Tesis a presentar miércoles 15:
3:00 p.m. a 3:30 p.m.
How information disclosure does affect liquidity? Evidence from an Emerging Market
Estudiante: Ignacio Arango Gómez
Cross-sectional models show that information disclosure increases stock liquidity, but dynamic models show an opposite relation in news releases days. We address this puzzle by studying the effects of information arrival on liquidity and its determinants. We use trade and quote data from Colombia for 2015 and 2016, along with the complete database of news releases as reported by firms to the regulator. Our results suggest that news releases marginally reduce volatility but increase both informed and uninformed trading. All in all, the time series negative effect of news releases on liquidity appears explained by the increase of asymmetric information.
3:30 p.m. a 4:00 p.m.
Backtesting Expected Shortfall: Una aplicación en acciones de energía tradicional y renovable
Estudiante: Daniel Velásquez Gaviria
En esta investigación se realiza una aplicación al del BackTesting del VaR y del ES a las ocho principales acciones de energía renovable y tradicional pertenecientes al índice WilderHill New Energy Global Innovation y al índice Bloomberg World Energy utilizando las metodologías propuestas por Acerbi y Székely (2014, 2017) y extendiendo sus resultados a la distribución t-student asimétrica. En general se encuentra que el ES resulta ser una medición del riesgo más conservadora que el VaR, pero que se ajusta de manera más rápida y adecuada a momentos de alta volatilidad y efecto apalancamiento. Los resultados también evidencian una mejor cuantificación del riesgo por parte de la distribución t-Student y la t-Student asimétrica comparado con los resultados de la distribución Normal.
4:00 p.m. a 4:30 p.m.
Focused Estimation in Portfolio Selection Problems
Estudiante: Rosember Isidoro Guerra Urzola
We introduce the idea of Minimum Expected Loss estimates to propose an estimator for portfolio selection problems. In particular, we propose MELO estimators for the Global Minimum-Variance portfolio, Tangency Portfolio, and Treynor-Black model. We develop the theoretical framework, perform simulation exercises, and develop an empirical study to compare our MELO proposal to other competing alternatives in term of predictive power, and profitability. Our proposed exhibits better finite sample properties compared with competing alternatives, especially when the Tangency portfolio is taken as asset allocation strategy.
4:30 p.m. a 5:00 p.m.
The Role of Cognitive Abilities on Financial Literacy: New Experimental Evidence
Estudiante: Ana Melisa Muñoz Murillo
Demographic factors appear to be an important determinant of financial literacy, but there is much less evidence on the link between cognitive ability and financial Literacy. We add new experimental evidence to the literature on financial literacy determinants. We provide evidence that supports the key role of cognitive ability for Financial literacy, which according to the theory it is one of the fundamental factors that influence financial literacy. Our experimental setting allows us to (a) overcome particular limitations of the traditional multiple-choice questions, (b) provide compatible incentives, and (c) use a primary measure of cognitive abilities. We show that individuals with higher cognitive abilities are more financial literate. The positive association of cognitive ability with financial literacy holds even after controlling for the main confounding factors (i.e. non-cognitive abilities, human capital factors, and parental characteristics. Unlike previous studies, we and no evidence to support a gender gap on financial literacy levels.
Tesis a presentar jueves 16:
3:00 p.m. a 3:30 p.m.
Modeling, Estimation and Valuation of Options on Hydrological Contributions in the Colombian Electric System from a Stochastic Process
Estudiante: Juan Pablo Pérez Monsalve
This research aims to develop a price model for a financial derivative on weather, where the underlying asset is an index built on hydrological contributions in the colombian electricity system and its accumulation. To achieve this objective, we determine the stochastic model that describes the daily dynamic behavior of hydrological contributions and we build a gaussian estimation technique to find the parameters that best define the mean reversion process with periodic functional tendency. Once the estimation technique has been defined and validated, we proceed to apply it on the sample, which allows us to find the internal and external parameters of the process to simulate different paths and make a forecast. Finally, considering the stochastic modeling, we develop the price model for a financial derivative on weather, and we make the valuation of three financial options types from the model, results that are contrasted with the simulation of Monte Carlo for different cases.
3:30 p.m. a 4:00 p.m.
Inefficiency and Banks' Failure: A Joint Bayesian Estimation of a Stochastic Frontier Model and a Hazards Model
Estudiante: Jim Sánchez Gonzalez
In modeling bank failure, estimating inefficiency separately from the hazards model results in inefficient, biased, and inconsistent estimators. We develop a method to simultaneously estimate a stochastic frontier model and a hazards model using Bayesian techniques. This method overcomes issues related to two-stage estimation methods, allows for computing the marginal effects of the inefficiency over the probability of failure, and facilitates statistical inference of the functions of the parameters such as elasticities, returns to scale, and individual inefficiencies.
Simulation exercises show that our proposal performs better than two-stage maximum likelihood, especially when the available data is scarce. In addition, we find that inefficiency plays a statistically and economically significant role in determining the time to failure of U.S. commercial banks during 2001 to 2010.
4:00 p.m. a 4:30 p.m.
An alternative Gaussian Estimation Scheme for a general class of Reduced-form electricity price models
Estudiante: Juan Sebastián Marín Delgado
Understanding and modeling the evolution in time of electricity prices becomes of extreme importance when pricing power derivatives. Performing this task becomes particularly complex in the case of electricity due to certain distinctive characteristics, such as mean reversion, seasonality and the presence of spikes. The class of stochastic processes that adequately account for these characteristics are the so-called jump-diffusion processes. In that sense, this paper provides a new approach for estimating jump-diffusion processes for modeling electricity prices based on the decomposition of the series into the diffusion component and the spikes component, each of the components is estimated using maximum likelihood. The method is general for a broad class of jump-diffusion processes; however, we make special emphasis on the estimation scheme for the geman and roncoroni's model due to its broad usage for modeling electricity prices. Our approach showed better performance in terms of the overall mse when compared to the estimation technique proposed by geman and roncoroni. Finally, we illustrate our approach by specifying a reduced form model for the colombian electricity market price dynamics and applying our estimation scheme; the estimated model shows that the electricity price has usual characteristics such as mean reversion, yearly and 6-month effects as well as a significant dependence on the occurrence of el niño and la niña meteorological phenomena.
4:30 p.m. a 5:00 p.m.
Proximidad y Productividad Laboral en la Banca Comercial: Evidencia para Colombia basada en Microdatos
Estudiante: Alejandro Henao Montoya
Este trabajo muestra como la proximidad espacial entre las oficinas centrales de un banco comercial y sus sucursales afecta la productividad laboral de éstas. Utilizando datos de cerca de 348 millones de transacciones realizadas diariamente entre enero de 2014 y diciembre de 2016, se encuentra un efecto positivo y estadísticamente significativo entre la proximidad de las sucursales a la dirección general y la productividad laboral de las sucursales. Sucursales ubicadas a mayor distancia de las oficinas centrales o hacia las cuales los tiempos de transporte son mayores, son menos productivas. Los resultados son robustos al utilizar diferentes medidas de distancia (e.g., geográfica y tiempo de desplazamiento) y al controlar por características propias de las sucursales (e.g., tipo, antigüedad, tamaño, ubicación, tamaño del mercado, entre otras). Los resultados soportan la hipótesis de que la proximidad espacial entre el principal y el agente afecta la efectividad de las actividades de monitoreo y el flujo de información entre el principal y el agente.
Mayores informes