Search

Poverty and health are two topics in the field of development economics that are of critical importance to both researchers and policy-makers. Despite advances in poverty alleviation and gains in health outcomes in many developing countries, many challenges remain. Two of these challenges include accurately measuring poverty and improving the quality of health care delivery systems. In this dissertation, I present three essays with theoretical, empirical, and policy-relevant insights into these two challenges. The first essay addresses the issue of accurate poverty measurement by developing a new asset index that captures long run household economic well-being. The accurate measurement of household well-being is necessary for measuring poverty levels and targeting poverty programs. However, since standard expenditure aggregates are costly to collect, relative well-being in developing countries is often measured using asset indices based on durable goods ownership. Although various methods exist to generate proxies for economic well-being (e.g., principal component analysis), the underlying theories associated with these methods have not been formalized. This makes it difficult to interpret the economic meaning of the resulting indices and can lead to inaccurate targeting and evaluation. In this paper, I develop a new asset index, the utility index, by modeling and structurally estimating household preferences over discrete assets. By drawing from economic theory, the utility index can be more directly interpreted as capturing long run household well-being. In contrast to existing asset indices, the utility index incorporates additional information on prices, demographics, and spatial and temporal variation and can therefore be used for policy simulations that are not otherwise possible. After developing the theoretical model, I describe a strategy to construct the utility index by structurally estimating the marginal utility associated with each asset. I then demonstrate how the utility index can be used by measuring changes in poverty in Nicaragua using data from the Living Standards Measurement Surveys. I also use the model to project changes in poverty under a constant income distribution but changing prices and find that about a third of the poverty decrease measured from 1998 to 2005 can be attributed to decreasing asset prices. In addition, I show through the empirical analysis that traditional asset indices are only moderate approximations for household well-being. Finally, I discuss and demonstrate the distinctions between asset and consumption measures, which point to the complementary nature of the two strands of measurement.The second essay presents an alternate approach for improving accurate poverty measurement in developing countries. Although the utility index developed in the first essay presents a method for measuring long run economic well-being, complementary measures of short run welfare are necessary for identifying households which are vulnerable to falling into transitory poverty. Again, given the expenses associated with collecting full consumption data, researchers have developed methods to construct wealth indices based on dichotomous asset and consumption indicators. This work provides guidance on generating such indices by comparing across various methods of construction and variable choices. Specifically, we assess the performance of alternate indices using data from the Living Standards Measurement Surveys in five countries in Sub-Saharan Africa--Ghana, Rwanda, Uganda, Tanzania, and Malawi. We compare indices against a benchmark of household per capita expenditure according to three criteria: rank correlation coefficients, sensitivity to identifying poor households, and accuracy of classifying households as poor or non-poor. Comparing across construction methods, we find that indices generated using principal components analysis correspond most closely with expenditure, though variation across construction methods is small. Comparing across variable inclusion groups, we find that indices generated using a combination of indicators drawn from the categories of staple food consumption, other food consumption, housing quality, semi-durables expenditure, and durables ownership tend to outperform indices generated using variables from only one or two categories. We also assess the various indices in urban and rural subsamples and in analyses of repeated cross-sections and find that index performance is similar to what we find in national, single wave analyses.The third essay turns to the challenge of improving the quality of health care delivery systems by looking at provider investment decisions. Pay-for-performance (P4P) programs, which aim to increase health service provision and quality using financial incentives, have been recently introduced in a number of developing countries. P4P programs contract directly on outputs without specifying the mechanisms for improvements, allowing providers to innovate and modify different aspects of health care delivery as needed. Characterizing these provider responses can help to identify successful mechanisms for quality improvement and enhance our understanding of the links between P4P and overall health systems strengthening. In this paper, we examine provider input responses to the Rwandan P4P program using facility-level data from the 2007 Demographic and Health Survey Service Provision Assessment (SPA) collected after the randomized program rollout to a subset of districts. We focus on facility-level incentives for institutional deliveries, which, as documented in earlier research, resulted in higher institutional delivery rates. Using the SPA facility data, we find that the program's effect on institutional delivery rates is comparable to results in previous studies that used household surveys. Comparing system inputs, we find positive treatment effects for a general management indicator and the daily presence of staff per capita providing maternity-related services. There are no differences in other delivery-specific and general health care delivery inputs. Additionally, we perform a mediation analysis to assess the link between inputs and outcomes and find that management and staffing differences explain a relatively small fraction of the P4P effect on institutional delivery rates. The small mediation effects indicate the potential importance of unobserved factors, such as recruitment effort, in the provider production function. Furthermore, the null results for the other analyzed inputs suggest a weaker link between P4P and overall health system strengthening.

