Aproximaci'on pr'actica a los m'etodos de selecci'on de portafolios de inversi'on

summary

The gist

This paper explores practical approaches to portfolio selection methods, delving into portfolio theory concepts such as expected return, variance, asset correlation, and opportunity sets.

In short

The study explored practical portfolio selection using Markowitz and Mean-Semivariance models, comparing their performance against real data from 95 companies. It found that the Mean-Semivariance model offered a better fit. Genetic algorithms were introduced as a computational method to find optimal portfolios, proving effective for complex problems involving transaction costs and integer constraints.

Key concepts

Markowitz Model
This model is used to find the efficient frontier by balancing expected return against risk (variance). It uses quadratic programming to determine portfolios that either maximize return for a set risk level or minimize risk for a target return.
Mean-Semivariance Model
An alternative model that goes beyond simple variance. It incorporates asset return distributions' skewness and kurtosis, providing a more comprehensive view of risk and return than the standard Mean-Variance approach.
Genetic Algorithms (AGs)
A computational method used to optimize portfolio selection by exploring large search spaces effectively. They are useful for finding global optima in complex problems that are hard for traditional deterministic methods to solve.
Integer Constraints
Practical limitations where the number of shares purchased must be whole numbers. These constraints, along with transaction costs, make portfolio optimization much more complex and necessitate heuristic methods like AGs.

Terminology used across episodes

This episode discusses

The paper

Aproximaci'on pr'actica a los m'etodos de selecci'on de portafolios de inversi'on · Read on arXiv

Centro de Estudios en Computaci펪vanzada (CECAv) Universidad Nacional Autona MX Centro de Investigaci펮n e Innovaci펬nologias de la InformaciÓn y Comunicaci펪vanzada (INFOTEC)

This paper explores the practical approach to portfolio selection methods for investments. The study delves into portfolio theory, discussing concepts such as expected return, variance, asset correlation, and opportunity sets. It also presents the efficient frontier and its application in the Markowitz model, which employs mean-variance optimization techniques. An alternative approach based on the mean-semivariance model is introduced. This model accounts for the skewness and kurtosis of the asset return distribution, providing a more comprehensive view of risk and return. The study also addresses the practical implementation of these models, including the use of genetic algorithms to optimize portfolio selection. Additionally, transaction costs and integer constraints in portfolio optimization are considered, demonstrating the applicability of the Markowitz model. -- Este documento explorar la aproximación práctica a los métodos de selección de portafolios para inversiones. El estudio profundiza en la teoría de los portafolios, discutiendo conceptos como el rendimiento esperado, la varianza, la correlación entre activos y los conjuntos de oportunidades. También presenta la frontera eficiente y su aplicación en el modelo de Markowitz, que utiliza técnicas de optimización media-varianza. Se introduce un enfoque alternativo basado en el modelo media-semivarianza. Este modelo tiene en cuenta la asimetría y la curtosis de la distribución de retornos de los activos, proporcionando una visión más completa de riesgo y rendimiento. El estudio también aborda la implementación práctica de estos modelos, incluyendo el uso de algoritmos genéticos para optimizar la selección de portafolios. Además, se consideran los costos de transacción y las restricciones enteras en la optimización del portafolio.

Transcript

Introduction to the show: ident: Quantitative Finance. The day in quantitative finance: models of markets, risk and prices.

Ruth: I'm Ruth.

Warren: And I'm Warren, here with Iris.

Ruth: We're looking at the paper: "Aproximaci'on pr'actica a los m'etodos de selecci'on de portafolios de inversi'on".

Warren: Here's the gist of the paper. This paper explores practical approaches to portfolio selection methods, delving into portfolio theory concepts such as expected return, variance, asset correlation, and opportunity sets.

Ruth: It presents the efficient frontier derived from the Markowitz model, which utilizes mean-variance optimization techniques.

Warren: An alternative approach based on the mean-semivariance model is introduced to provide a more comprehensive view of risk and return by accounting for skewness and kurtosis of asset return distributions.

Ruth: The study examines the practical implementation of these models, including transaction costs and integer constraints in portfolio optimization, demonstrating the applicability of the Markowitz model in real-world scenarios.

Paper summary: Ruth: So, we're talking about this paper now, "Aproximaci'on pr'actica a los m'etodos de selecci'on de portafolios de inversi'on." It looks like the main idea is taking all those classic portfolio selection ideas and making them actually work in the real world.

Warren: Right. This paper claims it looks at expected return, variance, asset correlation, and opportunity sets through these practical methods. It presents what we call the efficient frontier from the Markowitz model using mean-variance optimization techniques.

Iris: And they introduce a different way too, based on the mean-semivariance model which tries to account for skewness and kurtosis of asset return distributions to give a more complete picture of risk and return.

