Cette thèse propose des évaluations de réformes récentes du système de retraites français, en utilisant deux approches distinctes et complémentaires : l’évaluation ex post par microéconométrie et l’évaluation ex ante par microsimulation.La première partie de la thèse rassemble des travaux d’évaluation menés sur les données administratives de la Caisse nationaled’assurance vieillesse. Un chapitre introductif propose une revue des littératures des travaux évaluant l’effet du système de retraite sur les comportements d’activité. Ensuite, nous utilisons les réformes du dispositif de mise à la retraite d’office dans les années 2000, pour identifier un effet de la demande de travail sur les comportements de retrait dumarché du travail. Nous évaluons ensuite l’effet du relèvement de l’âge minimal de départ à la retraite instauré par la réforme de2003. Nous étudions les effets de la réforme sur l’emploi après 60 ans, mais également les effets de substitution potentiels vers d’autres dispositifs publics, en particulier l’assurance chômage.La deuxième partie de cette thèse, consacrée à l’évaluation par microsimulation, s’ouvre sur une présentation de l’approche appliquée à la retraite et de son usage dans les modèles français. Nous proposons d’abord une évaluation de l’augmentation de la durée d’assurance prévue par la réforme de 2003.Le principe de maintien constant du rapport entre durée en carrière et durée en retraite est interrogé, puis confronté aux évolutions projetées en simulation à l’aide du modèleDestinie de l’Insee. Enfin, nous utilisons le modèle Pensipp de l’Institut des politiques publiques pour simuler des options de réformes inédites permettant de réduire la dépendance à la croissance du système de retraite et l’incertitude sur l’équilibre financierqu’elle induit. / This thesis brings together different evaluations of recent reforms of the French pension system using two complementaryapproaches: ex post evaluation using microeconometrics techniques and ex ante evaluation built on microsimulationmodelling.The first part of the thesis is based on microestimations using administrative data of the French Social Security agency (Cnav). We first use changes in the rules of mandatory retirement to identify an effect of labordemand on retirement behavior. In a second contribution we estimate the effect of the increase in the early retirement age implemented by the 2010 reform. Besides its effect over employment, we study thepotential substitution effects of the reform towards alternative schemes like unemployment insurance.The second part of the thesis focuses on evaluation by microsimulation. A first study evaluates the effect of the increase in therequired insurance duration implemented by the 2003 reform. The principle of the reform - preserving a constant ratio between work and retirement duration - is discussed on theoretical grounds, and then confronted with the long-term evolutions projected by the Insee Destinie model. Finally, we use the Pensipp model of the Institute of public policies to simulate the effect of originalreforms reducing the growth dependence of the pension system and the uncertainty over financial balance it generates.

