Portfolio selection with parsimonious higher comoments estimation

Lassance, Nathan; Vrins, Frédéric

doi:10.1016/j.jbankfin.2021.106115

Cited by 20 publications

(4 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The skewness and kurtosis coefficients as well as the Jarque-Bera test indicate that in almost all the cases, the underlying systematic risk factors are not univariate normally distributed. As expected, given the theoretical construction of the four techniques, the underlying factors are uncorrelated with each other in almost all the cases in the four databases, as the corresponding correlation matrices show 20 . In most of the cases, the correlation was zero and we couldn't reject the null hypothesis of non-correlation at a 5% of statistical significance, except in the case of the ninth non-linear component extracted using NNPCA in the four databases, where we reject the null hypothesis of non-correlation; nevertheless, the correlation value of this component with the rest of them was negligible 21 .…”

Section: Underlying Systematic Risk Structure: Statistical and Graphical Analysissupporting

confidence: 75%

“…On the other hand, factors computed by FA and ICA in some periods of the observations present some similarities as well, but not at the same level as NNPCA and PCA, in points of high volatility, they behave quite differently. In addition, 20 The correlation matrices of the underlying systematic risk factors extracted by the four techniques in the four databases and the entire test window are not included in this document. However, they are available upon request to authors.…”

Section: Underlying Systematic Risk Structure: Statistical and Graphical Analysismentioning

confidence: 99%

See 1 more Smart Citation

Comparison of Statistical Underlying Systematic Risk Factors and Betas Driving Returns on Equities

Cortés

Porras

Monte

2021

REMEF

View full text Add to dashboard Cite

The objective of this paper is to compare four dimension reduction techniques used for extracting the underlying systematic risk factors driving returns on equities of the Mexican Market. The methodology used compares the results of estimation produced by Principal Component Analysis (PCA), Factor Analysis (FA), Independent Component Analysis (ICA), and Neural Networks Principal Component Analysis (NNPCA) under three different perspectives. The results showed that in general: PCA, FA, and ICA produced similar systematic risk factors and betas; NNPCA and ICA produced the greatest number of fully accepted models in the econometric contrast; and, the interpretation of systematic risk factors across the four techniques was not constant. Additional research testing alternative extraction techniques, econometric contrast, and interpretation methodologies are recommended, considering the limitations derived from the scope of this work. The originality and main contribution of this paper lie in the comparison of these four techniques in both the financial and Mexican contexts. The main conclusion is that depending on the purpose of the analysis, one technique will be more suitable than another.

show abstract

Section: Underlying Systematic Risk Structure: Statistical and Graphical Analysissupporting

confidence: 75%

Section: Underlying Systematic Risk Structure: Statistical and Graphical Analysismentioning

confidence: 99%

Comparison of Statistical Underlying Systematic Risk Factors and Betas Driving Returns on Equities

Cortés

Porras

Monte

2021

REMEF

View full text Add to dashboard Cite

show abstract

“…For instance, maximum entropy is a standard principle for pricing derivative securities in incomplete markets (Stutzer, 1996;Frittelli, 2000) or to analyze contagion in the interbank market (Mistrulli, 2011). It is also at the heart of the concepts of mutual information and independent component analysis (Hyvarinen et al, 2001), which have been used in recent portfolio optimization schemes (Lassance and Vrins, 2021a;Lassance et al, 2021;Lassance and Vrins, 2021b).…”

Section: Independent Variablesmentioning

confidence: 99%

The Role of CDS Spreads in Explaining Bond Recovery Rates

Barbagli,

Francois,

Gauthier

et al. 2024

SSRN Journal

View full text Add to dashboard Cite

“…Some of these approaches focus on modifying the constraints and/or the goal objective function, e.g. by imposing restrictions on portfolio weights (e.g., Michaud (1989), Levy and Levy (2014)) and incorporating higher moments and tail-risk measures (e.g., Harvey et al (2010), Lassance and Vrins (2021)). Other approaches target appropriate modifications to the input of the portfolio selection problem including robust optimization methods (e.g., DeMiguel and Nogales (2009), Huang et al (2010), Bayesian statistics (e.g., Pastor and Stambaugh (2000), Bodnar et al (2022)) and the Black-Litterman model Litterman (1991, 1992).…”

Section: Introductionmentioning

confidence: 99%

Integrating multiple sources of ordinal information in portfolio optimization

Çela¹,

Häfner²,

Mestel³

et al. 2022

Preprint

View full text Add to dashboard Cite

Active portfolio management tries to incorporate any source of meaningful information into the asset selection process. In this contribution we consider qualitative views specified as total orders of the expected asset returns and discuss two different approaches for incorporating this input in a mean-variance portfolio optimization model. In the robust optimization approach we first compute a posterior expectation of asset returns for every given total order by an extension of the Black-Litterman (BL) framework. Then these expected asset returns are considered as possible input scenarios for robust optimization variants of the mean-variance portfolio model (max-min robustness, min regret robustness and soft robustness). In the order aggregation approach rules from social choice theory (Borda, Footrule, Copeland, Best-of-k and MC4) are used to aggregate the total order in a single "consensus total order". Then expected asset returns are computed for this "consensus total order" by the extended BL framework mentioned above. Finally, these expectations are used as an input of the classical meanvariance optimization. Using data from EUROSTOXX 50 and S&P 100 we empirically compare the success of the two approaches in the context of portfolio performance analysis and observe that in general aggregating orders by social choice methods outperforms robust optimization based methods for both data sets.

show abstract

Portfolio selection with parsimonious higher comoments estimation

Cited by 20 publications

References 45 publications

Comparison of Statistical Underlying Systematic Risk Factors and Betas Driving Returns on Equities

Comparison of Statistical Underlying Systematic Risk Factors and Betas Driving Returns on Equities

The Role of CDS Spreads in Explaining Bond Recovery Rates

Integrating multiple sources of ordinal information in portfolio optimization

Contact Info

Product

Resources

About