The Aamjiwnaang First Nation is an Ojibwe Tribe at the junction of Lake Huron and the St. Clair River, and is surrounded by 58 petrochemical facilities collectively known as Canada’s Chemical Valley. To increase understanding of chemical contamination and exposures, this dissertation was conducted in partnership with the Aamjiwnaang First Nation Health and Environment Committee. Aim One was to study metals contamination of stream ecosystems on-Reserve (n=4), within the surrounding Chemical Valley (n=14), and a reference community (n=4) during three seasons (2010-2011). Aluminum (3-fold), cadmium (20-fold), lead (2-fold), and cobalt (1.5-fold) levels in streams on-Reserve were greater than the reference community (as indicated in brackets). Many sites were above screening levels for aluminum. Aim Two focused on mercury, motivated by knowledge of existing legacy and active mercury sources in the region. Mercury concentrations of sediments and soils were assessed at stream sites. Levels on-Reserve were higher than the reference community (soil=2.5-fold; sediment=6-fold). One on-Reserve site exceeded screening values in sediment and soil. Human exposures were assessed in 43 mother-child pairs. Mean hair (mothers=242.1±188.9 µg/kg; children=126.7±112.9 µg/kg) and blood levels (1.7±2.0 µg/L; 1.5±1.9 µg/L) were lower than the general public. This is likely due to low amounts of fish consumption. Urinary mercury (0.8±0.9 µg/L; 0.5±0.9 µg/L) exceeded values seen nationally. Aim Three was to increase understanding of multiple chemical exposures among Aamjiwnaang mothers and children. Blood, serum, and urine were used as biomarkers of metals (n=9 elements), perfluorinated compound (n=6 congeners), polycyclic aromatic hydrocarbon (n=12), organochlorine pesticide (n=2), polychlorinated biphenyl (n=5), and polybrominated biphenyl ether (n=5) exposures. Distributions of multiple chemicals were elevated in Aamjiwnaang participants. The outcome of this dissertation significantly increased understanding of multiple chemical exposures for the Aamjiwnaang First Nation. The disproportionate placement of industrialization depicts an environmental justice issue seen among many Tribes. The data will fill gaps in the literature, since First Nations Tribes are under-represented in environmental health research. In identifying key chemicals, future studies can be molded to better serve the community. The data will empower the Aamjiwnaang First Nation in taking public health actions to reduce exposures and improve health.

ReFace (Reality Enhancement Facial Approximation by Computational Estimation) is a prototype facial approximation software program developed by the Federal Bureau of Investigation (FBI) in conjunction with GE Global Research. The prototype extrapolates an “approximation” of a face from a skull using a database of computed tomography (CT) scans of living individuals. The test set consisted of CT scans of 53 articulated human skulls from the William M. Bass Donated Skeletal Collection and the William M. Bass Forensic Skeletal Collection, which are curated at the University of Tennessee in Knoxville. Through the Federal Bureau of Investigation’s Visiting Scientist Program, an educational opportunity administered by the Oak Ridge Institute of Science and Education (ORISE), the researcher conducted an independent validation of this software in two phases. Phase 1 tested and evaluated the software performance, resulting in improvements to the software and the development of standardized protocol for articulation, packaging, and preparation of human skulls for CT scans. Phase 2 validated the accuracy of the software in the production of facial approximations from human skulls using face pools and resemblance ratings. In Phase 2, computerized facial approximations were visually compared with antemortem photographs by four participant groups (N = 103). Ten test subjects of European ancestry (six females and four males) were selected for a photographic validation by face pool and resemblance rating validation tests. Participants were asked to choose the face pool photograph that most closely resembled the facial approximation produced by ReFace. In the second test, the same volunteers were asked to rate (on a scale of 1 to 5) how closely ReFace facial approximations of target subjects resembled an antemortem photograph. In the Face Pool Validation Test, nine out of ten target subjects were correctly identified above random chance, and the frequency distribution was statistically above chance expectations for nine out of ten target subjects (p < .01). The mean hit rate for all subjects was 24% (10% above random chance). There were no significant differences in the hit rates between male participants (67%) and females participants (33%), or between participant groups. All participants were non-experts. Male target subjects received higher numbers of correct responses than female target subjects. The overall ratings for the Resemblance Rating Validation Test were 13% none, 24% slight, 22% approximate, 25% close, and 16% strong. The majority of subjects were rated as close resemblance (six subjects), strong resemblance (one subject), approximate resemblance (one subject), and slight resemblance (one subject). The foil comparison received an equal number of ratings for no resemblance (30.5%) and slight resemblance (30.5%).