Tesis de Maestría en Economía

Permanent URI for this collection

https://repositorio.udesa.edu.ar/handle/10908/585

Browse

Selection of optimal window size and ensemble methods for rental price estimation with XGBoost: Evidence from Buenos Aires
(2024?) Sosa, Juan Bautista
In the context of forecasting rental prices of residential properties, the fact that the relationship between rents and property characteristics may be subject to significant changes over time can hinder the predictive performance of models that do not adjust to these changes. Previous studies have tested the accuracy of different statistical models in predicting rent prices, but none has focused on how the definition of the estimation window of observations used in the training stage can affect predictive performance. This paper explores the impact of selecting different window sizes in a forecasting experiment of rent prices in the City of Buenos Aires from 2020 to 2022 using the machine learning algorithm XGBoost. The results obtained for out-of-sample one-month-ahead forecasts indicate that the decision to use an expanding window or rolling windows is not trivial in terms of the achieved predictive performance. Among a set of rolling window sizes, the model with a size of 6 months yields the lowest error in terms of root mean square error and mean absolute error. Additionally, three methods that dynamically combine the predictions of models with different window sizes are tested. An ensemble method that outputs a weighted average of the predictions using time-varying weights based on the inverse of previous forecasting errors emerges as the superior strategy, producing a mean absolute percentage error of 19.7%. Overall, these results underscore the need to take into account the temporal dimension when selecting observations for training if the goal is to maximize out-of-sample predictive performance.
Beyond Pain Relief: Assessing OxyContin’s Impact on Foster Care New Entries
(2024-07) Bullano, Maria Emilia
In this study, I assess the causal impact of the early stages of the opioid crisis on new entries into foster care, focusing on the introduction of OxyContin in 1996 and the influence of Triplicate Prescription Programs (TPP) in the United States. My findings indicate a composition effect: in states without TPP, where OxyContin had greater market penetration, children entering foster care are more likely to come from households where parents abuse drugs. These results underscore the complex interplay between public health policies and child welfare outcomes, highlighting the need for integrated strategies addressing the broader implications of drug policy on vulnerable populations.
Effects of Weather Shocks on Labor Market Outcomes: Evidence from Argentina
(2024-07) de Marcos, Inés
How do weather shocks affect labor market outcomes in Argentina? This study examines the effect of five types of weather shocks —droughts, heatwaves, cold waves, and intense and excessive rainfall— on labor market outcomes in Argentina from 2001 to 2010. In the context of increasing climate variability, this study is particularly relevant for economies like Argentina, which are highly dependent on agriculture and have significant segments of the population vulnerable to these shocks. The study presents a unique data set that merges 50 years of high-resolution climatic data with the 2000 and 2010 national censuses. Additionally, it includes imputed income data derived from the 2005 and 2010 Household Survey (EPH) using a survey-to-census methodology. By employing a differences-in-differences approach, the study examines the effects of drought, heatwaves, cold waves, and intense and excessive rainfall on employment status, labor force participation, and labor income. Results show that drought, heatwaves, and cold waves have significant negative effects on labor market outcomes. Importantly, these effects are disproportionally adverse for poorer income quintiles and less-educated individuals, potentially exacerbating existing social vulnerabilities. The study contributes to the climate-economy literature, particularly on labor markets, by extending beyond the usual single weather approach and, methodologically, utilizing a survey-to-census income imputation technique. The findings highlight the critical importance of quantifying weather impacts to inform the design of public policies for increased resilience to weather shocks in middle-income countries like Argentina.
¿Quién es el siguiente? Impacto de suicidios presidenciales en la sociedad
(2024-06) Arispe Tejada, Samuel
La preocupación por el aumento de la tasa suicidios en diversos países se ve exacerbada por nuestra influencia mutua en un entorno cada vez más conectado, donde la información se propaga rápidamente a través de Internet y las redes sociales. Este estudio investiga cómo el suicidio de un ex presidente muy conocido en el Perú afecta la probabilidad de que ocurran suicidios en las fechas cercanas al evento, revelando un aumento del 71% en dicha probabilidad durante la primera semana después del suceso. Además, se identifican efectos heterogéneos con el nivel de cohesión social, religiosidad arraigo hacia la figura pública y acceso a Internet como posibles explicaciones de esta tendencia, lo que subraya la importancia de considerar estos elementos al desarrollar estrategias para prevenir los suicidios en nuestra sociedad.
