博士（経済学） / x, 171 p. / Hitotsubashi University

The landscape of South African National Development Finance Institutions (DFIs) is comprised of twelve entities. Their institutional objectives range from supporting farmers, financing industrialisation, infrastructural development, and promoting financial inclusion. These DFI objectives fall under the umbrella of Private Sector Development (PSD) interventions. Literature established that the success of PSD is contingent on effective impact evaluation. Consequently, the main research question explored in this dissertation is: In what ways, and using what tools and systems, do South African DFIs measure the development impact of their investments? In support of the main question, two sub-questions were are also investigated. Firstly, whether impact evaluation systems provide credible, timely and relevant information. Secondly, whether impact evaluation systems support evidence-based decision making and learning. In response to these questions, a qualitative case study of six National DFIs was carried out. Semi-structured interviews were conducted with DFI staff members involved in impact evaluation. This was supported by secondary data from annual reports and organisational websites. It was established that, firstly, DFIs use non-uniform impact evaluation systems and tools to measure the impact of their investments. Secondly, the systems lack qualitative detail and focus on measuring outputs instead of outcomes. Thus, much emphasis is placed on monitoring instead of impact evaluation. This renders the impact evaluation systems and tools highly ineffective. Finally, whilst the avowed objective of DFIs is development, financial viability takes precedence when selecting projects. Therefore, an emerging conclusion was that systems in place do not support development impact evidence-based decision-making. These findings generated recommendations for changing the development impact evaluation tools and systems used by South African National DFIs. It is expected that recommended changes will maximise DFI socio-economic benefits.

This research evaluates the post-program treatment effects of the Southwest Poverty Reduction Project (SWPRP), a large-scale ($463.55 million) rural development project jointly funded by the World Bank and the Chinese Government from 1995 to 2001. The SWPRP aimed at reducing poverty and increasing living standards for the absolute poor in southwest China. The treatment effects are measured by the changes in 21 indicators at the village level. The dataset for this research includes 327 project villages and 3887 non-project villages in Guangxi Zhuang Autonomous Region. Rigorous econometric methods are employed to remove selection bias. A probit model is established to investigate the selection rule of the project villages. In addition to the control function approach, different methods of propensity score matching such as nearest neighbor, caliper or radius, and kernel-based matching, are used to estimate the treatment effects, including the average treatment effect, average treatment effect on the treated, and average treatment effect on the untreated.The evidence from the treatment effect estimations shows that the SWPRP achieved its overall objective but not necessarily all specific objectives. The evidence supports a statement of significant impacts on farming, offarm employment, and infrastructure by the project investments, while there is no strong evidence to support a conclusion of significant impacts on primary education and rural healthcare services. The poverty rate in the project villages was reduced by about 3.0-3.3 percent and net income increased by about 24-26 Yuan. Further investigation of the specific treatment effects on individual villages expose that the treatment effects vary with land resources in the villages. Lastly, the project was successful in targeting the poorer villages but not necessarily the poorest.This research also reveals some findings of practical relevance for social program design. The approach of integrated policies proves to be effective in large-scale poverty reduction. However, designers should be aware that households may trade off one activity against another to maximize their utility rather than simply follow the whole package of integrated activities. In addition, the minimization of operational costs of the project agents should not be detrimental to the effectiveness of the project.

The two papers in this thesis are aimed at solving problems in Integrated Pest Management project, practice, and program evaluations. In the first paper, an evaluation design is constructed for the Onion ipmPIPE website, an onion pest information website. The Bayesian decision theoretic approach may not accurately model onion growers' pest management decisions throughout the season. Randomization of the treatment is possible, but an incomplete grower list proved to be a problem. The analysis shows that an instrumental variables approach may be the most appropriate method for estimating the impact of the Onion ipmPIPE website because its data needs are solved by using USDA-NASS surveying services. In the second paper, the challenge is to develop a practical method to measure benefits accruing to women from agricultural research using secondary data. Donors, governments and others are interested in determining how benefits from agricultural research accrue to women. We develop a three-step framework to identify women's crops and technologies. In step one, total potential benefits from research are estimated; step two allocates those benefits between men and women; step three, incorporates technology-specific parameters to refine the estimates of potential benefits. We apply this framework to Honduras and find that steps one and two provide the most information on the magnitude and distribution of benefits, but that refinements in step three can affect rankings of research program impacts on women. / Master of Science

No description available.

