Preliminary Study of the Impact of Land Use Change and Incidence of Malaria and Schistosomiasis on Household Productivity in Tigray, Ethiopia by

Several non-government organizations (NGOs) have introduced a major rural development program to increase agricultural productivity in Tigray, Ethiopia, by implementing water resource projects such as microdams and afforestation of surrounding watersheds. A potential negative impact from these projects is the spreading of diseases such as malaria and schistosomiasis from seasonal events to a year round phenomenon. This study analyzes the health effects of microdams and afforested watersheds on household production and consumption relationships. Preliminary results suggest that work time lost to illnesses have a significant impact on labor productivity and household consumption schedules. In addition, household location in relation to microdams and afforested lands may be an important factor in assessing future losses in labor productivity. * Donald L. Grebner is Assistant Professor, Department of Forestry, Forest and Wildlife Research Center, Box 9681, Mississippi State University, Mississippi State, MS 39762. Gregory S. Amacher is Associate Professor, Department of Forestry, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061. William Hyde is Professor, Department of Forestry, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061. Approved for publication as Journal Article No. FO 120 of the Forest and Wildlife Research Center, Mississippi State University. INTRODUCTION Tigray is an arid to semi-arid region in northern Ethiopia characterized by subsistence farm households raising predominantly cereal crops for local consumption. Crop production has declined because of recurrent draughts and serious soil erosion. In addition, the region has experienced an influx of new immigrants from the aftermath of the Ethiopian/Eritrean war. Several non-government organizations (NGOs) worldwide have introduced a major rural development program to improve agricultural productivity in Tigray. The program emphasizes water resource development in general and, specifically, the construction of microdams and afforestation of the watersheds surrounding the microdams. The problems associated with this development program lie in the potential negative effects that accompany the new resource activities. Irrigated agriculture may introduce longer-term salinization problems in the soil. The new water impoundments and afforestation systems may have even more serious side effects on local health. These effects may lead to the enhancement of environmental conditions favoring transmission of malaria and shistosomiasis. At present, outbreaks of malaria and shistosomiasis are seasonal events, but the concern is that these projects may result in endemic disease incidence and, in turn, might negatively affect labor productivity. PROJECT OBJECTIVES There are two main objectives for this study: 1. Measure household production and consumption gains and losses due to the construction of microdams and afforestation of watersheds. 2. Conduct a benefit/cost analysis of public investment in microdams. Only the preliminary results related to the first objective are reported herein. The benefit/cost analysis has not yet been completed. DATA In 1996, an enumerated survey instrument was used to collect data from 731 households across 30 villages in Tigray. Information was collected on production and consumption decisions made during the year by each household sampled. Detailed information was collected on the amount of animal capital, fertilizers, seeds, and labor employed in various production processes. Expenditure and market sales data were also collected. EMPIRICAL MODELS This section outlines the preliminary production and consumption decisions that we investigated. The decisions were simultaneous and non-separable because of the inclusion of resource and health variables (Amacher et al. 1996). All models applied to these decisions have a log-log functional form (Chung 1994). The first models described were the disease incidence functions. Models were estimated using adult male sick time, adult female sick time, and household children sick time as dependent variables. The second set of models specified and estimated were production for households. Production models were described for cereals, vegetables, fuelwood, and agricultural residues. Labor demand models were specified for hired adult males and females, household adult males, females and children by activity type. For instance, labor demand models were described for hired adult males for cereal, vegetable, fuelwood, and agricultural residue production. These models were also log-log in nature. In addition, these models included shadow prices for health care labor and the production of commodities (Jacoby 1993). The consumption model depicts the quantity consumed by a household. This consumption function includes purchased goods and goods that the household collects or produces that are not sold in the market. RESULTS AND DISCUSSION The results presented in this paper are preliminary in nature. Tables 1-4 present a subset of the results estimated for each of the models. The large number of independent variables in the models necessitates that only the most relevant be presented. To simplify the discussion only results for production, consumption, and labor demand of fuelwood will be presented. The disease prevalence equation describes how sick time depends on various market and environmental factors. It is important to determine how these factors affect both male and female household time allocations, because male and female laborers may be engaged in different activities (e.g., consider healthcare for sick children). Table 1 presents parameter estimates for disease incidence for household adult males, females, and children. The first model for male sick time indicates that the dummy variable depicting cases of malaria among adult men is positive, suggesting that a 1% increase in malaria incidence causes a 2% increase in days of work lost by adult men. The coefficient for households with mosquito nets is positive, suggesting that as net use increases by 1%, more males get sick. Although this result seems counter-intuitive, only 5 households over the 731 data points declared use of mosquito nets. This evidence suggests that the sign of the coefficient, despite being significant, is not reliable. The second model in Table 1 for sick time of adult females indicates that the dummy variable depicting cases of malaria among adult women is positive, suggesting that a 1% increase in malaria incidence causes a 1.89% increase in days of work lost by adult females. An important connection exists between household time and the presence of the microdams. The coefficient for the distance from the microdam is negative, implying that as a household is farther away, adult females lose fewer workdays due to illness. The parameter estimates for households using mosquito nets are also positive. As previously mentioned, these variables have few non-zero observations, making the interpretation of these results unreliable. In addition, the coefficient for the distance a household is from its drinking source is positive, suggesting that water access problems for daily use may be an important factor in days lost to illness among adult women. The third model for children sick time shows the dummy variable depicting cases of malaria and schistosomiasis among children is significant. This suggests that both diseases had an important impact on child labor. The coefficient for the distance to health care centers is positive, implying that children in households farther away are more likely to lose working time to illness. As with the sick time model for women, the coefficient for the distance a household is from its drinking source is positive, suggesting once again that water access problems for daily use may be important to days lost to illness among children. Table 1. Preliminary Disease Incidence Results.