<p>This dissertation asks whether frequency misspecification of a New Keynesian model</p><p>results in temporal aggregation bias of the Calvo parameter. First, when a</p><p>New Keynesian model is estimated at a quarterly frequency while the true</p><p>data generating process is the same but at a monthly frequency, the Calvo</p><p>parameter is upward biased and hence implies longer average price duration.</p><p>This suggests estimating a New Keynesian model at a monthly frequency may</p><p>yield different results. However, due to mixed frequency datasets in macro</p><p>time series recorded at quarterly and monthly intervals, an estimation</p><p>methodology is not straightforward. To accommodate mixed frequency datasets,</p><p>this paper proposes a data augmentation method borrowed from Bayesian</p><p>estimation literature by extending MCMC algorithm with</p><p>"Rao-Blackwellization" of the posterior density. Compared to two alternative</p><p>estimation methods in context of Bayesian estimation of DSGE models, this</p><p>augmentation method delivers lower root mean squared errors for parameters</p><p>of interest in New Keynesian model. Lastly, a medium scale New Keynesian</p><p>model is brought to the actual data, and the benchmark estimation, i.e. the</p><p>data augmentation method, finds that the average price duration implied by</p><p>the monthly model is 5 months while that by the quarterly model is 20.7</p><p>months.</p> / Dissertation

No description available.

Parameter estimation of respiratory infection surveillance dynamics commonly utilize data aggregated over space and time. However, estimates derived from aggregated data may fail to account for biologically meaningful spatiotemporal heterogeneity of effects or to identify where and when transmissions occur. This dissertation shows that high-resolution temporal and spatial data can improve our understanding of heterogeneity while producing more valid and precise estimates of transmission parameters (e.g., contagiousness), behavioral trends (e.g., face mask utilization), and intervention effects (e.g., at-home test distribution). In three projects, we evaluate spatiotemporal heterogeneity in the context of two major respiratory pathogens: Tuberculosis and SARSCoV-2. First, in project one, we identify disease transmission hotspots from a tuberculosis case surveillance system in Greater Vitória, Brazil. Utilizing a human mobility model and recently developed method to quantify disease transmission, we overcome multiple methodological constraints that often obscure spatially and temporally accurate transmission measurements. We estimate that two cities in Greater Vitória, Vila Velha (reproductive number = 1.05, 95%CI: 1.03–1.07) and Vitória (reproductive number = 1.04, 95%CI: 1.02–1.06), help sustain tuberculosis transmission in the entire region and may be effective targets for intervention, while Cariacica (reproductive number = 0.95, 95%CI: 0.94–0.97) fell below the critical threshold of 1 required to sustain transmission alone. Next, in project two, we utilize interrupted time series methods to estimate the effect of mask mandates on mask adherence using a nationally representative digital health survey on masking and a comprehensive database of pandemic-related government policies. The analysis focuses on improving previous attempts at measuring the effectiveness of mask mandates at the state level, by utilizing county-level exposure and outcome data. We find that mask mandates were associated with a large heterogeneity of effects, ranging from increasing masking approximately 8% in counties with low levels of prior masking to 1% or lower change in masking in places like the Northeast U.S. where masking levels were already high. Last, in project three, we leverage the same nationally representative digital health survey to understand at-home testing patterns in the United States. We utilize two different economic measures of resource allocation and a regression model with autoregressive integrated moving average errors to examine if the Covidtests.gov government program reduced at-home testing inequities. We show that Covidtest.gov did increase at-home testing across all demographics; however, income-, geographic- and race-based disparities in at-home test utilization were heightened during periods when the program was active. Specifically, the regression results estimate that Theil’s T, an economic metric used here to measure at-home testing disparities, was 53% (95%CI: 6%–121%) higher for household income, 214% (95%CI: 86%–429%) higher for race, and 90% (95%CI: 23%–193%) higher for geography during Covidtest.gov dissemination periods. Disparities were not elevated for age. Together, these three projects demonstrate the substantial role that high-resolution data can play in improving our understanding of respiratory infection surveillance and informing effective public health interventions.

