Investment portfolio construction using multi-criteria methodology

Santiago

Conferencia; 23rd Conference of the International Federation of Operational Research Societies; 2023

Universidad de Chile - Pontificia Universidad Católica de Chile

In order to support the investment decision analysis process on listed companies of the Buenos Aires Stock Market, we constructed medium to long-term portfolios using ratios constructed from the final financial statements published by the companies for the 2018 year. The procedure consisted of classifying companies according to their investment attractiveness into three categories, using the Basic Additive Model (DEA) in two stages. The given categories were modeled as a pre-order by applying the Additive Utility Discriminant Method (UTADIS), obtaining a piecewise linear additive utility function. Through a cross-validation procedure leaving 10% of the firms out, we were allowed to assign the firms to classes in a more robust way.Finally, considering the best-practice firms, we constructed investment portfolios with different numbers of assets, using the additive utility function to determine each firm share. Once the portfolios were constructed, their average annual returns were compared with those of the most important stock market indices in the local capital market: S&P Merval and S&P/BYMA Argentina General. The results show that the portfolios returns are higher than those of the general indices in all cases.