Shining a Light on Resilience: Overcoming Hurricane Odile’s Impact on Economic Activity
(2024-06) Pasman, Clara María
Climate change-induced disasters pose a significant threat to economic activity in developing regions, particularly those with inadequate infrastructure services. Understanding the impact of climate-related hazards, achieving resiliency, and mitigating the consequences for affected communities are critical challenges for these regions. In this study, we focus on Hurricane Odile, which struck Baja California Sur, Mexico, in September 2014, severely affecting economic activity across localities. By analyzing high-frequency and highly-disaggregated nighttime light data (NTL) from NASA, combined with census data, we characterize the recovery path of the economic activity following the hurricane’s impact. Our methodology allows us to characterize recovery timelines and identify distinctive recovery patterns across localities. We find that Hurricane Odile caused an immediate drop in luminosity levels, reflecting a downturn in overall activity, with localities experiencing varying degrees of impact and recovery. Overall, economic recovery was not achieved over the year after the hurricane. Recovery patterns differ across localities, emphasizing the need for tailored mitigation policies and aid plans.
Inflación y riesgo en las decisiones de inversión de los hogares
(2024-05) Tanzi, Horacio Nicolás
Este estudio cuantifica los efectos de estado estacionario de la inflación sobre el riesgo en las decisiones de inversión de los hogares. En una economía pequeña y abierta, utilizando un modelo neokeynesiano de agentes heterogéneos (HANK), se introduce un activo riesgoso, el capital k, y un activo libre de riesgo, los bonos b. El retorno del capital se modela a través de una variante de Calvo-pricing, vinculándo el riesgo en k a la inflación. Los resultados indican que la inclusión del efecto de la inflación sobre el riesgo del capital provoca una recomposición de la cartera de los hogares, disminuyendo el stock de capital y aumentando el stock de activos seguros. Esta recomposición implica una disminución de la producción en estado estacionario.
Transiciones laborales hacia la informalidad luego de un sudden stop: el caso de Argentina 2018
(2024-03) Alvarez Kuhnle, Tomás
Este trabajo busca analizar los efectos que tiene un shock internacional sobre la composición de los mercados de trabajo en países emergentes con alta incidencia de la informalidad. Para ello se toma el sudden stop ocurrido en Argentina comparando los cuartos trimestres de 2017 y 2018. Se utilizan los datos de panel de la EPH para analizar las transiciones y se emplea un modelo logit multinomial para identificar las características personales y del puesto de trabajo que influyen en la probabilidad de cambiar de categoría laboral. Se observa que los empleados en el sector no transable tuvieron una mayor probabilidad de haber cambiado a un trabajo informal, mientras que aquellos que ya estaban en la informalidad era más probable que perdieran su empleo en comparación con los trabajadores en el sector transable. Posteriormente, se utiliza un modelo DSGE que diferencia entre sectores transables y no transables, así como entre empleo formal e informal, para simular el impacto del shock y explicar los hechos estilizados observados. Se encuentra que el modelo base no logra capturar de manera precisa las dinámicas de la relación formal-informal dentro de los sectores transable y no transable. Sin embargo, luego se realizan extensiones en busca de entender cuáles son los supuestos que pueden generar las dinámicas esperadas.
