Role of Small Scale Fishing on the Livelihood Improvement of Haor Fishermen: An Empirical Evidence from Bangladesh

Corresponding Author: Md. Nur Mozahid Department of Agricultural Economics and Policy, Sylhet Agricultural University, Bangladesh Email: mozahid.aep@sau.ac.bd Abstract: The single most factor affecting the livelihood of haor people of Bangladesh is fishing. Data and information regarding this issue are lacking in Bangladesh, therefore, the study was conducted to assess the extent and determinants of livelihood step up of fishermen in haor area of Sunamganj district, Bangladesh. The present investigation showed that, the majority of the fishermen (60.0%) had small land ownership. Most of them were illiterate (56.3%) and belonged to (53.75%) middle income ($621.0-$915.0) and 6.25% had a small income ranging $305.0-$610. Among the fishermen, 45.0% were received credit from Mahajan and only 12.5% of fishermen had savings. Different livelihood assets were increased to a large extent due to small-scale fishing. Financial, human, social and physical capitals were increased from 32.0% to 76.67%, 45.0% to 75.33%, 42.0% to 62.5% and 30.0% to 73.37%, respectively due to engage in fishing. The logistic regression model revealed four significant variables i.e., family type, farm size, boat ownership and credit access were responsible for the livelihood improvement of haor fishermen. Furthermore, this study also found out constraints which were faced by the fishermen. Among all the constraints, the flash flood was reported to as a major problem by the fishermen.


Introduction
Within fisheries management and development policy, the importance of sustaining small-scale fisheries is increasingly recognized (Allison, 2001). In Bangladesh, fisheries sector, both inland and marine has a significant prospective to make generous contribution to national socio-economic development, economic revitalization and poverty reduction (Salagrama, 2006a). Fisheries products as an effective cash crop, has more potential to generate cash income comparable to agricultural products (Bene et al., 2007). This cash crop nature of fisheries product acts as a strong market stimulator and wealth generation with multiplier affect which continuously providing broader income and employment opportunity (FAO 2004a). This sector incorporates a diverse range of livelihood activities, 15.0% of total global employment from production and processing to marketing of fish product and ancillary functions (FAO, 2006) but many of the people engaged in these activities remain unrecognized as fish workers.
The fisheries sector plays a significant role in the national economy of Bangladesh contributing 3.69% of GDP and 22.60% to the agricultural GDP (FRSS, 2016). Additionally, in Bangladesh more than 17 million (including 1.4 million women) depend on fisheries sector (NFW, 2015) and another 11 million people are engaged in other related activities such as fish fry production, aquaculture and enhanced fisheries, fish trading, dry fish processing net/trap and boat making, fisheries labor, etc. especially the people of coastal and haor (Low Laying Wet Land) region of Bangladesh (Thilsted, 2014). Smallscale fishing is characterized by fishing craft, with nonmechanized force or low-horsepower outboard or inboard engines; use of passive fishing methods, manual operation of fishing gear and the absence of electronic fish finding and navigational devices (FAO, 2004b). It also frequently characterized as "the occupation of the last resort" (Smith, 1979;Panayotou, 1982;Christy, 1986). In particular, additional fishing gear and improved infrastructure are the key factors to enhanced productivity which would leads to improve wellbeing through income generation, reduce poverty and ensure food security (Bailey and Jentoft, 1990). These fishing operations are very common in haor area. The solution of poverty reduction have centered on the necessity to make small scale fisheries more economically efficient (Edward et al., 2001). However, small-scale fisheries are often neglected in development planning because their contribution does not take into account in socioeconomic influence (Thorpe et al., 2005). There are 96 haors covering an area of 1, 92,367 hectares located here and there, mostly lie in the district of Kishoregonj, Netrakona, Kushtia, Habigonj, Sunamganj, Moulvibazar and Sylhet district of Bangladesh (Minar et al., 2013) which has considerable economic and aesthetic value, greatly influencing the ultimate environment quality in a diversified way (Hossain, 2014). Haor is a highly productive natural source of livelihoods that support millions of poor people and plays a crucial role in supplying protein (FAO, 2010). Particularly, fishing communities secure their livelihoods from haor by capturing fish, fish trading, fish drying, aquatic life and net weaving (Iqbal et al., 2015). Notably, fishing community who are living hand to mouth are considered as the poorest of the poor (Kabir et al., 2012). Being an isolated community, these people are deprived of many amenities of life (Alam, 2010). Fishing communities are still the dominant communities of poor people inhabit coastal areas, especially in countries that are developing or third world (FAO, 2007). The fishermen of southeastern part of Bangladesh are belonging to hardcore poor (Kleih et al., 2003) and lack of adequate capital is their main constraint (Ali et al., 2008). Moreover, significant research has not yet been conducted on the haor fishermen of north eastern Bangladesh although it has a great ecological, commercial and socio-economic importance in the economy of Bangladesh. The purpose of this study was to document livelihood status of haor fishermen, identifying the factors that are affecting the livelihood improvement and figured out the constraints faced by the fishermen in haor areas of Bangladesh.

