FinNet · Backreaction of Financial Networks: Risk Estimation and Asset-Management

Backreaction of Financial Networks: Risk Estimation and Asset-Management

The financial market is deeply interconnected. Events, news or data shape the expectations of market participants. This influences their behaviour and may trigger in- or divestment decisions, consequently impacting market prices. A positive surprise in sales volumes for a company’s key product (like Apple and its iPhone) can, for example, push its stock price up, while a missed target could be the reason for a price drop. From this perspective, asset prices carry important information about how market participants perceive and assess each individual asset as a part of the financial market. However, a sector crisis would probably impact the price of various stocks, not just one. In the same way, the good performance of one company is able to also increase the market valuation of other companies, which are part of its supply chain. This collective behaviour shows effective interconnectivity between the financial actors or products. Understanding and monitoring the financial interconnectivity is therefore crucial for evaluating risks in the financial market. Investors, regulators and policymakers have risk at the foundation of their decision-making process. A misplaced understanding of financial risks can bring to disastrous scenarios such as the crisis of 2008, whose consequences had an impact on the entire world's economy and society. This project aims to utilize and develop cutting-edge technologies in data science and machine learning together with quantitative modelling to improve risk assessment and the decision-making process. Networks can mathematically represent and model the interconnectivity between actors, and when applied to the financial market, their dynamics unveils changes in the risk contribution of each node. The project allowed the introduction of a new framework that combines Explainable AI methods with traditional methods for testing investment strategies. The framework is able to improve transparency and understanding of the investment risks and performances. Furthermore, the project introduced and studied new network-based strategies that can provide more stable and reliable investment performances. The project could enhance awareness among academics and practitioners about the opportunity of using cutting-edge AI technologies in the finance industry, reinforcing the use of such tools in financial institutions such as in the project host.

Data: CORDIS, © European Union