Project

Efficient stock market theory in Latin-america

Goal: Evaluar la eficiencia de los mercados desde diferentes metodologías

Methods: Regression Model, MATLAB Simulation, GARCH, Economic Efficiency, Random Walks

Updates
0 new
4
Recommendations
0 new
6
Followers
0 new
7
Reads
0 new
117

Project log

Leonardo Talero
added 2 research items
En Colombia existen centros tradicionales de acopio donde se comercializan y distribuyen productos agrícolas. El precio de venta de estos productos es publicado por instituciones del Gobierno en informes semanales. Con el objetivo de determinar si las series de precios siguen una caminata aleatoria y si existen relaciones entre los centros, se aplican seis pruebas que evalúan la hipótesis de eficiencia débil y la prueba de causalidad de Granger. Para ello se analiza el precio histórico semanal de 28 productos agrícolas comercializados en seis mercados, durante la primera semana de 2013 a la última de 2017. Los principales resultados indican que los mercados tienden a la eficiencia, aunque no tienen el mismo nivel, ya que esta varía según el producto que se comercialice. Además, los centros en Manizales, Barranquilla y Villavicencio influyen sobre los precios de los mercados de Bogotá, Bucaramanga y Medellín
The main objective of this article is to develop a Cellular Automaton Model in which more than one type of stockbroker interact, and where the use and exchange of information between investors describe the complexity measured through the estimation of the Hurst exponent. This exponent represents an efficient or random market when it has a value equal to 0.5. Thanks to the various proposals, it can be determined in this investigation that a rational component must exist in the simulator in order to generate an efficient behavior. © 2017 Universidad Nacional Autónoma de México, Facultad de Contaduría y Administración.
Leonardo Talero
added a research item
In Colombia, there are traditional food supply centers to trade agricultural products. Government institutions publish sale prices of these products in weekly reports. In order to determine the random characteristics of these price series and if there are relationships between the centers, we use a battery of six different tests to examine the weak form of efficient market hypothesis and the Granger causality test. For this, this work collected the historical weekly price of 28 agricultural products, taking into account six markets, during the first week of 2013 to the last week of 2017. The main results indicate that markets tend towards efficiency, however, the efficiency depends on the product traded. In addition, the centers in Manizales, Barranquilla and Villavicencio influence the prices of the Bogotá, Bucaramanga and Medellín markets.
Leonardo Talero
added a research item
El objetivo principal del presente trabajo es evaluar las metodologías para la construcción de mercados bursátiles artificiales, determinando, primero, un modelo generador de precios para el mercado colombiano; segundo, simulando el modelo autómata conductual propuesto por Fan en 2009; y, tercero, aplicando un modelo racional propuesto mediante la adaptación de los Filtros de Alexander. Los modelos son contrastados mediante su consenso (volumen de compras, retenciones y ventas). Durante esta investigación se establece que el modelamiento de agentes racionales permite un comportamiento menos predictivo y un mayor impacto en el uso de la información, similar a la Hipótesis de Eficiencia del Mercado; además, los mercados bursátiles celulares conductuales y racionales no convergen; por ende, se plantea para futuras investigaciones la necesidad de construir modelos racional-conductuales con algoritmos de estimación de precios.
Juan Benjamín Duarte Duarte
added an update
Dear colleagues,
In this update, I share a Conference Paper about a comparative study between Colombia, the United States and Spain, focus on stock markets behavioral biases.
This paper is in Spanish.
Greetings!
Abstract:
The objective of this article is to evaluate the rationality in the stock markets based on behavioral biases, using the main indexes of the stock exchanges of Colombia (Colcap), United States (S & P 500) and Spain (Ibex 35), during the period between January 3, 2000 to August 17, 2016, divided into three sub periods: pre crisis (2000 - 2008), crisis (2008 - 2010) and postcrisis (2010 - 2016), taking as variables of study the returns and transaction volume of the mentioned indexes, using the model proposed by Dhaoui. The results obtained detect the presence of different biases in the markets analyzed in their different periods of study.
 
Juan Benjamín Duarte Duarte
added a research item
El objetivo del presente artículo es evaluar la racionalidad en los mercados bursátiles a partir de sesgos del comportamiento, utilizando los principales índices de las bolsas de valores de Colombia (Colcap), Estados Unidos (S&P 500) y España (Ibex 35), durante el periodo comprendido entre enero de 2000 y agosto de 2016, dividido en tres sub periodos entorno a la crisis subprime: pre crisis (2000 - 2008), crisis (2008 - 2010) y poscrisis (2010 - 2016), tomando como variables de estudio la rentabilidad y el volumen de transacción de los índices mencionados, utilizando el modelo propuesto por Dhaoui [1]. Los resultados obtenidos detectan la presencia de diferentes sesgos en los mercados analizados en sus distintos periodos de estudio.
Leonardo Talero
added a research item
El objetivo principal de este artículo es desarrollar un modelo autómata celular en el que interactúen más de un tipo de agentes bursátiles, donde el uso y el intercambio de información entre los inversores describen la complejidad medida a través de la estimación del coeficiente de Hurst, que representa un mercado eficiente o aleatorio al tener un valor igual a 0.5. Gracias a las variantes propuestas, en esta investigación se puede determinar que debe existir un componente racional en el simulador con el fin de generar un comportamiento eficiente.
Leonardo Talero
added an update
In our works, we use a Hurst estimation variant developed by Andrew Lo, in this update, we share a script created in MATLAB (file.m) whit a random walk generator and Andrew Lo Hurst estimation.
 