Cette thèse contribue à l’analyse des relations entre santé, revenu et politiques de santé. Il m’a semblé nécessaire de mener cette recherche pour chaque direction de la relation. Ainsi, le 1er article étudie la solvabilité financière de différentes politiques de lutte contre le SIDA grâce à un modèle de microsimulation. La santé y est vue comme un facteur de production ; son amélioration génère des gains de productivité. Cette analyse montre que les programmes de traitement peuvent générer un surplus économique et ce net des coûts engendrés. Le 4ème article étudie la relation entre santé infantile et milieu social d’origine à partir d’une enquête internationale. Les analyses montrent que si un gradient socio-économique de santé s’impose à tous à travers le globe (les enfants issus de milieux pauvres sont en moins bonne santé), ce gradient varie selon le niveau de revenu et d’offre de soin du pays considéré. Les deux autres articles s’intéressent aux problèmes de mesures liés à la mise sous traitement des agents malades : le poids de ces derniers dans la population augmente. Au niveau macroéconomique, le PIB par habitant pourrait diminuer si les traitements ne permettent pas de maintenir un niveau de productivité suffisant. Les analyses dans le cadre du VIH montrent que cet effet pervers ne supplante pas les effets positifs. Cette problématique est élargie à la mesure du bien être dans le 3ème article de la thèse. En ne s’attachant qu’aux populations vivantes, les indicateurs traditionnels ne tiennent pas compte du fait que certains agents auraient pu être maintenus en vie (avec un bien être moindre). Les comparaisons inter-pays pourraient ainsi être faussées. / This dissertation aims at improving our understanding of the links between health and wealth, and between health programs and macroeconomic outcomes. Because the former might be bi-directional, it seemed sensible to tackle this issue for each direction of the causality. In the 1st paper, I examine using microsimulation the financial solvability of alternative policies against HIV. Health improvements at the individual level generate productivity gains which translate into an economic surplus that outweighs programs’ costs. In the 4th paper, I examine the relationship between child health and social background using an international survey. Analysis reveals a substantial gradient in health: across the globe, poorer children have worse health. Yet, the effect of wealth is moderated by country-level income and health-supply variables. The two other papers focus on rather methodological issues raised by the fact that curative programs save lives but increase the prevalence of the disease. One study highlights that average income could fall if treatments cannot guarantee a sufficient level of productivity among sick workers. Despite this adverse effect, the microsimulation model demonstrates that treatment policies can raise per capita income in the context of HIV. The 3rd paper of the thesis extends this message to welfare measurement. By restricting attention to the living population, standard indicators of welfare ignore the fact that individuals who would otherwise be dead can be kept alive through treatment, but with a lower than average welfare. Cross-country comparisons based on indicators that are made invariant to the population size may therefore be biased.

Cette thèse propose une analyse de l'investissement éducatif au niveau individuel. Elle développe une perspective dynamique qui cherche à analyser les liens entre éducation et trajectoires sur le long terme. Pour ce faire, nous utilisons un modèle de microsimulation dynamique. La première partie de la thèse porte sur l'éducation initiale. Elle cherche à évaluer l'impact de la diversité des trajectoires sur l'hétérogénéité des gains monétaires associés à différents niveaux de diplôme. Nous y montrons tout d'abord que certains diplômés, en raison des trajectoires défavorables qu'ils connaissent sur le marché du travail, ont une probabilité non négligeable de ne pas valoriser financièrement leurs études. L'analyse des gains liés à l'éducation initiale est ensuite élargie aux comportements conjugaux des individus. En particulier, nous montrons que la prise en compte des revenus salariaux des conjoints réduit les inégalités entre les diplômés, et ce, malgré l'existence de phénomènes d'homogamie éducative. La seconde partie de la thèse est consacrée à la « formation continue » (FC) ou « postscolaire ». En tenant compte des facteurs familiaux et professionnels qui peuvent jouer sur l'accès à la FC, cette partie a pour objectif de mesurer le niveau des dépenses dont bénéficient les individus pour se former une fois leurs études initiales terminées. Les résultats mettent en lumière une forte hétérogénéité de ces dépenses et le rôle divergent des différents dispositifs de formation. Ils montrent également que les montants investis dans la formation postscolaire des individus demeurent faibles au regard de ceux engagés pour la formation initiale. / This thesis provides an analysis of investment in education from an individual point of view. Using a microsimulation model, we develop a dynamic perspective that aims to take into account the long-term relationships between education and individual trajectories. The first part of the thesis focuses on initial education. Its objective is to assess the impact of the diversity of trajectories on the heterogeneity of education benefits. We first show that some graduates, because of their vulnerable position on the labour market, are exposed to negative return to initial education. Our analysis of initial education benefits is then extended to marital behaviour. We more particularly show that taking into account spouse's labour market income reduces inequality between individuals despite the existence of educational homogamy. The second part of the thesis deals with further education and training (FET). Considering family and professional factors that may affect access to FET, we measure the level of expenditure received by individuals to train throughout their career. The results point out the heterogeneity of these spending and indicate that investments in FET are low compared to those invested in initial education.

