Open Risk White Papers

Open Risk is occasionally publishing White Papers covering a range of topics around risk management, both conceptual and technical (quantitative / development issues). We welcome also external collaborators who are interested in the areas where Open Risk is active.

Open Risk White Paper 7: Risk Capital for Non-Performing Loans (OpenRiskWP07_022616)

Risk capital for non-performing loans

We develop a conceptual framework for risk capital calculation for portfolios of non-performing loans. In general banking practice, loans that pass a threshold of delinquency are declared non-performing and are provisioned. Yet there is a residual risk that the provisioning is not sufficient. This risk must be covered by capital buffers. The literature for risk capital requirements for NPL portfolios is very limited, which implies that Stress Testing and Internal Capital Adequacy Assessment (ICAAP) requirements for non-performing loans are harder to meet. Our framework builds on tools used in portfolio credit risk modeling and provides a structured approach to address the risk profile that is specific to non-performing loans.

Open Risk White Paper 6: Stress Testing Methodology for FX Lending (OpenRiskWP06_012116)

We develop a simple methodology for stress testing portfolios of credit instruments classified as foreign exchange lending. Loans whose repayment schedule is denominated in a currency other that that of the borrower’s domestic currency are commonly seen in many jurisdictions and have a risk profile that is considerably more complicated than domestic currency loans. Yet the literature for credit risk assessment and stress testing of portfolios of such loans is very limited, which means that Stress Testing and Internal Capital Adequacy Assessment (ICAAP) requirements are harder to meet. Our methodology builds on existing standard tools used in portfolio credit risk modeling and enables obtaining insights into the additional risk factors embedded in foreign currency lending.

Open Risk White Paper 5: Identification Framework for Business Model Risks (OpenRiskWP05_010916)

The Business Model Canvas

We develop an analytical framework for the systematic identification of business model risks. The framework utilizes as a starting point a simplified business model schema known as the Business Model Canvas. We review each one of the elements of the schema in turn, identifying the main risk characteristics associated with each.

Open Risk White Paper 4: Open Risk Model Taxonomy (OpenRiskWP04_061415)

We develop a taxonomy for risk models that aims to support an open source risk models framework. The proposal builds on and extends some commonly used risk taxonomies within financial services firms but introduces some significant new elements.

We first review the motivation for risk taxonomies, the concepts and tools that are involved and some of the weaknesses of current schemes. We try also to clarify the link between risk models and risk taxonomies.

Open Risk White Paper 3: Introducing the Open Risk API (OpenRiskWP03_053115)

We develop a proposal for an open source application programming interface (API) that allows for the distributed development, deployment and use of financial risk models. The proposal aims to explore the following key question: how to integrate in a robust and trustworthy manner diverse risk modeling and risk data resources, contributed by multiple authors, using different technologies, and which very likely will evolve over time.

The proposal builds on two key modern technological frameworks, Semantic Data and RESTful API’s, which in turn are both examples of rapidly adopted and evolving Web technologies. We review the motivation for such an infrastructure, the concepts and tools that can enable such a design and various related initiatives. We describe in detail the current version of the API specification.

For definiteness we illustrate the concept with an open source implementation that takes a use case from the analysis of credit risk in loan portfolios. The implementation consists of demo model and data servers and clients implemented using Python and MongoDB.

Open Risk White Paper 2: Confidence Capital: The Principle (OpenRiskWP02_040815)

We review the structure of economic capital frameworks commonly used within financial institutions and identify why the derived capital metrics do not explicitly address the needs for maintaining ongoing confidence on the soundness of the firm. In the follow up to the financial crisis the need for more explicit such tests has been highlighted by regulatory stress testing methodologies.

The likelihood and severity of a future ratings downgrade (as opposed to a default within the risk horizon) are the two key new “risk appetite” inputs required for the framework. The temporal correlation of losses beyond the risk horizon with those within the horizon is one of the main new risk parameters that are highlighted by the framework.

We derive explicit formulas for implementing a confidence capital framework in a two period setup that can lead to tractable implementations. We include a brief quantitative study that addresses a very simplified case that is solvable in terms of simple formulas. We explore the relation of confidence capital to economic capital for various choices of risk appetite and inter-temporal loss correlations

Open Risk White Paper 1: Revisiting Simple Concentration Indexes (OpenRiskWP01_032705)

We review the definitions of widely used concentration metrics such as the concentration ratio, the HHI index and the Gini and clarify their meaning and relationships. This new analytic framework helps clarify the apparent arbitrariness of simple concentration indexes and brings to the fore the underlying unifying concept behind these metrics, thereby enabling their more informed use in portfolio and risk management applications. We also propose that the sensitivity of concentration indexes to growing concentration should be a defining criterion for adopting an index and explore the sensitivity of common indexes to changing portfolio concentrations. We show that this sensitivity can vary significantly between indexes for parametric families of portfolio distributions and hence selecting and using a simple concentration index should take this aspect carefully into consideration.