Essays in open economy macroeconomics

Research macroeconomists have witnessed remarkable methodological developments in mathematical, statistical, and computational tools during the last two decades. The three essays in this dissertation took advantage of these advances to analyze important macroeconomic issues. ^ The first essay, “ Habit Formation, Adjustments Costs, and International Business Cycle Puzzles” analyzes the extent to which incorporating habit formation and adjustment costs in investment in a one-good two-country general equilibrium model would help overcome some of the international business cycle puzzles. Unlike standard results in the literature, the model generates persistent, cyclical adjustment paths in response to shocks. It also yields positive cross-country correlations in consumption, employment, investment, and output. Cross-country correlations in output are higher than the ones in consumption. This is qualitatively consistent with the stylized facts. These results are particularly striking given the predicted negative correlations in investment, employment, and output that are typically found in the literature. ^ The second essay, “Comparison Utility, Endogenous Time Preference, and Economic Growth,” uses World War II as a natural experiment to analyze the degree to which a model where consumers' preferences exhibit comparison-based utility and endogenous discounting is able to improve upon existing models in mimicking the transitional dynamics of an economy after a shock that destroys part of its capital stock. The model outperforms existing ones in replicating the behavior of the saving rate (both on impact and along the transient paths) after this historical event. This result brings additional support to the endogenous rate of time preference being a crucial element in growth models. ^ The last essay, “Monetary Policy under Fear of Floating: Modeling the Dominican Economy,” presents a small scale macroeconomic model for a country (Dominican Republic) characterized by a strong presence of fear of floating (reluctance to have a flexible exchange rate regime) in the conduct of monetary policy. The dynamic responses of this economy to external shocks that are of interest for monetary policy purposes are analyzed under two alternative interest rate policy rules: One being the standard Taylor rule and another that responds explicitly to deviations of the exchange rate with respect to its long-term trend. ^

An investigation of the relationship between antemortem and postmortem drug concentrations in blood

In the field of postmortem toxicology, principles from pharmacology and toxicology are combined in order to determine if exogenous substances contributed to ones death. In order to make this determination postmortem and (whenever available) antemortem blood samples may be analyzed. This project focused on evaluating the relationship between postmortem and antemortem blood drug levels, in order to better define an interpretive framework for postmortem toxicology. To do this, it was imperative to evaluate the differences in antemortem and postmortem drug concentrations, determine the role microbial activity and evaluate drug stability. Microbial studies determined that the bacteria Escherichia coli and Pseudomonas aeruginosa could use the carbon structures of drugs as a source of food. This would suggest prior to sample collection, microbial activity could potentially affect drug levels. This process however would stop before toxicologic evaluation, as at autopsy blood samples are stored in tubes containing the antimicrobial agent sodium fluoride. Analysis of preserved blood determined that under the current storage conditions sodium fluoride effectively inhibited microbial growth. Nonetheless, in many instances inconsistent drug concentrations were identified. When comparing antemortem to postmortem results, diphenhydramine, morphine, codeine and methadone, all showed significantly increased postmortem drug levels. In many instances, increased postmortem concentrations correlated with extended postmortem intervals. Other drugs, such as alprazolam, were likely to have concentration discrepancies when short antemortem to death intervals were coupled with extended postmortem intervals. While still others, such as midazolam followed the expected pattern of metabolism and elimination, which often resulted in decreased postmortem concentrations. The importance of drug stability was displayed when reviewing the clonazepam/ 7-aminoclonazepam data, as the parent drug commonly converted to its metabolite even when stored in the presence of a preservative. In instances of decreasing postmortem drug concentrations the effect of refrigerated storage could not be ruled out. A stability experiment, which contained codeine, produced data that indicated concentrations could continue to decline under the current storage conditions. The cumulative data gathered for this experiment was used to identify concentration trends, which subsequently aided in the development of interpretive considerations for the specific analytes examined in the study.