Ruth: It's not just theory, though. They also tackle how to actually implement these models in practice, including things like transaction costs and integer constraints when you're making those portfolio selections.

Warren: That practical bit is key because in reality, you can't always do pure math; you have to deal with those costs and the fact that you can only buy whole shares. They show how the Markowitz model applies in real-world scenarios.

Iris: And they look at how you find those optimal portfolios, whether it’s maximizing return for a certain risk level or minimizing risk for a specific target return using quadratic programming.

Ruth: They use quadratic programming to maximize expected return subject to a risk constraint, and also to minimize risk while hitting a target return. They also look at different measures of central tendency besides the mean, like the median or mode, when assessing portfolio performance.

Warren: I saw them examining ninety-five companies over about three years, from January two thousand five to December two thousand seven. They show how you can achieve different risk levels by setting target expected daily returns and finding the minimum variance portfolios for those targets.

Iris: And they compare the results from the Mean-Variance model against the Mean-Semivariance model, which actually shows that the mean-semivariance model fits actual realized returns better, showing a reduced mean error in estimation for both expected return and underestimation error.

Ruth: They also say that when you plot both frontiers on the same graph, they practically look identical, especially at the extremes. It suggests they are essentially describing the same line when you look at it closely.

Warren: And then there's this part where they introduce genetic algorithms as a computational method for optimizing portfolio selection itself.

Iris: They develop a specific genetic algorithm to optimize the Markowitz model using an alternative formulation, which is defined by maximizing a function that balances expected return and variance with some lambda parameter.

Ruth: That's interesting because they compare the solution from that standard quadratic programming approach with what this genetic algorithm finds, and both methods come out very similar, with the algorithm finding a solution close to the global optimum.

Warren: The paper suggests genetic algorithms are good at finding global optima in complex problems because they can explore huge search spaces effectively and handle problems where things aren't totally predictable.

Iris: And they point out that when you add practical stuff like transaction costs and integer constraints, heuristic methods like genetic algorithms become really effective tools for solving those complicated optimization problems.

Ruth: Specifically, those integer constraints let you figure out the exact number of shares to buy from each company based on how much capital you have available, which changes what the final portfolio looks like.

Warren: It's a lot of practical detail there—lot sizes influence composition and return—showing that this model is usable even when you add those real-world limitations.

Iris: The comparison between the Mean-Variance and Mean-Semivariance models suggests the mean-semivariance model might actually produce better portfolios because it penalizes underperformance, even if the difference isn't huge in actual returns.

Ruth: So, to wrap up this summary of "Aproximaci'on pr'actica a los m'etodos de selecci'on de portafolios de inversi'on," the main point is that it takes the standard math and makes it practical by adding more realistic elements like risk distribution details and computational tools.

Warren: The authors are Carlos Minutti Martínez, and this work comes out in two thousand twenty-four as arXiv:2410 point 11070v1. It’s essentially a guide on how to apply portfolio theory when you have to deal with real-world frictions.

Iris: It really highlights how important it is to move beyond just the simplest models and look at more detailed risk measures like skewness and kurtosis when you're building something that needs to match actual market behavior.

Ruth: And for the world, this suggests that when we try to build these investment strategies, we need those sophisticated tools like genetic algorithms because the search space gets too big for just pure deterministic math to handle cleanly.

Warren: It means finding a truly optimal portfolio in a messy financial environment requires blending those exact mathematical models with these powerful approximation techniques.

Conclusion: Ruth: So we've been looking at how this paper, "Aproximaci'on pr'actica a los m'etodos de selecci'on de portafolios de inversi'on," tackles those tricky portfolio selection ideas in the real world.

Warren: It’s about taking the classic Markowitz model and making it actually work when you have to deal with things like transaction costs and having to buy whole shares.

Iris: They introduce a mean-semivariance model too, which tries to get a better picture of risk by looking at things like skewness and kurtosis in how asset returns are distributed.

Ruth: And they show that this more detailed model actually fits the real data better than the simpler mean-variance one, even if the difference between them looks tiny when you just plot them side by side.

Warren: That means when we're trying to build a strategy, using that slightly more complex risk measure might give us a better starting point for what we expect to happen.

Iris: But they also show how you can use genetic algorithms to find the best settings for that Markowitz model, which is useful when the math gets too messy.

Ruth: So it boils down to this paper showing that you need these advanced tools—the mean-semivariance idea and the genetic algorithms—when you move from a simple textbook problem to something that has real-world costs and constraints.

Warren: It suggests that for anyone actually managing money, just sticking to the simplest formulas isn't enough; you gotta account for all those practical frictions.

Iris: And they flag that while the genetic algorithms find good approximate solutions, we still need to be careful about what those approximations mean in terms of true risk.

Ruth: It opens up a new way of thinking about how we choose portfolios when the underlying data is messy and you have to deal with real-world limitations.

More episodes

← Home