My work centers on drawing economic insights about the macroeconomy based on disaggregated mechanisms and empirical patterns. In my first chapter, I study technology upgrading in the Chinese manufacturing sector and its dynamics after trade liberalization. I first document that Chinese firms often engage in capital substitution episodes, during which firm labor productivity increases, labor shares drop, and skill intensity increases. A model in which firms adopt new skill-intensive technology through investment in capital upgrading naturally rationalizes these facts, linking capital substitution events to technological change. Empirically, trade liberalization shocks reduce capital substitution at Chinese firms, raising the possibility that trade liberalization may delay short-run growth. I then build a quantitative GE model with heterogeneous firms, capital upgrading, and trade liberalization shocks. After liberalization in the model, strategically delayed capital upgrading by firms pushes technological and consumption gains further into the future, meaningfully expanding the horizon over which trade gains manifest themselves. In the second chapter, I exploit rich data on tens of millions of housing transactions from Zillow to document poor house price growth in manufacturing-heavy regions in the US. The chapter shows that manufacturing shares strongly predict dampened house price growth, mechanically contributing to a rise in housing wealth inequality across regions. However, this price growth difference is particularly strong for lower-priced houses, amplifying inequality within regions as well. Overall, I find that cross-sectional house price inequality has increased by around 10%, with around a third of this increase due to the relative decline of lower-value homes. In the third chapter, I combine empirical tools and structural modeling to measure the effect of monetary policy on consumption through housing. Exploiting quarterly US data, I estimate empirically that a 1% unexpected interest rate shock causes average house prices to drop by about 1.4% in two years. Feeding this empirical response into an incomplete markets model, I find that aggregate consumption shifts by around 0.3% in response to the shock. A lean-against-the-wind monetary policy can stabilize consumption dynamics along a transition path.

The fossorial Florida Sand Skink (Plestiodon [=Neoseps] reynoldsi) inhabits a restricted range of scrub and sandhill fragments on the ridges of central Florida. The high rate of urban and agricultural development in this area necessitates conservation strategies other than land acquisition and management because of the limited remaining Florida Sand Skink habitat available. This study tests the viability of translocation as a conservation strategy for this species and assesses which features of a recipient site contribute to the successful establishment of a population. In 2007, 300 individuals were collected and moved from an intact scrub habitat, individually marked, and moved to a nearby reclaimed site with no existing Florida Sand Skink population. Fifteen 20 m² enclosures were constructed at the recipient site, and 20 skinks were randomly assigned to each. These enclosures were divided among five treatments, which were represented the range of habitat types at the donor site and differed in the presence or absence of a shade-providing object and coarse woody debris. Translocated skinks were monitored for two years to measure survival and reproduction. While survival and reproduction were apparent in all treatments, survival was significantly greater in enclosures with no shade-providing object and low soil moisture, and reproduction was most evident in enclosures with less light intensity and soil compaction. Common measurement of environmental variables at the donor and recipient sites showed that all of the recipient site enclosures differed from the donor site in the amount of vegetative cover but contained the structural heterogeneity that is associated with Florida Sand Skink presence in the wild. This study indicates that translocation is a practical conservation strategy for this species, and my results can be used to inform protocol for future Florida Sand Skink translocation efforts.

No description available.

Rangelands are heterogeneous working landscapes capable of supporting livestock production and biodiversity conservation, and heterogeneity-based rangeland management balances the potentially opposing production and conservation goals in these working landscapes. Within fire-dependent ecosystems, patch-burn grazing aims to create landscape patterns analogous to pre-European rangelands. Little work has tested the efficacy of patch-burn grazing in northern US Great Plains. We investigated patch contrast in above and belowground ecosystem properties and processes during the summer grazing seasons from 2017 ? 2020 on three patch-burn pastures stocked with cow-calf pairs and three patch-burn pastures stocked with sheep. We focused on vegetation structure, plant community composition, forage nutritive value, grazer selection, livestock weight gain, soil nutrient pools, soil microbial community composition, and decomposition activity. We used mixed-effect models and ordinations to determine whether differences: along the time since fire intensity gradient, between ecological sites, and between grazer types existed. Despite no significant shifts in the plant community, structural heterogeneity increased over time as the number of time since fire patches increased and was higher than homogeneously managed grasslands. Grazing livestock preferred recently burned patches where the available forage had a higher nutritive value and lower available biomass than surrounding patches at a given point in time. With the exception of 2018, livestock weight gains were consistent. Soil nutrient pools and microbial abundances differed more by ecological site than by the time since fire intensity gradient, and ecological sites exhibited similar nutrient and microbial responses to the time since fire intensity gradient. That belowground response variables were mostly resistant to patch-burn grazing is supportive of further use of this management, especially given the desirable results with aboveground response variables.

