Traditional models for pricing catastrophe (CAT) bonds struggle to capture the complex, relational data inherent in these instruments. This paper introduces CATNet, a novel framework that applies a geometric deep learning architecture, the Relational Graph Convolutional Network (R-GCN), to model the CAT bond primary market as a graph, leveraging its underlying network structure for spread prediction. Our analysis reveals that the CAT bond market exhibits the characteristics of a scale-free network, a structure dominated by a few highly connected and influential hubs. CATNet demonstrates high predictive performance, significantly outperforming a strong Random Forest benchmark. The inclusion of topological centrality measures as features provides a further, significant boost in accuracy. Interpretability analysis confirms that these network features are not mere statistical artifacts; they are quantitative proxies for long-held industry intuition regarding issuer reputation, underwriter influence, and peril concentration. This research provides evidence that network connectivity is a key determinant of price, offering a new paradigm for risk assessment and proving that graph-based models can deliver both state-of-the-art accuracy and deeper, quantifiable market insights.

The global markets provide enterprises with selling opportunities and challenges in stabilizing operational strategies. From the perspective of production management, it is important to improve the profitability of an enterprise by exploiting the different timing of the selling season in different markets to develop an operational strategy that is optimized and configured on a global scale. This paper examines the above issue with an insightful model of selling the product to two markets (a primary and a secondary market) with multiple risks of changes in the market environment and nonoverlapping selling seasons. We refer to this problem as the "global robust newsvendor" problem. We provide closed-form solutions of the optimal operation strategy for demand-independent and demand-related scenarios for the above two market stochastic systems. The closed-form solutions fully reflect the influence of the relationship between supply and demand on strategy selection. We find that the demand correlation and the lack of demand information will not substantially affect the operation strategy, and the enterprise's industrial chain and supply chain remain stable. However, the reduction of inter-market tariffs or logistics costs will cause changes, and the existence of the secondary market will lead to more capacity planning in the primary market. In addition, our model explicitly considers the impact of exchange rate uncertainty on operating strategies.

Artificial Intelligence (AI) systems have had varying degrees of success when measured against humans. In problems such as chess, which succumb to sheer computational firepower, the machines have advanced greatly in a short time. Like chess, financial markets operate under the supposition of rational participants. Humans are supposed to be calculating the odds, maximizing return and minimizing risk. If this were indeed the case, artificial intelligence trading software would stand a good chance at bettering its human counterparts in making and capitalizing on market decisions.