Juan Benjamín Duarte Duarte
added a research item
This research investigates whether the major stock markets in Latin America (Brazil, Mexico, Chile, Colombia, Peru and Argentina) exhibited herd behavior over the period January 2, 2002 to June 30, 2014, using the variation in the returns overall and by sector in the most representative stock market index in each country, using the model proposed by Christie y Huang (1995). The results do not reveal any herd behavior in the total market, or in the sectors of the markets examined in the study.
Leonardo Talero
added a research item
La presente investigación busca evaluar el nivel de complejidad del mercado latinoamericano, mediante la construcción de un modelo autómata celular. Para ello se estudian seis índices bursátiles: COLCAP, IPSA, MERVAL, MEXBOL, SPBLPGPT e IBOV, en el periodo 2004-2016. Estas series son analizadas a partir de su comportamiento estadístico, el ajuste de retornos y la estimación de su grado de complejidad. Este último es contrastado posteriormente con el nivel de complejidad obtenido mediante la simulación de un mercado bursátil artificial, y se concluye que los mercados latinoamericanos, a pesar de presentar diferencias, suelen tener tendencias similares, ya que su grado de complejidad no puede ser pronosticado por un modelo autómata celular conductual basado netamente en la imitación.
Leonardo Talero
added a research item
In recent years, scholars, politicians, and analysts have wondered whether it makes sense for some countries to renounce their currencies in order to adopt the currency of another country or to create a regional currency (Edwards, 2006). According to Bichara (2001), in order to initiate an advanced macroeconomic policy coordination or a single monetary union, it is necessary to study the economic interdependence and the symmetry of shocks. Taking into account the formulated by this author and others such as Bayoumi (1992) and Kempa (2002), who make the case for the theory of Optimal Currency Areas. The present work focuses on evaluating one of the aspects generally posed in the literature, that is to say, the economic interdependence of the countries that compose a region, which in this case is the Pacific Alliance (Chile, Colombia, Peru, and Mexico), including the United States of America. Although the latter does not belong to the Pacific Alliance, may have some level of relation with the mentioned countries due to commercial and financial reasons. To achieve this, we used correlation, cross-correlation and causality tests, following the methodology of Granger (1969). The baseline data are the quarterly Gross Domestic Product (GDP) of the nations under study for the period from the first quarter of 2000 to the second quarter of 2016, including a pre- and post-Pacific Alliance analysis, which takes place before and after the first quarter of 2011.
Juan Benjamín Duarte Duarte
added an update
Project goal
Evaluar la eficiencia de los mercados desde diferentes metodologías
Background and motivation
A considerable increment of direct foreign investment (IED) has been observed in Latin America in the past few years. This increment is mainly the result of two synergic facts: on one hand, the rapid growth of natural resources exploitation and regional infrastructure by countries such as the USA (17%), The Netherlands (13%), China (9%), Canada (4%), and Spain (4%), among other [ CEPAL (2011)]. On the other hand, the fact that current economic crisis results from the fall of the subprime in the U.S. in the year 2008. This has encouraged foreign investors to look at Latin America as a promising scenario for the generation of wealth [IMF (2010)].
As a consequence of this economic boost, gross capital flows have increased which, in turn, has led to consolidation of the bidirectional process of integration in the international financial market [World Bank (2012)]. This stock exchange dynamics in Latin America justifies the conduction of empirical contrasts of the weak forms of stock market efficiency during the last 10.6 years considering that it is one of the basic premises of important asset valuation models such as the Capital Asset Pricing Model (CAPM) and the Pricing Theory Arbitration model or APT usually utilized in merging and business acquisition processes.
The definition of Efficient Market is associated to the concept of equity of conditions proposed by the Italian mathematician Cardano (1565) in “The Book on Games of Chance”, as a fundamental principal for gambling. Applying this concept to stock exchange markets, it could be stated that this is the idea underlying the theory on efficient stock exchange markets, particularly regarding the equity of conditions in the information sector, makes the stock exchange market a “fair game” in which no agent can obtain extraordinary profits on a systematic basis from any type of information.
The modern definition of 11 Efficient Market is based on the contributions from Bachelier (1900); Samuelson (1965); Mandelbrot (1963) and (1966); Malkiel (1992) and Fama (1965a), (1991) and (1998). These authors, in general terms, propose that a market is efficient if it has enough liquidity and economic rationality by the agents so any type of relevant information is absorbed by prices instantaneously, thus generating a random behaviour in them and making systematic prediction impossible. Taking into account that, this project focus on the Latin-America stock markets behavioral, testing hypothesis and analyzing economics drivers.
 