Many health geography studies, including the Ottawa Neighbourhood Study (ONS), have faced significant challenges uncovering local variation in patterns of community health in rural areas. This is due to the fact that sparsely populated rural areas make it difficult to define neighbourhoods that are representative of the social and resource utilization patterns of the individuals therein. Moreover, rural areas yield small samples from population-based regional health surveys and this leads to insufficient sample sizes for reliable estimation of health determinants and outcomes. In response to this issue this thesis combines geographical and statistical techniques which allow for the simulation of health variables within small areas and populations within rural Ottawa. This methodological approach combines the techniques of dasymetric mapping and statistical micro-simulation in an innovative way, which will allow health geography researchers to explore health determinants and health outcomes at small spatial scales in rural areas. Dasymetric mapping is used to generate a statistical population surface over Ottawa and then estimate socio-economic (SES) variables within small neighbourhood units within rural Ottawa. The estimated SES variables are then used as correlate variables to simulate health determinant and health outcome variables form the Canadian Community Health Survey (CCHS) using statistical micro-simulation. Through this methodology, simulations of specific health determinants and outcome can be investigated at small spatial scales within rural areas. Dasymetric mapping provided neighbourhood-level population estimates that were used to re-weight as set of SES variables that were correlates with those in the Canadian Community Health Survey (CCHS). These neighbourhood-level correlates allowed microsimulation and consequent spatial exploration of prevalence for smoking, binge drinking, obesity, self-rated mental health, and the presence of two or more chronic conditions. The methodology outlined in this paper, provides and innovative way of exploring health determinants and health outcomes in neighbourhoods for which population and health statistics are not traditionally collected at levels that would allow traditional statistical analyses of prevalence.

This thesis considers the extent to which tax and transfer policies in South Africa support children between 2000 and 2008. The analyses are carried out using a four-dimensional analytical framework which separates the dimensions of welfare ideology, policy aims, policy instruments and welfare outcomes. This approach is adopted in recognition of the fact that the extent to which tax and transfer policies support children is seen to vary according to the dimension of analysis. The analysis of welfare ideology, policy aims and policy instruments is undertaken by considering key legislative texts, including the Bill of Rights in the South African Constitution, budget speeches and policy documents. Welfare outcomes are analysed at the individual and household level using microsimulation modelling. A microsimulation model for South Africa, SAMOD, is developed specifically for these analyses. The findings of this thesis add conceptual and empirical understanding to the impact of tax and transfer policies on children. Children are found to be supported by policy to some extent, and have been prioritised in reforms to social assistance. However, recent reforms to tax policy have not benefited children and the analyses indicate that child poverty rates in South Africa could be lower than they are at present had the government pursued alternative policy reforms. The construction of the microsimulation model SAMOD is a valuable tool to facilitate future policy evaluation in South Africa. Further development of SAMOD is recommended to continue to progress and enhance debates on policy reforms. In addition, this thesis highlights some key areas for future research including developing further understanding of the patterns of inter and intra-household income allocation and the impact that this may have on poverty measures for different groups.