Esta tesis pertenece al campo de la macroeconomía, en concreto al estudio de la desigualdad económica, de renta y riqueza, y al análisis de los efectos macroeconómicos de las reformas fiscales. Para estudiar cuestiones concretas en este campo, la herramienta principal utilizada en esta tesis es el modelo Aiyagari-Huggett de agentes heterogéneos, mercados incompletos y riesgo idiosincrático, adaptado o extendido según el objetivo de cada capítulo. Usando esta metodología, cada capítulo realiza un análisis cuantitativo aplicado a la economía real de la cuestión de interés. El objetivo general de la tesis es examinar cómo distintas formas de heterogeneidad, tales como diferencias en evasión fiscal por ocupación o el nivel de desigualdad de riqueza, pueden afectar al efecto de reformas fiscales, y cómo mejorar la modelización de determinadas formas de heterogeneidad, en concreto, de la productividad individual de los trabajadores. El primer capítulo de la tesis incorpora la evidencia empírica para Estados Unidos sobre los altos niveles de evasión fiscal del sector empresarial para estudiar su efecto en el análisis de reformas fiscales. El capítulo desarrolla un modelo de agentes heterogéneos y riesgo idiosincrático con decisión de ocupación (entre iniciar o mantener una empresa, o trabajo asalariado), donde se incorpora además la decisión de evadir parte de los impuestos debidos. El equilibrio estacionario está calibrado para Estados Unidos, y es capaz de reflejar el nivel de evasión agregada, y las diferencias en evasión por ocupación. Usando este modelo, se analizan los efectos de largo plazo de una reforma fiscal consistente en reemplazar el impuesto progresivo de la renta por uno proporcional, y se comparan los resultados con los de un modelo similar sin posibilidad de evasión fiscal. Aunque los efectos agregados y distributivos están en la misma dirección cualitativamente, en términos cuantitativos hay grandes diferencias. Al incluir la evasión fiscal, los efectos positivos agregados son sustancialmente menores, y el aumento de desigualdad de riqueza es menos pronunciado que en el modelo sin evasión. El segundo capítulo, conjunto con Claudio Campanale, trata se centra en el tema de la modelización de la productividad individual. Distintos procesos estadísticos se usan en la literatura de modelos cuantitativos principalmente para replicar la distribución de salarios. La elección del proceso (algunos tienen mucho riesgo idiosincrático, otros menos) resulta ser determinante para algunos ejercicios de política fiscal, al afectar de distinta forma al comportamiento individual ante cambios en impuestos. En este capítulo, hacemos uso de reciente evidencia empírica para Estados Unidos con información muy detallada sobre la evolución salarial para los individuos en la parte alta de la distribución. En concreto, proponemos una modificación del proceso estándar AR(1) que añade crecimiento heterogéneo de la productividad, donde un pequeño porcentaje de los individuos experimenta un crecimiento extraordinario (un estado "superestrella") con persistencia decreciente durante la vida. Usando nuestro proceso en el marco de un modelo de ciclo vital, mostramos que la evolución salarial durante la carrera laboral es consistente con la evidencia empírica, y que además el modelo es capaz de reproducir estadísticos adicionales relacionados con la distribución de salarios, y el riesgo idiosincrático de los trabajadores con mayores salarios. Además mostramos que, en comparación con nuestro proceso, un modelo con un proceso de productividad estándar (con productividad superestrella permanente) obtiene resultados muy lejanos de la evidencia empírica. Por último, el tercer capítulo, coautorado con Claudio Campanale, estudia cómo el nivel de desigualdad de la riqueza, y distintos supuestos que afectan al comportamiento del consumidor respecto al ahorro afectan al estudio de reformas fiscales, en concreto, al cálculo de la curva de Laffer para la progresividad del impuesto de la renta. Consideramos un modelo base de ciclo vital, con heterogeneidad de productividad individual y riqueza, y añadimos redistribución de herencias, transmisión (imperfecta) de habilidad entre generaciones, y heterogeneidad en preferencias (factores de descuento). La versión final tiene una concentración de riqueza mucho más cercana a los datos de Estados Unidos que la versión básica. Encontramos que aumentar la concentración de riqueza apenas afecta a la localización del pico de la curva de Laffer, aunque sí afecta al nivel de recaudación extra. Además, analizamos en cada versión del modelo la "elasticity of taxable income" (ETI), y comprobamos que su magnitud en el top de la distribución puede ser sensible a los supuestos del modelo, o al nivel de concentración.

