Econometric analysis of the effects of livestock market participation on food security and welfare of smallholder farmers' in Ethiopia

In Ethiopia, though reducing poverty and achieving food security is significantly dependent on the development of livestock sector, the contribution of the sector to welfare and food security of the country is limited by limitied commercialization of the livestock sector. Hence, the first objective of this study was examining the factors affecting the decision and level of market participation in the livestock market and its subsequent effect on food security and welfare of households. The nationally representative cross sectional data was used in this study. The double hurdle model was used to examine factors affecting the decision and level of market participation in the livestock market. The examination of the effect of livestock market participation on food security and welfare of the households were executed using endogenous switching regression and the propensity score matching models. The two stage Tobit model was used to examine the effect of the level of market participation on food security and welfare of households. The results of the double hurdle model indicated that gender of household heads, proxies of transaction costs, access to extension services, herd size, and access to credit were significant determinants of the decision to participate in the livestock market. The level of market participation was significantly affected by age of the household head, proxies of transaction costs, access to extension services, and herd size. In propensity score matching model, results of the sensitivity analysis varying between 1.1-1.85 suggested that both observable and non-observable characteristics should be considered. Consequently, the endogenous switching regression model used to account for these problems indicated that the expected household dietary diversity score was 6.985 and 7.661 for participants and non-participants respectively. Similarly, percapita consumption expenditure used as measure of households' welfare increased by 31 % for the market participating households. The gain would be 67 % had the non-participatns made a decision of participating in the market. The results of the endogenous switching probit model also indicated that market participation reduces the probability of self-reported food insecurity by 32 percent for market participating households. The results also show that market participation reduces the probability of a negative change in diet by 49 %for market participant households. Similarly, participating in the livestock market have siginificantly reduced the negative change and the possibility of reducing food intake (71%). Furthermore, the results of two stage Tobit model indicated that, a unit increase in the level of market participation increases the food security and welfare of households by 8 and 2 % respectively. These results showed that market participation and level of market participation are critical for improving the food security and welfare of small-holder farmers. However, transforming the livestock production system into a more sustainable and market-oriented production system is critical to further improve the households' food security and welfare. These transformation of the production system of the households should include the gender dimension as one important policy intervention. These efforts to transform livestock production system into a market-oriented production system should also consider establishing the market place and developing and improving the road infrastructure as an important policy tool. Controlling animal diseases should be another important policy intervention to be used to increase herd size and encourage market orientation of small-scale households. Therefore, the effort to commercialize livestock production system in Ethiopia should integrate an animal health, production and marketing services system aimed at reducing liquidity constraints and improving marketing information access by small-scale households.