Publicação
Analysis of the cryptocurrency market: a network science approach
| Resumo: | The rapid growth and increasing prominence of cryptocurrencies in the global financial market have brought new challenges in risk management and asset allocation. The high volatility and interconnectedness of digital assets make understanding risk contagion crucial for investors, portfolio managers, and regulators. Network Science provides a powerful framework for studying these interdependencies by modeling relationships as networks where assets are connected based on various metrics, such as correlations or causality measures. The primary objective of this research is to identify price contagion among cryptocurrencies, using Network Science methodologies to analyze these transmission effects and offering practical insights for risk management and portfolio optimization. The methodology starts with a Network Science approach to model the relationships between cryptocurrencies. Correlation networks are created to visualize the connections between digital assets, indicating where strong relationships and potential contagion effects may occur. To enhance this analysis, Granger causality tests are applied to assess the directionality of these relationships, identifying predictive connections where the performance of one cryptocurrency may impact another. Finally, the Louvain algorithm, a community detection technique within Network Science, is used to cluster cryptocurrencies into groups based on the strength of their interconnections, providing insights into the structural composition of the cryptocurrency market. The network-based approach reveals significant interconnections among cryptocurrencies, with correlation networks indicating clusters of assets that share strong relationships. Granger causality analysis provides evidence of directional risk transmission, suggesting specific paths through which risk may propagate. The Louvain algorithm identifies groups of highly interconnected cryptocurrencies, offering insights into potential diversification strategies and highlighting areas where risk mitigation may be necessary. The results inform investors and portfolio managers on managing risk by identifying groups of cryptocurrencies with strong interdependencies, which may impact v diversification strategies. Additionally, the findings provide valuable insights for regulators aiming to monitor systemic risk in the cryptocurrency market. This study advances the understanding of risk contagion in the cryptocurrency market by integrating Network Science methodologies, including correlation networks, causality analysis, and community detection. It offers a comprehensive view of interdependencies and risk transmission, providing practical guidance for constructing more resilient cryptocurrency portfolios. |
|---|---|
| Autores principais: | Melo, Marta |
| Assunto: | Cryptocurrencies Network Science Community Finding Granger Causality Tests Correlation Network Criptomoedas Ciência das Redes Comunidades Testes de causalidade de Granger Rede de Correlação |
| Ano: | 2024 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório da Universidade de Lisboa |
| Resumo: | The rapid growth and increasing prominence of cryptocurrencies in the global financial market have brought new challenges in risk management and asset allocation. The high volatility and interconnectedness of digital assets make understanding risk contagion crucial for investors, portfolio managers, and regulators. Network Science provides a powerful framework for studying these interdependencies by modeling relationships as networks where assets are connected based on various metrics, such as correlations or causality measures. The primary objective of this research is to identify price contagion among cryptocurrencies, using Network Science methodologies to analyze these transmission effects and offering practical insights for risk management and portfolio optimization. The methodology starts with a Network Science approach to model the relationships between cryptocurrencies. Correlation networks are created to visualize the connections between digital assets, indicating where strong relationships and potential contagion effects may occur. To enhance this analysis, Granger causality tests are applied to assess the directionality of these relationships, identifying predictive connections where the performance of one cryptocurrency may impact another. Finally, the Louvain algorithm, a community detection technique within Network Science, is used to cluster cryptocurrencies into groups based on the strength of their interconnections, providing insights into the structural composition of the cryptocurrency market. The network-based approach reveals significant interconnections among cryptocurrencies, with correlation networks indicating clusters of assets that share strong relationships. Granger causality analysis provides evidence of directional risk transmission, suggesting specific paths through which risk may propagate. The Louvain algorithm identifies groups of highly interconnected cryptocurrencies, offering insights into potential diversification strategies and highlighting areas where risk mitigation may be necessary. The results inform investors and portfolio managers on managing risk by identifying groups of cryptocurrencies with strong interdependencies, which may impact v diversification strategies. Additionally, the findings provide valuable insights for regulators aiming to monitor systemic risk in the cryptocurrency market. This study advances the understanding of risk contagion in the cryptocurrency market by integrating Network Science methodologies, including correlation networks, causality analysis, and community detection. It offers a comprehensive view of interdependencies and risk transmission, providing practical guidance for constructing more resilient cryptocurrency portfolios. |
|---|