Heterogeneity within tumors has long been studied as a potential confounding factor for effective therapies, with recent studies pointing to heterogeneity resulting in distinct clonal subtypes, each with varying degrees of fitness and metastatic potential. Studies of heterogeneity have previously been limited to microscopy observations, immunohistochemistry, and flow cytometry. Recently, however, it has become possible to examine heterogeneity at a previously unexplored level: the transcriptome of individual cells. Osteosarcomas have been known to be highly heterogeneous, so we have selected osteosarcoma as our primary tumor to study as a proof-of-concept. Additionally, we have elected to create a murine patient derived xenograft (PDX) model from a primary osteosarcoma tumor and examine differences between the primary tumor and resulting xenograft at the single-cell level. Through this, we hope to better understand tumor heterogeneity and add to the current discussion in the scientific community regarding the relevance of PDX models for testing promising new therapies and personalized medicine. Through our examination of single-cell heterogeneity in osteosarcomas, we have confirmed the extensive heterogeneity previously reported, but this time at the level of mRNA. The osteosarcomas were so hetereogeneous that our resulting dataset of over 1,000 cells still did not have enough resolution to generate highly differentiated and separate groupings of cells. Upon examining inter-tumor heterogeneity, we observed the cells from different tumors to generally cluster separately. However, there were certain populations of cells from all tumors that clustered together. We also generated a PDX model and sequenced the resulting tumor, observing markedly reduced heterogeneity as compared to the original primary tumor. Importantly, the cells from the PDX model clustered within the larger group of cells from the original tumor, lending credence to the theory of clonal selection. This work presents evidence of extensive intra- and inter-tumor heterogeneity at the mRNA level within osteosarcoma tumors. This heterogeneity requires further single cell sampling to shed light on the biology of tumor diversity. Further, this heterogeneity is significantly reduced in a generated murine PDX model. This difference should serve as a potential warning about additional factors to take into account when evaluating therapies in PDX models, and suggests that further studies examining cause and effect of this observed heterogeneity are warranted.

Plant root system response to nutrient heterogeneity was tested in ten plant species of varying life form and successional status. All plants tested are native to the South Carolina coastal plain. Morphological responses of the root system (scale, precision and discrimination) and overall plant response (sensitivity) to increasing nutrient heterogeneity were tested. Ten individuals of each species were placed into four treatments which had varying nutrient distribution but the same overall nutrient addition. Plants were harvested when roots reached pot edge. I observed high variation in scale (mass and extent of a root system), precision (the ability to proliferate roots in nutrient patches) and sensitivity (growth benefits gained as nutrient heterogeneity increases; measured as total biomass). No significant discrimination responses were observed, although greatest mean root density occurred at intermediate fertility levels for all species. I tested the hypothesis that scale and precision would be negatively correlated, and I did not observe this relationship in these plant species. However, in herbaceous species scale and precision were positively correlated. Sensitivity was not closely related to precision indicating that proliferating roots in fertile patches does not always yield growth benefits in heterogeneous soils. Further, some sensitive species had very low precision suggesting that other characteristics lead to positive growth response in heterogeneous environments. Plasticity of root uptake rates and demography of roots are proposed as two other mechanisms which may play important roles in plant sensitivity responses. Scale was negatively correlated to sensitivity for herbaceous plants suggesting that plants that monopolize the most soil space are not able to gain benefits from nutrient patches within the soil matrix. There was no trend observed to suggest that plant life form was correlated with precision or sensitivity. However, scale was greater in herbs than in woody plants, possibly because the two life forms develop at different times. / Master of Science