Forecasting enrollment in differential assessment programs using cellular automata

Urban growth models have been used for decades to forecast urban development in metropolitan areas. Since the 1990s cellular automata, with simple computational rules and an explicitly spatial architecture, have been heavily utilized in this endeavor. One such cellular-automata-based model, SLEUTH, has been successfully applied around the world to better understand and forecast not only urban growth but also other forms of land-use and land-cover change, but like other models must be fed important information about which particular lands in the modeled area are available for development. Some of these lands are in categories for the purpose of excluding urban growth that are difficult to quantify since their function is dictated by policy. One such category includes voluntary differential assessment programs, whereby farmers agree not to develop their lands in exchange for significant tax breaks. Since they are voluntary, today’s excluded lands may be available for development at some point in the future. Mapping the shifting mosaic of parcels that are enrolled in such programs allows this information to be used in modeling and forecasting. In this study, we added information about California’s Williamson Act into SLEUTH’s excluded layer for Tulare County. Assumptions about the voluntary differential assessments were used to create a sophisticated excluded layer that was fed into SLEUTH’s urban growth forecasting routine. The results demonstrate not only a successful execution of this method but also yielded high goodness-of-fit metrics for both the calibration of enrollment termination as well as the urban growth modeling itself.

Essays in Open Economy Macroeconomics

Research macroeconomists have witnessed remarkable methodological developments in mathematical, statistical, and computational tools during the last two decades. The three essays in this dissertation took advantage of these advances to analyze important macroeconomic issues. The first essay, “ Habit Formation, Adjustments Costs, and International Business Cycle Puzzles” analyzes the extent to which incorporating habit formation and adjustment costs in investment in a one-good two-country general equilibrium model would help overcome some of the international business cycle puzzles. Unlike standard results in the literature, the model generates persistent, cyclical adjustment paths in response to shocks. It also yields positive cross-country correlations in consumption, employment, investment, and output. Cross-country correlations in output are higher than the ones in consumption. This is qualitatively consistent with the stylized facts. These results are particularly striking given the predicted negative correlations in investment, employment, and output that are typically found in the literature. The second essay, “Comparison Utility, Endogenous Time Preference, and Economic Growth,” uses World War II as a natural experiment to analyze the degree to which a model where consumers' preferences exhibit comparison-based utility and endogenous discounting is able to improve upon existing models in mimicking the transitional dynamics of an economy after a shock that destroys part of its capital stock. The model outperforms existing ones in replicating the behavior of the saving rate (both on impact and along the transient paths) after this historical event. This result brings additional support to the endogenous rate of time preference being a crucial element in growth models. The last essay, “Monetary Policy under Fear of Floating: Modeling the Dominican Economy,” presents a small scale macroeconomic model for a country (Dominican Republic) characterized by a strong presence of fear of floating (reluctance to have a flexible exchange rate regime) in the conduct of monetary policy. The dynamic responses of this economy to external shocks that are of interest for monetary policy purposes are analyzed under two alternative interest rate policy rules: One being the standard Taylor rule and another that responds explicitly to deviations of the exchange rate with respect to its long-term trend.

Macroalgae Decrease Growth and Alter Microbial Community Structure of the Reef-Building Coral, Porites astreoides

With the continued and unprecedented decline of coral reefs worldwide, evaluating the factors that contribute to coral demise is of critical importance. As coral cover declines, macroalgae are becoming more common on tropical reefs. Interactions between these macroalgae and corals may alter the coral microbiome, which is thought to play an important role in colony health and survival. Together, such changes in benthic macroalgae and in the coral microbiome may result in a feedback mechanism that contributes to additional coral cover loss. To determine if macroalgae alter the coral microbiome, we conducted a field-based experiment in which the coral Porites astreoides was placed in competition with five species of macroalgae. Macroalgal contact increased variance in the coral-associated microbial community, and two algal species significantly altered microbial community composition. All macroalgae caused the disappearance of a γ-proteobacterium previously hypothesized to be an important mutualist of P. astreoides. Macroalgal contact also triggered: 1) increases or 2) decreases in microbial taxa already present in corals, 3) establishment of new taxa to the coral microbiome, and 4) vectoring and growth of microbial taxa from the macroalgae to the coral. Furthermore, macroalgal competition decreased coral growth rates by an average of 36.8%. Overall, this study found that competition between corals and certain species of macroalgae leads to an altered coral microbiome, providing a potential mechanism by which macroalgae-coral interactions reduce coral health and lead to coral loss on impacted reefs.

Measuring the Impact of Melaleuca quinquenervia Biochar Application on Soil Quality, Plant Growth, and Microbial Gas Flux

Biochar has been heralded a mechanism for carbon sequestration and an ideal amendment for improving soil quality. Melaleuca quinquenervia is an aggressive and wide-spread invasive species in Florida. The purpose of this research was to convert M. quinquenervia biomass into biochar and measure how application at two rates (2% or 5% wt/wt) impacts soil quality, plant growth, and microbial gas flux in a greenhouse experiment using Phaseolus vulgaris L. and local soil. Plant growth was measured using height, biomass weight, specific leaf area, and root-shoot ratio. Soil quality was evaluated according to nutrient content and water holding capacity. Microbial respiration, as carbon dioxide (CO2), was measured using gas chromatography. Biochar addition at 5% significantly reduced available soil nutrients, while 2% biochar application increased almost all nutrients. Plant biomass was highest in the control group, p2 flux decreased significantly in both biochar groups, but reductions were not long term.