Shaking rivers and budget cycles : a comprehensive framework
(2024-03) Campabadal, Joaquín
The 1990s were host to a variety of political economy papers that tried to explain why and how incumbents would use fiscal resources strategically to get reelected. Under the main argument, a politician might be inclined to increase spending/reduce taxes when close to an election to increase their reelection probabilities, even if this goes against their preferences (i.e. a party increases expenditure even if it’s right-leaning) or if it harms the economy in the long run (by reducing the stream of resources in the future). Allegedly, this literature - called the political budget cycle literature - wanted to endow the experiences of the Reagan administration (among others) with a rationale, especially those near the end of his second term. In this thesis I will focus on two objectives. First, I will study the US case to try to understand whether these dynamics are present there, if they are heterogenous across parties and how they are affected by the likelihood of being reelected. Secondly, I will develop a theoretical framework in which parties use fiscal variables detrimentally to enhance their chances at reelection, and explore the interplay between election prospects and the behavior of the incumbent regarding fiscal variables.
Nowcasting economic activity in Argentina using newspaper articles
(2024-03) Romero, Juan Pablo
This paper explores leveraging unstructured textual data from Argentinian newspapers to nowcast the monthly Economic Activity Estimator (EMAE) of Argentina. Various economic uncertainty indexes are constructed by applying natural language processing techniques on local newspaper articles. The association between these indexes and the EMAE is analyzed through correlation analysis, Granger causality tests, and out-of-sample nowcasting exercises. The results suggest that the proposed indexes exhibit predictive value for nowcasting the EMAE across diverse data splits. While limitations exist regarding generalization across economic cycles, this study contributes in the viability of extracting valuable signals from news content to gain timely insights into economic trends, and highlights the potential for nowcasting key indicators from unstructured data as text mining capabilities and data availability continue expanding.
Should I stay or should I go? effects of being acquired on employee outcomes: evidence from Brazil
(2024-02) Liwski, Joaquín Matías
This paper examines the diverse impacts of corporate acquisitions on employees with data from Brazil’s formal employment sector (RAIS survey, 2007-2015) and the Thomson Reuters SDC Platinum Database. Employing a matched event studies design, I navigate labor market complexities, revealing heterogeneous effects at individual and firm levels. Notably, the analysis uncovers a significant short-term increase in log-wages for individuals in acquired firms, particularly in the immediate 0 to 1 years following events; however, separation outcomes predominantly occur in the same year as the acquisition. Job changes play a pivotal role, with incumbents and voluntary leavers experiencing positive outcomes, while involuntary terminations lead to declines, also in the long-run. The study explores age-related, educational, skill-based, and occupational differences, showing diverse impacts across these dimensions. Additionally, firm size emerges as a critical factor, influencing log-wages and separation outcomes. Ultimately, the change in firm productivity or firm-specific wage premiums drives observed effects, underlining one mechanismn of M&A impacts on the workforce.
Memory and Biased Reports: Evidence from a Pilot Experiment
(2024-02) Serramo, Martín Alexis
In certain situations, people’s earnings rely on their self-reports of private information to third-parties. Are self-reports more biased in a financially convenient direction when recalling the information is more challenging? I present evidence from the pilot of an experiment designed to investigate this question. In this experiment, I assess the effectiveness of an interference task in increasing the difficulty of remembering information to be reported later and test its impact on misreporting among participants incentivized to report what is convenient for them. I find no evidence of increased misreporting when participants engage in the interference task before making their reports. The interpretation of this result is complicated by the fact that the interference task reduces perceived memory accuracy but does not affect objective memory accuracy.
What Works to Strengthen Beekeeper Productivity and Climate Resilience? Evidence from México
(2024-02) Luca, Victoria
This paper studies the effectiveness of a technical assistance program aimed at improving the productivity and the climate resilience of beekeepers in rural Mexico. The program was provided free-of-charge to beekeepers who lost hives during tropical storm Cristobal in 2020. Program participants in 14 localities of Yucatan were offered four main program components: (i) basic inputs, (ii) a year-long training on beehive management, (iii) queen bees, and (iv) a specific training on queen breeding. Due to the difficulties of implementing the program during the COVID-19 Pandemic, participants in 3 other nearby localities were offered only basic inputs and serve as a control group. We use a difference-in-differences approach combined with entropy balance to control for pre-existing differences across groups. We take advantage of a detailed baseline and follow-up survey of 356 beekeepers who participated in the program. Treated beekeepers increased their number of hives, honey production, yields, and adoption of best practices substantially more than the control group, who only received basic inputs. While the provision of the queen bee asset had the largest impact on honey production and yield, the trainings were particularly important for the adoption of best practices, which bodes well for longer-term climate resilience.