Cette thèse a pour objectif de mieux comprendre quand et comment les organisations apprennent de leur expérience en matière d’innovation et de management opérationnel en étudiant les facteurs qui favorisent ou nuisent à l’apprentissage. Plus précisément, cette recherche étudie l’apprentissage à travers deux types d’expérience, l’échec et la réussite, et souligne le rôle du temps dans la variation des résultats d’apprentissage issus de ces expériences. Elle examine en outre les facteurs liés à la base de connaissance de l’entreprise et au contexte organisationnel qui influencent l'apprentissage organisationnel. Les rappels de médicaments constituent le cadre empirique pour examiner les échecs. Il est montré que l’apprentissage organisationnel par l’échec produit deux résultats différents : une réduction du temps passé à résoudre les échecs actuels d’une part, et une réduction du taux d'échecs futurs d’autre part. Le premier est le fruit de l'apprentissage dit « en simple boucle », tandis que le second relève de l'apprentissage « en double boucle ». L’apprentissage par l’échec s’avère alors à double tranchant : l’expérience de la gestion des échecs favorise la résolution rapide des difficultés qui y sont associées, mais cette rapidité génère aussi davantage d’échecs. De plus, l'hétérogénéité des raisons des échecs passés augmente le temps consacré à la résolution des difficultés liées aux échecs actuels. Ensuite, in est montré que le « timing » du succès (en termes d’introduction d’innovation de rupture) influence l’évolution de l’entreprise et sa performance, les entreprises connaissant rapidement des succès étant ensuite moins susceptibles de générer de nouvelles ruptures. L’expérience entrepreneuriale du fondateur et l’expérience technologique de l’inventeur atténuent cependant l’effet négatif du succès précoce. Il est par ailleurs montré que l’utilisation des connaissances distantes internes améliore les résultats d’apprentissage organisationnel en termes d’innovation. Cet effet est plus fort lorsque la structure de connaissances de l’entreprise est hautement décomposable - les éléments de connaissances sont étroitement liés les uns aux autres. En combinant la perspective de la courbe d’apprentissage avec les niveaux d’apprentissage (apprentissage en simple boucle et apprentissage double boucle), cette thèse étend la courbe d’apprentissage à la compréhension de l’apprentissage à partir d’échecs. Cela montre, en particulier qu’en cas d’échec, apprendre à être plus rapide n'est pas nécessairement favorable à des résultats organisationnels à long terme. Enfin, cette thèse révèle les micro-mécanismes associés à l’apprentissage par le succès et montrent le rôle du capital humain de l’entreprise pour échapper aux pièges de l’apprentissage. Enfin, en soulignant le rôle moteur du contexte organisationnel, elle démontre les conditions dans lesquelles l’exploitation des connaissances internes peut générer des avantages supérieurs à ceux généralement soulignés dans les études antérieures. / This dissertation aims to improve our understanding of when and how organizations learn from their experience in innovation and operational activities, by studying factors that enhance or hinder learning. Specifically, I investigate learning from specific types of experience—i.e., failure and success—and highlight the role of time for explaining some of the variance in learning outcomes that firms extract from these experiences. I further examine knowledge-related and organizational contextual factors that influence organizational learning from experience. The empirical setting for examining failures is drug recalls. I argue that in this context organizations may generate two different learning outcomes from failure: reduction of time spent on resolving current failure and reduction in the rate of future failure. The former is an outcome of single-loop learning while the latter is an outcome of double-loop learning. The findings of this study reveal the double-edged sword of learning from failure; with increasing failure experience organizations become faster in resolving failures but the faster they are on average in resolving past failures, the more failures they face in the future. Additionally, I find that heterogeneity in the reasons of past failures increases the time spent on resolving current failure. Next, I argue and demonstrate that the timing of success (in terms of first breakthrough invention) effects on firm’s evolution and performance such that firms that have success early in their lifetime, are likely to generate less breakthroughs in the future. In addition, I show that founders’ entrepreneurial experience and inventors’ technological experience attenuate the negative effect of early success. Finally, I find that building on internally generated distant knowledge improves organizational learning outcomes in terms of innovation impact and that this effect is stronger when the firm’s knowledge structure is highly decomposable—knowledge elements are tightly coupled to each other. By combining learning curve perspective with arguments on levels of learning (i.e., single-loop and double-loop), this thesis extends and adapts learning curve perspective for understanding learning from failures. It shows, in particular, that in the case of learning from failure being quicker is not necessarily better for long term organizational outcomes. Moreover, the results reveal the micro-mechanisms associated with learning from success and highlight the role of firm’s human capital for overcoming myopic learning traps. Finally, by emphasizing the enabling role of the organizational context my work provides insights on conditions under which building upon internal knowledge can yield to higher payoffs than is generally assumed by prior studies.

