television. 9% of all the sampled households owned a radio. While 37% households in Islampur owned a radio, only 0.8% households in Kashipur owned a radio. 56% of all sampled households owned cattle shed. 78% households in Rajnagar owned cattle shed while only 41% households in Islampur owned cattle shed. Around 8% households in the sample owned a non-farm building. While 14% households in Islampur owned a non-farm building, only 2% households in Kashipur owned a non-farm building. For the overall sample, the average value of nonfarm assets was higher than that of agricultural assets at INR 35,016. The value of nonfarm assets was lowest in Raghunathpur-I (INR 16,509) and highest in Silao (INR 59,695). The nonfarm assets were less unequally distributed compared to agricultural assets with a gini of 71.6%. These assets were least unequally distributed in Raghunathpur-I (Gini: 53.5%) and most unequally distributed in Silao (78.7%). In terms of total assets (sum of agricultural and nonagricultural assets), the average value of total assets was INR 39


56,727. The value of total assets was the least in Raghunathpur-I (INR 19,487) and highest in Silao (INR 1,13,079). The gini of total asset value was 70.6% which was lowest for Raghunathpur-I (54.6%) and highest for Silao (73.6%). Housing and access to electricity are also crucial assets held by households. Table 3.1.4 presents the details of housing and access to electricity for the sample households. Table 3.1.4. Housing and Access to Electricity of Households in the Study Sample (India) Nalanda Silao

Islampur

Birbhum All

Rajnagar

Nalhati-2

Purulia All

Kashipur

Raghunathpur

All Sample All

Pucca

21.6%

8.2%

15.0%

28.3%

30.7%

29.5%

19.6%

41.2%

30.4%

24.5%

Semi-Pucca

69.0%

73.3%

71.1%

20.9%

15.6%

18.2%

22.8%

16.4%

19.6%

38.0%

Kuccha

9.4%

18.5%

12.1%

50.8%

53.7%

52.3%

57.6%

42.4%

50.0%

37.5%

Electricity Access

76.0%

55.5%

65.8%

97.1%

97.6%

97.3%

90.4%

88.4%

89.4%

83.3%

From table 3.1.4, for the entire sample, we observe that one-fourth of the households were pucca households, while 38% of households were semi-pucca and kuccha households. In Nalanda, a high percentage of households (71%) were semi-pucca households while fewer households were pucca (15%) and kuchha houses (12%). Silao had more pucca households (22%) compared to Islampur (8%). In Birbhum, a high percentage of households (52%) were kuccha households. 30% households were pucca households while 18% were semi-pucca households. In Purulia, 50% households were kuccha households. 30% households had a pucca house and 20% households had semi-pucca houses. Kashipur had a higher percentage of kuccha house ownership (58%) compared to Raghunathpur-I (42%). In the financial assets and liability category, we would also like to understand the debt portfolios of the households. Table 3.1.5 presents the loan portfolio of the households

40


Table 3.1.5. Loan Portfolio of Households in the Study Sample (India) Nalanda Silao

Islampur

Birbhum All

Rajnagar

Nalhati-2

Purulia All

Kashipur

Raghunathpur

All Sample All

Loan from Bank

18.1%

18.1%

18.1%

27.9%

28.6%

28.2%

14.8%

11.2%

13.0%

19.7%

Loan from Family and Friends

10.8%

23.5%

17.1%

2.5%

13.9%

8.2%

12.0%

19.6%

15.8%

13.9%

Loan from MFI

11.1%

12.1%

11.6%

13.1%

4.1%

8.6%

2.4%

2.0%

2.2%

7.6%

Loan from Moneylenders

5.2%

3.6%

4.4%

4.9%

7.8%

6.3%

5.2%

4.8%

5.0%

5.4%

Loan from Others

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

34.8%

74.8%

54.8%

17.6%

Loan from Any Source

41.1%

51.2%

46.1%

41.8%

50.2%

46.0%

52.4%

84.0%

68.2%

53.2%

Crop Loan from Bank

7.0%

5.0%

6.0%

14.3%

20.8%

17.6%

11.2%

14.0%

12.6%

10.8%

Crop Loan from Family and Friends Crop Loan from MFI

2.8% 4.2%

3.2% 2.8%

3.0% 3.5%

0.4% 2.5%

4.9% 2.9%

2.7% 2.7%

1.6% 0.0%

6.4% 0.0%

4.0% 0.0%

3.2% 2.1%

Crop Loan from Moneylenders

0.3%

0.7%

0.5%

2.5%

1.6%

2.0%

3.2%

1.2%

2.2%

1.5%

Crop loan from Others

0.3%

0.0%

0.2%

0.8%

4.9%

2.9%

8.8%

3.2%

6.0%

2.9%

Crop loan from any source

12.9%

11.4%

12.1%

20.1%

33.9%

27.0%

22.8%

18.0%

20.4%

19.5%

Loan from Formal Sources

96375

53392

73937

28822

58285

42595

35321

59416

45741

55570

Loan from Informal Sources

45592

76973

64472

19333

25007

23596

20581

11367

14587

27257

Crop Loan from Formal Sources

54207

39364

47804

24400

31879

28827

22241

35800

28011

33569

15529 287

17250 281

16316 568

11333 244

24103 245

21079 489

17788 250

13674 250

15857 500

17615 1557

Crop Loan from Informal Sources N

41


From table 3.1.5, for the entire sample, we observe that 53% households have taken loans from informal and formal sources. Around 20% households have taken a loan from a bank while 18% households have taken a loan from other sources such as grocery stores and input dealers. 14% households have a loan from family and friends. Only 20% households have taken a crop loan. Around 11% households have a crop loan from bank while 3% have crop loans from family and friends and other sources. The average loan amount from formal sources (bank and MFI) was INR 55,570 while that from informal sources was INR 27,257. The average crop loan from informal sources was INR 33,569 while that from informal sources was INR 17,615. So, in general the loans for purposes other than crop loan were relatively higher. Also, the loan amounts from formal sources were almost double that from informal sources. In Nalanda, 46% of households had a loan from different sources while only 12% loans had a crop loan. 18.1% and 11.6% households had a loan from bank and MFI, while 17% and 4% had a loan from friends and families and moneylenders respectively. The average loan size in Nalanda was higher from both formal and informal sources (INR 73,937 and INR 64,472). In Silao, average loan amount from formal sources (INR 96,375) was more than double that from informal sources (INR 45,592), whereas in Islampur, the loan amount from informal sources (INR 76,973) was higher than the loan amount from formal sources (INR 53,392). The average crop loan size from formal sources was higher than those from informal sources in Nalanda and in both the blocks. In Birbhum too, 46% households had a loan from different sources. However, as compared to Nalanda, a higher share (28%) had a loan from banks. Nalhati-2 had a higher proportion of households (14%) that had received a loan from friends and family as compared to Rajnagar (3%) while there were a higher proportion of households (13%) with a loan from MFIs in Rajnagar as compared to Nalhati-2 (4%). The average size of loans and crop loans from formal and informal sources in Rajnagar was lesser than that in Nalhati-2. In Purulia, 68% households had a loan from different sources. In Raghunathpur-I, 84% percent households had a loan from different sources while only 53% households had a loan in Kashipur. Almost three-fourths of the households in Raghunathpur-I had a loan from other sources while 42


one-fifth of the households had a loan from friends and family. 5% households in RaghunathpurI had borrowed from a money-lender. Only 11% households and 2% had loan from formal sources such as bank and MFI in Raghunathpur-I respectively. The average loan size from formal sources in Raghunathpur-I was much higher than from informal sources.

43


Section 3.2. Physical and Financial Capital Endowments of Surveyed Households in Bangladesh 3.2.1. Land Holdings of the Sampled Households in Bangladesh On an average, the land owning households had 52.65 decimal of land. The land owners in Khulna had the highest average land holding (65.25 decimal), whereas the respondents in Patuakhali had the lowest average (43.73 decimal). In both the shrimp growing districts, (i.e., Khulna and Satkhira) households in high saline areas had more land than their counterparts living in relatively low saline areas. Difference in land holding was more notable in Khulna, where the average land owned by a household in the high saline areas had more than double the average land holding in relatively low saline areas. In contrast, in the non-shrimp growing district of Patuakhali, low and high saline areas had an average land ownership of 52.69 decimal and 34.27 decimal, respectively (Table 3.2.1). Table 3.2.1: Landholdings by the Sampled Households (decimal) N

Salinity Level District

Low

High

All

Khulna

41.37

89.28***

65.25

Patuakhali

52.69

34.27**

43.73

Satkhira

44.37

52.31

48.36

All

46.05

59.35**

52.65

Own Land (decimal) 315 294 314 923

Total Cultivated Crop Land Holding (decimal) for the Farming Households Khulna

72.57

171.90***

136.75

Patuakhali

177.39

157.93

170.44

Satkhira

92.43

62.86

87.57

All

132.59

153.25

140.44

65 126 67 258

Own Cultivated Crop Land Holding(decimal) for the Farming Households Khulna

54.89

156.00**

121.70

Patuakhali

85.84

97.21

89.63

Satkhira

86.42

57.65

80.43

All

80.83

120.94***

96.83

56 84 48 188

Note: ***, ** and * indicate mean differences of land areas between less and high saline areas are significant at 1%, 5% and 10% level, respectively.

In high saline areas households cultivated more land than their counterparts living in less saline areas. Specifically, in high saline areas, a farm household, on average, cultivated 153.25 decimal of land, whereas their counterparts in low saline areas cultivated 132.59 decimal of land. Here also, 44


in Khulna, the difference between low and high saline areas was the highest and significant. But in the other two districts, farmers in relatively less saline areas cultivated more land than farmers in high saline areas. In case of own cultivated land, significant differences exist between low and high saline areas. In high saline areas, a farm household owned 121 decimal of cultivable land which is almost 1.5 times higher than that in the low saline area. In Khulna high saline area farmers had around 3 times more own land than the farmers in the less saline areas (Table 3.2.1). By dividing the sampled households into five quintiles based on their land ownership, Table 3.2.2 reveals that there exist significant differences in ownership across different quintiles. Significant differences also exist in both the low and high saline areas. The highest and lowest quintile have 186.42 decimal and 3.42 decimal of land, and differ by a ratio of 60. In the high saline areas, the difference is even higher. Against 203.47 decimal of land owned by the highest quintile in the high saline areas, the lowest quintile have only 3.29 decimal, which gives a ratio of almost 70. Table 3.2.2. Land ownership (decimal) by different quintiles Salinity Level Land Ownership Quintile

Low

High

All

Q1

3.54

3.29

3.42

Q2

7.73

8.28

8.03

Q3

15.74

16.99

16.31

Q4

45.42

45.86

45.61

Q5

166.00

203.47

186.42

All

46.05***

59.35***

52.65***

Note: ***, ** and * indicate mean differences across different quintiles in less and high saline areas are significant at 1%, 5% and 10% level, respectively.

Following the Department of Agricultural Extension’s classification, the sampled households were classified into four categories: landless (less than 0.49 acre of land), marginal (0.50 acre to 1.49 acre of land), small (1.5 acre to 2.49 acre of land) and medium and large farmers (above 2.5 acre of land). Among the sampled farmers, nearly half (48.68%) were landless. In less saline areas, 53.98% of the farmers were landless, whereas the proportion was 38.20% in high saline areas. Among the three districts, Satkhira had the highest proportion (70.59%) of landless farmers. In Khulna and Patuakhali, proportions of farmers are almost identically distributed across different categories. This is a well acknowledged fact, as these two districts are geographically adjacent. Surprisingly, in Satkhira, only around 2% (of the sample) were small farmers whereas around 10% were medium & large farmers. In less saline areas, comparatively higher proportion of farmers fall 45


in the landless and marginal farmers groups, whereas in high saline areas proportions of small and medium & large farmers were around thrice and twice the same in less saline areas (Table 3.2.3). Table 3.2.3: Proportion of Farmers Belonging to Different Categories across Regions Salinity Level

Districts

Farmer’s Category

All

Khulna

Patuakhali

Satkhira

Less

High

Landless

40.45%

45.60%

70.59%

53.98%

38.20%

48.68%

Marginal

37.08%

35.20%

17.65%

34.09%

29.21%

32.45%

Small

13.48%

13.60%

1.96%

6.82%

20.22%

11.32%

Medium & Large

8.99%

5.60%

9.80%

5.11%

12.36%

7.55%

On average, a farm household owned 90.75 decimal of land, of which 59.15 decimal was cultivable land. By renting in additional land a farm household managed to cultivate on 127.17 decimal of land. Own cultivable land was only around 7% of the total cultivable land for the landless farmers, whereas the same shares for marginal and small farmers were 44% and 69% respectively. The medium & large farmers hardly had any rented lands. They had 309.79 decimal of total cultivable land, of which only 6% was rented in. Relatively low share of rented land for the bottom and top categories may have two different interpretations. The medium & large farmers may not need additional land, whereas the landless may not have the financial capability to rent additional land (Table 3.2.4). Table 3.2.4: Landholding (decimal) by Different Categories of Farmers Farm Category

Own Land

Total Cultivated Crop Land

Own Cultivated Crop Land

Landless

20.66

80.70

5.41

Marginal

89.25

125.31

55.71

Small

192.57

210.53

145.10

Medium & Large

386.04

309.79

291.64

All 90.75*** 127.17*** 59.15*** Note: ***, ** and * indicate mean differences across different categories in less and high saline areas are significant at 1%, 5% and 10% level, respectively.

Quantities and values of farmer’s own agricultural land across different salinity areas are presented in Table 3.2.5. Agricultural land includes not only crop land but also land for pond, livestock and gardening. In high saline areas, farmers owned 92.97 decimal of land which was valued at 754,463 tk (Bangladeshi currency is valued at about 80 tk per USD). Their counterparts in less saline areas had 54.64 decimal of land which they could sell at 558,051 tk. Though the top two categories of farmers in high saline areas had more land than their counterparts in less saline areas, the landless 46


and marginal farmers in less saline areas had more land (than those in the high saline areas) . On average, a sampled farmer in the study area had 71.20 decimal of land which had a market value of 640,158 tk. Table 3.2.5: Quantity (decimal) and Value (tk) of Agricultural Land Owned by Different Categories of Farmers Salinity Level

Landless

Less

Small

Medium & Large

All

71.00

167.68

285.99

54.64***

100358

577928

1986368

3301471

558051***

9.61

68.27

172.47

373.44

92.97***

Value

86772

561616

1528360

2832054

754463***

Quantity

10.57

69.74

170.72

337.56

71.20***

Value Quantity

High

Marginal

11.08

Quantity

All Value 95848 570533 1703236 3040684 640158*** Note: ***, ** and * indicate mean difference across different farm categories in less and high saline areas are significant at 1%, 5% and 10% level, respectively.

Table 3.2.6 presents inequality in land holdings through Gini coefficients. Among the three districts, inequality in own land holding was the highest in Khulna (0.692) followed by Satkhira (0.677) and Patuakhali (0.636). Satkhira evidenced highest inequality in case of both total and own cultivated land. Inequality was high in the high saline areas for own cultivated and total cultivated land categories, whereas in case of own cultivated land category, inequality was more in low saline areas. Inequality in own land holding and total cultivable land was higher among the households in nearest villages compared to the households in distant villages. However, households living far off from district centers in high saline areas had relatively higher inequality (0.495) than their counterparts living in nearest villages (0.467). Small farmers had the lowest level of inequality coefficient amongst own and cultivated land categories. Inequality was severe amongst the bottom (landless) and top (medium & large) categories of farmers. The differences in Gini coefficient value are notable across different farm categories. Table3.2.6: Gini Coefficient for Land Holding Own Land

Total Cultivated Land

Own Cultivated Land

Khulna

0.692

0.461

0.428

Patuakhali

0.636

0.442

0.495

Satkhira

0.677

0.463

0.559

Less

0.652

0.438

0.483

High

0.693

0.488

0.456

District

Salinity Level

47


Proximity Nearest

0.691

0.474

0.467

Distant

0.660

0.465

0.495

Landless

0.413

0.485

0.249

Marginal

0.172

0.466

0.272

Small

0.088

0.277

0.113

Medium & Large

0.242

0.314

0.312

Farm Category

Section 3.2.2. Types of House and Access to Utilities (Electricity and Water) Type and value of houses is an important indicator of household’s economic condition. Mud house, probably the cheapest one to make, was the most common type of house among the sampled households. Out of every five households sampled, around four live in mud houses. In Khulna, Patuakhali and Satkhira districts 68.75%, 89.06% and 76.56% had mud houses respectively. In low and high saline areas 75.21% and 81.04% lived in mud houses. Very few of the households (1.56%) had brick houses in Patuakhali. Relatively higher proportion of households in Khulna and Satkhira had brick houses, estimated at 13.13% and 17.81%, respectively (Table 3.2.7). Table 3.2.7: Type of Houses District

Types of Houses

Salinity Level

All

Khulna

Patuakhali

Satkhira

Low

High

Brick House

13.13%

1.56%

17.81%

9.38%

12.29%

10.83%

Semi-Brick House

18.13%

9.38%

5.63%

15.42%

6.67%

11.04%

Mud House

68.75%

89.06%

76.56%

75.21%

81.04%

78.13%

The average value of a house including land was 161,702 tk (or roughly 2000 USD for the entire sample). In Khulna district, respondents had the highest value for their houses estimated at 179,513 tk, followed by Satkhira (157,213 tk) and Patuakhali (148,380 tk) districts. In less saline areas, the sampled households had significantly higher value for their houses (175,624 tk) compared to the same in high saline areas (147,780 tk). The most notable and significant difference was observed in Patuakhali, where in less saline areas average value of the house was 187,314 tk against 109,447 tk for the high saline areas, indicating indirect linkage effects between salinity and market value of houses. In Satkhira, households were found to have higher value for their houses in the high 48


saline areas as compared to the same in less saline areas, though the difference was marginal and insignificant (Table 3.2.8). Table 3.2.8. Value of Houses (tk) in the three Districts Salinity Level District

Low

High

All

Khulna

185625

173400

179513

Patuakhali

187314

109447***

148380

Satkhira

153933

160494

157213

All

175624

147780**

161702

Note: ***, ** and * indicate mean differences between value of houses across low and high saline areas to be significant at 1%, 5% and 10% level, respectively.

Among the sampled households, 78.13% had electricity connection. This figure is much higher than those mentioned in secondary sources (59.6% of the population in Bangladesh has access to electricity).35 But many of these connections are part-time and partial through private sources such as solar energy providers. From government sources, 33.33% received electricity. Among the three districts, Khulna had the highest government sourced connection (65.42%). Patuakhali (9.27%) had surprisingly very low number of government provided connections. But when it comes to having connection irrespective of the source, all the districts show almost similar proportions. Compared to the low saline areas, in high saline areas marginally higher proportion of households have electricity connection, though in high saline areas households rely more on private sources for electricity (Table 3.2.9). Table 3.2.9: Households with Electricity Connection District Electricity Connection % of households with connection

Salinity Level

Khulna

Patuakhali

Satkhira

Low

High

All Areas

75.00%

80.94%

78.44%

75.42%

80.83%

78.13%

Government

65.42%

9.27%

27.49%

35.64%

31.19%

33.33%

Private

34.17%

87.64%

65.74%

58.84%

67.27%

63.20%

Others

0.42%

3.09%

6.77%

5.52%

1.55%

3.47%

electricity

Source of Electricity

Respondents in some of the surveyed areas struggled to access drinking water. Many respondents reported having multiple water sources. Around 16% of the households had to purchase drinking

35

http://data.worldbank.org/indicator/EG.ELC.ACCS.ZS

49


water t for least some months in a year. Interestingly, this proportion does not vary much across low and high saline areas. The most common form of drinking water source was public tube well (36.62%), followed by open water bodies (26.50%) and own tube well (17.61%). In less saline areas, higher proportion of households relied on free water sources such as public tube well and open water bodies. It is understandable that due to higher saline contamination in high saline areas, fewer households could use open water sources. Salinity also makes tube well installation difficult, costly and failure-risk prone in high saline areas. In these areas, tube wells also remain nonfunctioning when salinity level increases. For these reasons, relatively lower proportion of the households in high saline areas were found to be using tube wells. Even if they had access, they were not able to use tube well round the year (Table 3.2.10). Table 3.2.10: Household’s access to Various Drinking Water Sources Salinity Level Water Source

Low

High

All

Public Tube well

39.65%

33.57%

36.62%

Open Water Bodies

27.72%

25.27%

26.50%

Own Tube well

12.63%

22.61%

17.61%

Water Purchase

15.96%

16.43%

16.20%

Others

4.04%

2.12%

3.08%

Section 3.2.3. Livestock Possessed by the Sampled Households Tables 3.2.11 to 3.2.15 contain information regarding poultry and livestock assets owned by the households. On average, households had 13 poultry birds worth 2527tk. Compared to the households in low saline areas, both numbers as well as the value of poultry birds were higher for households living in high salinity affected areas, except for Patuakhali. But the difference is significant only in Satkhira (Table 3.2.11). Table 3.2.11: Number and Value of Poultry Birds owned by Households rearing poultry (tk.) Number of Poultry

N

Value of Poultry(tk) Salinity level

Districts

Low

High

All

Low

High

All

Khulna

13

16

15

2700

4125

3408

Patuakhali

18

17

18

3682

3410

3541

Satkhira

5

7***

6

1170

1973***

1577

All

12

13

13

2527

3185

2861

50

300 295 284 879

ACIAR Report ADP/2015/032 Note: ***, ** and * indicate mean differences in number and value of birds between less and high saline areas to be significant at 1%, 5% and 10% level, significantly.

Table 3.2.12 presents number and value of live animals (excluding poultry) owned by the households. In Khulna, a household in high salinity affected area had around 3 animals, which is significantly higher than that for the households in low saline areas (2 animals). On an average, a household had around 2 animals worth 19,126tk. It is interesting to note that the value of live animals held by an average household in the sample is around 7 times higher than that for poultry. Households would be expected to own more cattle as risk hedging measure, more so in high salinity areas. However, even though the households in highly saline areas had more livestock than their counterparts in less saline areas, the value of livestock was higher for households in less saline areas (Table 3.2.12). Table 3.2.12: Number and Value of Animals (excluding poultry) owned by the Households rearing livestock (tk) Number of Livestock

N

Value of Livestock(tk) Salinity Level

Districts

Low

High

All

Low

High

All

Khulna

2.07

2.76

2.41**

18362

21823

20081

Patuakhali

1.68

1.75

1.75

31532

19841

25468*

Satkhira

2.14

1.65

1.89

16852

6357

11531***

300 295 284

879 All 1.96 2.05 2.01*** 22193 16126 19126** Note: ***, ** and * indicate mean differences in number and value of livestock between low and high saline areas to be significant at 1%, 5% and 10% level, respectively.

Total value of livestock and poultry animals for the average household was 21,987 tk. Compared to livestock growers living in severely saline affected areas, the households in less saline areas had significantly higher value for livestock in Patuakhali and Satkhira (Table 3.2.13). Table 3.2.13: Value of Livestock and Poultry (tk.) owned by the Households N

Salinity Level District

Low

High

All

Khulna

21062

25948

23489

300

Patuakhali

35213

23250*

29009

295

Satkhira

18022

8330***

13108

284

All 24720 19334* 21987 879 Note: ***, ** and * indicate mean differences in number and value of livestock between low and high saline areas to be significant at 1%, 5% and 10% level, respectively.

