摘要：Banana production provides suitable options for subsistence and income generation in the mid and high elevation areas of East Africa, including Uganda. Limited access to factor markets (labour, land and credit), as well as critical biophysical factors (pests, diseases and soil degradation) have led to the decline of banana production in central Uganda and its rise in the southwest of the country. We formulate a farm production model to analyze farm household behavior in developing countries regarding resource allocation to crop production with specific reference to banana production. Findings have implications for policies to support sustainable agricultural production and growth, contributing to on-going debates about the separability of consumption and production decisions in developing economies and the response of poor households to price incentives. Our adapted model considers the non-separability of household production and consumption decisions. Perfect market conditions rarely exist in developing countries because of limited access to credit and the seasonal nature of crop production. Households often fail to satisfy annual cash income constraints, their expenditures exceeding revenue at certain periods of the year. We estimated a production function econometrically with a double log functional form to analyze output response to input use. A reduced form equation of labour demand was estimated to analyse the determinants of farmers' investment in banana production. Primary data was generated through a random sample that includes 660 households of which 533 were used. The sample was drawn from 33 villages located in major banana production systems, stratified by elevation and previous exposure to new technology. Elevation is highly correlated with differences in farm and biophysical characteristics such as soil fertility, incidence of pests and plant disease. "Exposure" captures the village-level effect on household decision-making of previous technology releases. Results from the production function showed positive and significant relationship between banana production and elevation, crop sanitation labour and natural pasture. Education of household head was negatively related to output, implying that improvements in education results to a withdraw labour from agriculture to other activities. Labour use in cooking banana responded negatively to wage rate but response to out price was not significant. Nonfarm self-employment was negatively related to labour use in cooking bananas implying withdraw of family labour from farm production to non-farm production. There was a negative relationship between distance to paved roads and labour use, which implies higher transaction costs for farmers staying far away from improved road network. Education of housewife was positively related to labour used in banana production in low altitude areas but not significant for high altitude areas implying that women have a big role in decisions regarding food crop production. Investment in education (improving farming skills) of women might increase food security in low input agricultural areas. The joint effect of household characteristics on labour use (output supply) was significant implying that the separability condition between production and consumption decisions among smallholder producers is not valid. The results indicate that, given the current environment constraints, investment in technology development and dissemination has positive implications for agricultural development in low input systems. Investment in human capital, especially in education of women, and providing an enabling environment for easy access to input markets play major roles in improving agricultural production.