Selection of Study Area and Sample Size
Dakshin (South) Sunamganj upazila (Sub-district) under Sunamganj district was purposely selected for the current study. Necessary information of fishermen was collected from three villages of Noapara, Jolklols and kaikker par through random sampling. Total 160 samples were interviewed from Dakshin Sunamganj upazila of which 80 were fishermen and another 80 were nonfishermen for attaining the objectives of the study.

Methods of Data Analysis
The double difference estimator was used to compare the changes in outcomes measured between treated (Fishermen) and controlled (Non-fishermen). During the impact study by Difference in Difference (DID) approach the following formula was used (Duflo et al., 2004). The formula of double difference estimator is DID = {(T 1 -C 1 )-(T 0 -C 0 )}, where, T implies treatment group (Fishermen) and C denotes control group (Non-fishermen). The rows distinguish between before and after the intervention (denoted by subscripts 0 and 1), (Table 1).

Binary Logistic Regression
To determine the factors responsible for livelihood improvement, Binary logistic regression model was used. Binary logistic regression estimates the probability that a characteristic is present (e.g., estimate probability of "success") given the values of explanatory variables, in this case a single categorical variable; π = Pr (Y = 1|X = x) (Gujrati, 2004).

Variables:
• Let Y be a binary response variable Y i = 1 if the trait is present in observation (person, unit, etc.) i Y i = 0 if the trait is not present in observation i • X = (X 1 , X 2 ,...,X k ) be a set of explanatory variables which can be discrete, continuous, or a combination.
x i is the observed value of the explanatory variables for observation i

Assumptions
The data Y 1 , Y 2 ,...,Y n are independently distributed, i.e., cases are independent. Distribution of Y i is Bin (n i , π i ), i.e., binary logistic regression model assumes binomial distribution of the response. The dependent variable does not need to be normally distributed, but it typically assumes a distribution from an exponential family (e.g., binomial, poisson, multinomial, normal). Does not assume a linear relationship between the dependent variable and the independent variables, but it does assume linear relationship between the logit of the response and the explanatory variables; logit (π) β 0 + β X . Independent (explanatory) variables can be even the power terms or some other nonlinear transformations of the original independent variables. The homogeneity of variance does not need to be satisfied. In fact, it is not even possible in many cases given the model structure. Errors need to be independent but not normally distributed. It uses Maximum Likelihood Estimation (MLE) rather than Ordinary Least Squares (OLS) to estimate the parameters and thus relies on large-sample approximations. Goodness-of-fit measures rely on sufficiently large samples, where a heuristic rule is that not more than 20% of the expected cells counts are less than 5 nonlinear transformations of the original independent variables. In this study, the outcome variable was dichotomous and there were a significant number of independent variables. In such case, the appropriate model was binary logistic regression model. The dependent variable was the livelihood improvement, which was categorized into two groups. Scores assigned as 1 and 0 if the response is "Yes" and "No" respectively. Ten variables were identified to be the major explanatory variables in this study area. These were family/household type, farm size, boat ownership, loan/credit accesses, age of the respondent, educational level, family size, fishing experiences, time of fishing and contact sell.

Constraint Facing Index (CFI)
The researcher identified the major problems faced by the fishermen in study area. An overall score of the problems faced fishermen were computed by adding their scores of the problems in all 9 selected problems. Each fishermen were asked to indicate the extent of difficulty caused by each of the problems by checking any of the four responses such as 'high', 'medium', 'low' and 'not at all' and weights were assigned to these responses as 3, 2, 1 and 0, respectively. The scores of Constraint Facing Index (CFI) for each selected problem were computed through using the subsequent formula: CFI = (C h ×3) + (C m ×2) + (C l ×1) + (C n ×0) Where CFI = Constraints Facing Index C h = Number of respondents having high constraints; C m = Number of respondents having medium constraints; C l = Number of respondents having low constraints; and C n = Number of respondents having no constraints. The problems were ranked according their CFI score which denoted their severity in fishing in haor area.