Inequalities in the value of livestock owned by the households are presented through Gini coefficients in Table 3.2.14. Among the three districts, inequality was highest in Khulna for poultry birds, whereas for livestock it was highest for Satkhira. Patuakhali has the highest inequality when 51


value of poultry and livestock was considered together. The difference in inequality among the three districts is hardly noticeable. Inequality is relatively more in high saline areas for both livestock and poultry. Table 3.2.14: Gini of Value of Livestock (including poultry) Poultry

Livestock

Livestock and Poultry

Khulna

0.693

0.454

0.624

Patuakhali

0.512

0.521

0.71

Satkhira

0.453

0.54

0.658

Low

0.591

0.521

0.665

High

0.593

0.541

0.697

District

Salinity Level

Values of livestock and poultry across different income quintiles are presented in Table 3.2.15. An interesting point to note is that the lowest income quintile owns the highest value of animals. The value of all live animals for the lowest quintile (Q1) is 26,886 tk, whereas it is 23,908 tk for the highest income quintile (Q5). This is true for less saline areas also. However, within the high saline areas, households in highest income quintiles (Q5) have the highest value for live animals, followed by the lowest quintile (Q1). But the difference in value of live animals across different income quintiles is not significant. The lowest income quintile may have invested more in animals as part of their risk hedging strategy against natural calamities. Here it should be noted that investing in animals is a risky decision, particularly in saline areas, where animals yield less than potential output, nevertheless the households are forced to do so given the lack of alternatives. Table 3.2.15.Value of Livestock and Poultry Owned by Different Income Quintiles in Low and High Saline Areas Salinity Level Income Quintile

All Low

High

Q1

32010

20559

26886

Q2

25220

15083

19126

Q3

21605

17011

19241

Q4

20656

20181

20432

Q5

23379

24419

23908

52


Section 3.2.4. Value and Quantity of Different Physical Assets Owned by the Sampled Households The sampled households own various types of assets. Proportion of households owning different types of assets along with their average values for asset across high and low saline areas are presented in Table 3.2.16. Among different assets, mobile phone was the most commonly owned (by 94.48% of the households). The average value of mobile phone for the owner households was 2,927 tk. The next common set of non-agricultural assets for the households were bicycles (29.69%), tube wells (19.58%) and television sets (17.29%). Motorcycle was the most valuable (amongst all assets) but owned only by 6.25% households. Average value of the motorcycle for the owner households was 83,900 tk. Proportions of households with different types of agricultural and non-agricultural assets across high and low saline areas are almost similar except for bicycles. In less saline areas 41.04% of the households had a bicycle, whereas in high saline areas only 18.33% had this asset. There also existed significant difference in value of bicycle owned by the households in less and high saline areas. In less saline areas, the average value of bicycles for the owner households was 3,299 tk, which was significantly higher than that in the high saline areas (2,206 tk). Significant differences also existed between the value of tube-wells owned by the households in less and high saline areas. Average value of tube wells for the households in less saline areas was 4,085 tk, which was more than 1.51 times higher than that of the households living in high saline areas (Table 3.2.16). Table 3.2.16: Quantity (no.) and Value (tk) of Different Assets Owned by the Sampled Households Low Saline Areas % of Households Value of with Asset Asset (tk)

Type of Assets

High Saline Areas % of Households Value of with Asset Asset (tk)

All Areas % of Households with Asset

Value of Asset (tk)

Non-Agricultural Mobile Phone

94.38%

2758

94.58%

3095

94.48%

2927

Refrigerator

1.04%

22000

2.08%

22100

1.56%

22067

Television

14.58%

5299

20.00%

5367

17.29%

5338

Radio

7.92%

605

6.46%

539

7.19%

575

Computer/Laptop

2.08%

23300

1.25%

32500

1.67%

26750

By-cycle

41.04%

3299

18.33%

2206***

29.69%

3054

Motor-cycle

6.04%

88448

6.46%

79645

6.25%

83900

Tube well

21.88%

4085

17.29%

2697***

19.58%

3416

Others

4.79%

16096

2.92%

8750

3.85%

13316

Well

5.83%

15924

1.46%

5950**

3.65%

13605

Agricultural Assets

53


Cattle Shed

37.71%

12103

17.50%

7035*

27.60%

10497

Diesel Engine

6.67%

11203

2.92%

10443

4.79%

10972

Electric Motor

0.63%

5000

3.54%

3147

2.08%

3425

Plough

6.25%

483

0.63%

433

3.44%

479

Sprayer

6.67%

1327

2.92%

1836

4.69%

1482

Fishing Net

55.63%

3464

58.33%

5400***

56.98%

4455

Boat

6.04%

17603

12.50%

23667

9.27%

21691

Others 1.88% 21794 0.63% 60000** 1.25% 31346 Note: ***, ** and * indicate mean differences between value of assets in low and high saline areas significant at 1%, 5% and 10% level, respectively.

Among the agricultural assets, fishing net was the most common and owned by 56.98% of the households. The average value of fishing nets for the owner households was 4,455 tk. The high saline area households had fishing nets worth of 5,400 tk, whereas for the less saline area households the average value was only 3,464 tk. This reveals a higher utility of fishing nets in high saline areas owing to higher concentration of shrimp farming. The next common agricultural asset was the cattle shed. Among the households, 27.60% had a cattle shed which they could have sold at 10,497 tk. In high and low saline areas 17.5% and 37.71% of the households had cattle sheds respectively. The average values of sheds for high and low saline area households were 7,035 tk and 12,103 tk respectively. This is consistent with the earlier section’s observation where households in less saline areas were found to own more livestock and had higher values as compared to their counterparts in high saline areas. Irrigation equipment such as diesel engines and electric motors were possessed by 6.87% households. Boats worth 2,1691 tk were owned by 9.27% households. In case of ownership and value, notable differences were observed across high and low saline areas. Proportion of owner households in less saline areas was higher for assets such as well, cattle sheds, diesel engine, plough and sprayer; whereas opposite was observed for electric motor, boat and fishing net. In case of value, except sprayer, fishing net and boat, households in less saline areas have higher valued assets for all other agricultural assets (Table 3.2.16). Table 3.2.17 presents average value of non-agricultural, agricultural and total assets owned by the households. The sampled households had total assets worth 21,467 tk, of which 13,447 tk and 11,516 tk were for agricultural and non-agricultural assets respectively. Compared to agricultural assets, value of non-agricultural assets was higher in all the districts except Patuakhali. Among the three districts, the residents in Khulna had the highest value for total assets (25,675 tk), followed 54


by Satkhira (21,289 tk) and Patuakhali (17,317 tk) residents. Compared to the residents in high saline areas, residents in less saline areas owned more valuable total assets, though the difference in value of total assets was significant only in Khulna. In Khulna, a household living in less saline areas had total assets worth 29,443 tk, whereas for their counterparts living in high saline areas, the value was 21,859 tk. Except Patuakhali, in the other two districts, households in less saline areas had higher number of agricultural assets. In case of non-agricultural assets, residents in less saline areas of Khulna and Patuakhali had higher value for assets. Finally, note that the value of total physical assets owned by the household is much lower compared to their livestock values. Table 3.2.17: Value of Assets (tk) Owned by the Households across Different Salinity Affected Areas Agricultural Asset Non- Agricultural Asset High All High All Districts Less Saline Saline Less Saline Saline Khulna 13008 8699 11022 19031 16206 17641 Patuakhali 18405 24474* 21053 7208 5060 6134 Satkhira 12166 9491 10719 12995 13225 13111 All 14221 12624 13447 13142 11516 12330 Note: ***, ** and * indicate mean differences in value of assets between less and high saline areas are respectively.

Less Saline 29443 18276 21563 23169 significant at

Total High Saline 21859* 16363 21021 19772 1%, 5% and

All 25675 17317 21289 21467 10% level,

On an average, a farm household owned agricultural assets worth 9,830 tk. Significant differences existed among different categories of farmers, and farmers with large land holdings had higher value of assets, whereas the marginal farmers had marginally lower value of assets than the landless farmers. The medium & large farmers had assets valued at 20,248 tk, which was around 2.5 times higher than that of the landless farmers (Table 3.2.18). Table 3.2.18: Value of Agricultural Assets (tk) Owned by Different Categories of Farmers Farm Category

All Areas

Landless

8533

Marginal

8455

Small

12403

Medium & Large

20248

All 9830** Note: ***, ** and * indicate mean differences across different farm categories to be significant at 1%, 5% and 10% level, respectively.

Table 3.2.19 compares inequality in asset holdings for the sampled households. In case of value of total assets and also for agricultural assets, inequality was highest in Satkhira district, though the differences in Gini coefficients across districts were not notable. Higher level of inequality existed across all districts. In Satkhira, the Gini coefficient for value of agricultural assets was as high as 0.726. Inequalities for both agricultural and non-agricultural assets were high in high saline areas compared to the low saline areas, indicating the role of salinity in reinforcing inequality. The 55


variation was relatively higher for agricultural assets than the non-agricultural assets. Similarly, the areas far from district headquarters had higher level of inequality for both agricultural and nonagricultural assets compared to the areas nearer to the headquarters. Table 3.2.19: Gini of Value of Asset Agricultural

Non-Agricultural

Total

Khulna

0.671

0.673

0.639

Patuakhali

0.552

0.685

0.644

Satkhira

0.726

0.682

0.647

Less

0.658

0.702

0.644

High

0.692

0.706

0.652

Nearest

0.655

0.682

0.631

Distant

0.690

0.728

0.664

District

Salinity Level

Proximity

Table 3.2.20 presents value of assets owned by different income quintiles. The table reveals that the highest income quintile had assets worth 34,947 tk which is around 2.4 times higher than the bottom income quintile. The top two quintiles had around 40% more assets than their preceding income quintiles. Table 3.2.20: Value of Assets Owned by Different Income Quintiles Income Quintiles

Agricultural Assets (tk)

Non-Agricultural Assets (tk)

Total Assets (tk)

Q1

9830

7910

14388

Q2

12262

7157

14474

Q3

11098

10521

17893

Q4

15071

14889

25228

Q5 18018** 20703*** 34947*** Note: ***, ** and * indicate mean differences across different income quintiles to be significant at 1%, 5% and 10% level, respectively.

Table 3.2.21 depicts the value of assets owned by different income quintiles in high and less saline areas. In both the areas, significant differences in value of assets exist across different income quintiles. In low and high saline areas, households in Q5 had 2.2 and 2.9 times higher value of assets than their respective Q1 quintile households. In less saline areas the Q2 quintile had 12% less value for assets than the Q1 quintile, whereas in high saline areas the Q2 quintile had 25% more assets than the Q1 quintile. The most notable change is observed in low saline areas within the Q4 quintile. This quintile had 51% higher value for assets than the preceding quintile. The top 56


quintile in less and high saline areas had around 36% and 43% more assets than their preceding quintiles, respectively. Table 3.2.21. Value of Total Assets Owned by Different Income Quintiles across Low and High Saline Areas Income Quintiles

Less Saline Areas

High Saline Areas

Q1

16751

11447

Q2

14686

14339

Q3

18134

17633

Q4

27403

22837

Q5 37182*** 32689*** Note: ***, ** and * indicate mean differences across different income quintiles to be significant at 1%, 5% and 10% level, respectively.

Section 3.2.5. Borrowing Capacity of the Sampled Households On average, the loan recipient households had 44,068 tk of outstanding loans. Compared to the households in less saline areas, high saline area households had taken more loans. This trend is true for all the districts, except Patuakhali. Most notable difference is observed in Khulna where an average household in high saline area had loan of 44,572 tk against 26,922 tk for the average household in less saline area. This again may suggest a higher level of livelihood stress in high saline areas. Farmers mostly took loans from institutional sources. Compared to non-institutional sources, farmers took almost double the amount of loan from institutional source (Table 3.2.22). This may be a result of higher interests charged by non-institutional sources. Table 3.2.22. Total Amount of Outstanding Loan (tk) with the Households Non-Institutional

Institutional

Total

Salinity Level Districts

Less

High

All

Less

High

All

Less

High

All

Khulna

7326

13881*

11180

19596

30691*

26118

26922

44572***

37298

Patuakhali

16743

24874

20728

40433

30359

35496

57176

55232

56223

Satkhira

16284

8829***

12208

18902

31163*

25605

35186

39992

37814

All 14187 15470 14878 27353 30762 29190 41540 46232 44068 Note: ***, ** and * indicate mean differences in amount of borrowed capital between less and high saline areas to be significant at 1%, 5% and 10% level, respectively.

Table 3.2.23 presents information on the amount of loan taken by different categories of farmers. On an average, a farmer took 53,346 tk of annual credit from institutional and non-institutional sources combined. Loans from institutional sources comprised around 57% of their total loan liabilities. For the medium and large farmers, institutional sources contributed around 73% of total loans. For other categories of farmers, the contribution was roughly around 52% to 56%. Among 57


different categories, the small farmers had the highest amount of outstanding loans (81,800 tk) followed by medium & large farmers (78,333 tk). In less saline areas, farmers collected relatively higher portion of their total credit (around 62%) from institutional sources, whereas in high saline areas, institutional sources contributed around 49% of total credit. Hence, the results here are indicating that in high saline areas, farmers rely more on non-institutional sources, as do the farmers with relatively low land holdings. Comparing total credit as a share of annual incomes of these households reveals that the landless and marginal farmers’ outstanding loan amounts were higher than their annual incomes. For the small farmers, credit accounted for more than 80 percent of their income share. This share is the lowest for medium and large farmers (around 50% of total income). Table 3.2.23. Amount of Loan (tk) taken by Different Categories of Farmers % of Different Categories of Farmers with Credit Farmer’s Category

Low

High

All

Low

High

All

Low

High

All

Total Credit as Share of Total Annual Income

Non-Institutional

Institutional

Total Credit

Salinity Level

Landless

60%

17762

27091

20235

28811

11795

24301

46574

38886

44536

1.14

Marginal

63%

19267

29250

23260

29827

30700

30176

49093

59950

53436

1.38

Small

58%

25714

50000

38667

42429

43750

43133

68143

93750

81800

0.83

Medium & Large

62%

25300

17857

20958

51700

61429

57375

77000

79286

78333

0.48

53346**

1.13

All 61% 19107 29930 22963 31144 29009*** 30383** 50250 58939** Note: ***, ** and * indicate mean difference between less and high saline areas significant at 1%, 5% and 10% level.

58


Chapter 4: Climatic Stress faced by households In this chapter we discuss the key climatic stressors that affect households’ livelihoods in the study areas as well as their coping strategies. These stressors are water scarcity for India and salinity in Bangladesh.

59


Section 4.1. Water Scarcity: India This section looks at different levels of water scarcities faced by the households in the India and Bangladesh sample households. We first start with the India sample households. On average, Nalanda is a district with relatively higher water levels and access to irrigation through groundwater. But, rainfall has not been regular in recent years in Nalanda. From 2009 to 2014, the state government declared drought in Nalanda for three years (2009, 2010 and 2013). Birbhum is a district with somewhat medium levels of water accessibility, while Purulia suffers from high levels of water scarcity. Table 4.1.1 presents some of the key indicators of water scarcity in the study sample.

60

ACIAR Report ADP/2015/032 Table 4.1.1. Indicators of Water Scarcity among Households in the Study Sample (India) Nalanda Silao

Islampur

Birbhum All

Rajnagar

Nalhati-2

Purulia All

Kashipur

Raghunathpur-I

All Sample All

Households cultivating in Kharif Area Irrigated (Kharif Major Crop) Households cultivating in Rabi Households cultivating more than one crop in Kharif Households cultivating more than one crop in Rabi Optimal No. of Irrigation (Kharif major crop) Applied No. of Irrigation (Kharif major crop)

97.6%

98.2%

97.9%

100.0%

99.2%

99.6%

98.8%

100.0%

99.4%

98.9%

97.7%

99.4%

98.5%

69.0%

96.8%

82.9%

32.6%

34.5%

33.6%

72.6%

97.6%

96.4%

97.0%

32.0%

85.7%

58.9%

12.0%

3.6%

7.8%

56.4%

6.6%

1.1%

3.9%

1.2%

0.8%

1.0%

0.8%

1.2%

1.0%

1.9%

60.6%

64.8%

62.7%

19.3%

21.6%

20.4%

2.4%

0.8%

1.6%

29.8%

7.3

6.7

7.0

4.2

4.9

4.6

4.1

3.6

3.8

5.2

5.2

5.4

5.3

2.0

3.7

2.9

1.2

1.1

1.1

3.2

Optimal No. of Irrigation (Rabi major crop) Applied No. of Irrigation (Rabi major crop) Cropping Intensity Water Scarcity Index (Kharif) Water Scarcity Index (Rabi) Paddy Yield (Qtls./acre) in Kharif Wheat Yield (Qtls./acre) in Rabi Number of times crop lost in last 5 years

3.3

3.4

3.3

4.6

10.8

9.1

6.1

3.3

5.5

5.3

2.6

2.8

2.7

4.1

9.4

8.0

4.8

2.3

4.2

4.5

180.6% 28.6%

175.4% 21.8%

178.0% 25.2%

107.5% 47.8%

168.4% 19.9%

138.0% 33.9%

102.4% 71.7%

100.2% 66.8%

101.3% 69.3%

140.8% 42.2%

16.9%

16.3%

16.6%

71.2%

21.3%

46.2%

90.1%

97.8%

94.0%

50.7%

12.9

12.1

12.5

13.7

16.4

15.0

4.5

3.1

3.8

10.5

10.5

11.2

10.9

11.3

12.0

11.6

7.2

5.0

6.1

10.9

1

2.9%

0.7%

1.8%

9.9%

45.6%

27.7%

8.0%

61.2%

34.7%

20.6%

2

23.4%

8.6%

16.0%

42.8%

46.5%

44.6%

56.6%

36.4%

46.5%

34.9%

3 or more Recent monsoon timing On time Earlier than usual Later than usual Random

73.7%

90.6%

82.2%

47.3%

6.2%

26.9%

35.3%

2.4%

18.8%

44.5%

1.4% 2.1% 5.6% 90.6%

0.7% 2.5% 1.8% 95.0%

1.1% 2.3% 3.7% 92.8%

11.1% 0.8% 2.0% 86.1%

31.8% 8.2% 3.3% 56.7%

21.5% 4.5% 2.7% 71.4%

3.2% 20.4% 7.6% 68.8%

3.6% 50.4% 0.8% 45.2%

3% 35% 4% 57%

8.2% 13.6% 3.5% 74.6%

61


When considering the entire sample, all but 17 households cultivated in the Kharif season. In terms of access to irrigation, an average of 73% of the cultivated area was under irrigation. In Silao, Islampur and Nalhati-2 blocks, more than 95% of cultivated area under Kharif major crop was irrigated. In Rajnagar, only 69% of the area cultivated under major crop in Kharif was irrigated, while in Kashipur and Raghunathpur-I, only 33% and 35% of the area cultivated under Kharif crop was irrigated. In Silao and Islampur, 98% and 96% households were able to cultivate in Rabi. In Nalhati-2 block, 86% households cultivated in Rabi season while in Rajnagar block only 32% households were involved in Rabi cultivation. In Kashipur and Raghunathpur-I blocks of Purulia, only 12% and 4% of the households cultivated in the Rabi season. The level of crop diversification was very low in the Kharif season. Except for Silao, where 7% households cultivated more than one crop in Kharif, only 1% cultivated more than one crop in Kharif in all the other blocks. Most of the households cultivated just paddy in Kharif season. Few households cultivated maize, pulses and other crops. There was more diversification in Rabi season. 63% households in Nalanda and 20% households in Birbhum cultivated more than one crop in Rabi season. Apart from wheat, which was cultivated by a majority of the households, a significant number of households cultivated crops such as potato, maize, paddy, mustard, gram, tomato, onion, coriander and other vegetables. Rabi cultivation was much more dependent on irrigation than Kharif cultivation, and only those households with access to irrigation were likely to be involved in cultivation in Rabi. This is also evident from observing the difference between optimal and applied irrigations by households that cultivated in Kharif and Rabi. In Kharif, while average optimal irrigation (according to the respondents) was 5.2, the average applied irrigation was 3.2. But, in Rabi, the average optimal irrigation was 5.3 and average applied irrigation was 4.5. For Nalhati-2 block households which cultivated paddy, mustard and potato in Rabi, the average optimal irrigation was 10.8 while optimal irrigation was 9.4. The cropping intensity was calculated as total area cultivated in Kharif and Rabi divided by area cultivated once by the household (maximum of area cultivated in Kharif and Rabi). This ratio was very high in Nalanda. The cropping intensity was 181% in Silao and 175% in Islampur, indicating that the land was cultivated almost twice over the year. The cropping intensity in Nalhati-2 block 62


was slightly lower than that in Nalanda at 168% but much higher than that of Rajnagar block (107%) in the same district. The average cropping intensity was 102% and 100% in Kashipur and Raghunathpur-I blocks in Purulia indicating single cultivation on the land among farmers in Purulia. We also used a simple index which serves as an indicator of water scarcity at farm level. This is calculated as the ratio of shortfall in applied irrigation to optimal irrigation. Mathematically, the formula used for water scarcity for each crop is: Water Scarcity Index = 1-

Min(Optimal Irrigation, Applied Irrigation) Optimal Irrigation

(4.1)

If the household has applied irrigation which was equal to or larger than (stated) optimal irrigation, the water scarcity index would be 0. The index takes a value of 0 if stated optimal irrigation is 0. It will be greater than 0 as and when the applied irrigation is less than optimal irrigation. This index was calculated for each crop in kharif and was weighted by the area on which the crop was cultivated. The same was done for rabi as well to derive an index of scarcity in the rabi season. For households that cultivated in kharif but did not do so in rabi, the index took value of 1 for rabi season. Based on this formulation, we find that the water scarcity index for all households in kharif was 42% while that in rabi was 51%. This meant that on average, households applied 42% lesser than optimal irrigation in kharif and 51% lesser than optimal irrigation in rabi. Among different blocks, blocks in Nalanda (Silao and Islampur) had lower water scarcity in rabi compared to kharif. This could be due to usage of groundwater in Rabi, which gives more assured irrigation. It could also be partly due to cultivation of crops requiring lesser irrigation in rabi. Overall, water scarcity was the least for Nalanda district and highest in the Purulia district. In rabi, water scarcity was as high as 98% in Raghunathpur-I block of Purulia district. In Rajnagar, water scarcity in kharif was only 48% while it was 71% in rabi. The water scarcity index in kharif for Kashipur at 72% was higher than water scarcity in rabi for any of the four blocks in the other two districts. Crop yield can also serve as a measure of water scarcity. The average paddy yield across households in the sample was 10.5 quintals per acre or 2.63 tonnes per hectare. This yield is

63


considered above average. Paddy yield36 was the highest in Nalhati-2 (16.4 quintals per acre or 4.1 tonnes per hectare) and lowest in Raghunathpur-I (3.1 quintals per acre or 0.78 tonnes per hectare). Any yield below 1 ton per hectare is considered a very low paddy yield. Raghunathpur-I had such low level of yields on average, while Kashipur had a yield that was slightly above 1 ton per hectare. This also indicates a high level of water scarcity in Purulia and a need to diversify into other livelihood activities for sustenance. The difference in the yield level indicator is particularly stark in the case of Raghunathpur-I when compared to blocks in Nalanda. All the above indicators seem to suggest a low level of water scarcity in Nalanda and Nalhati-2 block in Birbhum while a grim situation in Rajnagar block in Birbhum and a dire situation for Purulia. However, based upon the stated responses of the households, farm households in Nalanda seemed to have suffered more losses in the recent years. 74% households in Silao and 91% households in Islampur had suffered losses in 3 or more years in the last 5 years. 91% farmers in Silao and 95% in Islampur also felt that the monsoons in the recent years had been random. Given this background over heterogeneous levels of water scarcities within the study regions, we look into the responses to irregular rainfall in critical phases by households in the different blocks. Table 4.1.2 provides the typical responses to lack of rainfall in critical crop growth phases. Table 4.1.2 Response to Irregular Rainfall across Households in the Study Sample Nalanda Wait for Rainfall Buy water Change Crop Change Crop Variety Leave Cultivation Labour in same village Labour in other villages/urban areas Use own irrigation Water Buyers Water Sellers

36

Birbhum

Purulia

All Sample

Silao 24.0% 92.0% 1.7% 0.3%

Islampur 34.5% 93.2% 1.4% 0.0%

All 29.2% 94.4% 1.6% 0.2%

Rajnagar 95.5% 63.5% 0.0% 0.4%

Nalhati-2 89.0% 95.1% 2.0% 1.6%

All 92.2% 79.3% 1.0% 1.0%

Kashipur 92.0% 28.8% 0.0% 0.0%

Raghunathpur-I 97.6% 42.0% 0.0% 0.0%

All 94.8% 35.4% 0.0% 0.0%

70.1% 70.1% 0.9% 0.4%

1.4%

0.0%

0.7%

0.0%

0.0%

0.0%

0.4%

11.1%

8.0%

2.8%

1.4%

0.4%

0.9%

11.5%

6.9%

9.2%

39.6%

49.6%

44.6%

17.5%

0.3%

0.0%

0.2%

8.2%

2.0%

5.1%

45.6%

54.0%

49.8%

17.7%

14.6%

32.0%

27.0%

13.5%

2.9%

8.2%

12.0%

1.2%

6.6%

13.2%

92.3% 8.7%

94.7% 20.3%

93.5% 14.4%

63.1% 2.0%

95.9% 2.0%

79.6% 2.0%

29.2% 0.0%

42.0% 0.0%

35.6% 0.0%

70.5% 5.9%

http://drdpat.bih.nic.in/PA-Table-25-West%20Bengal.htm

64


The responses by farmers in different blocks were highly heterogeneous. In Nalanda, 30% of households said they would wait for rainfall to arrive in such situations, while 92% responded that they would buy water for irrigation. In Birbhum, 92% responded that they would wait for rainfall while 79% said they would buy water for irrigation. In Purulia, 98% households responded that they would wait for rainfall while only 35% said they would buy water for irrigation. This indicates a lack of option to buy water among a majority of households in Purulia. Very few households said they would change crop or crop variety in response to lack of rainfall. A few households were also likely to leave cultivation in response to lack of rainfall in critical phases. Only in Raghunathpur-I, 11% households said they would leave cultivation in such a situation. Very few households said they would go for labor in the same village in such situations in Birbhum and Nalanda. In Purulia, 40% in Kashipur and 50% in Raghunathpur-I said they would take to labor based income within the same village in such situations. Also, 46% and 54% of households in Kashipur and Raghunathpur-I said they would go for labor to other villages and urban areas in such situations. Using own irrigation was an option to households in Nalanda and in Rajnagar and Kashipur blocks. As expected, there were no water sellers in Purulia. Only 2% of sampled households in Birbhum were water sellers while 14% households in Nalanda were water sellers.