O tema deste trabalho compreende o estudo do impacto da educação no nível da renda e sua respectiva distribuição no Brasil ao longo dos anos (1993 a 2011). A metodologia adotada no estudo é um modelo de decomposição microeconométrica com simulações, no qual conseguimos avaliar o impacto da escolaridade sobre a pobreza e a desigualdade de renda. Destacamos dois principais resultados. Na simulação do efeito quantidade, temos que a educação explica 34% da queda no indicador de proporção de pobres entre 1993 e 2011, 43% da queda na proporção de extremamente pobres e 13% da queda no Gini (desigualdade de renda) no mesmo período. Para os anos entre 2001 e 2011, o efeito quantidade da educação explica 27% da queda na proporção de pobres e de extremamente pobres e apenas 1% da queda na desigualdade de renda ocorrida no período analisado. Já na simulação do efeito preço da educação, o impacto se mostrou relevante somente na queda na desigualdade de renda: a educação explica 15% da queda observada entre 1993 e 2011 e 33% da queda ocorrida entre 2001 e 2011. / The subject of this work is the impact of education on the level of income and its respective distribution in Brazil over the years (1993-2011). The methodology used in the study is a model of microeconometric decomposition with simulations in which we can isolate the impact of education on poverty and income inequality. We can highlight two main results. In the simulation of the quantity e ect, we have concluded that education explains 34% of the fall in the headcount index between the years 1993 and 2011, 43% of the fall in the headcount index using an indigence line and 13% of the fall in the Gini indicator (income inequality) in the same period . For the years between 2001 and 2011, the quantity e ect of education explains 27% of the fall in the headcount index (using both poverty and indigence lines) and only 1% of the fall in income inequality that happened during the period. In the simulation of the price e ect of education, the impact was also relevant but only in the fall in income inequality: education explains 15% of the fall observed between 1993 and 2011 and 33% of the fall that has occurred between 2001 and 2011.

Traffic simulation and signal timing optimization are classified in structure into two main categories: (i) Macroscopic or Microscopic; (ii) Deterministic or Stochastic. Performance of the optimized signal timing derived by any tool is influenced by the methodology used in how calculations are executed in a particular tool. In this study, the performance of the optimal signal timing plans developed by two of the most popular traffic analysis tools, HCS and Tru-Traffic, each of them has its inbuilt objective function(s) to optimize signal timing for intersection, is compared with an ideal and an existing timing plans (base case) for the area of study using the microsimulation software VISSIM. An urban arterial with 29 intersections and high traffic in Fort Lauderdale, Florida serves as the test bed. To eliminate unfair superiority in the results, all experiments were performed under identical geometry and traffic conditions in each tool. Comparison of the optimized plans is conducted on the basis of average delay, average stopped delay, average number of stops, number of vehicles completed trips, latent delay, and latent demand from the simulated vehicle network performance evaluation results in VISSIM. The results indicate that, overall, HCS with its overall delay objective and the Tru-Traffic programs produce signal timing with comparable quality that performed similar to the un-optimized base case for most of the performance measures. / Includes bibliography. / Thesis (M.S.)--Florida Atlantic University, 2019. / FAU Electronic Theses and Dissertations Collection

This thesis examines the benefits and costs of screening for colorectal cancer in the context of an organised population screening programme. It uses microsimulation modelling to derive an optimally cost-effective screening protocol for various combinations of the available screening tests. ¶ First a mathematical model for the natural history of colorectal cancer is derived, based on analyses of Australian population and hospital-based cancer registries combined with data from published studies. Then a model for population based screening is derived based mainly on data from published screening studies, including the four major published randomised controlled trials of faecal occult blood test (FOBT) screening. These two models are used to simulate the application of a screening programme to the Australian population. The simulations are applied to a period of 40 years following 1990 (the study’s base year), with both costs and benefits discounted back to the base year at an annual rate of 3%.¶ The models are applied to simulating a population screening programme based on FOBT with a colonoscopy follow up of positive tests. This simulation suggests that the optimal application of such a programme would be to offer annual screening to people aged 50 to 84 years. Such a programme would lead to a cumulative fall in years of life lost to colorectal cancer (YLL) of 28.5% at a cost per year of life saved (YLS) of $8,987. These costs and benefits are consistent with those arising from other currently funded health interventions. They are also consistent with the cost per YLS which Australian governments appear willing to pay for health interventions when justified on the basis of cost-effectiveness. The fall in colorectal cancer deaths from this screening programme should be first detectable by a national monitoring system after around three years of screening. However the full benefits from screening would not be realised before around 30 years of screening.¶ These simulations are based on the standard guaiac FOBT, but the results suggest that significant cost-effective gains could be made by using the newer immunochemical FOBT. Further cost-effect gains could be made by offering sigmoidoscopy every five years in addition to annual FOBT.¶ The models are then applied to simulating population screening programmes using colonoscopy and sigmoidoscopy as primary screening tools. Offering colonoscopy every ten years to all people aged from 45 to 85 leads to an overall fall in cumulative YLL of 37.6%, at a cost of $15,585 per YLS. Offering sigmoidoscopy every three years to all people aged 40 to 85 leads to an overall fall in cumulative YLL of 29.1%, at a cost of $4,862 per YLS. Both of these cost and benefit results are also consistent with the cost per YLS which Australian governments appear willing to pay. The fall in deaths with colonoscopy screening would also be detectable after three years of screening but the fall with sigmoidoscopy screening would not be detectable until after six years of screening. Sigmoidoscopy would need around 35 years of screening to reach its potential gains while colonoscopy screening would not reach its full potential during the 40 year screening period.¶ Finally the models are applied to targeting people at higher risk of cancer. The results show that offering colonoscopy every five years to people at higher risk because of a family history of colorectal cancer is a cost-effective addition to the annual FOBT screening programme.¶ An earlier version of chapter two of this thesis has been published as Stevenson CE 1995. Statistical models for cancer screening. Statistical Methods in Medical Research; 4: 19–23.¶ An expanded version of chapter two, along with parts of chapter one, has been published as Stevenson CE 1998. Models of screening. In: Encyclopedia of Biostatistics. Armitage P, Colton T, eds. John Wiley and Sons Ltd, pp 3999–4022.

