Md Sdique Rhmn, Mohmmd Miznul Hque Kzl, Shh Johir Ryhn,Shirjum Mnjir
aDepartment of Management and Finance, Sher-e-Bangla Agricultural University, Sher-e-Bangla Nagar, Dhaka, 1207, Bangladesh
bDepartment of Development and Poverty Studies, Sher-e-Bangla Agricultural University, Sher-e-Bangla Nagar, Dhaka, 1207, Bangladesh
Keywords:
Adoption
Aquaculture
Extension
Ordered probit
Productivity
A B S T R A C T
Adopting improved management practices for rearing carp can play a vital role in developing the aquaculture sector of Bangladesh.This study identifies the factors that facilitate the adoption of improved management practices and their relationship with productivity.Primary data was collected from 300 carp farmers through face-to-face interviews.Improved management practices were divided into three categories: simple, intermediate and complex.Findings indicated that high adopters (those who adopted at least one simple, one intermediate and one complex practice) were only around 10% of respondents.The adoption of improved management practices was influenced by spousal education level, training, extension contact, and off-farm income sources.Productivity of high adopters was 38% higher than that of non-adopters.Policy implications from this research included improving extension facilities from public agencies to facilitate adoption of improved practices and to increase and sustain productivity in carp farming.Modifying the current extension approach and diversifying income sources among farmers will facilitate widespread adoption of improved management practices.
Bangladesh freshwater fish production ranked third globally in 2017–2018, with a total production of 4.27 million metric tons, 33.57 percent of which were derived from carp species (DoF, 2018).Carp is the third most important cultured fish species in the world and carp producing countries are mainly located in Asia (Karnai & Szucs, 2018).Bangladesh features one of the largest and most active deltas in the world, fed by three powerful rivers: the Padma, Meghna, and Jamuna.This contributes to a high potential for freshwater culture fisheries,which consists mainly of pond polyculture of carp (Shamsuzzaman &Xu, 2017).In 2017–2018, pond fish production, dominated by carp species, was 1.9 million metric tons, which represents 44.43 percent of total overall production (DoF, 2018).The freshwater aquaculture industry as a whole contributes to the economy with expanding production capacity and high export opportunities.Pond polyculture contributes between 3% and 15% of total family income and 25%–50% of total fish consumption, making it critical for both household income and fish consumption (Belton & Azad, 2012).
Pond polyculture of carp in Bangladesh has been increasingly diversified over the last few decades through the introduction of improved management practices,1Improved management practices include: use of improved breeds of carp fingerlings, maintaining appropriate stocking density, providing food according to body weight of the fish, change pond water when necessary, apply lime according to soil pH, and having a proper drainage facility.developed by researchers at the Bangladesh Fisheries Research Institute (BFRI), to augment productivity and improve the livelihoods of carp farmers.Yet, average carp productivity in ponds is still much lower compared to other carp producing countries (Mohsin et al., 2012).Adoption of improved management practices can increase productivity substantially (Saha & Islam, 2005;Alam, 2011; Kobir et al., 2017; Khan, Guttormsen, & Roll, 2017).Adopting such improved technologies can play a key role in the development of aquaculture (Kumer et al., 2018).
However, the decision to adopt is complex and a number of factors may be influential.The identification of these factors can play a crucial role for further development in the sector.Investment in improving management practices can bring significant benefits to aquaculture producers and consumers (Kumar & Engle, 2016).There are few studies on management practices adoption in the aquaculture sector conducted in and outside of Bangladesh.Obiero et al.(2019) found that diversi fication in on-farm activities, farm size, and training significantly affected aquaculture technology adoption in Kenya.Prodhan and Khan (2018)have identified education and training as key determinants for scientific aquaculture management practices adoption in Bangladesh, suggesting that scientific management practices had an impact on fish productivity.Improving education, extension services, income, and credit facilities could augment the adoption of improved practices in other countries(Ofuoku et al., 2008; Amankwah & Quagrainie, 2019).
Considering the role of pond polyculture of carp in Bangladesh’s economy, it is important to understand the factors that have an effect on the adoption of improved management practices than can, in turn, increase productivity within Bangladesh.Extension activities will take more time and money to reach farmers if characteristics of eventual adopters are not identified.It is therefore important to identify the determinants for efficient use of extension resources.Identifying factors also addresses information gaps for policymakers, which may aid in the continued successful development of carp culture policies in Bangladesh.This study aims to identify the factors affecting adoption of improved management practices and their relationship to productivity in Bangladesh.