O objeto deste trabalho é o programa de formação continuada Letra e Vida, oferecido pela Secretaria Estadual de Educação de São Paulo a professores alfabetizadores da rede entre 2003 e 2006. Os objetivos da pesquisa foram avaliar os impactos do Programa nos resultados de desempenho dos alunos de 1ª série mensurados por meio do Sistema de Avaliação do Rendimento Escolar do estado de São Paulo (Saresp), considerada a aferição de 2007, e também seus possíveis efeitos sobre concepções e práticas de alfabetização dos professores que dele participaram. Para o desenvolvimento do estudo, aplicou-se o método misto, que integra abordagens quantitativa e qualitativa. Assim, realizaram-se entrevistas em três escolas, com coordenadores, diretores e professores de 1ª a 4ª série que tivessem participado do Programa, com prioridade dos professores do Ciclo 1. Os dados provenientes do Questionário para os Professores de 1ª e 2ª série do Ensino Fundamental e os resultados dos alunos na prova de 1ª série, ambos instrumentos do Saresp 2007, foram reorganizados e reinterpretados para as análises baseadas em técnicas de árvores de decisão e de regressão linear múltipla e para a comparação entre grupos de escolas com diferentes proporções de professores que participaram do Programa. Concluiu-se que o Letra e Vida influenciou o discurso dos docentes, sobretudo em sua concepção de alfabetização, embora nem sempre se tenham percebido, nas visitas às escolas, evidências da transposição desse discurso para a prática da sala de aula. No que tange aos impactos do Programa sobre o desempenho dos alunos, pode- -se afirmar que, em alguns contextos e em conjunto com outras características estudadas, o Letra e Vida se destacou na explicação dos resultados dos alunos, mas, sozinho, não teve impactos significativos. / This dissertation focuses on the continued literacy teacher education program called Letra e Vida, offered by the Sao Paulo state Department of Education during the period of 2003 to 2006. The research objectives were to assess the impact of the Program in the first grade students performance and also its possible effects on concepts and practices of teaching literacy from teachers who participated in it. To achieve these goals the São Paulo Education Achievement Assessment System (Saresp) was used, considering the evaluation of 2007 and a mixed method study was developed to integrate qualitative and quantitative approaches complementarily. Thus, interviews in three schools were conducted with coordinators, principals and 1st to 4th grade teachers with priority given to teachers of 1st and 2nd grades who attended the program. The data from the questionnaire for teachers of 1st and 2nd grades, and the results of students in the 1st grade, both collected by SARESP 2007, were reorganized and reinterpreted to be used in statistical analysis using multiple linear regression, decision trees and comparison between groups of schools with different proportions of teachers who participated in the Program. It was possible to conclude that the effects of the program in teachers discourse were considerable, although evidence of implementation of such discourse was not always possible to observe during visits to schools. With regard to the impact of the Program on student performance, it is clear that in some contexts the Letra e Vida in conjunction with other variables were able to offer some explanation to the student outcomes, but its effects are not significant when standing alone.

Esta pesquisa avalia o impacto do \"Programa de Garantia da Atividade Agropecuária\" para agricultores familiares, conhecido como Proagro Mais. A relevância do trabalho fundamenta-se no considerável tamanho do Programa dentro do contexto das políticas de gestão de risco agrícola no Brasil. Além disso, é a primeira pesquisa desse tipo na literatura científica do país. A amostra é formada por produtores de milho do Estado do Paraná, tendo como linha base o ano de 2003, uma vez que é o ano anterior ao lançamento do Proagro Mais, e o ano de 2005 como ano de impacto. A base de dados utilizada neste estudo foi fornecida pelo Tribunal de Contas da União (TCU), cujas variáveis relevantes incluem características da cultura e dos agricultores familiares, como área financiada, atividades agrícolas complementares, educação e rendimento esperado. Adicionalmente, a partir de outras fontes públicas, foram adicionadas variáveis meteorológicas e regionais para controlar a localização da fazenda. O objetivo da pesquisa é avaliar o impacto do Proagro Mais sobre o montante de crédito por hectare concedido aos beneficiários do Programa. As metodologias usadas incluem o Propensity Score Matching (PSM), a Diferença das Diferenças (DID) e dois estimadores condicionais do DID com PSM usando dados em painel e repeated cross-section. As estimativas econométricas mostram que o Efeito Médio do Tratamento nos Tratados (EMTT) teve sinal negativo na maioria dos modelos revelando que, após o período de perda de rendimento, o grupo de controle teve um valor médio mais elevado de crédito por hectare do que os beneficiários do Proagro Mais. Os resultados sugerem a existência de mecanismos que poderiam complementar ou substituir o Proagro Mais como instrumento de gestão de risco agrícola, mas também podem sugerir que o Programa avaliado não cubra todos os riscos do setor. / This research evaluates the impact of the \"Farm Activity Guarantee Program for Smallholders\" or Proagro Mais. The relevance of this research relies on the considerable size of the program within the context of agricultural risk management policies in Brazil. This research is also the first of its kind in the scientific literature in Brazil. It was used a sample of small-scale corn producers from the State of Paraná. It was assumed 2003 as the baseline since it is the year prior to the launch of Proagro Mais, and 2005 as the endline. The database used in this study was provided by the Federal Accounting Court of Brazil (TCU). The relevant variables include crop and growers characteristics such as area financed, complementary agricultural activities, education, and expected yield. It was also added meteorological and regional variables from other public sources to control farm location. The main objective of the research is to evaluate the impact of Proagro Mais on the amount of credit per hectare granted to the beneficiaries of the Program. The methodology includes Propensity Score Matching (PSM), Difference-in-Difference (DID), and two conditional DID along PSM estimators, using longitudinal and repeated cross-sections data. The econometric estimates show that the Average Treatment Effect on the Tread (ATET) was negative. This suggests that after the yield loss period, the control group got a higher average amount of credit per hectare than Proagro Mais beneficiaries. Thus, the question that arises is whether there may be other agricultural risk management mechanisms more suited for smallholders than Proagro Mais, or whether the evaluated program could not achieve its main goal because it does not cover all risks faced by its beneficiaries.