Les villes sont des systèmes complexes que les modèles urbains peuvent aider à comprendre. Des modèles les plus simplistes aux modèles les plus sophistiqués, la modélisation urbaine a permis de mieux comprendre la question urbaine et ses implications sociétales. Dans ce contexte, les modèles peuvent avoir une valeur-ajoutée appréciable dans le processus de décision publique. Encore faut-il que ces modèles deviennent pratiques et répondent aux contraintes opérationnelles de la chaîne de décision. Dans ce sens, peu de recherches s’est intéressée à la question de praticité des modèles urbains et leur utilisation en situation opérationnelle. À ce jour, les modèles urbains standard qui reposent sur une description agrégée de l’espace sont parmi les approches de modélisation les plus opérationnelles et aussi les plus répandues. De par sa relative praticité, cette approche standard est attractive et simple à mettre en oeuvre. Toutefois, l’agrégation spatiale peut aussi être une source de biais statistiques préjudiciables à la qualité de la modélisation. C’est en particulier, le cas des modèles intégrés Transport-Urbanisme ou des modèles de transport à quatre étapes.La présente thèse a pour objectif d’étudier la question de l’agrégation spatiale dans les modèles transport et plus particulièrement dans les modèles d’affectation des déplacements. Les modèles d’affectation servent à calculer les temps de parcours et les conditions de déplacement sous congestion, présents et futurs, des personnes et des marchandises. Ils servent aussi à calculer les accessibilités nécessaires aux modèles d’usage des sols dont les modèles de choix de localisation des ménages et des entreprises. Toute erreur ou biais dans l’affectation des déplacements peut compromettre la validité et la qualité globales de la modélisation. Dans ce cadre, une attention particulière doit être allouée au problème d’agrégation spatiale dans les modèles d’affectation. Dans ces modèles, l’agrégation spatiale consiste à regrouper les observations individuelles enutilisant une description agrégée de l’espace, i.e. des zones. Par nature, l’utilisation d’une description agrégée à la place d’une représentation continue engendre une omission de l’information et de sa variabilité et donc un biais statistique dans la modélisation. C’est le cas par exemple avec l’utilisation des connecteurs de zones ou avec l’omission des trafics intrazones dans les modèles d’affectation.En reposant sur les zones comme unité spatiale de base, les modèles de transport recourent à l’utilisation des connecteurs de zones pour relier les centroïdes de zones au réseau de transport. Les connecteurs sont des liens fictifs qui modélisent les conditions moyennes d’entrée et de sortie du réseau de transport. Pour ce faire, la majorité des modèles de transport reposent sur une méthode simpliste sujette au problème d’agrégation spatiale. La présente thèse examine en détail l’impact de cette description simpliste sur les résultats et la qualité d’un modèle d’affectation des déplacements en transports en commun. Cette thèse propose aussi une nouvelle méthode de modélisation des connecteurs de zones afin de s’affranchir partiellement du biaisd’agrégation spatiale dans la modélisation des conditions d’accès au réseau des transports en commun.L’utilisation des zones comme unité spatiale de base a aussi pour conséquence l’omission des trafics intrazones de l’affectation des déplacements. Les trafics intrazones ont pour origine et pour destination la même zone et de ce fait ne sont pas pris en compte par les modèles standard d’affectation. Cette omission a souvent été ignorée et son impact sur la qualité de la modélisation demeure non évalué. Cette thèse développe une méthode stochastique pour l’évaluation de cet impact... / Cities are complex systems that urban models can help to comprehend. From simplistic models to more sophisticated ones, urban models have pushed forward our understanding the urban phenomenon and its intricacies. In this context, models can be of great value to policy makers providing that these tools become practical. In this regard, research has put little emphasis on the practicality of urban models and their use under operational conditions.To date, urban models which rely on spatial aggregation are the closest possibility to come to practical models. For this reason, the spatially aggregated modeling framework is widely used. This framework is relatively practical when compared to other modeling frameworks like microsimulation. Nevertheless, spatial aggregation is a serious source of bias in these models. This is especially the case of Land-Use and Transport Interaction (LUTI) models and more particularly of Four Step Models.The current PhD is committed to the study of spatial aggregation issues in traffic assignment models. Traffic assignment is responsable for the computation of travel times and travel conditions of present and future travel demand. Accessibility measurement, which is at the core of LUTI models, is tightly dependent on traffic assignment modeling and outcomes. Any bias in traffic assignment is likely to corrupt the overall modeling framework. In this context, a special attention is to be paid to spatial aggregation in traffic assignment models.In traffic assignment, spatial aggregation consists in grouping observations using zones or traffic analysis zones instead of using a continuous representation of space. By design, aggregation bears an implicit omission in data variability and thus a potential bias if this omission is not random. This is the case with the definition of centroid connectors and the omission of intrazonal demand in traffic assignment. With the use of zones as the basic spatial units, transport models require the use of centroid connectors to attach zones to the transportation network. Centroid connectors are introduced to model average access and egress conditions to and from the network. Nevertheless, average accessibility conditions are found to be too crude to render accurately accessibility conditions as encountered by trip makers. The current PhD explores the extent of the impact of this spatial aggregation bias in the case of transit models and suggests a new modeling strategy to overcome such modeling errors.The use of zones as spatial units induces a loss of intrazonal data. The omission of intrazonal trips in traffic assignment models is an example of such omission. This research introduces an uncertainty framework to study the statistical impact of ignoring intrazonal trips in traffic assignment models. Findings from this research are used to design new assignment strategies that are more robust towards the omission bias and more generally towards the spatial aggregation bias.