Indiana University-Purdue University Indianapolis (IUPUI) / Subgroup analyses assess the heterogeneity of treatment effects in groups of patients defined by patients’ baseline characteristics. Identifying subgroup of patients with differential treatment effect is crucial for tailored therapeutics and personalized medicine. Model-based variable selection methods are well developed and widely applied to select significant treatment-by-covariate interactions for subgroup analyses. Machine learning and data-driven based methods for subgroup identification have also been developed. In this dissertation, I consider two different types of subgroup identification methods: one is nonparametric machine learning based and the other is model based. In the first part, the problem of subgroup identification was transferred to an optimization problem and a stochastic search technique was implemented to partition the whole population into disjoint subgroups with differential treatment effect. In the second approach, an integrative three-step model-based variable selection method was proposed for subgroup analyses in longitudinal data. Using this three steps variable selection framework, informative features and their interaction with the treatment indicator can be identified for subgroup analysis in longitudinal data. This method can be extended to longitudinal binary or categorical data. Simulation studies and real data examples were used to demonstrate the performance of the proposed methods. / 2022-05-06

Indiana University-Purdue University Indianapolis (IUPUI) / Acute kidney injury (AKI) affects millions of patients worldwide yet has few treatment options. There is a critical need to identify novel interventions for AKI, especially approaches targeting cell types that are central to the disease, such as neutrophils. Neutrophils are professional phagocytic cells that respond early to tissue injury. In rodent models of severe ischemic-reperfusion-injury AKI, neutrophils transiently infiltrate the injured kidney, appearing within 6 hours, and are gone by 72 hours. These infiltrating neutrophils are considered proinflammatory and harmful to tissue repair and recovery of kidney function. However, neutrophils can exhibit atypical activity such as antigen presentation and have a central role in recovery from myocardial ischemic injury. Furthermore, little is known of neutrophil polarization, atypical activity, or neutrophil diversity in AKI. Lastly, the kidney generated and renal-protective immunomodulatory protein uromodulin (Tamm-Horsfall Protein, THP) regulates granulopoiesis. In the absence of uromodulin, there is a systemic increase in neutrophils and mouse kidneys are sensitive to injury in AKI. To elucidate neutrophil diversity in AKI and their sensitivity to uromodulin, I performed a series of single-cell sequencing experiments to generate transcriptional profiles of neutrophils from the blood and kidneys of wild-type and THPknockout mice after renal ischemic-reperfusion-injury (IRI). Neutrophil diversity was detected following IRI of the mouse kidney in the blood and kidney. The distribution of subpopulations was sensitive to the kidney milieu. Within the kidney, this diversity and the transcriptional programs of neutrophil subpopulations was sensitive to the severity of ischemic injury. Lastly, Cxcl3 was uniquely upregulated in specific neutrophils after severe ischemic injury. Using single-cell sequencing of uromodulin knock-out mice, I detected the upregulation of toll-like receptor pathways and complement cascades across neutrophil subpopulations in a THP sensitive manner. Furthermore, CXCR2 ligand expression was a combination of moderate and severe injury in wild-type mice. This confirmed previously reported cytokine dysregulation in the uromodulin knock-out mouse after IRI and uncovers a novel role for Cxcl3. Thus, upon revisiting the well-studied neutrophil, I have uncovered novel neutrophil diversity that correlates with recovery of kidney function in AKI and suggests new roles for an old player.