Continuous property measurement techniques and physics based mathematical model for frost growth control

The development of a new set of frost property measurement techniques to be used in the control of frost growth and defrosting processes in refrigeration systems was investigated. Holographic interferometry and infrared thermometry were used to measure the temperature of the frost-air interface, while a beam element load sensor was used to obtain the weight of a deposited frost layer. The proposed measurement techniques were tested for the cases of natural and forced convection, and the characteristic charts were obtained for a set of operational conditions. ^ An improvement of existing frost growth mathematical models was also investigated. The early stage of frost nucleation was commonly not considered in these models and instead an initial value of layer thickness and porosity was regularly assumed. A nucleation model to obtain the droplet diameter and surface porosity at the end of the early frosting period was developed. The drop-wise early condensation in a cold flat plate under natural convection to a hot (room temperature) and humid air was modeled. A nucleation rate was found, and the relation of heat to mass transfer (Lewis number) was obtained. It was found that the Lewis number was much smaller than unity, which is the standard value usually assumed for most frosting numerical models. The nucleation model was validated against available experimental data for the early nucleation and full growth stages of the frosting process. ^ The combination of frost top temperature and weight variation signals can now be used to control the defrosting timing and the developed early nucleation model can now be used to simulate the entire process of frost growth in any surface material. ^

Developing performance measures of mangrove wetlands using simulation models of hydrology, nutrient biogeochemistry, and community dynamics

The goal of mangrove restoration projects should be to improve community structure and ecosystem function of degraded coastal landscapes. This requires the ability to forecast how mangrove structure and function will respond to prescribed changes in site conditions including hydrology, topography, and geophysical energies. There are global, regional, and local factors that can explain gradients of regulators (e.g., salinity, sulfides), resources (nutrients, light, water), and hydroperiod (frequency, duration of flooding) that collectively account for stressors that result in diverse patterns of mangrove properties across a variety of environmental settings. Simulation models of hydrology, nutrient biogeochemistry, and vegetation dynamics have been developed to forecast patterns in mangroves in the Florida Coastal Everglades. These models provide insight to mangrove response to specific restoration alternatives, testing causal mechanisms of system degradation. We propose that these models can also assist in selecting performance measures for monitoring programs that evaluate project effectiveness. This selection process in turn improves model development and calibration for forecasting mangrove response to restoration alternatives. Hydrologic performance measures include soil regulators, particularly soil salinity, surface topography of mangrove landscape, and hydroperiod, including both the frequency and duration of flooding. Estuarine performance measures should include salinity of the bay, tidal amplitude, and conditions of fresh water discharge (included in the salinity value). The most important performance measures from the mangrove biogeochemistry model should include soil resources (bulk density, total nitrogen, and phosphorus) and soil accretion. Mangrove ecology performance measures should include forest dimension analysis (transects and/or plots), sapling recruitment, leaf area index, and faunal relationships. Estuarine ecology performance measures should include the habitat function of mangroves, which can be evaluated with growth rate of key species, habitat suitability analysis, isotope abundance of indicator species, and bird census. The list of performance measures can be modified according to the model output that is used to define the scientific goals during the restoration planning process that reflect specific goals of the project.

Improving resource management in virtualized data centers using application performance models

The rapid growth of virtualized data centers and cloud hosting services is making the management of physical resources such as CPU, memory, and I/O bandwidth in data center servers increasingly important. Server management now involves dealing with multiple dissimilar applications with varying Service-Level-Agreements (SLAs) and multiple resource dimensions. The multiplicity and diversity of resources and applications are rendering administrative tasks more complex and challenging. This thesis aimed to develop a framework and techniques that would help substantially reduce data center management complexity.^ We specifically addressed two crucial data center operations. First, we precisely estimated capacity requirements of client virtual machines (VMs) while renting server space in cloud environment. Second, we proposed a systematic process to efficiently allocate physical resources to hosted VMs in a data center. To realize these dual objectives, accurately capturing the effects of resource allocations on application performance is vital. The benefits of accurate application performance modeling are multifold. Cloud users can size their VMs appropriately and pay only for the resources that they need; service providers can also offer a new charging model based on the VMs performance instead of their configured sizes. As a result, clients will pay exactly for the performance they are actually experiencing; on the other hand, administrators will be able to maximize their total revenue by utilizing application performance models and SLAs. ^ This thesis made the following contributions. First, we identified resource control parameters crucial for distributing physical resources and characterizing contention for virtualized applications in a shared hosting environment. Second, we explored several modeling techniques and confirmed the suitability of two machine learning tools, Artificial Neural Network and Support Vector Machine, to accurately model the performance of virtualized applications. Moreover, we suggested and evaluated modeling optimizations necessary to improve prediction accuracy when using these modeling tools. Third, we presented an approach to optimal VM sizing by employing the performance models we created. Finally, we proposed a revenue-driven resource allocation algorithm which maximizes the SLA-generated revenue for a data center.^

