Proceedings

Frank De Jonghe

Conceptually, a good model is free of biases and robust in the face of uncertainty. In reality, executing a model risk management framework can be complex to manage people, processes, and systems effectively. You’ll look at frequently encountered challenges and potential remedies for traditional models, and learn how to include AI and machine learning algorithms into the model risk management framework.

Paul Peeling, MathWorks

As the variety and velocity of model developments increase, and constraints around the explainability and responsible use of machine learning models emerge, model risk managers and model validation practitioners need a suite of tools to meet regulatory, operational, and business needs. In this talk, you’ll discover tools that support the automation, documentation, and traceability of model governance and model validation processes that MathWorks has developed for model risk management.

Heng Chen, HSBC

Operational risk modeling using the parametric models can lead to a counterintuitive estimate of value at risk at 99.9% as economic capital due to extreme events. To address this issue, a flexible semi-nonparametric (SNP) model is introduced using the change of variables technique to enrich the family of distributions that can be used for modeling extreme events. The SNP models are proven to have the same maximum domain of attraction (MDA) as the parametric kernels, and it follows that the SNP models are consistent with the extreme value theory and peaks over threshold method, but with different shape and scale parameters. By using the simulated data sets generated from a mixture of distributions with varying body-tail thresholds, the SNP models in the Fréchet and Gumbel MDAs fit the data sets by increasing the number of model parameters, resulting in similar quantile estimates at 99.9%. When applied to an actual operational risk loss data set from a major international bank, the SNP models yield economic capital estimates 2 to 2.5 times as large as the single largest loss event and exhibit a reasonable stability towards the change of loss history in the scenario analysis.

Nadége Lespagnol, Euler Hermes

Euler Hermes (EH) is the leading B2B credit risk business of the Allianz Group, helping customers protect themselves from bad debt.

EH has a strategic objective to centralize all credit assessment model calibration data, model design, and model monitoring processes in one common modeling platform, helping to meet the regulatory requirement for reconciliation and transparency for all credit assessment models. EH’s proprietary GRADE model is a probability of default (PD) model used in both the underwriting process and in the allocation of risk capital.

In 2020, EH launched a transformation project to migrate all credit risk models from a legacy infrastructure to MATLAB^® running on AWS^®. EH used components of the MATLAB Model Risk Management solution to develop and maintain the full suite of credit risk models, which are based on fuzzy logic approaches as well as tree-based algorithms.

In this presentation, learn how EH has built a new model design architecture with MATLAB and AWS that will allow many improvements in the process of building and testing future models.

Ian McKenna, MathWorks

The heart of any robust model risk management framework is the model inventory. However, many inventory systems are proving unable to keep up with the increased demands from regulators because they are effectively databases with a well-designed interface and lack the traceability and automation needs of a modern system.

While linking and drilling down on models, data, and documents is often the starting point for oversight discussions, the areas that are often overlooked are document management, interoperability, workflow automation, and customization through advanced analytics. These key areas provide enhanced cost reduction to the business by accelerating regulatory approval and eliminating inefficiencies.

In this session, you’ll see a live demonstration of the MathWorks Model Inventory, showcasing how several large global banks that adopted this solution addressed key challenges around model lineage, governance, automation, and review.

Ben Steiner, Columbia University
Oleh Khalayim, World Bank
Matt Barnes, HSBC
Stuart Kozola, MathWorks
Siddharth Sundar, MathWorks

Modeling and analytics can provide tremendous value to organizations across all industries and disciplines, and financial services is no exception. Financial institutions are investing significant time and capital to build modern analytics capabilities. While several of these institutions have teams with skills to build highly complex models, they face significant challenges in moving these models to production and getting buy-in from key stakeholders including business owners, internal compliance teams, and regulators.

In this panel session, learn how an agile ModelOps infrastructure can help organizations take their models from design to deployment in a cost-effective and timely manner. By effectively monitoring model performance to determine proactive actions, organizations can satisfy the needs of all stakeholders involved across the model lifecycle.

Oleh Khalayim, World Bank

Governments around the world try to attract private investors to infrastructure projects because public coffers cannot cover the costs and may not be suited for financing modern infrastructure solutions.