Section 4.2. Salinity and Water Scarcity (Bangladesh) Section 4.2.1. Salinity and Drinking Water It is quite unusual for an average rural Bangladeshi household to purchase drinking water. In rural Bangladesh, households mostly rely on free water sources such as tube wells and open water bodies. But in the saline areas, many households have to purchase drinking water as these water sources get contaminated by saline water, particularly during dry seasons. Refer to table 2.5.2 for households’ access to drinking water. Except West Lunda, in all the villages of Patuakhali all the interviewed households had to purchase drinking water at least for some months. They had to purchase water for around five months in a year. In Bethbunia, Shymnagar and Guhalbathan villages of Khulna district, almost all the households had to purchase drinking water. Drinking water scarcity in these three villages of Khulna was much higher than the villages 65


in other two districts, as villagers purchase water for more than seven months in a year. Dumuria and Parshemari were the two severely water scarce villages in Satkhira district. Mainuddin et al. (2014a) have also reported that the south west districts are affected by low water availability in the dry season. Salinity problems in surface and groundwater, along with groundwater arsenic contamination, has limited irrigation development in these regions.

Section 4.2.2. Salinity and Cropping Salinity severely constraints crop cultivation. Due to salinity, hardly any of the sampled farmers could cultivate paddy in the dry season. Kharif paddy known as Aman, is the major crop for the farmers in the study area. Kharif paddy is practically the only crop grown during the monsoon season. In the dry season, the crop fields in Khulna and Satkhira are turned into shrimp ponds, whereas in Patuakhali, farmers take to some other cash crops on a small scale. Section 4.2.3. Salinity and Crop Diversification The Herfindahl Index (HI) was calculated to derive the level of crop diversification in the sampled households. The index value ranges from 0 to 1, where a higher index value indicates a lower level of diversification and vice-versa. Unitary index value means perfect specialization. The estimated index values for the sampled farm households presented in the Table 3.2.3 indicates a very low level of crop diversification. The two shrimp producing districts (Khulna and Satkhira) had almost no diversification, whereas in Patuakhali, the index value was 0.659 indicating below moderate level of diversification. Due to salinity, farmers in the saline areas can cultivate paddy only during the monsoon season. They find it rational to devote all their available land to paddy in this season to ensure sufficient supply of the staple food for their households. Winter is the main cropping season in Bangladesh where farmers cultivate different types of vegetables, food and non-food cash crops along with paddy. The practice in the study area is quite different as farmers are constrained by salinity, and growing paddy in most cases is not possible. In Khulna and Satkhira, during the winter season, saline water turns the crop fields into fish ponds. When a nearby plot is transformed into fish pond with saline water, a farmer has no other alternative but to turn their own land into pond as well. But as saline water fishing is not possible in farm lands of Patuakhali, the farmers there cultivate different types of cash crops such as: mung, lentil, vegetables, etc.

66


Therefore, paddy in the Kharif season has become the only crop for farmers in Khulna and Satkhira, with almost no scope for cropping in winter. Table 3.2.3. Herfindahl Index for Crop Diversification Salinity Level Low

High

All

Khulna

0.919

0.979*

0.944

Patuakhali

0.645

0.683

0.659

Satkhira

0.978

0.959

0.976

Landless

0.891

0.769***

0.859

Marginal

0.753

0.769

0.758

Small

0.583

0.917***

0.783

Medium & Large

0.696

0.955**

0.838

District

Farm Category

All 0.813 0.822 Note: ***, ** and * indicate mean differences between less and high saline areas significant at 1%, 5% and 10% level.

0.816

The Herfindahl index is also calculated for different categories of farmers. The index value for all the sampled farm households ranges from 0.758 to 0.859 with no notable difference across different farm categories. When it comes to less saline areas, the index value is higher for upper landless and marginal farmers (indicating lower diversification). Contrary to this, in the high saline areas, diversification gradually decreases while moving from landless farmers to larger farmers. As farmers in less saline areas are more likely to have better opportunities for farming, the higher the land size the more likely are the farmers to exploit the potential. Alternatively, in high saline areas, there remains reduced farming potential, particularly during winter. As it is more likely for the larger farmers to be in possession of better quality land, they devote their land entirely to paddy. Whereas, farmers with relatively less land holdings are likely to have lower quality land, particularly high salinity prone land, which is not suitable for paddy. Consequently, they cultivate some saline tolerant local variety crops, for instance sona mung (a local variety of mung) and felon (a particular type of lentil that can be grown in saline land), which ultimately contributes towards a higher crop diversification. Section 4.2.4. Salinity and Crop Loss

67


In saline prone areas, sudden saline water intrusion or rise in salinity levels causes serious crop loss. Table 4.2.4 presents farmers’ perception over production losses due to salinity in the rabi and kharif seasons of 2015 and 2014. As is true for most saline areas, farmers can cultivate paddy only in the kharif season, they hardly grow any other crops along with paddy. Rainfall and availability of fresh water in the water bodies reduces salinity levels and enables farmers to cultivate paddy. On the other hand in the rabi season, salinity level increases due to less rainfall and reduced flow in the water bodies from the upper streams. Presence of salinity does not allow for deep water irrigation, and in the saline areas the upper aquifers are also saline. Deep irrigation is costly and sometimes not possible, which ultimately forces farmers to go for relatively less water dependent crops. Hence, to some extent, salinity may encourage crop diversification, but at the cost of reduced livelihood incomes. According to the sampled farmers, they lost one kilo out of every five kilos they produced in the kharif 2015 season due to salinity. In the same season, they lost one kilo out of every four kilos of vegetables they produced. No significant differences across farmer categories in productivity loss means that different categories of farmers experienced similar effects of salinity. In Rabi season of 2015, farmers lost around 30% of their total paddy production due to salinity. The loss in Rabi paddy significantly varied across different farm categories and was the highest for the landless farmers who lost 34.09% of their total production. In case of vegetables and watermelons too, the landless farmers sustained highest losses and the loss amount varied across different farm categories (Table 4.2.4). Table 4.2.4. Farmers’ Perception over Percentage of Production Lost due to Salinity Crops

Landless

Marginal

Small

Medium & Large

All

Paddy

34.09

25.48

25.10

29.23

30.33**

Wheat

11.71

15.77

14.50

17.50

14.00

Maize

10.97

9.52

8.75

10.00

10.15

Pulses

13.07

11.24

12.88

10.83

12.19

Vegetable

26.58

15.72

22.00

19.58

22.71**

Watermelon

25.93

11.29

20.00

21.11

21.11***

Paddy

19.53

18.82

24.44

21.39

20.01

Vegetable

19.19

29.09

30.00

30.00

25.35

% of production lost in Rabi 2014

% of production lost in Kharif 2015

Section 4.2.5. Perception over Changes in Salinity Levels and the Reasons for Salinity Increase 68


There exist considerable differences in farming households’ perception over whether salinity had increased or decreased over the past five years. Among the sampled respondents, 40.86% observed that salinity had increased during the last five years. Around half of the respondents (48.97%) reported salinity having increased in the less saline areas. This proportion is higher (around 16%) than the high salinity areas. In the less saline areas too, a relatively higher proportion of respondents (36.90%) reported salinity to have decreased when compared with similar responses in the high salinity areas. In high salinity areas, 47% of the respondents observed salinity levels having remained unchanged over time, whereas, in less saline areas only 14.14% reported unchanged salinity levels (Table 4.2.5). Table 4.2.5. Respondents Perception over Changes in Salinity Levels during the Past Five Years Area

N

Increased

Decreased

Unchanged

Less Saline Areas

48.97%

36.90%

14.14%

290

High Saline Areas

33.44%

19.56%

47.00%

317

All

40.86%

27.84%

31.30%

607

In response to reasons for salinity increases, majority of the respondents (54.17%) identified sea level rise as the cause. In high and low saline areas, around 61% and 50.50% of the responses reported this as the main reason. Man-made reasons such as mis-management of polder gates, mostly in the form of not closing them on time, was reported in less saline areas in 36.63% cases. This is mainly done to let in saline water so that shrimp cultivation can be done. Only 8.18% respondents in high saline areas regarded this as the source of increased salinity. Around 31% and 61% responses regarded broken polder gates and sea level rise as reasons behind salinity increase respectively. Table 4.2.6. Reasons for Increasing Salinity (multiple responses) Responses

Less Saline Areas

High Saline Areas

All

Mismanagement in polder gate

36.63%

8.18%

26.60%

Broken polder gate

12.87%

30.91%

19.23%

Sea level rise

50.50%

60.91%

54.17%

202

110

312

Number of Responses

Local people who have been involved in shrimp farming have played a key role in changing the polder gate management practice over time, which has ultimately resulted in increased salinity 69


levels. In 63.03% cases local people made this decision and implemented it. The local elite did this in 34.45% cases. Involvement of local government in such decisions appears to be negligible (Table 4.2.7). Table 4.2.7. Persons/Groups Changing Management Practices which Led to Increased Salinity (multiple response) Persons/Groups

Less Saline Areas

High Saline Areas

All

Local elite/shrimp farmers

32.04%

50.00%

34.45%

Local government

2.91%

0.00%

2.52%

Local people

65.05%

50.00%

63.03%

103

16

119

Number of Responses

As with the reason for increasing salinity, in the case of salinity reducing factors, around 44% responses mentioned heavy monsoon. Managing polder gates properly through opening and closing them on time was reported in 40.52% cases. This was the major reason mentioned in less saline areas (50.64% responses). Repair of polder gate in less saline areas was cited as a reason by only 1.92% respondents, whereas in high saline areas it was mentioned by34.51% respondents (Table 4.2.8). Table 4.2.8. Reasons for Reducing Salinity (multiple responses) Reasons

Less Saline Areas

Severe Saline Areas

All

Polder Management

50.64%

26.55%

40.52%

Repair Damaged Polder

1.92%

34.51%

15.61%

Heavy Monsoon

47.44%

38.94%

43.87%

Though government was found to be almost an invisible actor in contributing towards increasing salinity, it was reported as the major actor responsible for reducing salinity. Central and local government initiatives in 20.31% and 35.63% of the responses are reported to have reduced salinity levels, respectively. Particularly in high saline areas, central government’s role was reported in 39.83% of the responses. Government initiative was mainly in the form of repairing polder gates and excavating water bodies. NGOs and development organizations were mostly mentioned as a reason by the less saline prone area respondents. Hardly anyone in high saline areas reported their role in salinity reduction. A reason for this could be that initiatives in high saline areas are required mostly through polder gate repair and excavation of water bodies, which require

70


large amounts of funds. Therefore, making meaningful interventions in these areas could be beyond the NGOs’ and development organizations’ capacities (Table 4.2.9). Table 4.2.9. Perception over Role of Government and NGOs in Reducing Salinity Persons/groups

Less Saline Areas

Severe Saline Areas

All

Central Government

4.20%

39.83%

20.31%

Local Government

34.27%

37.29%

35.63%

NGO/Development Organizations

28.67%

0.85%

16.09%

Local People

32.87%

22.03%

27.97%

71


Chapter 5: Livelihood Profiles of households Basic livelihood profiles of the sampled households are described in this chapter. Households are profiled based upon whether they are into farming, dairy and poultry, fishing, enterprises, labor force participation, migration and salaried employment. Distributions of incomes across these profiles are categorized into quintiles and Gini coefficients provide an account of inequality across various groups.

72


Section 5.1. Livelihood Diversification Strategies (India) First, we look at the participation of the households into livelihood generating activities for the study sample. Table 5.1 presents the participation of households amongst various income generating activities.

73

ACIAR Report ADP/2015/032 Table 5.1 Participation of Sample Households in different Income Generating Activities Nalanda Silao

Islampur

Birbhum All

Rajnagar

Nalhati-2

Purulia All

Kashipur

Raghunathpur-I

All Sample All

Households involved in cultivation Households involved in dairy/poultry/fishing

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

98.0%

99.6%

98.8%

99.6%

34.1%

27.4%

30.8%

20.9%

10.6%

15.7%

31.2%

14.8%

23.0%

23.6%

Households involved in Agricultural Labour

11.1%

13.9%

12.5%

55.7%

42.9%

49.3%

43.6%

61.2%

52.4%

36.9%

Households involved in Casual Labour

50.2%

51.6%

50.9%

79.1%

50.2%

64.6%

83.6%

86.4%

85.0%

66.2%

Households undertaking Rural Migration

5.9%

2.5%

4.2%

9.8%

4.9%

7.4%

56.4%

2.0%

29.2%

13.2%

Households undertaking Urban Migration

30.0%

35.9%

32.9%

15.2%

17.6%

16.4%

20.0%

12.4%

16.2%

22.4%

Households earning salaries

15.0%

13.2%

14.1%

9.8%

7.8%

8.8%

10.8%

12.0%

11.4%

11.6%

Household taking up enterprises Households earning from other income sources

6.6%

6.0%

6.3%

15.6%

40.0%

27.8%

11.2%

19.6%

15.4%

16.0%

19.2%

26.0%

22.5%

9.8%

9.8%

9.8%

12.8%

6.8%

9.8%

14.5%

74


When considering the overall sample, all except 6 households in Purulia were engaged in cultivation. 30% households in Nalanda earned income from dairy, poultry or fishing. In Birbhum, 16% households earned income from livestock while 23% earned income from livestock in Purulia. Only 13% households in Nalanda were involved in agricultural labour while 50% households in Birbhum and 52% households in Purulia earned income from agricultural labour. Most of the households earned income from non-migrant casual labour. In Nalanda, 51% households earned income from casual labour while 65% earned income from casual labour in Birbhum and 85% earned income from casual labour in Purulia. 56% of households in Kashipur earned income from rural migration while less than 10% households undertook migration in all the other blocks. This could be partly due to lack of connectivity of Kashipur to different rural areas. So, those who were forced to earn from agricultural labour in other rural areas in kharif and rabi had to migrate. Raghunathpur-I was connected to nearby rural areas and farmers were able to travel for labour in nearby rural areas without having to migrate. One-third of the households in Nalanda undertook urban migration while one-sixth undertook urban migration in Birbhum and Purulia. 14% households in Nalanda earned income from salaried employment, 9% households in Birbhum earned from salaries while 11% households in Purulia earned from salaries. Apart from this, enterprise take-up was low in Nalanda. Only 6% households earned from enterprises in Nalanda. This was 16% in Rajnagar and 40% in Nalhati-2. Most of the households in Nalhati-2 were involved in sericulture. 11% households in Kashipur and 20% households in Raghunathpur-I were involved in enterprises. 14% households earned incomes from other income sources such as poultry, rent, interests, dividends, pensions, water sales, etc. Table 5.1.2 provides the number of income generating activities in which households were engaged.

75

ACIAR Report ADP/2015/032 Table 5.1.2. Number of Livelihood Activities Engaged in by the Household Nalanda

Birbhum

Purulia

All Sample

Silao

Islampur

All

Rajnagar

Nalhati-2

All

Kashipur

Raghunathpur-I

All

1

5.9%

3.2%

4.6%

3.3%

4.9%

4.1%

3.2%

0.0%

1.6%

3.5%

2

48.8%

44.8%

46.8%

20.1%

35.9%

28.0%

12.0%

24.0%

18.0%

31.7%

3

30.3%

38.1%

34.2%

44.3%

36.7%

40.5%

31.2%

45.6%

38.4%

37.5%

4

11.1%

11.4%

11.3%

25.0%

20.0%

22.5%

36.4%

27.6%

32.0%

21.5%

5

3.8%

2.5%

3.2%

6.6%

2.4%

4.5%

13.6%

2.8%

8.2%

5.2%

6

0.0%

0.0%

0.0%

0.8%

0.0%

0.4%

2.8%

0.0%

1.4%

0.6%

7

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

8.8%

0.0%

0.4%

0.1%

Average income generating activities of a household

2.6

2.7

2.6

3.1

2.8

3.0

3.6

3.1

3.3

3.0

76


From table 5.1.2, we can observe that households in more water scarce regions seemed to be engaged in higher number of livelihood activities. In Rajnagar, Kashipur and Raghunathpur-I, the average number of activities of households was 3.1, 3.6 and 3.1 respectively. In Silao, Islampur and Nalhati-2, households were on average involved in 2.6, 2.7 and 2.8 activities. What we find in water scarce regions is higher engagement in different activities to supplement their household incomes. The quality of employment in such activities would help us understand if they could invest in their human capital augmentation in the future. For the whole sample, very few households (4%) were involved in only one income generating activity. 32% households were involved in 2 activities, 38% in 3 activities, 22% households in 4 activities and 6% in 5 or more activities. In Nalanda, around 85% households were involved in 3 or lower number of activities, while 27% households in Birbhum and 30% households in Purulia were involved in more than 3 activities for income generation. Table 5.1.3 presents the average income from different activities in the study sample.

77

ACIAR Report ADP/2015/032 Table 5.1.3. Income from different livelihood activities among households in Study Sample (INR) Nalanda Silao

Islampur

Birbhum All

Rajnagar

Nalhati-2

Purulia All

Kashipur

All Sample

Raghunathpur-I

All

Income from Cultivation

19226 38286 9466

33615 5359

35975 7434

17572 2331

23107 453

20345 1390

908 5186

-2698 2216

-895 3701

4337

3530

4358

3939

6010

10762

8391

3156

5309

4232

5432

25100

25059

25080

19560

14236

16893

29291

34662

31976

24723

Income from Salaries

28418

20395

24449

9865

12468

11169

12259

20692

16476

17718

Income from Rural Migration Income from Urban Migration Income from Enterprises

3262

1537

2409

3715

1148

2429

7980

848

4414

3059

27042

36101

31524

7637

10884

9264

11377

6564

8971

17290

3899

4246

4070

6655

10989

8827

5228

8180

6704

6410

8453

5792

7136

4021

3832

3927

7688

6371

7029

Income from Dairy/Poultry/Fishing Income from Agricultural Labour Income from Casual Labour

Income from Other Sources

6094

Total Income Gini of Household Income

147456

136461

142016

77367

87880

82634

83073

82143

82608

104289

43.9%

40.7%

42.4%

32.2%

35.1%

33.4%

37.9%

37.7%

37.8%

41.4%

78


From table 5.1.3, we observe that average household income for the entire sample is INR 1,04,289 or INR 8,691 per month. The households earned INR 19,226 from cultivation and 24,723 from casual labour. They earned INR 17,718 from salaried employment and INR 17,290 from urban migration, INR 6,410 from enterprises, INR 5,432 from agricultural labour, INR 4,337 from livestock, INR 3,059 from rural migration and INR 6,094 from other sources such as forestry, rent, interests, dividends, pensions, water sales, machine hiring and government benefits. The average household income was the highest in Silao block, where households earned INR 1,47,456 or INR 12,288 per month. Among the different sources, cultivation based income was the highest (INR 38,286) followed by salaries (INR 28,418), urban migration (INR 27,042) and casual labour (INR 25,100). They also earned INR 9,466 from livestock and around INR 3,000 from each of rural migration, enterprises and agricultural labour categories. In Islampur too, households earned INR 1,36,461 or INR 11,372 per month on average. Here, urban migration income contributed the highest share (INR 36,101). It was followed by cultivation (INR 33,615), casual labour (INR 25,059) and salaried (INR 20,395). On average, households in Islampur earned INR 5,359 from livestock, INR 4,358 from agricultural labour, INR 4,246 from enterprises, and INR 1,537 from rural migration. In Nalhati-2 block, on average, households earned INR 87,880 or INR 7,323 per month. Cultivation contributed the highest with INR 23,107. Casual labour, salaries, enterprises, urban migration and agricultural labour, contributed somewhat equally at INR 14,236, INR 12,468, INR 10,989, INR 10,884 and INR 10,762 respectively. Households earned INR 1,148 from rural migration and INR 453 from livestock. In Rajnagar block of Birbhum, households earned INR 77,367 or INR 6,447 per month on average. Casual Labour was the main source of income in this block with households earning INR 19,560 on average from it. Households earned INR 17,572 from cultivation. They earned on average, INR 7,637 from urban migration, INR 6,655 from enterprises, INR 6,010 from agricultural labour, INR 3,715 from rural migration and INR 2,331 from livestock. In Raghunathpur-I block of Purulia, households on average earned slightly higher (INR 82,143) than those in Rajnagar block. But here, households on average earned negative net incomes from cultivation (INR -2,698). Casual labor (INR 34,662) was the major source of income for the 79


households sampled in the block. The households earned INR 20,692 from salaries, INR 8,180 from enterprises, INR 6,564 from urban migration, INR 5,309 from agricultural labour, INR 2,216 from dairy and INR 848 from rural migration. In Kashipur block of Purulia, on average, households earned INR 83,073 or INR 6,923 per month. Households earned the most from casual labour (INR 29,291) followed by salaries (INR 12,259), urban migration (INR 11,377), rural migration (INR 7,980), enterprises (INR 5,228), livestock (INR 5,186) and agricultural labour (INR 3,156). In terms of distribution of income, the gini was moderately high for the whole sample at 41%. The gini was highest for the district of Nalanda at 42% and lowest for Birbhum at 33%. To further examine the diversification patterns of the households, we calculated the average shares of different sources of income. There were many households with negative cultivation incomes. While calculating the share of a particular source of income in the household income, these values presented an incorrect picture. For instance, a household in Raghunathpur-I earned INR –ve 65,000 from cultivation, INR 7,000 from agricultural labour and INR 66,000 from casual labour. The total net income of the household was INR 8,000. If we calculated the share of cultivation, agricultural labour and casual labour to total household income for this household, the shares were -812.5%, 87.5%, and 825% respectively. Such figures distorted average shares. Therefore, for understanding the diversification patterns, we excluded those households with negative net incomes from crop cultivation. Figure 5.1 below indicates the income diversification patterns of the households. Figure 5.1. Income Diversification Pattern of Households (overall sample) Enterprise 6%

Other 4%

Urban Migration 13%

Cultivation 24%

Rural Migration 4%

Livestock 4%

Salary 7% Casual Labour 30%

80

Agricultural Labour 8%


From figure 5.1, we observe that the most important source of income for the sampled households is casual labour, which, on average, contributes 30% of the household income. Crop cultivation contributes 24% of the household income. Urban migration contributes 13% towards household income. Agricultural labour, salary and enterprise contribute 8%, 7% and 6% to household income respectively. Rural migration, livestock incomes and other income, on average, contribute 4% each towards household income. To analyze the distribution of income further, we examine the income from different sources across different quintiles. Table 5.1.4 presents the income from different livelihood sources across 5 quintiles. Table 5.1.4 Annual Income Distribution across Quintiles

Quintiles

Q1

Cultivatio n Income

Livestock Income

Agricultural Labour Income

Casual Labour Income

Salaried Income

1893

836

4130

19070

Q2

7290

1695

6066

Q3

12989

4552

5882

Q4

23233

5964

Q5

50802 26.8

Q5/Q1

Rural Migratio n Income

583

2064

25085

897

24404

2801

5988

27061

8648

5094

10.4

1.2

Urban Migration Income

Enterprise Income

Other Income

Total Income

1577

2298

763

33213

2779

5129

3971

1621

54534

3803

10977

6888

1285

73581

7148

2844

19648

7952

5016

104855

28016

77262

3806

49191

10952

21817

255588

1.5

132.4

1.8

31.2

4.8

28.6

7.7

From table 5.1.4, we can observe that the income of the households belonging to fifth quintile is around 8 times the income of the households in the first quintile. Among different sources of incomes, the difference is most stark for salaried income where the households in fifth quintile earn 132 times that of incomes in the first quintile. Income from urban migration for the fifth quintile was 31 times the same for the first quintile, and cultivation income of fifth quintile was 27 times that of first quintile. The inter-quintile ratios for livestock income, enterprise income, rural migration, casual labour and agricultural labour income were 10.4, 4.8, 1.8, 1.5 and 1.2 respectively. To understand whether the sources of incomes are inequality decreasing or inequality increasing, we use a gini decomposition technique to estimate the elasticity of a source of income to gini. Table 5.1.5 presents the results of gini decomposition for the complete sample.

