Moniruzzaman
Senior Scientific Officer
Agricultural Economics Division, BARI, Gazipur
M. A. Monayem Miah
Senior Scientific Officer
Agricultural Economics Division, BARI, Gazipur
Sahadat Hossain
Chief Scientific Officer and Head
Agricultural Economics Division, BARI, Gazipur
Rural Household, Food Security, livesock, borrow money, fisherman, Salinity problem
Khulna, Satkhira, Bagherhat, Bhola, Patuakhali, Barguna, Noakhali, Laxmipur and Cox’s Bazar district
Food Safety and Security
Price behavior
Study area The study was conducted three districts namely Khulna, Bagherhat and Satkhira based on the highest coastal areas of Bangladesh during the period 2011-2012. To address the food security and an indicator of environmental sustainability of the coastal region, a total of 9 (nine) upazilas of the coastal region in the three districts of Bangladesh were selected. Sampling technique and Sample size The sample of rural household’s was selected randomly for this study. A total sample size of 450 household’s was selected for this study. Among the 450 household’s 150 respondents was collected from each district. Analytical Techniques After collecting the first hand information from the study areas, data were edited, summarized and tabulated for satisfying the objectives of the study. Tabulated data were analyzed and condensed by using averages, percentages, combination etc to attain the results. Estimation of calorie and nutrition intake: The crops, animal products and other food items consumed from own production and those purchased from market by the sample households were taken into consideration for estimating the per capita per day calorie and nutrient intake of the coastal households of the study areas. For this purpose, household consumption data for the last three days were collected through interviewing female household members. In total, data on eighteen types of food items were collected and considered them for analysis. The quantities of crops, animal products and other food items produced and purchased in kilogram were recorded and calculated for the energy and nutrient values (i.e. protein, calcium, iron, and fat). Then this value is divided by the adjusted household size to obtain the calorie and nutrient intake per capita per day by a household member. Irrespective of male and female, two children under six years of old were considered one adult member in this study (Omotesho et al., 2006). The tables of nutrient composition of Bangladeshi foods (Darnton-Hill, et al., 1988) were used to calculate the calorie and nutrient values of the foods. Based on observation, a certain percentage was deducted from each of the produced and purchased food item in calculating the actual edible part. The deducted percentages were 20% for fish and papaya, 25% for orange and bitter plum (jujube), 10% for sweet gourd (yellow pumpkin) and bottle gourd, 5% for potato, brinjal (egg plant), cauliflower, cabbage, leafy vegetables and plantain stem. Construction of food security index: In order to measure food security, a household food security index was constructed by defining a minimum level of nutrition necessary to maintain a healthy living. It also indicates the ‘food security line’ for the population under study (Omotesho et al., 2006). Any household above this line was classified as food-secured. The food security line used in this study was measured using average recommended level of calorie intake of 2400 kcal as the desirable and cut off point (FAO, 2002). A similar approach was used by Olayemi (1998) which was 2260 kcal as a daily recommended level of calorie intake. The calorie content of both the produced and purchased food items were used to estimate the dietary energy availability in the household. The food security index was calculated using the following formula. Where X/Z=K0 Thus, for a household to be food secured K0 must be greater than or equal to one (Ko ≥1) otherwise, the household is food-insecure. Determinants of household level food security: The logit regression model was used to identify the determinants of food security among the coastal people of study areas. The logit regression model is one of the binary choice regression models in which a dichotomous regression variable is considered as the dependent variable. The logit model was chosen for this study instead of the linear probability and probit models because according to Gujarati (1995), the logit model guarantees that the estimated probabilities lie in between 0-1 and that they are not linearly related to the explanatory variables. This is an advantage over the linear probability model. In addition, it is easier and more convenient to compute than the probit model. The logit model is based on the cumulative logistic distribution function expressed below. Pi = E(Y=1/Xi)= α+βi X (1) For ease of exposition Zi =α + β1X1 + β2X2...................... βnXn Where Pi= Probability of being food –secured. The log of odds ratio or logit (Li) In order to obtain the value of Zi, the likelihood of observing the sample needs to be formed by introducing a dichotomous response variable Yi (dependent variable). The dependent variable is food security. Households whose per capita per day calorie intake were found to be greater than food security line were regarded as being food-secured and were assigned a value of 1, while households experiencing a calorie intake less than food security line were regarded as food insecure and they were assigned a value of 0. At first nine explanatory variables such as farm land size, adjusted household size, dependency ratio, household crop production, annual household income, off-farm income, input cost, crop diversification index, and education of the household’s head were hypothesized to be major determinants of household food security among the coastal people of Bangladesh. After testing multicolinearity among variables and the level of significance, seven variables were finally included in the model to determine the probability of food security among coastal households. The independent variables are specified as follows: X1 = Farm land size (decimal) X2 = Dependency ratio (No. of non-working children and adult/Household size) X3 = Annual farm income of household in taka X4 = Annual off-farm income of household in taka X5 = Household annual crop from own production, in kg grain equivalent X6 = Annual input cost of production in taka X7 = Educational of the household’s head u= error term of factors determining food security and the elasticity of the probability of food security were estimated based on expressions derived from the logit model as: __ Ep=βi Xi (1-Pi) (4) Where, βi = estimated logit regression coefficient with respect to the ith factor Pi=estimated probability of household food security status. __ Xi= arithmetic mean of household ith attribute Ep= elasticity of probability of food security
750000
Report/Proceedings