On out of sample errors in portfolio optimization : a study of the portuguese market

• Autor(a) principal: Bento, João Fernando Henriques
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.14/39234
• topic: Covariance matrix; Out of sample errors; Portfolio optimization; Risk management; Erros fora da amostra; Gestão de risco; Matriz de covariância; Otimização de portfólios; Domínio/Área Científica::Ciências Sociais::Economia e Gestão

Motivated by a long-standing notion of the challenge that is investing successfully in the Portuguese stock exchange, this thesis studies the implementation in Portugal of a new model that tackles one of portfolio optimization’s greatest obstacles: estimation errors. The Galton model, originally developed and tested in the United States market, uses a discovered pattern in out of sample errors, the difference between forecasted and realized values in out of sample tests, to correct historical inputs. These errors are mostly disregarded by other optimization efforts. With the corrected inputs, the model delivers a performance upgrade all around - more sensible weighting of the assets, improved return performance, returns concentrated around its mean avoiding extreme occurrences, less turnover and, above all, accurate forecasting of the portfolios’ volatility, in the same line as the models’ conclusions in the American market.