Conselho Nacional de Desenvolvimento CientÃfico e TecnolÃgico / This work aims to provide non-biased estimates of the speed of reversion of deviations from the PPP for 11 Brazilian cities, between 1991 and 2013, using the methodology proposed by Choi, Mark and Sul (2006), which makes use of a panel estimation method with correction for three possible sources of bias, those being: the bias of inappropriate grouping of cross-section units with heterogeneous coefficients, the Nickell bias and the bias arising from the temporal aggregation of price indexes. The half-lives obtained are of the order of 4.41 and 3.18 years with Brazil and the Average as references, respectively, and median half-life of 3.13 years, when considering all Brazilian cities analyzed as the numeraire. The half-lives found were also substantially lower than those obtained for American cities. Furthermore, 33.33 % of the half-lives obtained were inferior to the consensus range suggested by Rogoff (1996) of 3-5 years, and none surpassed that range. / O presente estudo se propÃe a prover estimativas nÃo viesadas da velocidade de reversÃo dos desvios da PPC para 11 cidades brasileiras entre 1991 e 2013 atravÃs da metodologia proposta por Choi, Mark e Sul (2006) que usam um mÃtodo de estimaÃÃo em painel com correÃÃo para trÃs possÃveis fontes de viÃs, quais sejam, viÃs de agrupamento inapropriado de unidades cross-sections com coeficientes heterogÃneos, viÃs de Nickell e o viÃs oriundo da agregaÃÃo temporal dos Ãndices de preÃos. As meias-vidas obtidas sÃo da ordem de 4.41 e 3.18 anos tendo Brasil e MÃdia como referÃncia, respectivamente, e meia-vida mediana de 3.13 anos considerando todas as cidades brasileiras analisadas como numerÃrio. As meias-vidas encontradas tambÃm se mostraram substancialmente inferiores Ãquelas obtidas para as cidades americanas. AlÃm disso, 33.33% das meias-vidas aqui obtidas se mostraram inferiores ao intervalo consensual proposto por Rogoff (1996) de 3 a 5 anos, e nenhuma o ultrapassou.

Die Beiträge der vorliegenden Dissertationsschrift untersuchen zum einen, inwieweit Wetterindexversicherungen einen Beitrag zur Stabilisierung von wetterbedingten Einkommensschwankungen in der Landwirtschaft leisten können. In der Landwirtschaft ist trotz bedeutender wetterbedingter Einkommensschwankungen bisher nur ein sehr verhaltener Einsatz von Indexversicherungen zu beobachten. Allerdings gibt es bislang kaum Studien, die Möglichkeiten zur Reduzierung des mit dem Einsatz von Wetterindexversicherungen verbundenen Basisrisikos untersuchen. Zum anderen wird untersucht, inwieweit Wetterrisiken das Rückzahlungsverhalten landwirtschaftlicher Mikrokreditnehmer beeinflussen. Das Risiko bei der Kreditvergabe an landwirtschaftliche Klein-Betriebe in Entwicklungs- und Schwellenländern gilt aus Bankensicht aufgrund der vergleichsweise hohen Einkommensschwankungen in der landwirtschaftlichen Produktion als besonders hoch. In der Literatur wird das wetterbedingte Einkommensrisiko als wesentlicher Grund für das vergleichsweise hohe Kreditrisiko von landwirtschaftlichen Mikrokrediten angeführt. Allerdings wurde dies bislang noch nicht empirisch verifiziert. Die Dissertationsschrift widmet sich dem Thema in vier Beiträgen, die unterschiedliche Aspekte der übergeordneten Problematik behandeln.

No description available.