81

ACIAR Report ADP/2015/032 Table 5.1.5. Gini Decomposition Source

Sk

Gk

Rk

Share

% Change

Cultivation Income

18%

82%

63%

23%

4.4%

Livestock Income

4%

88%

43%

4%

-0.4%

Agricultural Labour Income

5%

82%

4%

0%

-4.8%

Casual Labour Income

24%

55%

12%

4%

-19.9%

Salaried Income

17%

94%

87%

33%

16.4%

Rural Migration Income

3%

93%

9%

1%

-2.4%

Urban Migration Income

17%

87%

63%

22%

5.3%

Enterprise Income

6%

91%

33%

4%

-1.7%

Other Income

6%

95%

66%

9%

3.0%

Total Income

41%

From table 5.1.5, we observe that casual labour income (24%) has the highest share in total income followed by cultivation income (18%), salaries (17%) and urban migration income (17%). Gini coefficients of different sources of incomes are quite high except casual labour income which indicates a comparatively lower gini of 55%. The gini correlation (Rk) indicates the correlation between the income from a particular source and the cumulative distribution of total income. A higher value would indicate that higher quintiles earn more from that particular income source while lower value would indicate that the lower quintiles earn more from the particular source. The gini correlation for agricultural labour (4%), rural migration labour income (9%) and casual labour income (12%) is relatively lower than other income sources. The product of share (Sk), gini (Gk) and gini correlation (Rk) for a particular source is the share of gini of household income. Here, we find that salaried income contributes the highest (33%) to gini, followed by cultivation (23%) and urban migration income (22%). These three sources of income are inequality increasing with 1% increase in share of these sources that could lead to 16.4%, 4.4% and 5.3% in gini respectively. Casual labour is the most inequality decreasing with 1% increase in share that could reduce gini by 19.9%. So, casual labour along with contributing the highest to household income is also inequality lowering. Agricultural labour income (-4.8%), rural migration (-2.4%) and enterprise income (-1.7%) are also inequality lowering.

82


Section 5.2.: Livelihood Strategies in Climate Stressed Areas (Bangladesh) Section 5.2.1. Households’ Involvement in Different Income Generating Activities Table 5.2.1 shows proportion of households involved in different types of income generating activities. Casual labour supply (67.1%) followed by poultry farming (50.2%) were the two most common income sources for the households. In high saline areas, relatively higher proportion of the households (70.6%) were involved in casual labour supply as compared to the less saline areas (63.5%), particularly in Khulna and Patuakhali districts. In Patuakhali district, as opposed to 59.4% in less saline areas, 80.6% households in high saline areas were supplying casual labour. Except Satkhira, in the other two districts, relatively higher proportion of the households had earnings from poultry farming. Another 15.9% of the households made some earnings from livestock through milk selling and beef fattening, with relatively more proportion of households involved in such activities in the less saline areas. Around 28% of the households were involved in crop farming. Similar proportion was involved in fisheries. According to national statistics, 54.5% of the rural population is involved in agriculture, forestry and fisheries (BBS, 2011)37. Quite noticeably, compared to high saline areas, almost twice the proportion of households in less saline areas were involved in crop farming. The difference is most notable in Satkhira, where around 3% and 29% in high and less saline areas were into farming. Among the three districts, Khulna had the highest proportion of households (estimated at 45.6%) involved in fisheries against the sample average of 27.8%. The households in Khulna and Satkhira appear to have taken advantage of the highly saline water and cultivated shrimp and other saline water fish. But as the practice of turning crop fields into fish ponds is not prominent in Patuakhali, very low proportion of households (6.3%) in high saline areas were found to be doing fisheries. In less saline areas of Patuakhali, 23.8% of the households, which is comparable to the proportion in Satkhira, were involved in fisheries. Fish collected from open water bodies contributed to 16.6% of the household’s earnings.

37

BBS (2011). Report on Labour Force Survey 2010. Bangladesh Bureau of Statistics, Statistics Division, Ministry of Planning, Government of the People’s Republic of Bangladesh, Dhaka.

83


For the entire sample, 15.3% households were involved in businesses or enterprises. In Khulna and Satkhira districts, almost similar proportions of households had some kind of enterprises, whereas in Patuakhali the proportion was relatively lower. Involvement with enterprises did not vary with salinity levels. 31% of the sampled households got some financial assistance from different social safety net programmes, mostly in the form of education stipend and relief. Around 10% of households had members into salaried jobs. At least one out of every ten households had income from other sources that included remittances, forestry, rent, gardening, tree selling, pension, asset selling, etc. Here it is noteworthy mentioning that compared to households in less saline areas, very few households in high saline areas were involved with activities such as gardening and tree selling. Alternatively, collection of wood, honey, leaf from forests as income sources was popular only with high saline area residents of Satkhira, who live near the Sundarbans forests.

84

ACIAR Report ADP/2015/032 Table 5.2.1 Proportion of Households Involved in Different Income Generating Activities Khulna Income Generating Activities

Patuakhali

Satkhira

All

Low

High

All

Low

High

All

Low

High

All

Low

High

All

Crop Farming

32.5%

23.1%

27.8%

48.8%

29.4%

39.1%

28.8%

3.1%

15.9%

36.7%

18.5%

27.6%

Poultry

50.6%

41.3%

45.9%

52.5%

46.3%

49.4%

55.6%

55.0%

55.3%

52.9%

47.5%

50.2%

Livestock Rearing

20.6%

12.5%

16.6%

17.5%

20.0%

18.8%

21.3%

3.8%

12.5%

19.8%

12.1%

15.9%

Fisheries

48.1%

43.1%

45.6%

23.8%

6.3%

15.0%

18.1%

27.5%

22.8%

30.0%

25.6%

27.8%

Fish Collection

25.0%

10.0%

17.5%

15.6%

16.3%

15.9%

8.8%

23.8%

16.3%

16.5%

16.7%

16.6%

Enterprise

16.9%

14.4%

15.6%

9.4%

15.6%

12.5%

18.8%

16.9%

17.8%

15.0%

15.6%

15.3%

Salaried Job

10.6%

10.6%

10.6%

13.8%

8.8%

11.3%

10.0%

7.5%

8.8%

11.5%

9.0%

10.2%

Casual Labour

58.8%

63.8%

61.3%

59.4%

80.6%

70.0%

72.5%

67.5%

70.0%

63.5%

70.6%

67.1%

Migration

12.5%

19.4%

15.9%

5.6%

5.6%

5.6%

18.8%

13.8%

16.3%

12.3%

12.9%

12.6%

SSNP

14.4%

15.6%

15.0%

42.5%

30.6%

36.6%

37.5%

45.6%

41.6%

31.5%

30.6%

31.0%

Other

24.4%

1.9%

13.1%

9.4%

0.6%

5.0%

17.5%

15.6%

16.6%

17.1%

6.0%

11.6%

85

ACIAR Report ADP/2015/032 Table 5.2.2. Annual Income (tk) from Different Sources Khulna

Patuakhali

Satkhira

Entire Sample

Income Sources

Low

High

All

Low

High

All

Low

High

All

Low

High

All

Crop Farming

-11

-787

-399

3024

-1216

904**

604

111

357

1205

-631

287**

Livestock & Poultry

5926

2186

4056***

4805

6413

5609

3100

1934

2517**

4610

3511

4061

Fish Farming

18417

36539

27478***

2857

542

1699***

7218

23122

15170***

9497

20068

14782***

Fish Collection

3012

8319

5665

4594

4789

4692

606

2452

1529***

2737

5187

3962*

Salary & Pension

13188

8899

11043

15288

8681

11984

7925

7003

7464

12133

8194

10164*

Casual Labour

35613

44278

39945

50746

40122

45434**

35895

40813

38354

40751

41738

41244

Enterprise

10813

13969

12391

12033

15513

13773

13784

11314

12549

12210

13598

12904

SSNP

417

365

391

830

449

639***

910

1044

977

719

619

669

Other

5486

2095

3790*

8721

265

4493***

2062

5851

3956***

5423

2737

4080**

Total Income

92857

115862

104360**

102896

75559

89227***

72103

93643

82873***

89285

95021

92153

Per Capita Income

23473

29777

26625**

23164

16998

20081***

17621

22255

19938***

21420

23010

22215

Gini of Total Income

0.378

0.405

0.398

0.332

0.367

0.368

0.353

0.309

0.336

0.377

0.369

0.374

0.373

0.384

0.39

0.388

Gini of Per Capita Income 0.385 0.399 0.399 0.339 0.354 0.364 0.391 0.347 Note: ***, ** and * indicate mean differences in income between less and high saline areas are significant at 1%, 5% and 10% level, respectively.

86


Section 5.2.2. Household’s Earnings from Different Sources Annually, an average sampled household earned 92,153 tk from different sources. The high saline area households earned 5, 465 tk more than the less saline area households. In Khulna and Satkhira, the high saline area households earned significantly higher than their counterparts in less saline areas, whereas in Patuakhali the scenario is completely the opposite. The difference is mainly due to difference in earnings from fisheries and salaried jobs. In Khulna, households in high saline areas earned almost double from fish farming compared to their counterparts in less saline areas, whereas in Satkhira, the households in high saline areas earned more than three times. Compared to Khulna, households in Satkhira earned less from fisheries. In Patuakhali, households earned negligible incomes from fisheries. A household earned annually 14,782 tk from fisheries. In less and high saline areas, households earned 9,497 tk and 20,068 tk respectively from fisheries. From salaries and pensions, a household earned 12,133 tk and 8,194 tk in less and high saline areas, respectively. The major contributor to a household’s income was casual labour supply. Through casual labour supply, on an average, a household earned 41,738 tk with almost equal earnings in less and high saline areas. In high saline areas of Khulna and Satkhira, households earned more from casual labour supplying than the households in less saline areas, the scenario is opposite and significantly different in Patuakhali. Through entrepreneurship activities, a household earned 12,664 tk (for the entire sample). A household in high saline areas earned more than less saline area households through enterprise uptake. Even though among the sampled households, 27.6% were involved in crop farming, earnings from this sector contributed hardly anything towards total household income. Moreover, net income from the sector was negative in Khulna and in the high saline areas of Patuakhali. Here it is noteworthy to mention that net income from crop farming is estimated using cash cost only. Doing farming even in presence of negative incomes means subsistence farming in underdeveloped rural settings where a farmer does not have other alternatives. Even though crop income is negative, it ensures food security for the household. Through crop farming, a farmer can ensure supply of food at a cheaper than the market price. Perhaps for some of the farmers it is the only livelihood avenue that they can engage in.

87


The estimated income Gini shows less than moderate level of income inequality, though inequality increases when per capita income is considered. The average total income Gini for the households is 0.374, with slightly higher value for less saline area households. Khulna district has the highest inequality estimated at almost 40%, followed by Patuakhali (0.368) and Satkhira (0.336). Except Satkhira, in the other two districts, income inequality is higher in high saline areas compared to the less saline areas. Table 5.2.3 presents annual income distribution of the households across the five quintiles. The highest income quintile earned around 11 times higher than the lowest income quintile. The incomes for the top and bottom income quintiles are 208,080 tk and 20,074 tk respectively. Compared to the fourth income quintile, the top income quintile earned more than double. Entrepreneurship activities made the difference between top and other quintiles. From enterprises, the households in the top quintile earned 41,595 tk, which is around 3.6 times and 22.5 times higher than the income of the fourth and first quintile from the same livelihood category. From fisheries, fish collection and other sources, the households in top quintile earned around 3.5 times higher than their preceding quintile. From casual labour supplying and social safety net programs, the households in Q5 earned less than the Q4 quintile. Table 5.2.3.Annual Income Distribution across Income Quintiles Income Quintiles

Income Sources Q1

Q2

Q3

Q4

Q5

Crop Farming

-2664

522

-36

1029

2735

Livestock & Poultry

2949

2023

3932

4401

6955

Fish Farming

-611

5539

5602

14274

49396

Fish Collection

275

740

2347

3673

12797

Salary & Pension

1563

1024

5427

11063

31717

Casual Labour

14474

41192

48795

52601

50413

Enterprise

1848

2328

7288

11484

41595

SSNP

620

709

565

864

591

Other

1621

1618

1911

3367

11881

Total Income

20074

55696

75832

102755

208080

Section 5.2.3. Income through Casual Labour Supplying Table 5.2.3.1 depicts the proportion of casual labour supplying households earning through migration. Nearly one out of every five households (18.79%) migrated in order to sell labour. 88


Almost similar proportion of households from less and high saline areas migrated. Compared to Patuakhali, migration is more common in Khulna and Satkhira. From Patuakhali, only 8.04% of the casual labour supplying households migrated, whereas 26.02% and 23.21% from Khulna and Satkhira migrated, respectively. No certain pattern in migration between high and less saline areas across districts is observed. For instance, in Khulna, higher proportion of households from less saline areas migrated, whereas in Patuakhali and Satkhira, more migration is observed from less saline areas. Refer to Appendix B3 for details based upon proximity criteria. Table 5.2.3.1.Proportion of Casual Labour Supplying Households Earning Through Migration Salinity Level District

Low

High

All

Khulna

21.28%

30.39%

26.02%

Patuakhali

9.47%

6.98%

8.04%

Satkhira

25.86%

20.37%

23.21%

All

19.34%

18.29%

18.79%

On average, a casual labour supplying household earned 61,482 tk through labour (see Table 5.2.3.2). In less and high saline areas, casual labour supplying households earned 64,133 tk and 59,097 tk, respectively. In Khulna and Satkhira, high saline area households earned more than the less saline area households. Contrary to this, less saline area households in Patuakhali earned significantly higher than the high saline area households. Households earned a lion’s share of their casual labour earnings from working in their own village. From within their own village, a household annually earned 54,527 tk, which is around 89% of their total casual labour supply based earnings. Migration contributed the remaining portion which is only 6,955 tk annually. Among the three districts, households in Khulna had the highest contribution from migration income, which is around 16% of total earnings through casual labour supplying. The relatively low Gini coefficient values for earnings from own village compared to migration earnings means that income from own village is more equally distributed than migration income across districts. Income is also more equally distributed in less saline areas than high saline areas, but in case of migration income the scenario is different.

89

ACIAR Report ADP/2015/032 Table 5.2.3.2. Annual Income Earned by the Causal Labour Supplying Households Salinity Level District

Low

High

All

Khulna

53234

56811

55095

Patuakhali

77057

49012

60906***

Satkhira

41604

54148

47652***

All

56231

52995

54527

Gini Coefficient

0.294

0.340

0.321

Khulna

7383

12645

10121

Patuakhali

8411

752

4000**

Satkhira

7906

6315

7139

All

7902

6103

6955

Gini Coefficient

0.491

0.404

0.453

Khulna

60617

69456

65217

Patuakhali

85467

49764

64906***

Satkhira

49510

60463

54791***

All

64133

59097

61482*

Gini Coefficient

0.280

0.310

0.298

Income Earned from Own Village

Income Earned through Migration

Total Income from Casual Labour Supplying

Note: ***, ** and * indicate mean differences in earnings between less and high saline areas are significant at 1%, 5% and 10% level of significance, respectively.

To further investigate the importance of casual labour supplying and inequality in earnings from the different sources, earnings by different quintiles and the corresponding Gini coefficients were estimated. Compared to the lowest income quintile (Q1), the highest income quintile (Q5) earned around 3.5 times higher through causal labour supplying. This may appear quite surprising, as one would assume higher income quintiles to rely less on casual labour supplying. However, the finding here is typical of rural settings where scope for alternative livelihoods is highly limited. In the case of labour migration also, the top two quintiles had more than 4 times higher earnings than the bottom quintile (Table 5.2.3.3). Table 5.2.3.3. Earnings from Causal Labour supplying by Different Income Quintiles Own Village Income Quintile

Migration

Total

Income (tk)

Gini

Income (tk)

Gini

Income (tk)

Gini

Q1

24005

0.374

2932

0.311

26937

0.317

Q2

44596

0.177

4673

0.342

49269

0.143

Q3

58650

0.351

4227

0.258

62877

0.135

90


Q4

64338

0.250

12172

0.307

76510

0.339

Q5

83025***

0.172

11870***

0.560

94894***

0.193

Note: ***, ** and * indicate mean differences in incomes across different income quintiles are significant at 1%, 5% and 10% level of significance, respectively.

Section 5.2.4. Income from Fish Farming Farmers simultaneously cultivate different types of fish in the same pond. They decide over one major type of fish and choose others which can be grown with the major type. As different types of fish utilize different layers of the pond, they do not compete for space, oxygen and food. This is also a part of a farmer’s risk management strategy. Mainly, three types of fishing practices were observed among the sampled farmers, namely: crab dominated, shrimp dominated and fresh water fisheries. The first two types require saline water. The three districts’ average statistics shows that among the fish growers, 64.04% and 28.84% were involved with shrimp and crab dominated fish farming. Around 32% cultivated different types of fresh water fish. Shrimp and crab dominated farming is more popular in Khulna and Satkhira, particularly in the high saline areas of these two districts. Around 80% and 74% of the fish growers in Khulna and Satkhira were engaged in shrimp dominated fishing practices. In the high saline areas of Khulna and Satkhira, there were hardly any farmers who did not do shrimp farming. But in Patuakhali, only 2.08% of the fish growers did shrimp farming. Crab farming, which is relatively new, was practiced by around 38% and 30% of the fish farmers in Khulna and Satkhira, respectively. Among the Patuakhali fish growers, 97.92% were involved in fresh water fisheries, whereas in Khulna and Satkhira 13.70% and 26.10% did fresh water fisheries, respectively (Table 5.2.4.1). Table 5.2.4.1. Proportion of Fish Farmers Involved with Different types of Fishing Practices Different Types of Fisheries Practices District

Khulna

Patuakhali

Satkhira

Salinity Level

Shrimp dominated

Crab dominated

Fresh water fisheries

Less

62.34%

32.47%

24.68%

High

98.55%

43.48%

1.45%

All

79.45%

37.67%

13.70%

Less

0.00%

0.00%

100.00%

High

10.00%

0.00%

90.00%

All

2.08%

0.00%

97.92%

Less

44.83%

6.90%

55.17%

High

93.18%

45.45%

6.82%

91


All

All

73.97%

30.14%

26.03%

Less

42.36%

18.75%

50.69%

High

89.43%

40.65%

10.57%

All

64.04%

28.84%

32.21%

On average, a fish farmer did farming on 102.72 decimal of land. In high and low saline areas, farmers did fish farming on 146.84 decimal and 65.03 decimal of land, respectively. Compared to Patuakhali, farmers of Khulna and Satkhira had more land for their fish farms. In high saline areas of Khulna and Satkhira, farmers did fish farming on significantly higher amounts of land than their counterparts who were in less saline areas. On the other hand, Patuakhali farmers in low and high saline areas did fish farming on almost similar sizes of lands, which were estimated to be 14.38 and 14.90 decimals, respectively. In high and low saline areas of Khulna, a farmer did fish farming on 162.25 and 85.18 decimals of land, respectively. In Satkhira, high and low saline area farmers did fish farming on 152.66 and 77.88 decimals of land, respectively. Fish farmers in these two districts devoted majority of their lands for shrimp farming. Khulna farmers did shrimp farming on 125.53 decimal of land, whereas in Satkhira farmers grew shrimp on 131.93 decimal of land. In Khulna, compared to less saline areas, high saline area farmers did shrimp farming on significantly higher amounts of lands, whereas in Satkhira, the difference was not notable. On average, a fish farmer in the entire study area did shrimp, crab and fresh water fisheries on 125.65, 49.85 and 24.42 decimal of land, respectively (Table 5.2.4.2). Table 5.2.4.2. Area (decimal) under Different types of Fishing Practices District

Khulna

Patuakhali

Satkhira

All

Different Types of Fishery Practices

Salinity Level Shrimp dominated

Crab dominated

Fresh water fisheries

All

Less

95.90

47.04

41.05

85.18

High

143.04

47.83

33.00

162.25

All

123.53**

47.47

40.65

121.60***

Less

0.00

0.00

14.38

14.38

High

32.00

0.00

13.00

14.90

All

32.00

0.00

14.12

14.49

Less

121.42

85.00

31.88

77.88

High

135.27

52.85

38.00

152.66

All

131.93

55.77

32.84

122.95***

Less

101.33

49.86

15.16

65.03

139.14

49.84

20.31

146.84

125.65**

49.85

24.42

102.72***

High All

92

ACIAR Report ADP/2015/032 Note: ***, ** and * indicate mean differences in area between less and high saline areas are significant at 1%, 5% and 10% level of significance, respectively.