Reassessing a cocaine shock : a contrary narrative from Rabo de Peixe
(2024-02) Cabral, Juan Andrés
This work seeks to analyze the effects of an exogenous shock increase in the supply of cocaine using São Miguel Island, Portugal as a case study. Located in the Azores archipelago, the island is politically organized as an autonomous region. In January 2001 half a ton of cocaine unexpectedly washed ashore on Rabo de Peixe, São Miguel, leading to media reports of various alterations on the island, such as a surge in drug dealers and a significant reduction in the drug's price. These events even formed the basis for a documentary and a Netflix series. However, my comprehensive analysis utilizing a Synthetic Control approach paints a contrasting picture. I aimed to examine the short and medium-term repercussions of the increased accessibility of cocaine on São Miguel's crime rates, unemployment, and deaths. Contrary to the narrative constructed by the media and popular culture, I found no significant effects on these outcomes. This study challenges existing media narratives and provides a robust, data-driven insight into the actual impact of this incident on São Miguel.
Federalismo fiscal y resultados electorales : Argentina 1985-2019
(2024-01) Cardozo, José Ignacio
Entender cómo funciona el federalismo fiscal y cuáles son sus implicancias es un tema de gran interés a la hora de estudiar economía política. Después de las reformas efectuadas en Argentina en la década de las ´90, sus efectos fueron intensificados y estudiados en la literatura. Este trabajo viene a aportar evidencia sobre el impacto de las transferencias discrecionales en el resultado de las elecciones a diputado nacional del partido incumbente. Mediante un panel dinámico y empleando el método Arellano-Bower con errores robustos, se analizaron las elecciones en los 24 distritos (23 provincias más CABA), desde 1985 hasta 2019 y se encontró que las transferencias corrientes afectan positivamente el resultado de estas elecciones. Estos resultados brindarían robustez al argumento de los votantes como liberales fiscales (Jones et al, 2012), puesto que los votantes premiarían, en forma de apoyo al candidato a legislador del partido incumbente, la capacidad del gobernador de conseguir fondos nacionales. También se aporta evidencia al razonamiento que sostiene intercambios informales entre gobernadores y presidentes (Ardanaz et al., 2014), e igualmente se valida, en cierta forma, el dominio que ejercen los caudillos provinciales sobre los legisladores con objetivo de aumentar su poder (Jones et al, 2002).
El rol de los salarios en el pass-through de tipo de cambio a precios al consumidor en Argentina
(2024-01) de la Vega, Diego Eduardo
Recientes trabajos de investigación han destacado la importancia de la rigidez salarial efectiva como modulador de la política monetaria y, en consecuencia, de los resultados económicos. El análisis de los salarios como canal de transmisión es relevante para el caso argentino dadas la inestabilidad macroeconómica imperante en las últimas décadas, que ha posibilitado la aparición de regímenes salariales con numerosas renegociaciones intraanuales y mecanismos de ajustes automáticos. En esta línea, el presente trabajo analiza el rol de los salarios como canal de transmisión de shocks cambiarios, haciendo uso de la heterogeneidad de precios y salarios evidenciada por las diferentes ramas de actividad económica que forman parte del Valor Agregado Bruto de Argentina. Para realizar este análisis se aplicó una metodología novedosa, basada en el trabajo de Cloyne et al. (2021), que incorpora al enfoque de Proyecciones Locales la descomposición Kitagawa-Blinder-Oaxaca, lo que permite evaluar en qué medida la respuesta de los precios ante un shock cambiario obedece a ajustes salariales. Los resultados muestran que los salarios juegan un rol significativo al modular la velocidad del traspaso a los precios de un shock cambiario.