2.1.Data sources
Three districts2Administrative unit in Bangladesh.were selected on the basis of their leading contributions to carp production in Bangladesh.From each district, two highly concentrated carp polyculture technology sub-districts were surveyed to collect data.The following formula was used to determine the appropriate sample size for this study (Kanyenji et al., 2020):
wherenois the sample size,z2is 95% confidence interval, p is the estimated proportion of an attribute that is present in the population, q is 1-p and e is the desired precision level.The local fisheries of fice provided a list of carp farming households in each sub-district.From that list, 50 carp farmers were randomly selected for each sub-district.A total of 300 carp farmers were surveyed.Out of the 300 carp farmers, 2 farmers did not provide any production-related data due to severe loss of production and were dropped from the analysis.The whole sample was divided into four groups: low adopters, medium adopters, high adopters,and non-adopters based on the complexity of improved management practices.All 6 improved practices were classified as simple, intermediate, or complex practices (Table 1).Improved breeds of carp fingerling and proper drainage facility were classified as simple practices due to farm-level availability.Intermediate techniques included appropriate stocking density and change of pond water.According to local fisheriesof fices, a good number of farmers were trained on the need of maintaining stocking density and water exchange.These two practices are well-known among carp farmers in the study areas, and they have begun to use them.Feed according to body weight and application of lime according to soil pH were considered as complex practices due to the time and expertise necessary to implement these two practices.A carp farmer was considered as low adopter if he adopted at least one simple practice.The farmer who has adopted at least one simple and one intermediate practice was considered as medium adopter, whereas a farmer was considered as high adopter if he adopted at least one simple,one intermediate, and one complex practice.The farmer who did not adopt any of the practices was considered as non-adopter.
Table 1Improved management practices of carp farming.
Data were collected through face-to-face interviews in 2018.Three enumerators were hired and trained before data collection.Interviews consisted of questions regarding socio-economic characteristics of the respondents, income, and productivity of carp farms.STATA econometric software was used to analyze the data.
2.2.Analytical techniques
The carp farmer’s decision to adopt improved management practices was analyzed using a random utility framework (Abebaw & Haile,2013).Under this framework, preference of a carp farmer among the available alternatives (in our study improved management practices)can be explained by a utility function.Carp farmer will choose the alternative with highest utility.
The selection between the discrete choice and the ordered probability models was difficult, as both models have advantages and disadvantages (Fountas & Anastasopoulas, 2017).Discrete choice models enable the estimation of distinguished sets of explanatory variables for each level of adoption.A number of studies (Bosma et al., 2012; Rahman et al., 2018) used a discrete choice model to identify the factors affecting adoption of a technology.However, a discreate choice model, with values of 0 or 1, is not suitable when the response variable has more than two possible values.Improved management practices are a package of multiple different measures.Thus, labeling two farmers as adopters is inappropriate because they may display varying levels of adoption.
The ordered probability models (ordered probit model) represent the ordinal nature of the data.However, it assumes that the same set of explanatory variables affect all levels of adoption (Abegaz et al., 2014;Forbes & Habib, 2015; Fountas & Anastasopoulos, 2017).The explanation of the intermediate categories in the ordered probit model is influenced by the thresholds which are restricted to befixed across the observations (Fountas & Anastasopoulas, 2017).In order to take this limitation into account, a random parameter ordered probit model was developed to capture unobserved heterogeneity, which was achieved by adding a randomly distributed error term (Shao et al., 2020).In addition, the generalized ordered probit model also relaxes the assumption of the ordered probit model, which assumes that the same set of explanatory variables influences all levels of adoption (Kanyenji et al.,2020).However, it can occasionally predict negative probabilities.The hierarchical ordered probit model accounts for this limitation (Fountas& Anastasopoulas, 2017).However, a hierarchical ordered probit model is proposed for data with group structure and an ordinal response variable.The existence of micro-observations integrated with regional context defines the group structure.Therefore, we could not use the hierarchical ordered probit model in this study.Jiang et al.(2017)suggested a zero-inflated ordered probit approach to modelling the data.Zero-inflated models can be used when the data exhibit a high fraction of observations in the lowest category, but this was not the case in our study, where most of our responses were in medium or high category.