Changing Bacterial Growth Efficiencies across a Natural Nutrient Gradient in an Oligotrophic Estuary

Recent studies have characterized coastal estuarine systems as important components of the global carbon cycle. This study investigated carbon cycling through the microbial loop of Florida Bay by use of bacterial growth efficiency calculations. Bacterial production, bacterial respiration, and other environmental parameters were measured at three sites located along a historic phosphorus-limitation gradient in Florida Bay and compared to a relatively nutrient enriched site in Biscayne Bay. A new method for measuring bacterial respiration in oligotrophic waters involving tracing respiration of 13C-glucose was developed. The results of the study indicate that 13C tracer assays may provide a better means of measuring bacterial respiration in low nutrient environments than traditional dissolved oxygen consumption-based methods due to strong correlations between incubation length and δ13C values. Results also suggest that overall bacterial growth efficiency may be lower at the most nutrient limited sites.

Quorum Sensing and Microbial Interactions in Coral Black Band Disease and Coral-Associated Bacteria

The black band disease (BBD) microbial consortium often causes mortality of reef-building corals. Microbial chemical interactions (i.e., quorum sensing (QS) and antimicrobial production) may be involved in the BBD disease process. Culture filtrates (CFs) from over 150 bacterial isolates from BBD and the surface mucopolysaccharide layer (SML) of healthy and diseased corals were screened for acyl homoserine lactone (AHL) and Autoinducer-2 (AI-2) QS signals using bacterial reporter strains. AHLs were detected in all BBD mat samples and nine CFs. More than half of the CFs (~55%) tested positive for AI-2. Approximately 27% of growth challenges conducted among 19 isolates showed significant growth inhibition. These findings demonstrate that QS is actively occurring within the BBD microbial mat and that culturable bacteria from BBD and the coral SML are able to produce QS signals and antimicrobial compounds. This is the first study to identify AHL production in association with active coral disease.

Improving Resource Management in Virtualized Data Centers using Application Performance Models

The rapid growth of virtualized data centers and cloud hosting services is making the management of physical resources such as CPU, memory, and I/O bandwidth in data center servers increasingly important. Server management now involves dealing with multiple dissimilar applications with varying Service-Level-Agreements (SLAs) and multiple resource dimensions. The multiplicity and diversity of resources and applications are rendering administrative tasks more complex and challenging. This thesis aimed to develop a framework and techniques that would help substantially reduce data center management complexity. We specifically addressed two crucial data center operations. First, we precisely estimated capacity requirements of client virtual machines (VMs) while renting server space in cloud environment. Second, we proposed a systematic process to efficiently allocate physical resources to hosted VMs in a data center. To realize these dual objectives, accurately capturing the effects of resource allocations on application performance is vital. The benefits of accurate application performance modeling are multifold. Cloud users can size their VMs appropriately and pay only for the resources that they need; service providers can also offer a new charging model based on the VMs performance instead of their configured sizes. As a result, clients will pay exactly for the performance they are actually experiencing; on the other hand, administrators will be able to maximize their total revenue by utilizing application performance models and SLAs. This thesis made the following contributions. First, we identified resource control parameters crucial for distributing physical resources and characterizing contention for virtualized applications in a shared hosting environment. Second, we explored several modeling techniques and confirmed the suitability of two machine learning tools, Artificial Neural Network and Support Vector Machine, to accurately model the performance of virtualized applications. Moreover, we suggested and evaluated modeling optimizations necessary to improve prediction accuracy when using these modeling tools. Third, we presented an approach to optimal VM sizing by employing the performance models we created. Finally, we proposed a revenue-driven resource allocation algorithm which maximizes the SLA-generated revenue for a data center.