Investors finance infrastructure mostly through fixed-income debt instruments. It is easier to find private investors for infrastructure sectors like telecoms, airports, and ports. Government support is needed for regulated utilities and social infrastructure, such as educational and medical facilities and sewage and solid waste management.

Learn about a tool implemented in MATLAB^® for simulating cash flows across a variety of scenarios that support decisions for investing in infrastructure projects by the World Bank.

Sri Krishnamurthy, Quant University

With complexity in data science pipelines growing, organizations are redesigning tooling and infrastructure to build agile processes, sandboxes for experimentation, and integrations with multiple tools to meet the needs of distributed teams. In addition, the cloud has made high-performant, scalable, and elastic computing accessible to data scientists and quant modelers without needing to plan elaborate hardware and software setups.

In this talk, you’ll learn about QuSandbox, a rapid prototyping platform that makes access to data, modeling tools, and compute infrastructure accessible to modelers for building large-scale quant and data science applications in the cloud. QuSandbox supports multiple data integrations and modeling tools including the MATLAB Online™ server to enable quants and data scientists to learn by doing in a sandbox environment. You’ll hear about a recent use case where team of quants learned to build full-fledged data pipelines with the QuSandbox and prime the environment for analysis on the MATLAB Online server.

You’ll also see a case study where data from EDGAR was scraped, cleaned, and annotated and a sentiment analysis model was built using the MATLAB Online server. We will also illustrate how Amazon S3 was used for data staging and how MLFlow was used for tracking experiments and how the entire data pipeline was orchestrated using the QuSandbox.

Matt Barnes, HSBC

Discover how HSBC focused on implementing a MLOps process within their global risk analytics team. You’ll hear a summary of the history of model development practices and the pitfalls encountered along the way, and also explore:

• The need for change and a new focus on engineering
• The implementation challenge and story so far
• The push to cloud and unleashing the possibilities of scale

Scott Nicholas, MathWorks

Containers are the universal building block for cloud and computing solutions that scale and adapt as needed. Learn about the basics of obtaining, building, running, and licensing MATLAB^® in a Docker container.

Stuart Kozola, MathWorks

Digital transformation has exponentially increased with the shift to remote work during the pandemic, and when coupled with AI adoption as a strategic initiative across organizations, has created an environment of rapid change and great potential for innovation. With this move to digitize, automate, and integrate AI into financial workflows, organizations are struggling to maintain a robust and responsible modeling pipeline that is agile in meeting increasingly shorter and more demanding business timelines, leaving shorter intervals for research and innovation. In this talk, you’ll hear an overview of the area of “responsible AI” and best practices for building an agile and responsible AI pipeline to deliver research to production efficiently.

Emilio Llorente, Recognition AMS

In this talk, learn how systematic investing brings about the new face of wealth management, marrying human and computer intelligence.

Discuss market analysis using data science, risk analysis before extreme market events, and portfolio construction under cutting-edge optimization methods. Discover real-life trading and portfolio management from the perspective of the technology and scientific-oriented professional. You’ll also see how the set of modules that integrate the investment process are produced and ensembled in MATLAB^®, showing the end-to-end capabilities of the different toolboxes used to offer a complete solution to wealth advisors and discretionary institutional portfolio managers.

Lawrence Johny, MathWorks

Natural language processing (NLP) is a rapidly growing area of interest in the financial services industry as quants, risk managers, and financial analysts are all interested in deriving new alpha and insights from speech and text data. Common NLP models do not consider context-specific vocabulary, but transformer models that are pretrained to recognize financial jargon and nuance, such as FinBERT, can provide far better results.

In this session, you will learn how MATLAB^® adds value and simplifies various fine-tuning tasks with apps such as Experiment Manager, Classification Learner, and Deep Network Designer.