In considering all of these factors, we analyzed the data using an ordered probit model, which permits the response variable to have more than two discrete values (Boz & Akbay, 2005; Moniruzzaman, Rahman& Sujan, 2021).We tested the validity of a parallel regression assumption using the likelihood ratio (LR) test.The findings of the LR test (χ2=6.09, p-value 0.1623) indicated that the ordered probit was more appropriate for our data compared to the generalized ordered probit.The response variable represented ordered responses related to the categories of the improved practices adopted by a farmer.The positive sign of the parameter estimates implies that they have a positive influence on the likelihood of adoption, whilst the negative sign suggests that they have a negative effect on adoption.The response variable was coded as 0 =non-adopters, 1 =low adopters, 2 =medium adopters, and 3 =high adopters.The ordered probit model is expressed as;
whereY＊ is the response variable andμrepresents the threshold values or cutoff points that would indicate the levels of inclination to adopt improved practices.The farmers with similar characteristics were expected to have similar cutoff points (Chen et al., 2002; Maddala, 1983).The log-likelihood function was estimated as:
where,jis the group category (0–3), Φ is the cumulative standard function of a standard normal distribution.Two models have been estimated: Model 1 contains the findings of a traditional ordered probit model using all the explanatory variables.In Model 2, only significant variables from Model 1 were used to check for changes in effect size due to the exclusion of insignificant variables.The marginal effect was calculated for only significant variables.
2.3.Explanatory variables
The selection of the explanatory variables for this study was based on previous studies and priori expectations (Ofuoku et al., 2008; Kumar et al., 2018; Kazal et al., 2020; Rahman et al., 2020).Descriptions of the explanatory variables used in our model are given in Table 2.
Table 2Descriptions of the explanatory variables.
3.1.Descriptive statistics
Table 3 provides a summary of descriptive statistics of the socioeconomic characteristics of the respondents.Selected characteristics are similar in terms of family members, farm size, pond ownership status, and access to credit between adopters and non-adopters.The average schooling year was much higher for adopter groups (around 9 years) compared to non-adopters (6.90).High adopters spent more days in trainings (8.62) relative to other groups, which informed their adoption decision.The adopter groups were significantly distinguishable in terms of extension contact.On average, 82% of the high adopters had contacts with extension staff, while only 35% non-adopters had contacts with extension staff.
Table 3Descriptive statistics of the explanatory variables.
3.2.Adoption status of improved management practices
The highest percentage (54%) of respondents were medium adopters, while only around 10% of the respondents were high adopters,implying that lack of awareness and skills hinders the adoption of improved practices (Table 4).Approximately 7% of the respondents did not adopt any of the improved practices.The findings also indicated that carp farmers are reluctant to adopt all of the recommended practices.
Table 4Adoption status of different improved practices.
3.3.Factors affecting adoption
The results of ordered probit model are presented in Table 5.The model χ2(50) was statistically significant at the 1% level.The calculated cutoff values were positive and statistically significant, indicating that the four groups had a natural order.To test whether or not there was multicollinearity, we estimated the variance inflation factor (VIF).The estimated VIF (1.15) was much lower than the threshold value of 10(Maddala, 1992).Of the 11 explanatory variables, 5 had a positive influence on the adoption decision.Spousal education and training were significant at 1% level, whereas extension contact was significant at 5% level.Societal membership and off-farm income source were significant at 10% level (Table 5).
Table 5Factors affecting adoption decision via ordered probit estimates.
The marginal effect for significant explanatory variables is presented in Table 6.The marginal effect for spousal education indicated that a one-year increase in education of respondent’s spouse would increase the likelihood of being medium and high adopters by 0.1% and 0.8%,respectively.The marginal effect for spousal education also indicated that a one-year increase in education of the spouse would decrease the likelihood of being low and non-adopters by 1.2%, and 0.6%, respectively.Findings also indicated that a single extra day spent in training increases the likelihood of being a high adopter by 0.4%, whereas it decreases the likelihood of being a non-adopter by 0.3%.The marginal effect analysis of extension contact suggests that extension contact increases the likelihood of being a high adopter while decreasing the likelihood of being a non-adopter.Similarly, farmers with off-farm income sources have a higher likelihood of being a high adopter (Table 6).