Table 5.2.4.3 presents net incomes from different types of fish farming. On an average, a fish farmer earned annually 53,132 tk from fish farming. The fish farmers in high saline areas earned 78,313 tk annually, which is significantly higher than 31,622 tk earned by the less saline area farmers. The farmers in Satkhira earned most (66,500 tk) from fish farming, followed by the farmers of Khulna (60,225 tk). Compared to these two districts, earnings for Patuakhali farmers was very low. Shrimp farming is the major contributor to income from fishing. A farm doing shrimp dominated fishing practices earned annually 66,268 tk, whereas the crab and fresh water fish farms earned 16,556 tk and 12,896 tk, respectively. Except Patuakhali, in the other two districts, high saline area farmers earn more compared to less saline area farmers. Table 5.2.4.3. Net Annual Income Earned by the Fish Growers from Different Types of Fish Farming District

Khulna

Patuakhali

Satkhira

All

Income from Different Types of Fishery Practices

Salinity Level Shrimp Dominated

Crab Dominated

Fresh Water Fisheries

All

Less

42741

25349

8456

38268

High

73692

21353

14150

84728

All

60839**

23090

8727

60225***

Less

Na

Na

11541

11895

High

24980

Na

6866

8678

All

83653.75

Na

10664

11225

Less

67110

-1013

17845

39824

High

83654

1863

51385

84081

All

79743

1544

23140**

66500***

Less

47202

22713

12063

31622

High

76993

13425

17700

78313

All

66268*** 16556* 12896 53132*** Note: ***, ** and * indicate mean difference in earnings between less and high saline areas are significant at 1%, 5% and 10% level of significance, respectively.

Undoubtedly, saline water fish farming positively contributes to a household’s incomes. However it has also led to some ecological controversies related to adverse effects of shrimp farming. Additionally, it has also been claimed that saline water fisheries are dominated by the large local farmers and other elites who appropriate most of the beneficial opportunities for farming. To investigate this, the net income by different farm categories and related Gini coefficients are estimated and presented in Tables5.2.4.4 and 5.2.4.5, respectively. While moving from low income quintile to upper quintiles, one can observe an increase in net income from shrimp up to the small farmer’s category. The medium & large farmers earned slightly less than the small farmers. Both 93


the categories of small and medium & large farmers earned more than double from shrimp farming than the landless farmers. In case of crab dominated farming, farms in higher categories earned more than the lower categories, and medium & large farms earned almost 2.5 times higher than the landless farmer’s groups. The total income from fisheries was highest for the small farms (92,792 tk) followed by medium & large farms (80,360 tk). The landless farmers earned the lowest among the four categories. Table 5.2.4.4. Annual Income from Fisheries by Different Categories of Farmers Different Types of Fisheries Practices Farm Category

Shrimp Dominated

Crab Dominated


All

Landless

43413

10197

11209

31807

Marginal

64520

17007

13492

52361

Small

93557

19509

7830

92792

92434***

25173

13036

80360

66268

16556

12896

53132

Medium & Large All

Inequality in earnings from fisheries by different farm categories is presented in Table 5.2.4.5. Moderate level of inequality existed (0.543) for earnings from fisheries. Among different categories, the landless group exhibited highest inequality estimated at 0.596. Inequality decreases as one moves along higher land categories. But this pattern differs across different production practices. Table 5.2.4.5. Income Gini from Fisheries by Different Farm Categories Different Types of Fishery Practices Farm Category

Shrimp Dominated

Crab Dominated


All

Landless

0.509

0.593

0.408

0.596

Marginal

0.399

0.609

0.464

0.521

Small

0.255

0.510

0.437

0.405

Medium & Large

0.349

0.412

0.627

0.397

All

0.422

0.570

0.482

0.543

94


Chapter 6: Sample Characteristics of Enterprise Take-up by Farming Households This chapter focuses specifically on the enterprise undertaking households across various demographic categories as well as across various levels of climatic stress. Details relating to the motivations for starting enterprise are also discussed.

95


Section 6.1: Sample Characteristics of Enterprising Households (India) Among the various livelihood activities available to farming households, our study focuses particularly on enterprise take-up. We find that around 16% households in the entire sample are involved in enterprise based activities to supplement their incomes. Irrespective of whether the households get into these activities due to push or pull factors, there will be certain key determinants of enterprise take-up and some key determinants of enterprise success. Our study focuses on these two aspects. Before we look into determinants of enterprise take-up and its impact on households’ livelihoods, we first examine the differences in characteristics of the households that take-up enterprise and those which do not. From table 5.1, we observed that while only 6% of households in Islampur take-up enterprise based activities, 40% of households in Nalhati-2 are into entrepreneurial activities. Almost half of the enterprise based households in Nalhati-2 have taken up sericulture. The list of other activities in the sample is presented in Appendix A. Table 6.1.1 compares the characteristics of households that take-up enterprises and those which do not.

96

ACIAR Report ADP/2015/032 Table 6.1.1. Comparing characteristics of households with and without enterprises (overall sample) Sample Characteristic

Households undertaking Enterprises

Households with no Enterprises

Difference

Caste SC

30%

23%

7%**

ST

2%

7%

-5%***

OBC

24%

37%

-13%***

Others

42%

32%

10%***

86%

88%

-2%

Hindu Education of Household Head Illiterate

7%

16%

-9%***

Informally Literate

4%

4%

0%

Primary

20%

21%

-1%

Secondary

49%

43%

5%

Higher Secondary

12%

9%

3%

Degree/Diploma

9%

6%

2%

Marginal

74%

75%

-1%

Small

20%

19%

1%

Medium

5%

6%

0%

Land Holding

1.9

2.1

-0.1

Number of Members in the Household

Average Land Holding

5.46

5.98

-0.52***

Dependency Ratio(Working Age)

0.27

0.24

0.04**

Dependency Ratio (Working Members)

0.53

0.54

-0.01

Cropping Intensity

1.39

1.41

-0.02

Water Buyers

70%

71%

-1%

Water Sellers

4%

6%

-2%*

Agricultural Assets

15098

22969

-7872

Non-agricultural Assets

62580

29769

32811***

Livestock

17769

23412

-5643***

Cultivation Income Household Income Note: *,**,*** indicates statistical significance at 10%, 5% and 1% confidence level.

97

15411

19952

-4541

115616

102132

13484*


From table 6.1.1, we observe that among those households that have taken up enterprises, there is a large proportion that belongs to Scheduled Castes (SCs) and forward castes. The percentage of Hindus among households that take-up enterprises and those that do not is the same. The proportion of illiterates among households that take-up enterprises is much less than those households that do not take up enterprise, and this difference is statistically significant at 1% confidence level. The proportion of marginal, small and medium landholding households is similar irrespective of whether households take-up enterprise or not. Households which take-up enterprise have lesser household members (5.46) compared to those who do not take-up an enterprise (5.98). The dependency ratio based on working age criteria is higher for households that take-up enterprises than those that do not by 0.03. The dependency ratio based on working members is similar for households that take-up enterprise and those that do not. The cropping intensity is also similar for both categories of households. Households that take-up enterprises have fewer agricultural assets (in terms of value) than those that do not, but this difference (INR 7,000) is not statistically significant at 10% level. The households that take-up enterprises have non-agricultural assets which are much higher in value than those that do not takeup enterprises. While the average value of nonagricultural assets for households involved in entrepreneurial activities was INR 62,580, the same for those who did not own any enterprise was only INR 29,769. Households that did not take-up enterprises earned slightly higher (by approximately INR 4000) than those that undertook enterprise based activities, but the difference was not statistically significant at 10% confidence level. However, the average household income of households that took up enterprises was INR 13,500 higher than that of households that did not take-up enterprises. We also posed a few questions only to those households who had taken up enterprise based activities. Table 6.1.2 summarizes some of their relevant responses.

98

ACIAR Report ADP/2015/032 Table 6.1.2 Responses by Households with Enterprise Nalanda Islampur Combined

Silao Reasons for starting business High Profitability Low Capital Requirement Low Risks Low Operating Costs Low Labour Requirements High Demand Influence of friends/relatives Influence of Water Scarcity on Starting Business Large Small None Initial Investment (in INR) Location of Business At home Within Same Village Highway Other Village/City Strategies to Increase Business Income Undertake additional enterprise Increase scale of business Increase scope of business Looking for newer markets Important Problems while operating Business Insufficient Capital Unavailability of Raw Materials Marketing Labour Shortage Labor Costs Expanding Scope of Business Competing with large producers High Interest Rate of Credit N

Rajnagar

Birbhum Nalhati-2

Combined

Purulia Raghunathpur-II

Kashipur

Combined

All

68% 37% 63% 37% 63% 42% 16%

71% 41% 59% 59% 76% 35% 12%

69% 39% 61% 47% 69% 39% 14%

63% 26% 37% 39% 24% 11% 8%

60% 56% 56% 65% 35% 13% 30%

61% 48% 51% 58% 32% 13% 24%

46% 54% 21% 68% 32% 11% 7%

65% 29% 47% 63% 47% 22% 24%

58% 38% 38% 65% 42% 18% 17%

61% 43% 48% 59% 40% 18% 20%

53% 42% 5% 108421

6% 69% 25% 66729

31% 54% 14% 88733

24% 26% 50% 32521

15% 63% 21% 24879

18% 53% 29% 27030

18% 43% 39% 15936

20% 43% 37% 10953

19% 43% 38% 12765

20% 50% 30% 31558

32% 53% 16% 0%

24% 35% 24% 18%

28% 44% 19% 8%

32% 47% 11% 11%

59% 27% 4% 10%

51% 32% 6% 10%

50% 29% 7% 14%

73% 8% 0% 18%

65% 16% 3% 17%

52% 29% 7% 12%

26% 63% 16% 11%

12% 76% 35% 12%

19% 69% 25% 11%

11% 63% 26% 13%

21% 78% 57% 17%

18% 74% 49% 16%

14% 79% 54% 4%

4% 92% 47% 2%

8% 86% 49% 3%

15% 77% 45% 11%

3.8 2.1 3.2 4.6 2.5 3.5 2.5 3.7 19

3.6 1.6 3.0 2.1 2.2 2.8 2.2 3.2 17

4.4 4.0 4.2 3.8 3.9 4.1 3.9 4.3 36

4.3 4.0 4.0 3.8 3.7 3.9 3.7 3.9 38

4.4 4.0 4.3 3.7 4.1 4.2 3.9 4.4 98

3.7 1.8 3.1 3.4 2.4 3.1 2.4 3.5 136

3.9 3.7 3.7 3.1 3.4 3.9 3.8 3.4 28

3.7 3.6 3.9 3.2 3.3 3.7 3.6 3.9 49

3.8 3.7 3.8 3.2 3.3 3.8 3.7 3.7 77

4.1 3.6 3.9 3.5 3.5 3.9 3.6 4.0 249

99


From table 6.1.2, we can observe that the most important reason for starting an enterprise for the households was its high profitability. 61% of the households with enterprises started the business because they felt it was highly profitable. 59% households started the business because of low operating costs associated with setting up the enterprise. Low risk was one of the reasons for starting a business for 48% households. Low capital requirements and low labor requirements were one of the reasons for starting the business for 43% and 40% households. Influence of friends and relatives was a reason for starting a business for only 20% households, and similarly high demand was also a factor for only 18% households. In Nalanda, more households (69%) were influenced by high profitability and low labour requirements, while very few households were influenced by friends/relatives (14%). In Birbhum, high profitability (61%) and low operating costs (58%) were the reasons for starting business for most of the households, while very few households were influenced by high demand (13%). In Purulia, low operating costs was the reason for starting business for 65% households closely followed by high profitability, as a reason for 58% households. When asked about the influence of water scarcity in taking up enterprise, 20% responded saying it influenced the decision to a large extent while 50% responded saying it had an influence to some extent. Only 30% said they were not influenced by water scarcity, while making the decision to take-up an enterprise. The average initial investment in the enterprise was INR 31,558. It was highest for Silao block in Nalanda (INR 1,08,421), while it was the lowest for Raghunathpur-I block in Purulia (INR 10,953). Most of the businesses were located at home (52%). In Raghunathpur-I and Nalhati-2, very high percentage of households (73% and 59%) were located at home. In Raghunathpur-I, vegetable vending was dominant, and the enterprise in this case was located at home. Similarly, in Nalhati-2, a large number of households were involved in sericulture, most of which was located at home. 29% of enterprises were located in the same village, 7% located on a highway and 12% located in other village or city. On being asked about strategies to increase business incomes, most enterprises (77%) felt increasing scale of business was the way ahead. Fewer households (45%) intended to increase scope of business. Very few households that took up enterprise wanted to undertake additional enterprises (15%) or look for newer markets (11%). On problems faced by enterprises, insufficient 100


capital (average 4.1 on a scale of 5) was rated very highly by most enterprises, closely followed by high interest rate on credit (4 on a scale of 5). Labor shortage and labor costs were considered relatively lesser problems by households (3.5 on a scale of 5). We also surveyed a few enterprises in the towns near the surveyed villages to understand how profitability of town enterprises compared with village enterprises and whether the determinants of enterprise uptake in towns varied significantly from their counterparts in the villages. Table 6.1.3 presents the summary of enterprises in the towns.

101

ACIAR Report ADP/2015/032 Table 6.1.3. Characteristics and Responses of Town Enterprises Nalanda Monthly Sales (INR) Monthly Profit (INR) Initial Investment (INR) Registered Businesses (%) Location of business (%) Highway Market Total Value of Business Assets (INR) % of Firms that believe they bring innovation in operation, marketing and sales % of Firms who believe they have taken risks in the past

Birbhum

All

Retail

Mfrg

All

Service

Retail

Mfrg

All

Service

Retail

Mfrg

All

45614

107676

68607

85017

61609

84620

72643

75580

134840

66238

152500

93740

84794

17486

13880

15057

15069

15313

13717

14857

14387

26740

12881

15333

16640

15347

108229

128880

88286

118121

54129

71815

50429

63161

78854

43158

70458

55358

81333

5.7

22.5

21.4

17.5

59.4

27.8

64.3

43

44.0

17.5

66.7

30.0

29.4

14.3 85.7 363529

9.9 83.1 132303

7.1 85.7 127071

10.8 84.2 199133

15.6 78.1 557100

13 75.9 394018

21.4 64.3 442857

15 75 453042

44.0 48.0 381240

50.8 41.3 156416

25.0 58.3 294583

46.0 45.0 229202

23.1 69.1 287876

74.3

73.2

75.0

73.8

50

38.9

85.7

49

48.0

61.9

75.0

60.0

61.7

44.1

38.0

42.9

40.4

31.3

38.9

28.6

35

32.0

22.2

16.7

24.0

33.6

Service

Retail

Mfrg

All

Service

Retail

Mfrg

All

Service

Retail

Mfrg

876074

1222332

704107

1060881

744956

1004813

668857

874625

551484

297494

432658

377211

789029

32.35

12.31

14.29

18.39

6.67

3.7

7.14

5.13

25.00

7.94

100.00

23.25

15.76

85.29

90.14

71.43

86.54

84.38

94.44

100

92

92.00

100.00

91.67

97.00

91.52

29.41

31.43

28.57

30.51

15.63

16.67

35.71

19

16.00

9.52

0.00

10.00

20.50

77.14

92.96

85.71

87.50

96.88

100

100

99

96.00

93.65

100.00

95.00

93.44

37.50 35.00

49.30 71.00

57.14 14.00

46.77 120.00

21.88 32

22.22 54

28.57 14

23 100

32.00 25.00

14.29 63.00

50.00 12.00

23.00 100.00

31.92 320.00

Nalanda

Total Value of Personal Assets (INR) % of Firms that use internet for business % of Firms that use mobile for business % of Firms with contacts in NGO or Government % of Firms with Bank Account % of Firms with Mobile Banking N

Purulia

Service

Birbhum

102

Purulia

All


Among all the firms, firms in Purulia had highest sales and profits while those in Birbhum had the lowest sales and profits. The average monthly sales for the entire sample of 320 firms was INR 84,974 and the net profit was INR 15,347. The initial investment amount was the least for Purulia and the highest for Nalanda. Around 30% of the firms surveyed were registered. 43% of firms in Birbhum were registered, while only 17% firms in Nalanda were registered. A high number of enterprises (46%) in Purulia were located on the highway compared to Nalanda (11%) and Birbhum (15%). This could be one of the reasons for the high profitability of enterprises in Purulia. The value of business assets was highest in Birbhum, while it was the lowest in Nalanda. A high proportion of enterprises in Nalanda (74%) felt they adopted innovations, while a very low proportion in Birbhum (49%) felt similarly. While 40% of the enterprises in Nalanda believed they had taken risks in business, only 24% of enterprises in Purulia felt the same way. Total value of personal assets was highest in Nalanda (INR 10,60,881) and lowest in Purulia (INR 3,77,211). While 23% of enterprises in Purulia used internet for business purposes, only 5% in Birbhum used internet. All the 12 manufacturing firms in Purulia used internet for business purposes. While 31% enterprises in Nalanda had a contact with an NGO or Government official, it was only 10% in case of Purulia. 88% of enterprises in Nalanda, 99% in Birbhum and 95% in Purulia had a bank account. 47% enterprises in Nalanda used mobile banking, while only 23% of enterprises in Birbhum and Purulia used mobile banking. Though there were many households that took up enterprises in Birbhum, the enterprises in the towns of Birbhum seemed to lag behind in many aspects. This could be due to lack of connectivity to cities. Similarly, since towns in Purulia were well connected to cities nearby, they performed well on many indicators. Nalanda, which is very near to Bihar’s capital seems to have lagged behind because of institutional factors. It is strange to find that only 17% of the firms in towns of Nalanda were registered given the proximity of these towns to Patna. The basic analysis so far seems to indicate that different factors could influence profitability of town enterprises as compared to those in the villages. Also, nearness to city could have better returns for enterprises in towns compared to those in villages. But, it can be said with more confidence that having enterprises located on highways seems to increase profits for all enterprises irrespective of their type.

103


Section 6.2. Enterprise Uptake in Bangladesh About 14% and 15% of the households in the less and high saline areas were involved with enterprises. The types of enterprise undertaken by the sampled households is neither new nor innovative, rather it is of traditional kind. Grocery shops were the most common form of enterprise. About one out of every three entrepreneur households had a grocery shop. The next common form was fish and fisheries related business. 26% of the entrepreneur households were doing business related to fish and fishery based products. They were selling fish, fingerlings, dried fish, etc. Tea stall as an enterprise was undertaken by 16.67% of the households. Other types of businesses, including fuel, medicine, computer shops, wooden furniture, selling bakery products, etc. were undertaken by 15.22% of the households. Proportion of entrepreneurs with different types of enterprises did not vary much across high and low saline areas. Table 6.2.1. Types of Enterprises undertaken by High and Low Saline Area Households (% of Entrepreneur Households) Types of Enterprise Salinity Level Tea Stalls

Fish & Fisheries

Crop and Livestock Products

Grocery Shop

Others

Low

17.91%

23.88%

5.97%

35.82%

16.42%

High

15.49%

28.17%

8.45%

33.80%

14.08%

All

16.67%

26.09%

7.25%

34.78%

15.22%

Some comparison of socio-economic and demographic characteristics between the entrepreneur and non-entrepreneur households is presented in Table 6.2.2. The average year of schooling for the heads of the entrepreneur households was significantly higher than the non-entrepreneur households. Relatively lower proportion of entrepreneur household heads were illiterate. Though, almost equal proportion of the heads from both categories had studied up to the primary level, relatively higher proportion of entrepreneur household heads had completed primary and further education. Much difference does not exist in terms of head’s age and number of family members between two groups. Dependency ratio, calculated, based on age and working members, reveals that entrepreneur households had relatively fewer dependent members. Compared to the entrepreneurs, non-entrepreneurs had more land and they practiced higher level of crop diversification. The entrepreneur households had significantly higher value of assets. An entrepreneur household possessed total assets worth of 29,005 tk, whereas for non-entrepreneur households, this value was only 19,605 tk. The difference is due to a significantly higher value of 104


non-agricultural assets owned by the entrepreneur households. The non-entrepreneurs owned a higher total value of livestock and poultry at 20, 858 tk, as compared to the entrepreneurs (16,181) tk. Table 6.2.2. Comparing characteristics of households with and without enterprises Sample Characteristics Socio-economic & Demographic Characteristics Education of Household Head¥

Non-entrepreneurs

Entrepreneurs

Illiterate Primary Secondary Higher Secondary & Above Head’s Average Year of Schooling Age of the HH Number of Members in the Household Dependency Ratio (Working Age) Dependency Ratio (Working Members) Average Land Holding (decimal)

26.57 36.16 31.73 5.54 4.55 44.76 4.48 0.304 0.686 51.76

18.37 34.69 36.05 10.88 5.53*** 44.79 4.41 0.277 0.656** 44.32

Value of Agricultural Assets (tk) Value of Non-agricultural Assets (tk) Value of Total Asset (tk) Value of Livestock & Poultry (tk)

9111 10495 19605 20858

9254 19751*** 29005*** 16181

Cropping Diversification Index Annual Net Income from Crop Farming (tk)

0.809 1316

0.863 1159

Assets

Cropping

Annual Income Salary & Pension (tk) 11295 4004** Net Income from Fisheries (tk) 14044 18802 Casual Labour (tk) 45463 18281*** Income from Migration (tk) 5300 1210** Total Household Income (tk) 83175 141024*** Note: ¥The conducted λ2 test showed that difference across different educational categories between entrepreneur and non-entrepreneur households was significant at 5% confidence level. *,**,*** indicate differences between entrepreneur and non-entrepreneur households to be significant at 10%, 5% and 1% confidence level, respective.

With respect to income, the non-entrepreneur households earned higher incomes from crop farming and salary & pension, and less from fisheries. Non-entrepreneur households’ income from salary & pension was almost 3 times higher than the entrepreneur households. Though the income from crop is not notable in the total income, the non-entrepreneur households earned marginally higher from crops than the entrepreneurs. From casual labour, the non-entrepreneur households earned significantly higher than the entrepreneur households. Annually, entrepreneur and nonentrepreneur households earned 18,281 tk and 45,463 tk from casual labour supplying. Income through migration is also significantly higher for the non-entrepreneurs. However, when it comes to total income, the entrepreneur households earned significantly higher than the nonentrepreneurs. Whereas an entrepreneur household earned 141,028 tk annually, the nonentrepreneur household earned only 83,175 tk annually (Table 6.2.2).

105


The initial investment made by the entrepreneurs to their businesses was 55,544 tk. Initial investment was highest for the Patuakhali entrepreneurs (69,000 tk) and lowest for the Khulna entrepreneurs (45,730 tk). In Patuakhali and Satkhira, the less saline area entrepreneurs made more initial investments than their counterparts in high saline areas, whereas in Khulna, the entrepreneurs in less saline area made significantly lower investments. Entrepreneurs managed around 50% of their initial investment through credit.

Annually, an entrepreneur earned 83,140 tk from their enterprise. Combined data for the three districts shows that, earnings from enterprises in low and high saline areas do not vary much. In Khulna, high saline area entrepreneurs earned around 1.5 times higher than their less saline area counterparts. But in Satkhira and Patuakhali, the less saline area entrepreneurs earned more than high saline areas. In none of the districts, profit difference between high and less saline areas was statistically significant at 10% confidence level (Table 6.2.3).