Assessing the EMBI Impact of IMF Agreements : An Event Study
(2024) Arias, Julián Nicolás
The International Monetary Fund (IMF) plays a significant role on a global scale, functioning as a lender of last resort to evade financial or balance of payments crises in developing countries. This paper aims to evaluate the effects of various key dates of such programs. High frequency EMBI data is combined with a novel database including the exact days of mission initiation, plan size, and signature of the Stand-By Arrangement (SBA) and Extended Fund Facility (EFF) programs are employed to answer this question. In the years before 2008, the signing of the plan had a detrimental effect on the EMBI, while the beginning of the mission appeared to have no impact. However, from 2008 onwards, the start day of the mission begins to play a significant role in determining the future expectations of the program. This start day, in particular, has a negative impact on the EMBI, whereas the day of the plan size announcement or plan signing does not seem to have any effect.
Assessing the Effectiveness of ECB’s LSAPs in Mitigating Euro Area Fragmentation: Evidence from the Covid Crisis
(2024) Gotsis, Juan Manuel
This paper evaluates the effectiveness of Large-Scale Asset Purchases (LSAPs) implemented by the European Central Bank (ECB) during the Covid crisis in the Euro Area. Using intraday changes in sovereign bond yields and an event-study methodology centered around ECB monetary policy decisions, I assess the impact of LSAPs on long-term sovereign bond yields of seven Euro Area countries. The analysis reveals that LSAPs contributed to reducing risk premia in the Euro Area periphery, mitigating fragmentation, and facilitating the transmission of monetary policy across the region.
It’s a matter of trust : corruption scandals and political discontent in pandemic times
(2023?) Amaya, Elard
We study how corruption scandals lead to different forms of political discontent. We take advantage of the fact that the largest corruption scandal during the pandemic in Peru (the vacungate scandal) erupted during the fieldwork of a public opinion survey and use the quasi-exogenous variation in exposure to the corruption scandal arising from differences in interview dates across respondent. We find that the vacunagate scandal triggered a sizable increase in the perception of corruption, reduced trust in both national and local governments, and changed the perception of democracy as a suitable political system. These short-run effects are significantly higher among those directly affected by this disease. We also show that our results are robust to placebo test, changes in the bandwidth used, and an alternative model specification.
Curb your enthusiasm on sign-in bonuses : evidence from Coinbase’s Super Bowl campaign
(2023?) Tinti, Bernardo
This thesis explores the impact of a sign-in bonus campaign on mobile app usage, using Coinbase’s 2022 Super Bowl marketing campaign as a case study. Marketing campaigns can affect user behavior over different stages of their “customer journey”: getting to know the app, downloading and engaging with it. Based on user activity daily data and using the synthetic control method for rigorous impact evaluation, we find that the campaign had a significant short-term effect on app downloads, but no effect on posterior user engagement. These results cast doubts on the profitability of such marketing strategies.
Limitaciones computacionales de la econometría espacial en la era de big-data : comportamiento de modelos y tests espaciales
(2023-12) Quiroga, Adrian Ezequiel
Debido al constante crecimiento en el volumen y la granularidad de los datos espaciales, el campo de la econometría espacial necesita comprender y dimensionar los límites de sus metodologías y procesos para adaptarse a la era del big data. Este artículo expone los resultados de un estudio de simulación sobre los límites al momento de estimar modelos espaciales con grandes conjuntos de datos, extendiendo el estudio realizado por Arbia et al. (2019) con los algoritmos actuales y agregando un análisis sobre los límites de los test de detección espacial en este contexto. Para hacerlo se simula un modelo de rezago espacial (SLM), el cual se estima utilizando el método de máxima verosimilitud (MV) y el de mínimos cuadrados en dos etapas (MC2E). Observamos avances significativos, con mejoras de hasta un 35% en los tiempos de estimación de los modelos en comparación con los datos de 2019. Adicionalmente, llevamos a cabo un análisis de los requerimientos computacionales de estos métodos y profundizamos nuestra investigación con un análisis empírico, evidenciado la viabilidad de trabajar con muestras de hasta 784,000 observaciones. Estos resultados subrayan la adaptabilidad y pertinencia de la econometría espacial ante las exigencias actuales.

Browse

Browsing Tesis de Maestría en Economía by Issue Date

Results Per Page

Sort Options