Determinants of Wheat Productivity among Smallholder Farmers in Ngororero District, Rwanda

Purpose: This study investigated the factors influencing wheat productivity in Ngororero district of Rwanda. Wheat is a major food security crop in many countries and has a significant role in poverty reduction, food security and income generation among households in many countries. In Rwanda wheat is a source of food and income and ranks second to maize. Despite its importance, wheat productivity in Rwanda remains low. Methodology: Simple random sampling method was used to select a sample of 100 wheat farmers and an interview schedule was adopted to collect the data from the respondents. The collected data were first entered in MS Excel and then exported to SPSS for editing, coding, classification and then analysed with relevant statistical analysis tools. In addition to SPSS, Frontier 4.1 software was also used to generate outputs necessary for the data analysis. The data were analyzed using a linear regression model and stochastic frontier model. Findings: The ordinary least square estimates revealed that seeds, labour, organic fertilizer and education were significant and influenced wheat productivity among the smallholder farmers in Ngororero district. The maximum likelihood estimates indicated that inputs such as farm size, seed, labour, and organic fertilizer influenced wheat productivity positively while inorganic fertilizer influenced it negatively. Socio-economic factors such as age and farmers' group membership reduced inefficiency while education and household size increased inefficiency among smallholder wheat farmers in Ngororero district. The estimated technical efficiency scores (TE) ranged from 3% to 100% with an average of 49%. Unique Contribution to Theory, Practice and Policy: Since the calculated likelihood ratio value (23.107) was less than the critical value (26.22) of at 12degrees of freedom at 1% level of significance, it means that the null hypothesis is accepted and all the coefficients of the second order in the translog function were equal to zero. Thus, the Cobb-Douglas frontier production function adequately captured the production pattern of wheat farmers in Ngororero district. The estimated value of gamma parameter ( ) which is the ratio of the variance output to variance of error was 0.999 and highly significant at 1% level. It is in accordance with the theory that true -value should be greater than zero. The estimated value of is significantly different from zero indicating that random error is playing a significant role in explaining the variation in wheat production and this is evident especially in case of agriculture where uncertainty is assumed to be a main source of variation. However, it should be noted that 99.9 percent variation in output was due to differences in technical inefficiency and the remaining 0.1 percent was due to stochastic random error. The likelihood ratio test is significant at 1% implying that the inefficiency effects are highly significant in the stochastic frontier model. Based on the findings, the study recommends that there is need to enforce and improve mechanisms to avail the required amount of inputs to farmers and at the right time, and to encourage a big number of wheat farmers to join different farmer associations and the cooperatives available in the study area as it was found that being a member of a group enhanced technical efficiency and therefore enhanced the productivity capacity of the farmers through acquisition of inputs such as fertilizer or certified seeds.