O objetivo dessa dissertação é avaliar o impacto de um programa de treinamento voltado para trabalhadores, utilizando o propensity score matching, mas com dois tipos de abordagem, uma não-paramétrica e a outra semi-paramétrica. Para estimação não paramétrica foi utilizado um método proposto por Li, Racine e Wooldridge (2009) e para estimação semi-paramétrica, o modelo utilizado foi o Generalized Additive Model proposto por Hastie e Tibshirani (1990). Os resultados obtidos indicam que os dois métodos utilizados apresentam estimativas tão boas ou melhores do que quando estimadas paramétricamente. / The goal of this thesis is to evaluate the impact of a job training program using propensity score matching methods with two types of approaches: a nonparametric e another semiparametric. For non-parametric estimation was used a method proposed by Li, Racine and Wooldridge (2009) and for the semiparametric model the Generalized Additive Model proposed by Hastie and Tibshirani (1990). The results indicate that both methods provide estimates as good or better than when parametrically estimated.

Dans ma thèse, j'utilise des méthodes récentes de microéconométrie appliquée pour évaluer les impacts des événements majeurs (catastrophe de Fukushima Daiichi, Jeux Olympiques), de l'infrastructure (utilisation des terrains urbains, éoliennes) et des institutions dans les systèmes éducatifs) sur le bien-être individuel, la santé et le comportement. Tout au long de mes articles, j'utilise des données longitudinales sur les ménages, en partie fusionnées avec des données spatiales très détaillées, tout en accordant une attention particulière à l'identification des effets causaux. / In my dissertation, I am using recent methods in applied microeconometrics for policy and programme evaluation to evaluate the impacts of major events (the Fukushima Daiichi disaster, the Olympic Games), infrastructure (urban land use, wind turbines), and institutions (instructional time in education systems) on individual well-being, health, and behaviour. Throughout my papers, I am using longitudinal household data, partly merged with highly detailed spatial data, while paying particular attention at identifying causal effects.

O Fundo de Financiamento ao Estudantil (Fies), programa de financiamento estudantil a taxas de juros subsidiadas, foi criado em 1999, mas passou por uma forte expansão após 2010, ano em que ocorreu a reformulação das condições de financiamento, como a redução dos juros para 3,4% a.a., ampliação dos prazos de carência e de pagamento e a dispensa da exigência de fiador para estudantes de menor renda. A partir de então, mais de 1,1 milhão de estudantes formalizaram contratos com o Fies, envolvendo recursos orçamentários da ordem de R$ 14 bilhões. A partir do Censo da Educação Superior de 2009 a 2012 e dos dados do Enade de 2009 a 2010, e utilizando a metodologia de diferenças em diferenças, este estudo avalia o impacto do Fies na quantidade de matrículas e nas taxas de evasão nos cursos presenciais de instituições privadas brasileiras de ensino superior. Como resultados, foram encontrados efeitos do Fies sobre as matrículas de 6,3% a 9% e sobre a taxa de evasão de -2 a -2,6 pontos percentuais. Como forma de testar a robustez dos resultados realizaram-se testes de placebo, em que os resultados foram replicados para a rede pública, para cursos não habilitados ao Fies e para o período pré-tratamento. Não foram encontrados efeitos estatisticamente significativos para nenhum dos casos, exceto para o teste de placebo para a taxa de evasão utilizando os dados anteriores ao início do tratamento, e estes significativos a 10%. / The Financing Fund the Higher Education Student (FIES), a student financing program of subsidized interest rates was established in 1999, but passed by a strong expansion after 2010 when there was an overhaul of the financing conditions, as the reduction of interest to 3.4% per year, expansion of grace period and payment deadlines to dispenses a guarantor requirement paragraph students of lower income. From then on, over 1.1 million students formalized contracts with Fies involving budgetary resources of the order of R$ 14 billion. At from the Census of the Higher Education from 2009 to 2012 and Enade data from 2009 and 2010, this study uses differences in differences methodology with the objective of evaluating the impact of Fies on enrollment amount and evasion rates in courses of brazilian higher education private institutions. The results suggest increase on enrollment from 6.3% to 9% and the dropout rate of -2 to -2.6 percentage points. In order to test the robustness of the results were performed placebo tests where the results were replicated to the public courses, for courses not entitled to Fies and for the pre-treatment period. No statistically significant effects were found for any of the cases, except for the placebo test for the dropout rate using data prior to the start of treatment, and these significant 10%.