The purpose of this thesis was to perform microsimulation analyses on intersections near Highway-Railroad Grade Crossings (HRGCs) to determine if controlling mean train speed and train speed variability would improve safety and reduce delays. This research focused on three specific areas. First, average vehicle delay was examined, and this delay was compared for seven specific train speed distributions, including existing conditions. Furthermore, each distribution was associated with train detectors that were placed at the distance the fastest train could travel during the given warning time. Second, pedestrian cutoffs were investigated. These cutoffs represented an occasion when the pedestrian phases were truncated or shortened due to railroad signal preemption. Finally, vehicle emissions were analyzed using a modal emissions model. A microscopic simulation model of the Wellborn Corridor in College Station, Texas was created using VISSIM. The model was run twenty times in each train speed distribution for each of three train lengths. Average vehicle delay was collected for three intersections, and delays were compared using the Pooled t-test with a 95% confidence interval. Comparisons were made between the distributions, and generally, distributions with higher mean train speeds were associated with lower average delay, and train length was not a significant factor. Unfortunately, pedestrian cutoffs were not specifically controlled in this project; therefore, no statistical conclusions can be made with respect to the pedestrian cutoff problem. However, example cases were devised to demonstrate how these cutoffs could be avoided. In addition, vehicle emissions were examined using the vehicle data from VISSIM as inputs for CMEM (Comprehensive Modal Emissions Model). For individual vehicles, as power (defined as the product of velocity and acceleration) increased, emissions increased. When comparing emissions from different train speed distributions, few significant differences were found. However, a scenario with no train was tested, and it was shown to have significantly higher emissions than three of the distributions with trains. Ultimately, this thesis shows that average vehicle delay and vehicle emissions could be lowered by specific train speed distributions. Also, work could be done to investigate the pedestrian cutoff problem.

Tax estimation is a fundamental prerequisite for a sustainable fiscal policy. This paper uses the Georgian Household Survey and s simple microsimulation model in order to describe the household incomes in Georgia for the year 2005, their structure and regional distribution within eleven historical regions. Based on a thorough analysis of the existing taxable incomes and following the documentation of the applied model both a tax allowance and three percent raise of the income tax are estimated with respect to tax revenue and distributional effects. The paper comes to the conclusion that the poor income situation of most Georgian households can be mitigated by a tax allowance but is very difficult to be financed because of expected revenue losses. In spite of some progressive distributional effects of an increase of the tax burden, most households will find a very hard to cope with additional tax liabilities.