106

ACIAR Report ADP/2015/032 Table 6.2.3. Basic Business Profiles and Cost-Return from Business by Region Khulna

Patuakhali

Satkhira

All Areas

Salinity Level Less

High

All

Less

High

All

Less

High

All

Less

High

All

Age of Enterprise (years)

10.19

10.81

10.47

11.63

9.56

10.37

8.74

10.63

9.67

9.91

10.37

10.14

Initial Investment (tk)

30982

64500

45730*

90167

55635

69000

68442

46517

56882

56500

54643

55544

Loan as Proportion of Initial Investment

46.00

58.06

52.58

43.15

51.81

48.34

63.22

47.78

55.50

49.03

53.40

51.47

Annual Net Return (tk)

64074

93125

77745

120325

99280

107493

73513

67046

70450

80282

85885

83140

107


Among different income quintiles, the top income quintile made the highest level of initial investment (Table 6.2.4). The entrepreneurs belonging to this quintile invested 72,536 tk (around 11 years back), which is about 2.7 times higher than that of the Q1 quintile. Except for the Q1 quintile, other quintiles managed around 50% of their initial investments through loan. For the Q1 quintile, loan contributed around 44% of their initial investments. The Q5 quintile earned 114,797 tk which is about 4.5 times the Q1 quintile incomes. While moving from lower income to upper income quantiles, one can observe a significant increase in net returns from business. Inequality (based on Gini coefficient) was the highest in top quantile (0.335) and the lowest in the third (0.17) income quintile. Table 6.2.4: Business Profits and Investment Deciphered through Income Quintiles Age of Enterprise

Initial Investment (tk)

Loan as % of Initial Investment

Annual Net Return (tk)

Gini of Income

Q1

9.88

27088

44.00

25108

0.294

Q2

10.58

24350

50.83

38273

0.279

Q3

10.55

55591

50.82

63602

0.170

Q4

8.62

50484

51.00

66818

0.281

Q5

10.78

72536

54.08

114797***

0.335

Income Quintile

108


Chapter 7: Regression Analysis In this section we perform econometric analysis to test key research hypotheses with respect to the study regions in India and Bangladesh. In particular, the factors influencing probabilities of farming households taking up key livelihood options (enterprise, rural labor, urban migration and salaried employment) other than farming are determined. The contributions of these livelihood profiles to total household incomes are discussed. The issue of selection bias is addressed and sensitivity analysis is performed with respect to fixed effects as well as the choice of instrumental variables. This section also presents the determinants of incomes for enterprises that are located in towns, and which are not necessarily owned by the surveyed farming households. This helps provide an additional dimension of livelihood opportunities or constraints present for farming households that are actively seeking alternate avenues to augment their household incomes.

109


Section 7: Regression Analysis: India Section 7.1. Explaining factors that influence enterprise uptake by farming households Using a probit model to explain the probability of enterprise uptake (refer to the marginal effects table 7.1.1 that follows the main regression table 7.1), it is found that the cropping intensity is positively related to enterprise uptake probability (increases uptake by about 8 percent, keeping other things constant), whereas a marginal increase in water scarcity in the Kharif season has a negative and significant impact (of about 9 percent) on enterprise uptake probability. Also, household crop income is negatively associated with enterprise uptake. That is, those households that face significant water scarcity and have low crop incomes are less likely to start an enterprise. The intuition behind this result could be that frequent disruptions in income due to adverse impacts of water scarcity on crop productivity leave little resources for these households to accumulate necessary capital (financial, human or other) for enterprise uptake. Households that stated a higher ‘minimum drought resilience’ (stated by the respondent as the number of months that the household could survive in a drought-like situation with only their current income) were also less likely to start an enterprise. This gives an indication that for a majority of households, the decision to start an enterprise could be more of ‘push’ or distress related one. Households from the ST community and other backward caste category were relatively less likely to start an enterprise compared to scheduled caste households. Households, that had a head with secondary and higher secondary education, were also more likely to start an enterprise as compared to those with illiterate household heads (by about 5 and 7 percent respectively). Those with a higher level of education, such as diplomas or graduates, were also more likely to start an enterprise. The propensity to start an enterprise does not seem to be influenced by the size of the land holdings of the farmers, nor is it influenced by the value of their agricultural assets. However, the value of non-agricultural assets is positively and significantly associated with the probability of starting an enterprise. The marginal probability being about 2 percent, that is, a unit increase in the value of non-agricultural assets leads to an increase in 2 percent probability of starting an enterprise. It also suggests that (when accounting for endogeneity) those who take to enterprises are able to shift their focus away from accumulating farming assets and accumulate non-agricultural assets. Total value of livestock has a negative influence on the probability to start an enterprise, suggesting that those with more livestock cope 110


better with low crop incomes and water scarcity without needing to start any enterprise. It could also be that maintaining livestock may not allow them time to engage in any enterprise activity. Ability to borrow money from their social networks or being a part of the SHG is positively associated with enterprise uptake. Similarly, those households that stated their preference for enterprise as compared to migration and agriculture as livelihood avenues were also more likely to take to enterprise. Considering district level fixed effects, relative to Birbhum in West Bengal, households in Purulia (which is the most water scarce of the three districts) were less likely to start an enterprise whereas those in Nalanda in Bihar were more likely to start an enterprise. Table 7.1: Probit Model of Enterprise Uptake (Overall Data) Dependent Variable = Enterprise take-up of the household (=1 if household has taken up an enterprise) (1) (2) With Water With Scarcity Demographics Indicators

-0.149 (0.96)

-0.085 (0.53)

(3) With Assets, Loan Social Capital and Preference for Migration and Enterprise -0.032 (0.19)

Water Buyer Dummy (=1 if the household is a water buyer)

-0.288** (2.35)

-0.291** (2.31)

-0.300** (2.23)

-0.292** (2.16)

Water Seller Dummy (=1 if the household is a water seller)

-0.341* (1.77)

-0.233 (1.17)

-0.146 (0.70)

0.058 (0.27)

-0.561*** (3.52)

-0.484*** (2.96)

-0.556*** (3.29)

-0.462*** (2.71)

Water Scarcity Indicator for Rabi Season

-0.137 (0.90)

-0.139 (0.89)

-0.232 (1.41)

-0.095 (0.56)

Minimum Drought Resilience

-0.080* (1.82)

-0.063 (1.52)

-0.099** (2.41)

-0.077* (1.75)

Household Crop Income

-0.016* (1.67)

-0.019* (1.83)

-0.031** (2.22)

-0.039** (2.24)

-0.144 (1.16)

0.024 (0.18)

0.402*** (2.74)

Scheduled Tribe

-0.748*** (3.25)

-0.575** (2.40)

-0.638*** (2.66)

Other Backward Caste

-0.394*** (3.39)

-0.296** (2.43)

-0.034 (0.26)

Forward Caste

-0.128 (1.23)

-0.108 (1.00)

-0.005 (0.05)


0.019 (0.91)

0.013 (0.57)

0.006 (0.29)

Square of Age of the Household Head

-0.000 (0.79)

-0.000 (0.51)

-0.000 (0.23)

Cropping Intensity (Gross Area Cultivated/ Area Cultivated Once)

Water Scarcity Indicator for Kharif Season

Religion = Hindu

(4) District Fixed Effects

0.374** (2.00)

Caste of the household (Base = Scheduled Caste)

111

ACIAR Report ADP/2015/032 Dependent Variable = Enterprise take-up of the household (=1 if household has taken up an enterprise) (1) (2) With Water With Scarcity Demographics Indicators

(3) With Assets, Loan Social Capital and Preference for Migration and Enterprise

(4) District Fixed Effects

0.304** (2.19)

0.286** (1.97)

0.200 (1.37)

0.439*** (3.50)

0.298** (2.29)

0.232* (1.75)

0.565*** (3.41)

0.387** (2.27)

0.344* (1.93)

0.581*** (3.15)

0.362* (1.91)

0.326 (1.64)

Number of Members in the Household

-0.029 (1.53)

-0.047** (2.26)

-0.002 (0.10)

Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) Landholding Size Class (Base = Marginal (2.5 acres and 5 acres)

-0.080 (0.38)

-0.010 (0.05)

Ln (Total Value of Agricultural Assets)

-0.017 (0.85)

0.009 (0.39)

0.115*** (3.18)

0.104*** (2.85)

-0.018 (1.56)

-0.027** (2.18)

0.023** (2.51) 0.089*** (4.57)

0.016* (1.65) 0.073*** (3.44)

0.447*** (4.43)

0.305*** (2.91)

-0.048 (1.40)

-0.034 (0.98)

0.190***

0.173***

(4.22)

(3.74)

Education of the Household Head (Base = Illiterate or Informally Literate) Primary(Years of Schooling >0 &5 &10 &59/ Number of members in the household) Landholding Size Class (Base = Marginal (0 &5 &10 &2.5 acres and 5 acres)

Ln (Total Value of Agricultural Assets)

127

ACIAR Report ADP/2015/032 Dependent Variable = Urban Migration (=1 if household has urban migration income) (1) With Water Scarcity Indicators

(2) With Demographics

Ln (Total Value of Non-Agricultural Assets)

Ln (Total Value of Livestock)

Ln (Total Loan Amount)

Ln (Total Amount that the household can borrow from its Network)

SHG member dummy (=1 if any household member is a member of SHG) Migration is essential for survival (1-5)

Migration is forced and there are few other options

(3) With Assets, Loan Social Capital and Preference for Migration and Enterprise (0.27)

(4) Block Fixed Effects

(0.62)

0.005

0.006

(0.86)

(1.02)

-0.007***

-0.007**

(2.77)

(2.51)

-0.000

0.001

(0.22)

(0.36)

-0.001

0.000

(0.40)

(0.08)

-0.017

-0.008

(0.83)

(0.38)

0.051***

0.053***

(6.31)

(6.41)

0.029***

0.028***

(3.61)

(3.42)

District (Base = Birbhum) 0.094**

Nalanda

(2.22) Purulia

-0.044 (1.41)

N

1538

Absolute t statistics in parentheses* p0 &5 &10) 0.6315** (0.2617) 5.6779* (3.2945) Age of the Household Head -0.0096** (0.0046) -0.0184 (0.0579) Dependency Ratio Based on Age (Number of members in age group 00.3415 (0.3188) -4.2071 (3.8058) 14 and >59/ Number of members in the household) Total Household Member 0.0524 (0.0412) 0.7874 (0.5207) Total Value of Assets 0.000003* (0.000002) 0.00004** (0.00002) Own land (decimal) 0.0008 (0.0010) -0.0531*** (0.0131) Net Return from Crop Farming 0.00001 (0.00001) 0.00001 (0.0001) Ln (Annual Income from Causal Labour Supplying) -0.0351*** (0.0133) -1.6476*** (0.1871) Ln (Net Return from Fisheries) -0.0250 (0.0183) -0.2985 (0.1923) Annual Income from Salary & Pension -0.00001** (0.000003) -0.0002*** (0.00004) Total Value of Livestock 0.000002 (0.000003) -0.00005* (0.00003) Total Loan Amount 0.000002** (0.000001) 0.0001*** (0.00001) NGO member dummy (=1 if any household member is a member of -0.0127 (0.1321) 1.2066 (1.5661) NGO) + Migration is better than farming in terms of income (index value) -0.1185 (0.2223) -4.3852* (2.6182) Migration dummy (=1 if any of the household members undertook -0.2641 (0.2688) -1.1118 (2.7372) migration) District (Base = Patuakhali) Khulna -0.4245* (0.2348) 6.3329** (2.8956) Satkhira -0.5921** (0.2891) 6.7977** (3.4459) Dummy for distant village (=1 if the village is a distant one) -0.4686*** (0.1839) -1.5031 (2.3487) Dummy for saline village (=1 if the village has high salinity) -0.0110 (0.2194) 6.4373** (2.7454) Constant 11.3350*** (0.3917) -4.7556 (4.7065) N 960 960 Log likelihood -462.73 -799.3942 Prob> chi2 0.0000 0.0000 Test statistics of the regression models AIC 1039.46 1656.79 BIC 1316.88 1797.93 LR χ2 (27) 673.33 Prob> χ2 0.0000 Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

Though respondents in Khulna and Satkhira had a higher probability of taking up entrepreneurial activities compared to the respondents in Patuakhali, entrepreneur households in these districts earned 42% and 59% less from enterprise compared to those in Patuakhali, respectively. This is consistent with findings mentioned earlier in the Table 6.2.3. Entrepreneurs in distant villages earned around 47% less than their counterparts in nearby villages (Table 7.9.1.3). Section 7.9.2. Participation in Casual Labour Market and Explaining Variation in Income from Casual Labour Supplying Section 7.9.2.1. Participation in Casual Labour Market 158


The determinants for participation in casual labour market are presented in Table 7.9.2.1, whereas the marginal effects of the variables used in the probit models are presented in Table 7.9.2.2. Individual or household level salinity indicators, such as number of months a household purchased drinking water and faced salinity, were found to have negative associations with the probability of casual labour supplying. Among the demographic variables, the impact of educational variables is clear and consistent across models. Household head’s education was negatively associated with probability of casual labour supplying. Compared to the category of illiterate heads, the category of primary educated heads had 5% lower probability of selling casual labour. For the middle and top educational categories the estimated probabilities are 13% and 24% lower than the base category, respectively. Households with more family members were likely to participate more in the casual labour market. As the larger families have more members to feed and also more earning hands, they are more likely to earn their livelihoods through casual labour supplying. Table 7.9.2.1. Probit Model of Participation in the Casual Labour Market Dependent Variable = Participation in the labour market (=1 if household has any member participating in the casual labour market) (M1) (M2) (M3) (M4) With Salinity With With Assets, Loan, District Fixed Indicators Demographics Social Capital, and Effects Variables Preference for Farming Coefficient (SE) -.0229* -.0564*** -.0830*** -.0670*** No of months hh had to purchase drinking water (.0136) (.0148) (.0176) (.0211) -.0394** -.0260 -.0330 .0116 No of months hh faced salinity (.0199) (.0212) (.0248) (.0339) Dummy household perceived salinity increasing over the past -.2965*** -.2418** .0105 .0208 5 years(=1 if the hh observed salinity increase) (.1148) (.1215) (.1447) (.1474) Product of (Dummy for crop growers) and (Maximum -.0592*** -.0511*** -.0047 .0032 Resilience¥) (.0122) (.0125) (.0151) (.0154) (Village affected by high salinity) X (Number of months in a .0104 .0083 .0126 -.0798*** year hh faced high salinity) (.0154) (.0164) (.0191) (.0293) (Dummy for Village far from district headquarter) X (Number .0348*** .0484*** .0667*** .0115 of months in a year hh faced high salinity) (.0126) (.0135) (.0162) (.0250) Education of the Household Head (Base = Illiterate or Informally Literate) -.2317** -.1841 -.2594* Primary (Years of Schooling >0 &5 &10) (.2116) (.2626) (.2718) -.0142*** -.0072* -.0093*** Age of the Household Head (.0034) (.0041) (.0042) Dependency Ratio Based on Age (Number of members in age .2046 -.1742 -.1731 group 0-14 and >59/ Number of members in the household) (.2347) (.2752) (.2829) .0408 .1339*** .1287*** Total Family Members (.0313) (.0380) (.0392) -.0069 -.0020 Ln (Total Value of Assets) (.0327) (.0338) -.0031*** .0033*** Own land (decimal) (.0006) (.0006) .000002 .000005 Net return from crop farming (.000004) (.000004) Ln (Annual Net Return from Enterprise) -.1324*** -.1382***

159


Ln (Net Return from Fisheries) Ln (Annual Income from Salary & Pension) Total Value of Livestock Ln (Total Loan Amount) NGO member dummy (=1 if any household member is a member of NGO) Migration is better than farming in terms of income+ (index value) District (Base = Patuakhali)

(.0126) -.0694*** (.0121) -.1624*** (.0163) -.000004*** (.000002) .0082 (.0109) -.1576 (0.1132 .4928*** (.1608)

(.0130) -.0729*** (.0126) -.1633*** (.0170) .000004*** (.000002) 0.0116 (.0114) -0.2249** (.1171) 0.4425*** (.1662)

1.5507*** (.3973) 960 .000

0.3734** (.2009) 0.3903* (.2434) .4878*** (.1769) 0.8529*** (.2027) 1.003*** (.4216) 960 .000

Khulna Satkhira Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity) Constant N Prob>chi2

.7773*** (.0842) 960 .000

1.655*** (.2409) 960 .000

Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

In general, with increasing asset values, households are less likely to participate in the labor market. Total asset, own land and value of livestock had negative signs across the models. Generally, the relatively rich households own more land and livestock. As land and livestock are a means of alternative income sources for them, they tend to participate less in the labour market (Table 7.9.2.1 and 7.9.2.2). All the three income sources (enterprise, fisheries and salaried employment) had negative associations with the probability of participating in the labour market. The negative association here means that households with more income from these sources are less likely to prefer casual labour supplying compared to these sources. As NGOs offer different income generating avenues to its members, the NGO member households are found to have around 5% lower probability of becoming a daily labour. The households who preferred migration over agriculture had a higher probability of becoming a labour supplying household (Table 7.9.2.1 and 7.9.2.2).

Table 7.9.2.2. Marginal Effects of the Variables on a Household's Probability of Participating in the Labour Market Dependent Variable = Participation in the labour market (=1 if household has any member participate in the casual labour market) (M1) (M2) (M3) (M4) Variables With Salinity With With Assets, Loan, District Indicators Demographics Social Capital, and Fixed Effects

160

ACIAR Report ADP/2015/032 Preference for Farming -.0184*** -.0073

No of months hh had to purchase drinking water -.0079* -.0173*** -.0142*** No of months hh faced salinity -.0137** -.0080 .0024 Dummy household perceived salinity increasing over the past -.1031*** -.0745** .0023 .0044 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum -.0205*** -.0157*** -.0010 .00069 Resilience¥) (Village affected by high salinity) X (Number of months in a .0036 .0025 .0027 -.0170*** year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number .0121*** .0149*** .0148*** .0024 of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate) Primary (Years of Schooling >0 &5 &10) -.5447*** -.2133*** -.2387*** Age of the Household Head -.0044*** -.0016*** -.0019*** Dependency Ratio Based on Age (Number of members in age .0630 -.0386 -.0368 group 0-14 and >59/ Number of members in the household) Total Family Members .0125 .0296*** .0274*** Ln (Total Value of Assets) -.0015 -.00043 Own land (decimal) -.00069*** -.0007*** Net return from crop farming .0000003 .000001 Ln (Total Income from entrepreneurship activities) -.0293*** -.0294*** Ln (Total Income from Fisheries) -.0153*** -.0155*** Ln (Annual Income from Salary & Pension) .0360*** -.0348*** Total Value of Livestock -.000001*** -.000001*** Ln (Total Loan Amount) .0018 .0024 NGO member dummy (=1 if any household member is a -.0349 -.0479** member of NGO) Migration is better than farming in terms of income+ (index .1092*** .0943*** value) District (Base = Patuakhali) Khulna .0795** Satkhira .0831* Dummy for distant village (=1 if the village is a distant one) .1039*** Dummy for saline village (=1 if the village has high salinity) .1817*** Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

Compared to Patuakhali, in Khulna and Satkhira a household had around 80% and 83% higher probability of supplying labour. This is not surprising as compared to Patuakhali, opportunities in the other two districts are higher as these have saline water fisheries. Furthermore, Khulna is an industrial city with a seaport. Satkhira is nearby the Indian border, where the local economy is more vibrant. Hence, these two districts are more likely to have opportunities for daily labourers. Compared to the nearby villages, households in distant villages had around 10% higher probability to have a member participating in the labour market. A household in the high saline villages had around 18% higher probability to supply labour compared to those living in less saline villages (Table 7.9.2.2). Section 7.9.2.2. Explaining Variation in Earnings from Casual Labour Supplying The LR test statistics and estimated AIC and BIC recommend the Cragg model over the Tobit one. As was observed with the probit model M4, the salinity and demographic variables had little role 161


in explaining income variation. With an additional family member, a household’s income from casual labour increases by around 12%. Households with a higher dependency ratio were likely to have lower casual labour income. Households with more assets were more likely to earn from casual labour supplying. Income from enterprise and salaried employment were negatively correlated with income from casual labour supplying. Profit from enterprise and salary both were inversely correlated with annual wage earned by the households. Even though, those involved with NGOs had a lower probability of becoming casual labour supplying households, when they got involved with the labour market they earned more. Casual labour supplying households in Khulna and Satkhira had around 30% and 64% lower incomes from labour market than the households in Patuakhali. This is in line with the finding in Table 5.2.2, where the Patuakhali households were estimated to earn more from labour market than the households in other two districts. Households in distant villages and high saline villages earned around 27% and 38% less as wages than their counterparts living in the nearby and less saline villages, respectively (Table 7.9.2.3). This is consistent with the observation in Table 5.2.3.2 that the casual labour supplying households in less saline areas earned significantly higher than their counterparts living in high saline areas. Table 7.9.2.3. Cragg and Tobit Model Estimates for Determinants of Income from Casual Labour supplying Dependent Variable: Log of Annual Income from Casual Labour Supplying Variables No of months hh had to purchase drinking water No of months hh faced salinity Dummy household perceived salinity increasing over the past 5 years(=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum Resilience¥) (Village affected by high salinity) X (Number of months in a year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate) Primary (Years of Schooling >0 &5 &10) Age of the Household Head Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) Total Households Members Total Value of Assets Own land (decimal) Net income from crop farming

162

Tier 2 of Cragg Model Tobit Model Coefficient (SE) -.0004 -.2289*** (.0093) (.0737) .0021 .0459 (.0155) (.1203) .1177* -.1169 (.0704) (.5394) -.0230*** .0373 (.0092) (.0630) .4705*** -.2077** (.0130) (.1013) .0117 .0498 (.0109) (.0861) .0833 (.0598) -.0401 (.0660) .1417 (.1824) -.0018 (.0019) -.2593** (.1317) .1145*** (.0179) .000002** (.000001) .00028 (.00039) -.000001

-.6842 (.4853) -2.103*** (.5227) -5.230*** (1.104) -.0384*** (.0149) -.3884 (1.010) .6227*** (.1374) -.000001 (.00001) -.0162*** (.0028) .000023


Annual Net Return from Enterprise Ln (Annual Income from Fisheries) Annual income from salary & pension Value of Livestock Total Loan Amount NGO member dummy (=1 if any household member is a member of NGO) Migration is better than farming in terms of income+ (index value)

(.000002) -.000002** (.000001) -.0084 (.0072) -.000003** (.000002) -.000001 (.000001) -.000001 (.000001) .0943* (.0545) .0563 (.0707)

(.000016) -.00007*** (.00001) -.3042*** (.0506) -.00009*** (.00001) -.000016*** (.00001) -.000003 (.000005) -.4468 (.4200) 1.5011*** (.5641)

-.2988*** (.0921) -.6470*** (.1062) -.2708*** (.0770) -.3762*** (.0823) 10.8913*** (.1455)

.9600 (.7137) .5655 (.8451) 1.491*** (.6088) 2.232*** (.6596) 7.962*** (1.133)

960

960

District (Base = Patuakhali) Khulna Satkhira Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity) Constant N

Log likelihood -955.325 -2237.6934 Prob> chi2 .0000 .0000 Test statistics of the regression models AIC 2020.651 4531.387 BIC 2288.333 4667.661 LR χ2 (27) Prob> χ2 Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

Section 7.9.3. Determinants of Saline Water Fisheries Production Practices and Explaining Income Variation Section 7.9.3.1. Determinants of Saline Water Fisheries Production Uptake The coefficients and the marginal effects for the determinants of the saline water fisheries based production are presented in Tables 7.9.3.1 and 7.9.3.2, respectively. The salinity indicator variables had significant effect across models. With an additional month of drinking water purchase, a household exhibits 1% to 2% lower probability of involving in saline water fisheries (refer to M1 through M4 in Table 7.9.3.2). However, their probability of adopting saline water fisheries increased when they perceived salinity having increased over time and when facing an additional month of salinity. As salinity is a precondition for doing shrimp and other saline water fisheries, increasing level of salinity creates further opportunities for households to take to fisheries. The positive coefficient for the product of crop growers and maximum resilience means that crop farmers who were resilient to adverse effects of salinity for longer period of time, were