Socio-Economic Profile of the Respondents
The basic information about the fishermen is represented in Table 2. It is seen that majority (89.0%) of fishermen responded to the survey is middle aged, ranging from 15 to 50 years old (Table 2). Relative high percentages (60.0%) of fishermen have medium family size living with joint family type (53.80%). Results shows that a high percentage of fishermen were illiterate (56.3%). Most of the respondent (87.5%) reported that they have no savings with annual income ranging from $621.0 to $915.0 which goes to the medium income group (Table 2). Fishermen (70.0%) in the study area are willing to catch fish in open water through current net as it is easy to operate with low maintenance cost and good harvesting record. The secondary occupations are negligible. There was no definite occupation in the dry period. Fishing is their primary occupations where highest numbers, 45% household heads had agriculture labor as secondary occupation and the second highest (22.5%) were having a piece of land for cultivation confirming the previous findings regarding socio economic condition of haor people (Sarif et al., 2016). Interestingly, about three quarters (75%) of the fishermen use sanitary latrine and it came possible for the willingness of different NGO's such as ASA, HILIP and fishermen organization in the study area.

Impact of Small Scale Fishing on Livelihood Improvement
The asset pentagon is an imperative component of sustainable livelihoods framework developed by DFID represents the inter-relationships among various asset of individual and group of a society. A change in asset status i.e., increases or decrease in access to livelihood assets may indicate improvement or no improvement of livelihood (Darwis et al., 2015). The key trends affecting the livelihoods of the poor in the haor fishing communities in Sunamganj district range across the whole spectrum of "assets" -i.e., the natural, physical, social, human and financial -and contribute to changes in terms of availability as well as access to the assets for the poorer stakeholders is a measure of livelihood improvement. The overall wellbeing of haor fishermen is associated with different types of livelihood assets as shown in Table 3. The results confirmed that, natural capital, human capital, physical capital, financial capital and social capital were increased by 47.5%, 75.33%, 73.3%, 76.3% and 62.5%, respectively. The obvious reason for highest percentages in financial capital is fishing; an important economic and business activity solely in haor areas. The lowest percentages were found in case of natural capital (47.55%) due to poor conservation method by the community people. The percentage of earthen and straw roof was decreased by 58.8 and 72.5 percent, respectively. This simultaneous trend indicates improving housing condition for all fishermen. After involving in fishing, about 88.8 and 68.8 percent of fishermen are capable of using mobile phone and toilet, respectively. Many fishermen were using modern amenities too. Uses of radio, watch and bicycle have increased tremendously for all the fishermen. Table 3 revealed that 65, 70 and 52.5 percent fishermen reported that their decision-making ability, women empowerment and participation in social activities were increased. In the present study, it was found that more organizations are now formally or informally working than before in the study areas to promote cooperation between people, coping distress and other awareness build up process.

Changes in Overall Livelihood Asset (Capital) by Fishermen
Overall change of assets is built through five core livelihood assets. A mixture of transforming structures and processes among these assets helps to obtain desirable livelihood outputs. Radar diagrams made of livelihood assets provide overall understanding of fishing impact and its resource endowments and sustainability ( Fig. 1-3). Livelihood assets and their variables have been scaled up to the boundary of these diagrams. Full access to or highest performing variables or assets assume periphery and no access to or lowest performing variables or assets assume to the center of these diagram. Thus, higher degree of robustness of the diagram indicates higher impact on livelihood and its capabilities. The development process of Bangladesh is closely linked with the development of haor area. The changes in overall Livelihood Asset (Capital) by fishermen are shown in the Table 4. Table 4 shows, overall livelihood situation of natural, financial, human, social and physical assets of fishermen and non-fishermen whether these were increased, decreased or remained unchanged. The highest increased responding fishermen were found in case of financial capital (32% to 76.67%) and for nonfishermen it was calculated 32% to 55% and lowest increased responding fishermen were found in case of natural capital 40% to 47.5% and for non-fishermen it was calculated 13% to 10%. It is because; fishing is one addition source of income for fishermen compared to non-fishermen. In case of human capital responding increased from 45% to 75.33% for fishermen, 36% to 56% for non-fishermen, 42% to 62.5% social capital for fishermen, 40% to 53% for non-fishermen and physical capital for fishermen 30% to 73.37% for fishermen, 22% to 45% for non-fishermen, respectively ( Fig. 1). The highest percentage of unchanged capital responding fishermen were found in case of natural capital 18% to 22.93% and for non-fishermen it was also calculated for same capital 56% to 59% (Fig. 2). The highest decreased responding fishermen were found in case of human capital (26% to 5.03%) and for non-fishermen it was calculated in case of financial capital 24% to 23% and lowest decreased responding fishermen were found in case of natural capital 25% to 29.6% and for non-fishermen it was calculated in case of physical capital 37% to 36%. The lowest decreased for fishermen were found because fishermen in haor area can diversify their livelihood easily compared to non-fishermen in this area (Fig. 3).

