Livelihood strategies of farmers in Bolivar, Ecuador: asset distribution, activity selection and income generation decisions in rural households
Households in rural Ecuador face several challenges. One of them is the severe deprivation that reaches alarming percentages in the countryside. Unequal distribution and limited assets constrain households from improving their economic conditions. These factors induce households to overexploit natural resources. Poor households engage in a variety of livelihood strategies. Livelihood strategies are characterized by the allocation of assets (natural, physical, financial, public, social and human), income-earning activities (on farm, off farm), and outcomes (food, income, security). Together these determine the well-being attained by an individual or households. We used data collected by INIAP as part of the SANREM-CRSP project to identify livelihood strategies, their determinants, and well-being implications of adopting a particular livelihood. These data were from a comprehensive survey of 286 households collected during September and November, 2006. Livelihood strategies for the Chimbo watershed were identified using qualitative and quantitative methods. The methods provide similar results and identified four main livelihoods: households engaged in diversified activities, agricultural markets, non-farm activities, and agricultural wage work. Most households are engaged in agricultural markets followed by households in diversified activities. Households engaged in agricultural markets own higher amounts of natural and physical resources, while households engaged in non-farm activities have, on average, more human capital. Households participating in agricultural wage work are mainly from the down-stream watershed and posses less natural, physical and human assets. Factors influencing the selection of livelihood strategies were examined using a multinomial logit model. Variables such as access to irrigation, amount of farm surface and value of physical assets were statistically significant determinants of livelihood selection. Households with higher endowments of natural and physical assets are more likely to engage in agricultural markets and less likely to participate in non-farm activities. Secondary education tends to decrease participation in the agricultural sector while increasing engagement in non-farm activities. Several geographic variables like watershed location, altitude, and distance to rivers and cities are statistically significant determinants of livelihood strategies. The well-being associated with each livelihood strategy was estimated using least squares corrected for selection bias. Since participation in each livelihood is endogenously selected it was necessary to correct for selection. We use the Dubin- McFadden (1984) correction, based on the multinomial logit model. In our models of well-being few variables were statistically significant; this may be due to data limitations. Credit is statistically significant and has a positive effect on wellbeing. A similar positive effect is shown by education but the variable is not statistically significant. An odd result was found in the coefficient of irrigation access. This coefficient appears to decrease household well-being for those engaged in agricultural markets. This result is hard to explain, as we would expect that irrigation would be positively associated with well-being. The lack of access to water in irrigation systems in the region (noted by many respondents) might explain this negative effect. Most households that access irrigation do not have enough water, and access to irrigation does not provide the advantages that it might otherwise. The selection models were used to estimate the amount of well-being that households currently engaged in other livelihoods might receive if they selected a different livelihood. For example, what level of wellbeing would be attained by households currently engaged in agricultural markets if they instead engaged in non-farm activities. Results indicate that most households might achieve higher well-being if they engaged in non-farm activities. However households that want to engage in this sector require special skills or assets that are not easy to obtain; thus there are constraining barriers to diversification in the watershed. Several policy changes were simulated to determine their impacts on livelihood choice and household well-being. First a policy change that provides wider education to households in the region was assumed, with more education livelihood strategy selection moves towards the non-farm sector and away from agricultural wage work. These changes generate positive effects on household well-being. The second policy change was creating wider access to irrigation. This change moves livelihood strategies towards agricultural production and away from diversification and non-farm activities, and it had the effect of decreasing household well-being. This was unexpected but it is explained by the negative coefficient of irrigation access in the well-being model. These two policy changes were made to variables that are not statistically significant determinants in the well-being models but were highly significant determinants of livelihood strategies. The third and final policy was wider access to formal credit. Although credit is not a variable that affects the selection of livelihood strategies, it has an important effect on well-being. This policy change generates the highest increment in average well-being. However even though credit is available, if it is not used for productive purposes, it might represent an unnecessary cost for the households instead of being beneficial.