![Probability & Partners](https://i0.wp.com/probability.nl/wp-content/uploads/2014/11/pp-logo3.png?fit=340%2C156&ssl=1 "Probability & Partners\"){kind=logo}
The EBA has published its roadmap for implementing the EU banking package (CRR3/CRD6), which encompasses around 140 regulatory products related to credit risk, operational risk, market risk, supervisory reporting, and Pillar 3 disclosure, among others. Learn more about the changes.
Met oog op de komst van DORA heeft DNB een nieuwe Good Practice Informatiebeveiliging uitgebracht. Wat zijn de belangrijkste wijzigingen en waar kan je op letten?
By the end of this month banks need to have implemented Credit Spread Risk in the Banking Book. We notice that many banks are still developing methods to measure CSRBB amidst limited regulatory guidance. To validate or challenge your internal discussions, our colleague Maurits van den Oever has investigated a simple and intuitive method to quantify systemic credit spread shocks. He calculates shocks for government bonds with different ratings and maturities which can be easily implemented in your IRRBB framework. Interested in further expanding your CSRBB framework in 2024? Then don’t hesitate to get in touch with our IRRBB lead Corné Ruwaard to exchange thoughts.
Assessing Solvency II interest rate shocks in various interest rate environments using Hull-White, Nelson-Siegel, and PCA for comparison.
In this paper, I give a bird’s-eye view of Large Language Models and outline the most significant issues related to their applications in financial services. I will discuss potential use cases, LLMs’ limitations, and the challenges associated with their applications. The objective is to provide the reader with understanding of various aspects of LLMs, placing them in the context of financial institutions. Additionally, I will discuss ways of implementing LLMs in finance-related areas, outline potential dangers and pitfalls, and explore emerging strategies of overcoming these challenges.
Large Language Models are reshaping Financial Services.
In November 2021, we organized a roundtable regarding the application of expert judgment and overrides in behavioral modeling of prepayments and savings, including interest rates, COVID-19, and elasticity of savings.
In this paper we describe AI fairness from a quantitative perspective, on the example of credit decision making – which candidates should receive a loan and which not – but the principles we will describe hold more generally in financial services. We explore the roots of bias in AI systems and present popular definitions of fairness adopted by the industry.
With this paper, Probability and Partners would like to assist financial institutions in understanding the implications the ECB guide has for them, as well as help them prepare for the future where climate-related
and environmental risks are a part of everyday reporting.
We have identified six key questions for decision-making with respect to using datasets from external providers. The approach focuses on six key questions regarding definition of default, granularity, scope of application, data sample, number of defaults in the dataset and the availability and quality of the explanatory variables.