Determinants of Livelihood Improvement in Haor Area
A binary logistic regression model was fitted to elicit the factors influencing the livelihood status of haor household. Ten variables were identified to be the major explanatory variables in this study area. All these factors expected to have positive impact on livelihood status of household. For comparing observed and expected frequencies of events and non-events to assess how well the model fits the data, Hosmer-Lemes how goodnessof-fit test were used. The p-value was greater than 0.005 (Table 5) so we could not reject the null hypothesis. Hence there was no difference between observed and predicted variables values. Finally for summarizing the proportion of variance in the dependent variable associated with the predictor (independent) variables Cox and Snell R 2 along with its correction (Nagelkerke Pseudo-R 2 ) was used which revealed that this model was being able to explain about 70 percent of the variation in the data. The result of binary logistic regression model is presented in Table 6. The result shows that model was suitable for explaining the determinants of livelihood status of haor household. Among all the variables considered in model four variables revealed significant. These variables are family type, farm size, boat ownership and loan access. Respondents belonging to joint family were 0.019 times significantly less likely to improve their livelihood than the respondents from nuclear family. Odd ratio of farm size coefficient is 1.032 indicated that, holding other variables as a fixed value, we will see 3.2% increase in the odds of getting a respondent experienced improved livelihood for a one decimal increase in farm size.  High (3) Medium (2)  Note: CFI score of fishermen (Flash flood) = (55×3) + (16×2) + (6×1) + (3×0) =203 The regression results also suggested that respondents without having own boat were 0.044 times significantly less likely to be in improved livelihood than the respondents having their own boat. In this study, the result revealed that respondents, who were not having loan, were 0.110 times significantly less likely to improve their livelihood than the respondents having loan.

Constraints Faced by the Fishermen
The problems related to fishing were poor communication and transportation facilities, flash flood, lack of marketing facilities, lack of scientific knowledge and technology, theft of fishing gear, low price of fish, lack of capital, high price of fishing gear and lack of institutional credits. In particular the problem rank was made according to the following kruskal wallis (H) test. The test statistic is given by The CFI for this problem was calculated at 196 which ranked as 2 nd problem. Inadequate capital seemed one more problem with CFI score 172 (3 rd rank) for the fishermen followed by 4 th rank problem theft of fishing gear (CFI score 158). Other problems like, lack of transportation and communication facilities, high price of fishing gear, lack of marketing facilities, lack of institutional credits and lack of scientific and technological knowledge were ranked as 5 th (CFI score 140), 6 th (CFI score 123), 7 th (CFI score 110), 8 th (CFI score 107) and 9 th (CFI score 103), respectively.

Discussion
Based on the empirical evidence emanating from the logistic regression, we can opine that livelihood of haor fishermen is improving with the increases in farm size, boat ownership, credit access and breaking of joint family. Joint family might increase the expenditure of family and thus resulting with lots of constraints to improve their livelihood comparing with nuclear family. Furthermore, within small families there is a marked preference towards shifting children from fishing into other occupations-preferably service-oriented. This arises from recognition of un-sustainability of fishing as livelihood, as well as from a desire for the upward mobility that white collar-employment is supposed to bestow. Once a family moves away from a primary sector livelihood based on an open-access regime, the importance of large family diminishes. Families that have more children-due to lack of awareness or religious/social/cultural reasons but quiet for economic reasons are generally poor (Salagrama, 2006b). The boat ownership conveys stability and helps enable group formation, even for economic migrants, in a way that is lacking for labors. In one village, a group of small boat owners argued "if you own your own boat, machine and gear you cannot be classed as poor anymore" (Rothschild and Beamish, 2009). Furthermore, credit systems in the fishing sector were introduced to support the diffusion of new technologies rather than to support and encourage existing, more equitable system of operation (Salagrama, 2006c). Within this broader framework, particular attention has been placed on one of the five assets identified as constitutive of livelihood strategies: Financial capital as the impact analysis revealed highest increase in financial capital along with highest decrease in human capital by the dint of fishing. The haor fishing community's standard of living has improved in terms of per head annual earnings and savings after meeting expenditure requirement from fishing. Additionally, the Socio-economic analysis displays that more than half of the respondent (56.3%) is illiterate and most of them (91.25%) had no training in the selected research area. Hence, basic education and training program on processing of the fish product, extraction of oil from dry fish might help to improve the livelihood of haor fishermen.