Sergey Paltsev, MIT

Climate change poses financial risks that arise from shifts in the political, technological, social, and economic landscape that are likely to occur during the transition to a low-carbon economy. One of the global community’s most significant contemporary challenges is the need to satisfy growing energy and food demand while simultaneously achieving very significant reductions in the greenhouse gas emissions and sustainable development. In pursuing this goal, decision makers need to make strategic choices that address both physical risks (damage from extreme events such as fires, floods, droughts, and sea-level rise) and transition risks (financially consequential shifts in political, technological, social, and economic landscapes in the transition to a low‑carbon future). Energy-economic models can be used to support decision makers in quantifying these risks by integrating across systems, sectors, and scales. Learn about a framework for addressing climate-related financial risks where scenario analysis plays a key role in climate risk management.

Lawrence Johny, MathWorks

The 2015 “Paris Agreement” places a binding obligation on the world’s governments to “make finance flows consistent with a pathway towards low greenhouse gas emissions and climate-resilient development.” Regulators, customers, investors, and other stakeholders are driving financial institutions to do their part to transition to a low-carbon economy and manage exposure to climate-related risks. They’re using new data sources and developing new types of models, often leveraging methods from other scientific and engineering fields. Practitioners need software that provides a broad range of modeling functionality, flexible interfaces, rich visualization capabilities, collaboration, and review features to keep up with the pace of change in this area.

Learn how MATLAB^® can get you started modeling both physical and transition climate risks. In a live demonstration, you will learn how to:

• Visualize flooding risk within a city (physical risk)
• Understand the impact of policies aimed at increasing the energy efficiency of buildings (transition risk)
• Model the impact of these risks on a portfolio of mortgages

Abhishek Gupta, MathWorks

Technological and regulatory changes create pressure for the financial services sector to evolve. Companies are responding to trends such as cloud computing, artificial intelligence, and climate change. The workforce is upskilling to keep pace with these trends.

As an educator or employer, how are you training students and employees to tackle present and future business needs? In this talk, you’ll explore the typical challenges customers encounter in developing financial thinking and resources they use to their advantage.

Michael Robbins, Columbia University
Malin Ortenblad, Columbia University
Yao Shang, Columbia University
Richard Wang, Columbia University

Using examples from his courses at Columbia University and his book, Quantitative Asset Management, Michael Robbins applies machine learning and factor investing to asset allocation. He demonstrates how leading institutions manage multibillion-dollar portfolios and includes real-world details like currency controls, market impact, and taxes. Learn about the entire investing process, from designing goals to planning, research, implementation, testing, and management.

Stefano Battiston, University of Zurich

Financial institutions face growing demand from investors and regulators to assess climate risk on their portfolios. Financial authorities recommend basing these assessments on NGFS scenarios (future trajectories of carbon prices and production output across the high- and low-carbon sectors of the economy, depending on future introductions of climate policies).

Even with these assessments, several challenges remain to integrate climate risk into risk management. The circularity between materialization of risk and risk perception can cause uncertainty about which scenario may occur. While risk is endogenous to finance in other contexts, in climate risk it impairs the use of standard risk assessment tools.

This talk is a collaboration with scientists of the IAM community involved in the development of NGFS scenarios and explains how a new framework can provide scenarios that assess the role of investors in the realizations of climate mitigation pathways.

Alessia Paccagnini, University College Dublin

The big data era is creating a lot of exciting opportunities for new developments in econometrics, economics, and finance. The recent improvements in computer science and the internet are making data collection easier. Meanwhile, the analysis of large data sets poses methodological challenges (for example, high-frequency observations, unstructured data, and new large datasets) where MATLAB^® can help and support the researchers. In this talk, explore the recent advancements in macroeconometrics and empirical financial analysis to deal with big data. See how machine learning techniques can be implemented in MATLAB to estimate macro and financial data for horserace forecasting and structural analysis.

Paul McNelis, Fordham University

Explore alternative measures of systemic risk for the financial sector based on forecast error variance decomposition as well as covariance-at-risk. You’ll see how to assess the relative importance of divergences in the implied volatility measures for swap-options for government securities issued in Hong Kong and the United States. In previous work on the mainland Chinese banking system, the effect of global and US economic uncertainty indices was transmitted to mainland China through divergences between the onshore CNY exchange rate and the offshore CNH exchange rate. Since Hong Kong has a hard peg with the dollar, discover how divergences in swap volatility dynamics affect the transmission of uncertainty indices to the Hong Kong financial system.