163


more likely to adopt saline water fisheries. In the distant villages, the households who faced salinity for more months were more likely adopt saline water fisheries (Table 7.9.3.1 and 7.9.3.2). Table 7.9.3.1. Probit Model Estimates for uptake of Saline Water Fisheries Dependent Variable = Adoption of Saline Water Fisheries Production (=1 if the household has saline water fisheries) (M1) (M2) (M3) With Salinity With With Assets, Loan, Indicators Demographics Social Capital, and Variables Preference for Farming Coefficient (SE) -0.089*** -0.076*** -0.107*** No of months hh had to purchase drinking water (0.017) (0.018) (0.021) 0.064*** 0.065*** 0.046* No of months hh faced salinity (0.025) (0.026) (0.029) Dummy household perceived salinity increasing over the past 0.680*** 0.666*** 0.515*** 5 years (=1 if the hh observed salinity increase) (0.127) (0.132) (0.154) Product of (Dummy for crop growers) and (Maximum 0.085*** 0.078*** 0.062*** Resilience¥) (0.015) (0.015) (0.016) (Village affected by high salinity) X (Number of months in a 0.022 0.022 0.031 year hh faced high salinity) (0.017) (0.018) (0.019) (Dummy for Village far from district headquarter) X (Number 0.037*** 0.029** 0.055*** of months in a year hh faced high salinity) (0.014) (0.014) (0.016) Education of the Household Head (Base = Illiterate or Informally Literate) 0.249* 0.135 Primary (Years of Schooling >0 &5 &10) (0.218) (0.257) 0.016*** 0.010** Age of the Household Head (0.004) (0.005) Dependency Ratio Based on Age (Number of members in age -0.281 -0.139 group 0-14 and >59/ Number of members in the household) (0.284) (0.311) 0.071** 0.045 Total Family Members (0.036) (0.041) 0.232*** Ln (Total Value of Assets) (0.050) 0.002*** Own land (decimal) (0.001) -0.000002 Net return from crop farming (0.000005) -0.051*** Ln (Total Income from entrepreneurship activities) (0.015) -0.084*** Ln (Total Income from Casual Labour supplying) (0.013) -0.074*** Ln (Total annual income from salary & pension) (0.018) 0.00000004 Total Value of Livestock (0.000002) 0.006 Ln (Total Loan Amount) (0.012) NGO member dummy (=1 if any household member is a 0.066 member of NGO) (0.121) Migration is better than farming in terms of income+ (index -0.010 value) (0.183) Doing a business is not prestigious and I would prefer to be in -0.086 faming despite low profits+ (index value) (0.178) District (Base = Patuakhali)

(M4) District Fixed Effects

-0.088*** (0.027) -0.006 (0.042) 0.485*** (0.166) 0.055*** (0.020) -0.046 (0.047) 0.095*** (0.037)

0.081 (0.187) 0.145 (0.188) -0.272 (0.286) 0.004 (0.005) 0.068 (0.341) 0.125*** (0.047) 0.209*** (0.055) 0.002*** (0.001) -0.000002 (0.00001) -0.061*** (0.017) -0.103*** (0.015) -0.077*** (0.020) 0.000001 (0.000002) 0.029** (0.014) -0.065 (0.139) -0.275 (0.210) -0.289 (0.204) 2.749*** (0.466) 1.836*** (0.503) -0.365 (0.310)

Khulna Satkhira Dummy for distant village (=1 if the village is a distant one)

164

ACIAR Report ADP/2015/032 0.843** (0.414) -1.601*** -2.975*** -3.866*** -5.195*** Constant (0.120) (0.314) (0.576) (0.736) N 960 960 960 960 Prob> chi2 0.000 0.000 0.000 0.000 Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value. Dummy for saline village (=1 if the village has high salinity)

The positive association between the educational variables and adoption means that compared to the illiterate households, the literate households had a higher probability of fishery business adoption. With increasing number of family members, household’s adoption probability also increases. With increasing assets and land size, farmers were more likely to adopt. Credit also increased adoption probability. Income from enterprise, casual labour supplying and salaried employment reduced a household’s probability to adopt saline water fisheries. As expected, compared to the households in Patuakhali, households in Khulna and Satkhira had higher adoption probabilities. In Khulna and Satkhira, households had 42% and 28% higher adoption probability than their counterparts in Patuakhali. Compared to low saline areas, households in high saline areas had around 13% higher adoption probability (Table 7.9.3.1 and 7.9.3.2). Table 7.9.3.2.Marginal Effect of the Variables in probit model on adoption of Saline Water Fisheries Dependent Variable = Adoption of Saline Water Fisheries Production Practice (=1 if the household has saline water fisheries) (M1) (M2) (M3) With Salinity With With Assets, Loan, Variables Indicators Demographics Social Capital, and Preference for Farming No of months hh had to purchase drinking water -0.020*** -0.016*** -0.019*** No of months hh faced salinity 0.015*** 0.014*** 0.008* Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) 0.154*** 0.142*** 0.093*** Product of (Dummy for crop growers) and (Maximum Resilience¥) 0.019*** 0.017*** 0.011*** (Village affected by high salinity) X (Number of months in a year hh faced high salinity) 0.005 0.005 0.006 (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) 0.009*** 0.006** 0.010*** Education of the Household Head (Base = Illiterate or Informally Literate) Primary (Years of Schooling >0 &5 &10) 0.154*** 0.027 Age of the Household Head 0.003*** 0.002*** Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) -0.060 -0.025 Total Family Members 0.015** 0.008 Ln (Total Value of Assets) 0.042*** Own land (decimal) 0.0004*** Net return from crop farming -0.0000003 Ln (Total Income from entrepreneurship activities) -0.009*** Ln (Total Income from Casual Labour supplying) -0.015*** Ln (Total annual income from salary & pension) -0.014*** Total Value of Livestock 0.00000001 Ln (Total Loan Amount) 0.001 NGO member dummy (=1 if any household member is a member of NGO) 0.012

165


-0.013*** -0.001 0.074*** 0.008*** -0.007 0.014***

0.012 0.022 -0.041 0.001 0.010 0.019*** 0.032*** 0.0003*** -0.0000003 -0.009*** -0.016*** -0.012**** 0.0000002 0.004** -0.010

ACIAR Report ADP/2015/032 Migration is better than farming in terms of income+ (index value) -0.002 -0.042 Doing a business is not prestigious and I would prefer to be in faming despite low profits+ (index value) -0.016 -0.044 District (Base = Patuakhali) Khulna 0.417*** Satkhira 0.278*** Dummy for distant village (=1 if the village is a distant one) -0.055 Dummy for saline village (=1 if the village has high salinity) 0.128** Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

Section 7.9.3.2. Explaining Variation in Income from Saline Water Fisheries As the Cragg model was found superior over the Tobit model through the LR test and associated BIC and AIC values, the discussion explaining variation in incomes from saline water fisheries will be based upon the Cragg model. Though the households that perceived salinity to have increased over time had a higher adoption probability, they earned lower than their counterparts who did not perceive salinity to have increased. Salinity is essential, but excess salinity may not always be good for fish health and growth. None of the demographic factors had a significant impact on fishery based income. Higher size of own land and value of assets not only increased adoption probability, they also positively contributed towards higher income. Compared to the fish growers in far off villages, the growers in nearby villages earned more. The fish producers in high saline areas earned significantly higher than their counterparts living in less saline areas (Table 7.9.3.3). It was observed earlier that fish farming in high saline areas generated significantly higher incomes than those in low saline areas (Table 5.2.2 and Table 5.2.4.3). Table 7.9.3.3. Cragg and Tobit Model Estimates for Determinants of Income from Saline Water Fisheries Dependent Variable: Log of Annual Income from Saline Water Fisheries (tk) Variables No of months hh had to purchase drinking water No of months hh faced salinity Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum Resilience¥) (Village affected by high salinity) X (Number of months in a year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate)

Tier 2 of Cragg Model Tobit Model Coefficient (SE) 0.004 -0.508*** (0.035) (0.226) 0.089 0.227 (0.071) (0.398) -0.564*** 4.612*** (0.203) (1.486) 0.013 0.403*** (0.017) (0.154) -0.146*** -0.674 (0.067) (0.431) 0.116*** 0.772*** (0.053) (0.346) -0.119 (0.239) -0.164 (0.243) -0.513 (0.321) -0.006 (0.007)

Primary (Years of Schooling >0 &5 &10) Age of the Household Head

166

2.099 (1.718) 3.029* (1.725) 0.341 (2.564) 0.038 (0.049)

ACIAR Report ADP/2015/032 Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) Total Family Members Total Value of Assets Own land (decimal) Net return from crop farming Annual Net Return from Enterprise Annual Income from Casual Labour Supplying Annual Income from Salary & Pension Total Value of Livestock Total Loan Amount NGO member dummy (=1 if any household member is a member of NGO) Migration is better than farming in terms of income+ (index value) Doing a business is not prestigious and I would prefer to be in faming despite low profits+ (index value) District (Base = Patuakhali) Khulna Satkhira Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity) Constant

0.133 (0.431) -0.048 (0.058) 0.000003* (0.000002) 0.004*** (0.001) -0.000006 (0.00001) 0.000000 (0.000002) -0.000002 (0.000001) -0.000001 (0.000003) -0.000003 (0.000002) 0.000001 (0.000002) 0.183 (0.159) -0.183 (0.258) 0.326 (0.229)

0.042 (3.177) 0.885** (0.436) 0.00003** (0.00001) 0.031*** (0.006) 0.00001 (0.00007) -0.00004*** (0.00002) -0.00007*** (0.00001) -0.00006*** (0.00002) 0.00003 (0.00002) 0.00004*** (0.00001) -0.114 (1.244) -2.877 (1.899) -2.698 (1.799)

0.861 (0.964) 0.912 (0.995) -1.181*** (0.442) 1.296*** (0.524) 9.676*** (1.093) 938 -454.07 0.000

26.335*** (4.512) 18.977*** (4.798) -4.162 (2.899) 8.301*** (3.705) -38.360*** (5.990) 938 -792.21 0.000

N Log likelihood Prob> chi2 Test statistics of the regression models AIC 1022.132 1642.429 BIC 1298.225 1782.898 LR χ2 (27) 634.99 Prob> χ2 0.000 Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

Section 7.9.4. Adoption of Fresh Water Fisheries and Explaining Income Variation Section 7.9.4.1. Determinants of Fresh Water Fisheries Adoption Though saline water fisheries dominated the occupations of fishery based households, fresh water fisheries were also common in the study areas, particularly in Patuakhali district. During the monsoon season, in low saline areas of Khulna and Satkhira, farmers grew different fresh water fish. Adoption of fresh water fisheries was highly influenced by salinity indicators. When a household had to purchase drinking water or faced salinity for an additional month in a year, its probability of adoption decreased. Similarly, households that observed salinity to have increased during the past five years had a lower adoption probability. Compared to the illiterate households, 167


the households with literate heads had higher adoption probabilities. Adoption was more likely for the category having heads with secondary level of education. Households belonging to this category had around 5% higher adoption probability than the households with illiterate heads. With increasing family members, a household was more likely to adopt fresh water fisheries, as additional members could be employed in the fish production, reducing cash cost burden (Table 7.9.4.1 & 7. 9.4.2). Table 7.9.4.1. Probit Model Estimates for Adoption of Fresh Water Fisheries Dependent Variable = Involvement with fresh water fisheries (=1 if household was engaged in fresh water fisheries) (M1) (M2) (M3) With Salinity With With Assets, Loan, Indicators Demographics Social Capital, and Variables Preference for Farming Coefficient (SE) -0.034 -0.030 -0.067*** No of months hh had to purchase drinking water (0.023) (0.024) (0.027) -0.055* -0.056* -0.057 No of months hh faced salinity (0.033) (0.034) (0.037) Dummy household perceived salinity increasing over the past -0.447** -0.476*** -0.470** 5 years (=1 if the hh observed salinity increase) (0.219) (0.225) (0.241) Product of (Dummy for crop growers) and (Maximum 0.091*** 0.083*** 0.080*** Resilience¥) (0.013) (0.014) (0.015) (Village affected by high salinity) X (Number of months in a -0.042 -0.050* -0.070*** year hh faced high salinity) (0.029) (0.030) (0.033) (Dummy for Village far from district headquarter) X 0.029 0.022 0.019 (Number of months in a year hh faced high salinity) (0.027) (0.028) (0.030) Education of the Household Head (Base = Illiterate or Informally Literate) 0.050 0.081 Primary (Years of Schooling >0 &5 &10) (0.279) (0.317) 0.008 0.005 Age of the Household Head (0.005) (0.005) Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the 0.277 0.318 household) (0.331) (0.354) 0.095*** 0.105*** Total Family Members (0.042) (0.046) -0.036 Ln (Total Value of Assets) (0.039) 0.002*** Own land (decimal) (0.001) 0.000003 Net return from crop farming (0.000004) -0.015 Ln (Total Income from entrepreneurship activities) (0.018) -0.025* Ln (Total Income from Casual Labour Supplying) (0.015) -0.003 Ln (Total annual income from salary & pension) (0.019) 0.0000004 Total Value of Livestock (0.000002) 0.053*** Ln (Total Loan Amount) (0.014) NGO member dummy (=1 if any household member is a -0.242 member of NGO) (0.151)

168


-0.106*** (0.045) 0.004 (0.050) -0.329 (0.251) 0.081*** (0.018) 0.081* (0.046) -0.091*** (0.041)

0.073 (0.202) 0.419** (0.206) 0.482 (0.338) 0.006 (0.006) 0.220 (0.371) 0.100** (0.049) 0.00008 (0.044) 0.002*** (0.001) 0.0000005 (0.000004) -0.006 (0.020) -0.023 (0.016) -0.001 (0.020) 0.0000004 (0.000002) 0.051*** (0.016) -0.356*** (0.163)

ACIAR Report ADP/2015/032 Migration is better than farming in terms of income+ (index value) Doing a business is not prestigious and I would prefer to be in faming despite low profits+ (index value)

-0.137 (0.191) -0.292 (0.189)

-0.128 (0.208) -0.390** (0.209)

District (Base = Patuakhali) -0.872*** (0.283) -0.665** (0.341) 0.916*** (0.227) -1.275*** (0.257)

Khulna Satkhira Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity) -1.128*** Constant

-2.174*** (0.334)

-1.620*** (0.511)

-1.794*** (0.575) N 960 960 960 960 Prob> chi2 0.000 0.000 0.000 0.000 Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value. (0.110)

Own land positively and significantly contributed to fresh water fisheries adoption. Among the various income sources, significant impact is observed only for income from casual labour supplying. A one percent increase in income from casual labour supplying will reduce probability of adopting fresh water fisheries by 0.00003%. All the four district level fixed effect variables significantly influenced adoption probability. Fish growers in Khulna and Satkhira had 10% and 8% lower adoption probabilities compared to Patuakhali growers, respectively. Compared to the households living in nearby villages, households in distant villages had around 11% higher adoption probability. The households in high saline villages had 15% lower adoption probability than their counterparts living in less saline villages (Table 7.9.4.1 & 7. 9.4.2). Table 7.9.4.2. Marginal Effects of the Variables in the Probit Model on a Household's Probability of Fresh Water Fisheries Uptake Dependent Variable = Involvement with fresh water fisheries (=1 if household is engaged in fresh water fisheries) (M1) (M2) (M3) With Salinity With With Assets, Loan, Variables Indicators Demographics Social Capital, and Preference for Farming No of months hh had to purchase drinking water -0.0049 -0.0042 -0.009*** No of months hh faced salinity -0.008* -0.0077* -0.007 Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) -0.064** -0.0651*** -0.061** Product of (Dummy for crop growers) and (Maximum Resilience¥) 0.013*** 0.0114*** 0.010*** (Village affected by high salinity) X (Number of months in a year hh faced high salinity) -0.006 -0.0068* -0.009*** (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) 0.004 0.0030 0.002 Education of the Household Head (Base = Illiterate or Informally Literate) Primary (Years of Schooling >0 &5 &10) 0.0750** 0.057 Age of the Household Head 0.0010 0.001

169


-0.0122*** 0.0004 -0.0380 0.0093*** 0.0093* -0.0106***

0.0084 0.0484** 0.0556 0.0007

ACIAR Report ADP/2015/032 Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) 0.0379 0.041 0.0254 Total Family Members 0.0130*** 0.014*** 0.0115** Ln (Total Value of Assets) -0.005 0.00001 Own land (decimal) 0.00026*** 0.0003*** Net return from crop farming 0.0000004 0.0000001 Ln (Total Income from entrepreneurship activities) -0.002 -0.0007 Ln (Total Income from Casual Labour Supplying) -0.003* -0.0027 Ln (Total annual income from salary & pension) -0.0004 -0.0001 Total Value of Livestock 0.0000001 0.0000001 Ln (Total Loan Amount) 0.007*** 0.0059*** NGO member dummy (=1 if any household member is a member of NGO) -0.031 -0.0411*** Migration is better than farming in terms of income+ (index value) -0.018 -0.0147 Doing a business is not prestigious and I would prefer to be in faming despite low profits+ (index value) -0.038 -0.0450** District (Base = Patuakhali) Khulna -0.1006*** Satkhira -0.0768** Dummy for distant village (=1 if the village is a distant one) 0.1056*** Dummy for saline village (=1 if the village has high salinity) -0.1471*** Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

Section 7.9.4.2. Determinants of Income from Fresh Water Fisheries Though the AIC value for the Cragg model was higher than the Tobit one, based on the LR test and relatively lower BIC for Cragg model we continue our explanation based on the Cragg model. An additional month of salinity decreased income from fresh water fisheries by around 54%. Income from crop farming was positively associated with income from fresh water fisheries, whereas earnings from salaried employment had negative correlation with fresh water fisheries income. Households in Satkhira earned less than their counterparts in Patuakhali district (Table 7.9.4.3). Table 7.9.4.3. Cragg and Tobit Model Estimates for Determinants of Income from Fresh Water Fisheries Dependent Variable: Log of Annual Income from Fresh Water Fisheries (tk) Variables No of months hh had to purchase drinking water No of months hh faced salinity Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum Resilience¥) (Village affected by high salinity) X (Number of months in a year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate)

Tier 2 of Cragg Model Coefficient (SE) 0.024 (0.117) -0.537*** (0.152) 0.321 (0.571) 0.034 (0.028) 0.413*** (0.126) 0.186 (0.124) -0.191 (0.311) -0.083 (0.309) 0.896 (0.573)

Primary (Years of Schooling >0 &5 &10)

170

Tobit Model -1.271*** (0.545) 0.585 (0.585) -7.071*** (0.034) 0.756*** (0.194) 0.272 (0.561) -1.289*** (0.522) -0.207 (2.214) 3.062 (2.210) 4.652 (3.867)


Age of the Household Head Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) Total Family Members Total Value of Assets Own land (decimal) Net return from crop farming Annual Net Return from Enterprise Annual Income from Casual Labour Supplying Annual income from salary & pension Total Value of Livestock Total Loan Amount NGO member dummy (=1 if any household member is a member of NGO) Migration is better than farming in terms of income+ (index value) Doing a business is not prestigious and I would prefer to be in faming despite low profits+ (index value) District (Base = Patuakhali) Khulna Satkhira Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity) Constant N Log likelihood Prob> chi2 Test statistics of the regression models AIC BIC LR χ2 (27)

-0.002 (0.010) 0.082 (0.606) -0.099 (0.084) 0.000004 (0.000004) 0.00022 (0.0014) 0.00001* (0.000005) 0.000002 (0.000004) 0.000002 (0.000002) -0.00001* (0.000003) -0.0000020 (0.000002) -0.0000021 (0.000002) -0.326 (0.261) 0.406 (0.324) -0.585* (0.353)

-0.010 (0.064) 4.328 (4.226) 1.577*** (0.576) -0.00004 (0.00003) 0.028*** (0.009) 0.00003 (0.00004) -0.00002 (0.000023) -0.00003 (0.00002) 0.000004 (0.000022) -0.00001 (0.000015) 0.00006*** (0.00002) -4.125*** (1.877) -1.262 (2.320) -3.574 (2.319)

-0.489 (0.516) 2.304*** (0.813) -0.458 (0.396) -0.141 (0.485) 10.342*** (0.757) 960 -289.35 0.000

-8.613*** (3.132) -7.398** (4.056) 10.428*** (2.552) -14.362*** (3.098) -17.609*** (5.247) 960 -444.59 0.000

692.71 970.12

947.18 1088.32 92.90

Prob> χ2

0.000


Section 7.9.5. Identifying Determinants of Fisheries (both saline and fresh water) and Explaining Variation in Income Section 7.9.5.1. Factors Influencing Fisheries Uptake Tables 7.9.5.1 and 7.9.5.2 present fisheries based enterprise (including both saline and fresh water) adoption probabilities for the sampled households. Table 7.9.5.3 presents analysis of income variation from fisheries business. An additional month of drinking water purchase may reduce a household’s probability of adopting fisheries by around 2%. Households that observed salinity to have increased during the past five years were less likely to adopt a fishery based enterprise. 171


Households that observed salinity to have increased during the past five years had 9% to 16% lower adoption probability than their counterparts. With increasing resilience against salinity, a crop grower’s adoption probability increased. Adoption probability increased among the households in distant villages if they faced an additional month of salinity (refer to M1 to M4 in Tables 7.9.5.1 & 7.9.5.2). Households with more educated heads were more likely to adopt fisheries. According to M2 in Table 7.9.5.2 compared to the households with illiterate heads, the primary, secondary and higher secondary & above educated heads had around 6%, 19.6% and 20.5% higher adoption probabilities for their households. The middle category, i.e., the category of households with secondary level of education had significantly higher adoption probabilities across all models. The positive but insignificant coefficient for the highest education category may mean that, with more education fewer households remain interested in agriculture or fishing. With increasing family members, households were more likely to get involved with fisheries. Depending on model specification, adoption probability for an additional family member may increase by 2% to 3%. Table 7.9.5.1. Probit Model Estimates for Adopting Fisheries Dependent Variable = Involvement with fisheries (=1 if household is engaged in fisheries) (M1) (M2) With Salinity With Indicators Demographics Variables

No of months hh had to purchase drinking water No of months hh faced salinity Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum Resilience¥) (Village affected by high salinity) X (Number of months in a year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate)

-0.067*** (0.015) 0.014 (0.022) 0.545*** (0.123) 0.144*** (0.016) 0.013 (0.017) 0.036*** (0.013)

Primary (Years of Schooling >0 &5 &10) Age of the Household Head Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) Total Family Members Ln (Total Value of Assets)

172

(M3) With Assets, Loan, Social Capital, and Preference for Farming Coefficient (SE) -0.055*** -0.087*** (0.016) (0.018) 0.008 -0.013 (0.023) (0.025) 0.532*** 0.376*** (0.126) (0.143) 0.133*** 0.107*** (0.017) (0.018) 0.015 0.013 (0.017) (0.018) 0.029*** 0.049*** (0.014) (0.015)


-0.112*** (0.023) -0.003 (0.033) 0.406*** (0.151) 0.087*** (0.018) 0.086*** (0.031) 0.005 (0.026)

0.234* (0.133) 0.723*** (0.131) 0.757*** (0.201) 0.012*** (0.004)

0.187 (0.142) 0.481*** (0.144) 0.305 (0.233) 0.007* (0.004)

0.119 (0.149) 0.336*** (0.152) 0.086 (0.246) 0.004 (0.004)

-0.124 (0.254)

-0.045 (0.277)

0.061 (0.288)

0.097*** (0.033)

0.103*** (0.036) 0.117***

0.150*** (0.039) 0.115***

ACIAR Report ADP/2015/032 (0.038) 0.003*** (0.001) 0.000001 (0.000004) -0.031*** (0.014) -0.071*** (0.012) -0.060*** (0.016) -0.000001 (0.000002) 0.028*** (0.011) -0.067 (0.109) -0.127 (0.157) -0.037 (0.148)

Own land (decimal) Net return from crop farming Ln (Total Income from entrepreneurship activities) Ln (Total Income from Casual Labour Supplying) Ln (Total annual income from salary & pension) Total Value of Livestock Ln (Total Loan Amount) NGO member dummy (=1 if any household member is a member of NGO) Migration is better than farming in terms of income+ (index value) Doing a business is not prestigious and I would prefer to be in faming despite low profits+ (index value) District (Base = Patuakhali)