Juan Benjamín Duarte Duarte
added 8 research items
This paper assesses the existence of the size effect on the most important stock markets in Latin America (Argentina, Brazil, Chile, Colombia, Mexico and Peru) for the period between 2002 and 2012, using the cross-section contrast methodology of the size effect in the CAPM context. Results show that there is reversed effect in some of the Latin American markets.
In this paper we evaluate the Mexico stock market efficiency from empirical testing of anomalies, both the Price and Quotes Index (IPC) and their three main actions (AMXL, WALMEXV and FEMSAUBD). The methodology focuses on estimating binary multivariate regression models for the period between January 2002 and April 2013. This work differs from previous research because subperiods are analyzed according to market trends upward or downward, in order to evaluate simultaneously the evolution of these anomalies through time and their behavior in periods of boom and crisis.
The efficient market hypothesis states that financial asset returns follow a random walk and depend on the information made available to the market instantly, therefore they can not be predicted. On the other hand, the Fractal market hypothesis says that prices depend of each behavior investor and his investment horizon, producing chaotic behavior in the markets. This paper tests the existence of chaotic behavior in major financial series of the Colombian stock market using the Hurst coefficient, whose estimation can be affected by autocorrelation. Therefore, the first part of the methodology focuses on removing the autocorrelations by ARIMA and GARCH filters, while the second part corresponds to the detectection of behavior by calculating the Hurst coefficient. The results reveal that the Colombian financial assets are persistent.
Juan Benjamín Duarte Duarte
added a research item
In recent years the behavior of the agents and what motivates them to make their investment decisions; ithas been the focus of study by many researchers in the fields of economics, finance and related fields. Withthis in mind, this article seeks to test empirically one of these behaviors, the herd effect, using the modelsproposed by Christie y Huang (1995) and Chang, Cheng y Khorana (2000), in the most representative indexand the sectors that compose it, of the main markets in Latin America (Brazil, Mexico, Chile, Colombia,Peru and Argentina) in the period 2002-2014, as well as in the sub periods pre and post crisis caused by thesubprime mortgage. The findings of this research show that the herd effect is present in the most representativestock index in Colombia, Chile and Peru and in some of the sectors that comprise it, either in the total periodand/or in sub periods pre and post crisis; in the Brazilian stock market, the effect is present in the bankingsector in the total period and in the sub period of pre-crisis; Argentina stock market that effect is present inthe banking sector and in the oil and gas sector, and Mexico stock market there is no evidence of this effect.
Juan Benjamín Duarte Duarte
added 11 research items
In this paper it fits ARIMA and GARCH models, for the major financial series Colombian stock market,until those which do not exhibit significant autocorrelation in the residuals. This, in order to allow the use of verification tools chaotic behavior in stock markets, which implies non-linearity in prices and dependence in the short term compared to its historical values, allowing the prediction, based in endogenous information. The methodology is to test in the first instance, the stationarity in the returns of the stock series, using the Aumengted Dickey-Fuller. Then, using tests such as QLB and Barttlet determine whether the series exhibit significant autocorrelations in some of its lags, thus justifying the use of the methodology of Box-Jenkins to generate ARIMA and / or GARCH adjustments, to obtain models with residuals not autocorrelated. Finally, we choose the best fit for each stock assets, using the Akaike criterion.This as a preamble to test the hypothesis of chaotic behavior in the Colombian financial series.
In this document are proposed a procedure to estimate the systematic risk coefficient (beta) for sectors, industries and companies listed or not in financial markets in emerging countries. The exposed methodology is a combination of the Capital Asset Pricing Model (CAPM) and the relative valuation. The main contribution of this paper, suggests the use of order procedures sectorial and individual, arising as an alternative solution to the endogeneity problem that can occur when using the traditional model (market model) to estimate the Beta coefficient in an emerging country.
The efficient stock market hypothesis is one of the basic assumptions of asset pricing models such as the Capital Asset Pricing Model and Arbitrage Pricing Theory and states that financial asset prices can not be predicted since they behave randomly. Conversely, the fractal market hypothesis states that prices have chaotic structure and could be predicted with nonlinear models, rejecting the efficient market hypothesis and invalidating the assumptions of asset pricing models. This paper seeks to evaluate the chaotic behavior in the Colombian stock market in order to reject or accept the efficient market hypothesis, using tools such as price charts, recurrence plots, correlation dimensión, Hurst coefficient, Lyapunov exponent and BDS test. The results show that the Colombian financial assets reveal chaotical behavior for “bull” market subperiods and randomized for “bear” market subperiods, supporting the fractal market hypothesis. These findings may support the use of nonlinear models for predicting prices in “bull” market episodes and reject the Colombian stock market efficiency.
Juan Benjamín Duarte Duarte
added a project goal
Evaluar la eficiencia de los mercados desde diferentes metodologías