| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 489.02 KB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A metodologia DEA (Data Envelopment Analysis) avalia a eficiência relativa de um conjunto de entidades homogéneas, vulgarmente designadas de DMUs (Decision Making Units), que produzem múltiplos outputs usando múltiplos inputs. Na sua forma clássica, cada DMU escolhe os pesos associados aos inputs e aos outputs que maximizam a sua eficiência. Esta flexibilidade conduz a dois problemas, estreitamente relacionados, o da existência de várias DMUs consideradas eficientes e o da distribuição irrealista dos pesos. Para ultrapassar estes dois problemas, vários métodos e modelos têm sido propostos, entre os quais o método da eficiência ou avaliação cruzada, considerado um dos mais populares.
O presente trabalho destina-se a apresentar um estudo sobre o método da avaliação cruzada na DEA com enfoque em modelos de objetivo secundário que o método utiliza para ultrapassar o problema das soluções ótimas alternativas fornecidas pelos modelos clássicos, que podem levar a obter rankings diferentes em função do software usado. Este estudo inclui a aplicação desse método e de alguns desses modelos à avaliação e determinação do ranking de dezassete seguradoras que operavam em Portugal no ano de 2020.
The methodology DEA (Data Envelopment Analysis) evaluates the efficiency related to a group of homogenous entities, commonly designated by DMUs (Decision Making Units), which produce multiple outputs using multiple inputs. In its classical form, each DMU chooses the weights associated to the inputs and outputs which maximize its efficiency. This flexibility leads to two problems, closely related: one is the existence of several DMUs considered efficient; and the other is the unrealistic distribution of the weights. To overcome these two problems several methods and models have been proposed, among which the method of efficiency or cross evaluation, considered one of the most popular. This work aims to present a study on the method of cross evaluation in DEA with focus on models of secondary object which the method uses to overcome the problem of the alternative excellent solutions supplied by the classical models, that may lead to obtain different rankings according to the used software. This study includes the application of that method and some of these models to the evaluation and determination of the ranking of seventeen insurance companies which operated in Portugal in the year of 2020.
The methodology DEA (Data Envelopment Analysis) evaluates the efficiency related to a group of homogenous entities, commonly designated by DMUs (Decision Making Units), which produce multiple outputs using multiple inputs. In its classical form, each DMU chooses the weights associated to the inputs and outputs which maximize its efficiency. This flexibility leads to two problems, closely related: one is the existence of several DMUs considered efficient; and the other is the unrealistic distribution of the weights. To overcome these two problems several methods and models have been proposed, among which the method of efficiency or cross evaluation, considered one of the most popular. This work aims to present a study on the method of cross evaluation in DEA with focus on models of secondary object which the method uses to overcome the problem of the alternative excellent solutions supplied by the classical models, that may lead to obtain different rankings according to the used software. This study includes the application of that method and some of these models to the evaluation and determination of the ranking of seventeen insurance companies which operated in Portugal in the year of 2020.
Description
Mestrado em Contabilidade e Gestão das Instituições Financeiras
Keywords
Data envelopment analysis (DEA) DMUs Avaliação cruzada Ranking Setor segurador Cross efficiency Insurance sector
Citation
Dias, P. T. (2023) Modelos de objetivo secundário na avaliação cruzada da DEA. Aplicação ao setor segurador. (Dissertação de mestrado não publicada). Instituto Politécnico de Lisboa, Instituto Superior de Contabilidade e Administração de Lisboa. Disponível em http://hdl.handle.net/10400.21/16162