This dissertation addresses the empirical issues pertaining to technology adoption decisions, agricultural commodity price volatility and the effects of remittances on recipient households combined with the motivation of migration decisions in low-income countries such as Nepal under the theories of incomplete and imperfect markets. This dissertation contains three substantive essays applying a number of econometric models to test a number of the hypotheses using both panel and cross-section data from the Ne-pal Living Standard Surveys and time series data for commodity prices and farm yields. Summaries of these essays are presented as follows.
The first paper examines factors affecting the adoption of improved seeds and in-organic fertilizers. I consider the adoptions of both these technologies as a joint decision and estimate over two repeated cross-section data from NLSSs. Both probit GMM with the moment restrictions and Linear Probability Models for period 2 (2004) combined with reduced form probit models for both periods and Tobit models were applied to con-trol for plot level, household characteristics, and other factors. The result weakly favours the hypothesis of joint decision. The results show significant effects on adoption decisions for farm technologies from four variables: the factor markets for credit and for labour, agricultural extension services, and household labour endowment. Proximity to road transport and access to markets also increase the adoption rate of improved seeds and inorganic fertilizers. Positive effects were associated with the increasing age and education of household heads with some exceptions. The results from Tobit models were also consistent with the reduced form and structural models with some exceptions. Well-functioning factor markets and well-developed infrastructure emerge as the precondition for agricultural-led growth in Nepal.
The second paper explores how price shocks affect the stability of farmers’ in-come at different levels across different regions of Nepal, using a recent theoretical model that allows examination of the household income variance through combination of household data sets with price and yield time series under the scenarios of actual, full and no exposure to Indian markets. Agricultural income variability is found to be higher among the farmers with higher share of agricultural products (more than 65 percent) in the total household income, followed by 30 to 65 percent share of agricultural products. The results show relatively high income variability in the poor than the non-poor farm households, but their difference is low. The increased income variability of agricultural households, observed in almost all belts and regions, and at all income levels, is attributable to the domestic shocks. In general, the degree of market integration with Indian prices seems to be widely affected by the geographical heterogeneity in Nepal. Granger-causality tests show a higher integration between border markets of both countries, revealing that Nepalese commodity prices follow Indian prices with the exception of some commodities in some border markets.
Finally, the third paper analyses the effect of remittance income on the hours of work in remittance-receiving households using panel data from the Nepal Living Standard Surveys. The study applies a number of econometric models to explain the impact of remittance income on the hours of work in different sectors (i.e. on-farm, self-employment, off-farm and hired labour) taking into account various methodological is-sues (endogeneity and selection bias) for migration decision and remittances. I first use a Zero Inflated Poisson model to examine the factors motivating migration. I then apply random effects model and instrumental variable Tobit models for estimating the impact of remittances on the household work hours both for different sectors and separately for working age men and women. Evidence shows that rural people with larger family size and higher per capita income without remittances have higher probability to go migrate. Remittances decrease work hours in a number of sectors, but increases work hours of hired labour in remittance-receiving households. Remittance income seems to be a substitute of non-labour income for remittance-receiving households. No significant effects on off-farm and self-employment activities were observed in the sample households. In contrast, non labour income appears to increase work hours of household members. Moreover, demographic characteristics seem to be an influential factor for the allocation of household work hours, implying that higher family size leads to higher work hours, and a larger number of children leads to a reduction of work hours of females, but not of males. Educated people are also more likely to increase their work hours.