Table 6Factors affecting adoption via marginal effect.
3.4.Relationship between adoption and productivity
The productivity of high adopters was 38.32% higher than nonadopters (Table 7), supporting previous findings that the adoption of improved management practices increased productivity (Thompson &Ma fimisci, 2014; Prodhan & Khan, 2018).These findings also indicate that farmers who have adopted complex practices have received higher productivity relative to their counterparts.
Table 7Adoption level and productivity of carp.
According to descriptive statistics, adopters of improved management practices received more training and communicate with extension staffs more frequently.According to the findings, a small portion of the farmers adopted complex practices.Earlier studies also suggested that farmers in Bangladesh rarely adopted complex agricultural technologies(Materu et al., 2016; Rahman et al., 2018).More awareness-building programs through the Department of Fisheries (DoF)3Apex public extension body in Bangladesh dealing with aquaculture technologies.and Non-Government Organizations (NGOs) are warranted in order to encourage the adoption of improved management practices.
According to the adoption analysis, education assists farmers in understanding the benefits of a new technology (Alene & Manyong, 2007).Carp farmers who have an educated spouse can confer with them about the advantages and disadvantages of carp polyculture, which can enhance adoption.Carp farmers benefit from training because it broadens their knowledge and encourages them to adopt improved practices.Farmers with more days in training adopt practices more readily.Training is one way of providing farmers with the knowledge they need to improve their farming performance.Adoption of improved practices necessitates a certain amount of technical expertise, and direct interaction with extension agents promotes knowledge acquisition(DeGraft-Johnson et al., 2014; Mensah-Bonsu et al., 2017).Thus, efforts are needed to increase the extension services, such as field days and demonstrations, in order to encourage adoption of these improved practices.Societal membership also has a positive effect on the adoption decision.Farmers who are members of different societal organizations have the privilege of meeting and sharing their knowledge with different types of people, which may increase adoption.Societal membership has had a positive effect on technology adoption in previous instances(Abebaw & Haile, 2013; Gabremichael, 2014).Positive and significant value of off-farm income indicated that adoption of new technology requires additional cost.The farmers who have income sources other than carp culture may use extra earnings to adopt new technologies,which in turn may also increase carp productivity.
The adoption of improved management practices is linked to increased productivity (Belay et al., 2014; Prodhan & Khan, 2018).Higher productivity indicates higher incomes, which may improve the livelihood status of carp farmers in Bangladesh.In comparison to conventional carp farming, polyculture technology combined with improved management practices can give a larger percentage of household income.More fish can be directed to on-farm household consumption if improved management practices are adopted, potentially improving nutritional status (Ahmed & Lorica, 2002).
This study identifies factors that facilitate the adoption of improved carp polyculture management practices in Bangladesh by employing an ordered probit model.The results suggest that carp farmers are reluctant to adopt all the recommended practices, with high adopters comprising only around 10% of respondents.Spousal education, training, extension contact, and off-farm income source played an important part in adoption and the productivity of high adopters was higher than that of nonadopters.The farmers who adopted complex practices (high adopters)received higher productivity than those who adopted simple and intermediate practices.Extension facilities, such as small group discussion,and training, are needed to increase the awareness and, therefore,adoption.This could be achieved by providing in-house training, and recruiting field level extension workers.Providing extension services to the family members, particularly the spouse of the respondent, may also help with widespread adoption of improved management practices.More visits to villages by the extension staff may ease the process of adopting improved management practices.Closer coordination across research, governmental, and non-governmental organizations, as well as farmer-to-farmer visits, might help spread improved management techniques.Diversifying sources of income may also play a decisive role in adoption.
CRediT authorship contribution statement
Md Sadique Rahman: Conceptualization, Data collection, Formal analysis, Writing – original draft, Writing – review & editing.Mohammad Mizanul Haque Kazal: Conceptualization, Data collection,Writing – review & editing, Supervision.Shah Johir Rayhan: Data collection, Writing – review & editing.Shirajum Manjira: Writing –review & editing.
Acknowledgements
The authors appreciate the financial assistance provided by the National Agricultural Technology Programme (NATP) - Phase II Project.Thanks, and appreciation are extended to the respondents and enumerators for their outstanding assistance during data collection.
推荐访问:Management practices Improved