0.550*** (0.200) -0.374 (0.242) 0.505*** (0.192)

Khulna Satkhira Dummy for distant village (=1 if the village is a distant one)

-0.705*** (0.223)

Dummy for saline village (=1 if the village has high salinity)

Constant

(0.039) 0.003*** (0.001) 0.000001 (0.000004) -0.030*** (0.014) -0.075*** (0.012) -0.052*** (0.017) -0.000001 (0.000002) 0.044*** (0.011) -0.153 (0.115) -0.263 (0166) -0.062 (0.156)

1.041*** (0.096)

-2.366*** (0.266)

-2.519*** (0.457)

-2.594*** (0.497)

N 960 960 960 960 Prob> chi2 0.000 0.000 0.000 0.000 Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

As was observed for other income sources, income and assets had consistent and robust effect on adoption. With increasing own land and value of assets households adopted more. Credit significantly and positively contributed to adoption. Income from enterprise, casual labour supplying and salaried employment had negative association with adoption. According to M4, a 1% increase in income from enterprise, casual labour supplying and salaried employment may reduce adoption probability by 0.7%, 1.7% and 1.1%, respectively. Compared to households in Patuakhali, in Khulna households had 12% higher adoption probability. In distant villages, households had 11% higher adoption probability than the nearby villages. Compared to the households in high saline villages, in low saline villages, households had 15.4% lower adoption probability (Table 7.9.5.2). Table 7.9.5.2.Marginal Effects of Variables on a Household's Probability of Adopting Fisheries Dependent Variable = Involvement with fresh water fisheries (=1 if household is doing fisheries) (M1) (M2) Variables With Salinity With Indicators Demographics

173

(M3) With Assets, Loan, Social Capital, and


ACIAR Report ADP/2015/032 Preference for Farming -0.0208*** -0.0030

No of months hh had to purchase drinking water -0.019*** -0.015*** -0.024*** No of months hh faced salinity 0.004 0.002 -0.001 Dummy household perceived salinity increasing over the past 0.158*** 0.144*** 0.0895*** 0.089*** 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum 0.042*** 0.036*** 0.0254*** 0.019*** Resilience¥) (Village affected by high salinity) X (Number of months in a 0.004 0.004*** 0.0031 0.019*** year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number 0.010*** 0.008*** 0.0117*** 0.001 of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate) Primary (Years of Schooling >0 &5 &10) 0.205*** 0.0725 0.019 Age of the Household Head 0.003*** 0.0018* 0.001 Dependency Ratio Based on Age (Number of members in age -0.034 -0.0107 0.013 group 0-14 and >59/ Number of members in the household) Total Family Members 0.026*** 0.0244*** 0.033*** Ln (Total Value of Assets) 0.0278*** 0.025*** Own land (decimal) 0.0007*** 0.001*** Net return from crop farming 0.0000003 0.0000002 Ln (Total Income from entrepreneurship activities) -0.0074*** -0.007*** Ln (Total Income from Casual Labour Supplying) -0.0169*** -0.017*** Ln (Total annual income from salary & pension) -0.0142*** -0.011*** Total Value of Livestock -0.0000002 -0.0000003 Ln (Total Loan Amount) 0.0066*** 0.010*** NGO member dummy (=1 if any household member is a -0.0159 -0.034 member of NGO) + Migration is better than farming in terms of income (index -0.0301 -0.058 value) Doing a business is not prestigious and I would prefer to be in -0.0089 -0.014 faming despite low profits+ (index value) District (Base = Patuakhali) Khulna 0.120*** Satkhira -0.082 Dummy for distant village (=1 if the village is a distant one) 0.110*** Dummy for saline village (=1 if the village has high salinity) -0.154*** Note: +On a scale of 1 to 5, with 1 implying completely disagrees and 5 completely agrees, the respondents were asked to rate the statement. The responses were then converted into an index value.

Section 7.9.5.2. Factors Influencing Variation in Income from Fisheries As with the earlier sections, the Cragg model is found superior to the Tobit model in explaining income from fisheries based upon the LR test and lower AIC and BIC values. None of the salinity factors are found to have significant association with fisheries income. Among the demographic factors, only head’s age is found to be negatively associated with income. Among the variables used to represent household’s economic condition, only net income from crop farming is positively associated with income from fisheries. In the saline areas, it is more likely for the larger farmers to earn more from crops and fisheries simultaneously (Table 7.9.5.3). Table 7.9.5.3. Cragg and Tobit Model Estimates for Determinants of Income from Saline Water Fisheries Dependent Variable: Log of Annual Income from Fisheries (tk) Tier 2 of Cragg Model Tobit Model Coefficient (SE) 0.152 -0.752***

Variables No of months hh had to purchase drinking water

174


No of months hh faced salinity Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum Resilience¥) (Village affected by high salinity) X (Number of months in a year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate) Primary (Years of Schooling >0 &5 &10) Age of the Household Head Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) Total Family Members Total Value of Assets Own land (decimal) Net Return from Crop Farming Annual Net Return from Enterprise Annual Income from Casual Labour supplying Annual Income from Salary & Pension Total Value of Livestock Total Loan Amount NGO member dummy (=1 if any household member is a member of NGO) Migration is better than farming in terms of income+ (index value) Doing a business is not prestigious and I would prefer to be in faming despite low profits+ (index value) District (Base = Patuakhali) Khulna Satkira Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity) Constant N Log likelihood Prob> chi2 Test statistics of the regression models AIC BIC LR χ2 (27) Prob> χ2

(0.098) 0.112 (0.165) -0.285 (0.577) 0.014 (0.043) -0.212 (0.149) -0.113 (0.117)

(0.181) 0.213 (0.276) 3.144*** (1.176) 0.548*** (0.112) 0.691*** (0.251) -0.059 (0.212)

0.485 (0.620) 0.011 (0.612) 0.184 (0.886) -0.035*** (0.017) 0.250 (1.105) 0.012 (0.143) 0.000002 (0.000005) 0.003 (0.002) 0.00003** (0.00001) 0.000003 (0.000005) 0.000001 (0.000004) -0.000004 (0.000006) 0.000006 (0.000004) 0.000003 (0.000004) 0.180 (0.423) -0.226 (0.641) 0.525 (0.525)

1.255 (1.200) 2.955*** (1.210) 1.335 (1.942) 0.029 (0.034) 0.937 (2.284) 0.984*** (0.304) 0.00001 (0.00001) 0.031*** (0.005) 0.00004 (0.00003) -0.00002** (0.00001) -0.00004*** (0.00001) -0.00003*** (0.00001) -0.0000002 (0.00001) 0.00004*** (0.00001) -0.155 (0.905) -2.788*** (0.313) 0.016 (1.229)

1.597** (0.804) 2.670*** (0.992) 0.994 (0.809) 2.366*** (0.029) 6.773*** (1.451) 960 -1052.28 0.000

4.325*** (1.587) -2.832 (1.941) 3.876*** (1.540) -7.044*** (1.844) -15.045*** (2.874) 960 -1224.47 0.000

2218.56 2495.97

2506.95 2648.09 352.45 0.000


175


The district level fixed effect variables had more robust relationship with income from fisheries. Compared to fish farmers in Patuakhali, the farmers in Khulna and Satkhira had notably higher incomes. Earlier it was found, in Table 5.2.4.3, that fish farmers in Khulna and Satkhira had around 5.37 and 5.92 times higher incomes than the Patuakhali Farmers. Similarly, fisheries income was significantly higher in high saline villages than the less saline villages.

Section 7.10. Explaining Variation in Annual Incomes Earned by the Sampled Households Section 7.10.1. Explaining Household Income Variation To explain variation in annual household incomes we develop four different types of OLS regressions. The first one has only salinity related indicators (M1). The second has demographic variables along with the salinity related indicators (M2). The third and fourth are further extended with variables such as assets, loans, social capitals, perceptions and preferences for income generating activities (M3); and district level fixed effect variables (M4). The estimated coefficients for these variables are presented in Table 7.10.1. Among the salinity indicators, number of months a household had to purchase drinking water was positively and significantly associated with annual income across all models. The households who perceived salinity to have increased over the last 5 years had higher incomes than their counterparts who did not observe salinity increases. But the variable was significant only in M1 and M2. In high salinity affected villages, households that faced additional month(s) with high salinity had higher incomes. Though household head’s education was positively correlated with income, the effect was found significant only for the category of higher secondary and above education in M2. An additional dependent member in the household negatively affected annual income, whereas with an additional family member the households tended to earn more (Table 7.10.1). Table 7.10.1. OLS Regression for Explaining Household Level Income Variation Dependent Variable = Ln (Annual Household Income in tk) (M1) With Salinity Indicators Variables

No of months hh had to purchase drinking water No of months hh faced salinity

0.0193** (0.0098) -0.0010 (0.0140)

176

(M2) With Demographics

(M3) With Assets, Loan, Social Capital, and Preference for Migration and Enterprise Coefficient (SE) 0.0159* 0.0364*** (0.0096) (0.0096) 0.0018 -0.0040 (0.0144) (0.0126)


0.0203** (0.0109) -0.0040 (0.0164)

ACIAR Report ADP/2015/032 Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum Resilience¥) (Village affected by high salinity) X (Number of months in a year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate)

0.1398* (0.0812) 0.0133 (0.0090) 0.0178* (0.0105) -0.0086 (0.0086)

Primary (Years of Schooling >0 &5 &10) Age of the Household Head Dependency Ratio Based on Age (Number of members in age group 0-14 and >59/ Number of members in the household) Total Household Members

0.1423* (0.0812) 0.0020 (0.0090) 0.0166 (0.0106) -0.0109 (0.0087)

0.0128 (0.0751) 0.0136 (0.0093) 0.0166** (0.0089) -0.0184*** (0.0077)

0.0288 (0.0742) 0.0090 (0.0091) 0.0325*** (0.0141) 0.0125 (0.0107)

0.0654 (0.0685) 0.0879 (0.0815) 0.3323*** (0.1461) -0.0002 (0.0022)

0.0035 (0.0619) -0.0777 (0.0776) 0.0044 (0.1391) -0.0015 (0.0020)

0.0261 (0.0618) -0.0372 (0.0744) 0.0906 (0.1449) -0.00004 (0.0020)

-0.7590*** (0.2678) 0.1584*** (0.0317)

-0.4798*** (0.2270) 0.0995*** (0.0292) 0.0464*** (0.0142) 0.0013*** (0.0005) 0.000002*** (0.000001) -0.0065 (0.0055) 0.9446*** (0.1038) 0.8007*** (0.1426) 0.5890*** (0.0821) -0.4743*** (0.0940) 1.0718*** (0.1266) 0.0957 (0.0613) -0.0509 (0.0619)

-0.4368** (0.2238) 0.0957*** (0.0287) 0.0461*** (0.0148) 0.0013*** (0.0005) 0.000002*** (0.000001) -0.0068*** (0.0055) 0.9649*** (0.1032) 0.8471*** (0.1479) 0.6161*** (0.0988) -0.4398*** (0.0952) 1.0621*** (0.1246) 0.1423*** (0.0610) -0.0257 (0.0635)

Ln (Total Value of Assets) Own land (decimal) Total Value of Livestock Ln (Total Loan Amount) Dummy for Enterprise Dummy for causal labour Dummy for Fisheries Dummy for Households doing Crop Farming Dummy for Households with Salaried Employment Dummy for Migrating Households NGO member dummy (=1 if any household member is a member of NGO) District (Base = Patuakhali)

11.0127***

10.7815***

9.7016***

-0.2935*** (0.1217) -0.3642*** (0.1144) -0.2723*** (0.0968) -0.1840** (0.0939) 9.9122***

(0.0674)

(0.1720)

(0.2097)

(0.2069)

Khulna Satkira Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity)

Constant

939

N Adjusted R

2

0.0074

939

939

939

0.0570

0.2685

0.2831

Value of assets and livestock were positively correlated with total income. A 1% increase in total value of assets would increase annual income by around 0.05% (refer to M3 and M4). A one decimal of additional land would increase annual household income by around 0.13% (refer to M3 177


and M4). Credit, which was found to positively contribute to incomes from different livelihood options, its contribution to total income is negative. All the dummies, except dummy for crop income, used to represent income from different sources had positive coefficient. The positive coefficients imply that households involved with enterprise, casual labour supplying, fisheries and salaried employment earned significantly higher than their counterparts who were not involved in these particular sources. Most notable effect is observed for salaried income group; followed by those taking to any enterprise, casual labour supplying and fisheries, respectively. As was observed earlier, net return from crop farming was very marginal or negative for most households (Table 5.2.2), the negative coefficient for the dummy is not unexpected (Table 7.10.1). All the four fixed effect variables showed significant association with income. Compared to the households in Patuakhali, the households in Khulna and Satkhira had their incomes lower by around 29% and 36%, respectively. The households in distant villages earned significantly lower than their counterparts living in nearby villages. Compared to households in high saline villages, a household in less saline village had significantly lower annual income (Table 7.10.1).

Section 7.10.2. Further Investigating the Role of Enterprise in Explaining Household Income Variation when Controlling for Selection Bias

Admitting the possible existence of self-selection bias with the enterprise dummy variable, we followed Cerulli (2011)41 and ran an endogenous switching regression model.42 Through a correlation analysis, four variables were selected as instrumental variables for the regression. Among these, two were households’ perceptions based on the statements - ‘Doing a business is not prestigious and I would prefer to be in farming despite low profits’ and ‘Migration is mostly caused by Salinity and I would not migrate if there was no salinity’. Household’s responses were sought on a scale of 1 to 5 where 1 meant completely disagreed and 5 meant completely agreed. The perceptions were then converted into index values. The other two instrumental variables were year of schooling for the highest educated member in the household and square of own land, measured in decimal units.

41

Cerulli (2011): ivtreatreg: a new STATA routine for estimating binary treatment models with heterogeneous response to treatment under observable and unobservable selection, 8th Italian STATA users’ Group Meeting. 42 Please refer to Section 7.3 for detailed discussion about self-selection bias.

178


The top part of Table 7.10.2 provides the selection bias corrected estimates (or the average treatment effect, ATE) of the impact of enterprise uptake on annual household income. The coefficients of the enterprise uptake dummy is robust across models but the magnitude of effect is much higher in the Heckit and Direct-2SLS models than was in the OLS models, whereas it is similar in the Probit-2SLS model. We conducted sensitivity analysis of the ATE with various combinations of the instruments. The enterprise dummy variable had positive effect across models implying that the entrepreneurs earned significantly higher than the non-entrepreneurs (Table 7.10.3).

Table 7.10.2 Explaining Contribution of Enterprise to Household Incomes after Controlling for Selection Bias Dependent Variable = Ln (Annual Household Income in tk)

Enterprise dummy (=1 if household undertakes an enterprise activity)¥ Predicted probability for (w=1|x,z) No of months hh had to purchase drinking water No of months hh faced salinity Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum Resilience¥) (Village affected by high salinity) X (Number of months in a year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate) Primary (Years of Schooling >0 &5 &10) Age of the Household Head Dependency Ratio (Ratio of Working Member no to Total Family Member) Total Household Members Ln (Total Value of Assets) Own land (decimal) Total Value of Livestock Ln (Total Loan Amount) Dummy for Causal Labour Dummy for Fisheries Dummy for Households doing Crop Farming

(1) Heckit 1.4281*** (0.2869)

(2) Probit -2SLS 0.9445*** (0.3020)

(3) Direct-2SLS 1.7654*** (0.6473)

(4) Probit - OLS

0.0234** (0.0115) -0.0013 (0.0175) 0.0234 (0.0799) 0.0036 (0.0131) 0.0355** (0.0150) 0.0109 (0.0127)

0.0201* (0.0115) -0.0042 (0.0175) 0.0290 (0.0799) 0.0092 (0.0131) 0.0323** (0.0150) 0.0126 (0.0127)

0.0256** (0.0126) 0.0007 (0.0187) 0.0195 (0.0839) -0.0002 (0.0152) 0.0377** (0.0162) 0.0096 (0.0134)

1.0011*** (0.3404) 0.0204* (0.0123) -0.0033 (0.0187) 0.0283 (0.0849) 0.0097 (0.0139) 0.0333** (0.0160) 0.0121 (0.0135)

0.0144 (0.0737) -0.0544 (0.0808) 0.0442 (0.1402) -0.0002 (0.0023) -0.3418** (0.1718) 0.0879*** (0.0217) 0.0387** (0.0172) 0.0017*** (0.0005) 0.000002** (0.000001) -0.0108* (0.0064) 1.0150*** (0.1275) 0.6471*** (0.0792) -0.4090*** (0.1051)

0.0267 (0.0737) -0.0365 (0.0808) 0.0926 (0.1404) -0.00004 (0.0023) -0.4409*** (0.1728) 0.0961*** (0.0217) 0.0464*** (0.0173) 0.0013*** (0.0005) 0.000002* (0.000001) -0.0067 (0.0064) 0.8397*** (0.1319) 0.6147*** (0.0794) -0.4412*** (0.1052)

0.0059 (0.0785) -0.0669 (0.0872) 0.0105 (0.1575) -0.0003 (0.0024) -0.2727 (0.2148) 0.0822*** (0.0247) 0.0333* (0.0202) 0.0019*** (0.0006) 0.000002** (0.000001) -0.0138* (0.0083) 1.1372*** (0.2469) 0.6696*** (0.0914) -0.3866*** (0.1164)

0.0268 (0.0784) -0.0401 (0.0862) 0.0914 (0.1494) 0.00002 (0.0024) -0.4420** (0.1836) 0.0956*** (0.0232) 0.0449*** (0.0185) 0.0014*** (0.0005) 0.000002* (0.000001) -0.0070 (0.0069) 0.8585*** (0.1456) 0.6111*** (0.0842) -0.4407*** (0.1119)

179


Dummy for Households with Salaried Employment Dummy for Migrating Households NGO member dummy (=1 if any household member is a member of NGO) District (Base = Patuakhali) Khulna Satkira Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity) Constant

1.2163*** (0.1400) 0.1528* (0.0932) -0.0356 (0.0624)

1.0553*** (0.1434) 0.1419 (0.0931) -0.0252 (0.0624)

1.3287*** (0.2407) 0.1605* (0.0984) -0.0428 (0.0665)

1.0651*** (0.1549) 0.1439 (0.0991) -0.0207 (0.0662)

-0.3201*** (0.1078) -0.3989*** (0.1273) -0.2745*** (0.0902) -0.2178** (0.1002) 9.7689*** (0.2458)

-0.2923*** (0.1078) -0.3627*** (0.1274) -0.2722*** (0.0902) -0.1825* (0.1003) 9.9186*** (0.2473)

-0.3395*** (0.1175) -0.4242*** (0.1400) -0.2761*** (0.0945) -0.2424** (0.1129) 9.6645*** (0.3125)

-0.3053*** (0.1155) -0.3738*** (0.1362) -0.2725*** (0.0959) -0.1923* (0.1075) 9.9144*** (0.2636)

Dependent Variable = Enterprise take-up of the household (=1 if household has taken up an enterprise) No of months hh had to purchase drinking water No of months hh faced salinity Dummy household perceived salinity increasing over the past 5 years (=1 if the hh observed salinity increase) Product of (Dummy for crop growers) and (Maximum Resilience¥) (Village affected by high salinity) X (Number of months in a year hh faced high salinity) (Dummy for Village far from district headquarter) X (Number of months in a year hh faced high salinity) Education of the Household Head (Base = Illiterate or Informally Literate) Primary (Years of Schooling >0 &5 &10) Age of the Household Head Dependency Ratio (Ratio of Working Member no to Total Family Member) Total Household Members Ln (Total Value of Assets) Own land (decimal) Total Value of Livestock Ln (Total Loan Amount) Dummy for causal labour Dummy for Fisheries Dummy for Households doing Crop Farming Dummy for Households with Salaried Employment Dummy for Migrating Households NGO member dummy (=1 if any household member is a member of NGO) District (Base = Patuakhali) Khulna Satkhira

-0.0385 (0.0256) -0.0471 (0.0384) 0.1344 (0.1667) 0.0607*** (0.0242) -0.0348 (0.0325) 0.0229 (0.0274)

0.1074 (0.1652) 0.1641 (0.1898) 0.4915 (0.2990) 0.0016 (0.0048) -0.9105*** (0.2701) 0.0895** (0.0457) 0.0884** (0.0429) -0.0073*** (0.0020) -0.000003 (0.000002) 0.0535*** (0.0131) -1.6254*** (0.1551) -0.2345 (0.1660) -0.3898* (0.2217) -1.6550*** (0.2530) -0.0447 (0.2218) 0.0324 (0.1323) 0.4619* (0.2425) 0.5844**

180


Dummy for distant village (=1 if the village is a distant one) Dummy for saline village (=1 if the village has high salinity) Doing a business is not prestigious and I would prefer to be in faming despite low profits + Migration is mostly caused by Salinity and I would not migrate if there was no salinity + Year of Schooling of the Highest Education Household Member+ Square of Own Land Quantity+ Constant Hazard Lambda N Rho Sigma Adjusted R2 Note: ¥ Instrumented Variable. + Instrumental Variables.

(0.2928) -0.0040 (0.1968) 0.3819* (0.2244) -0.4483*** (0.1759) -0.2893 (0.1903) 0.0106 (0.0265) 0.00001*** (0.000005) -0.6189 (0.5462) -0.2753* (0.1625) 939 -0.3278 0.8399

939

939

939

0.2831

0.2131

0.1891

Table 7.10.3. Sensitivity Analysis of the ATE With Respect to the Choice of Instruments (1) With Ins1 only Dummy for Enterprise Observations Lambda

1.4542*** (0.3020) 939 -0.2889* (0.1760) Rho -0.3431 Sigma 0.8414 Absolute t statistics in parentheses * p0 &5 &10 &59/Number of members in the household) (0.21)

(0.38)

(0.85)

(0.07)

(0.96)

(2.28)

-0.289**

-0.132

-0.102

-0.090

-0.264*

-0.202

(2.11)

(0.98)

(0.78)

(0.69)

(1.81)

(1.48)

Medium (>5 acres)

-0.835** (2.46)

-0.402 (1.26)

-0.631** (2.01)

-0.450 (1.57)

-1.291*** (4.35)

0.441 (1.38)

Ln total value of agricultural assets

0.141***

0.104***

0.007

0.061***

0.015

0.025

(5.51)

(4.02)

(0.27)

(2.58)

(0.54)

(1.04)

-0.051

0.011

0.005

0.024

-0.095**

-0.008

(1.51)

(0.31)

(0.14)

(0.74)

(1.98)

(0.21)

0.022 (1.36)

0.005 (0.34)

0.038** (2.41)

0.024 (1.58)

0.075*** (4.30)

-0.011 (0.69)

-0.019* (1.66)

-0.003 (0.28)

-0.007 (0.61)

-0.008 (0.78)

0.026** (2.05)

-0.019* (1.66)

0.003

-0.046

-0.014

0.039

0.144***

0.155***

(0.09)

(1.27)

(0.41)

(1.06)

(3.54)

(4.30)

Ln total income of household

0.246** (2.35)

0.131 (1.30)

0.125 (1.30)

0.227** (2.30)

-0.077 (0.67)

-0.088 (0.84)

Enterprise take-up of a household (=1 if household has taken up an enterprise)

-0.210*

0.063

-0.104

-0.147

0.090

0.104

(1.96)

(0.60)

(1.02)

(1.46)

(0.73)

(0.98)

0.181

-0.126

0.249

-0.215

0.503***

0.459***

(1.04)

(0.76)

(1.53)

(1.33)

(2.79)

(2.71)

0.248

0.273

0.482*

0.816***

-0.236

-0.607**

(0.79)

(0.93)

(1.71)

(2.72)

(0.67)

(2.06)

-0.041

0.086

0.196

0.323

-0.184

-0.985**

Landholding size class (Base = Marginal (2.5 acres and