Stress Test Scenarios Research Articles

High-dimensional macroeconomic stress testing of corporate recovery rate

We investigate macroeconomic stress testing frameworks for corporate bond recovery rate analysis using machine learning techniques. In doing so, we simulate the macroeconomic effect of a broad range of 182 macroeconomic variables extracting key factors with methods such as (sparse) principal component analysis and sparse group least absolute selection and shrinkage operation (LASSO). Using the adverse stress testing scenario from the US Federal Reserve as the benchmark, we demonstrate that our least squares-support vector regression model produces sensible and potentially valuable risk measures such as value-at-risk and conditional value-at-risk for recovery rates during periods of macroeconomic stress.

Variable selection in macroeconomic stress test: a Bayesian quantile regression approach

AbstractThe key assumption in stress test scenarios is that selected risk factors are useful in predicting banks’ tail risks under severe economic conditions. We argue that high-dimensional Bayesian quantile regression models with shrinkage priors are ideal for identifying those factors. We illustrate our methods by identifying key drivers for banks with different asset sizes from a high-dimensional database. We found that leverage indicators, asset prices, and labor market measures are the best predictors of banks’ performance. The usefulness of our methods is further demonstrated by a forecast comparison between the selected variables and those used in the regulatory stress tests.

The calibration of initial shocks in bank stress test scenarios: An outlier detection based approach

The shock scenarios used in the practice of stress testing are often based on ad hoc hypotheses concerning the evolution of risk factors. Consequently, they have been criticized for reflecting mild and temporary shocks maintained over very short periods. The aim of this article is to propose methodological improvements to the design and calibration of initial shocks based on the detection of outliers and structural breaks in (macro)economic and financial data. Our results of the real-world implementation of outlier detection algorithms show that the dynamics of shocks, the length of the stress horizons, and their magnitude are sensitive to the type and nature of the considered risk factor. The inferred stress horizon is longer (one to two years) for macroeconomic variables like GDP, real estate or oil prices, than for interest rate variables and the slope of the yield curve (six months).

Risk-Sensitive Investment Management: A Guide for Quants

Risk-sensitive investment management (RSIM) has emerged as a powerful approach that leverages the mathematics of stochastic control to address dynamic investment management problems. This note discusses RSIM and examines three practical implications of RSIM for investment managers. Firstly, RSIM models serve as a valuable tool for understanding and explaining the outcomes of realistic dynamic models. For example, they offer closed-form approximations of stochastic programming models. Secondly, RSIM sheds light on crucial investment management questions. RSIM enables practitioners to analyze the impact of various factors on investment performance, unravel the dynamics between active and passive management strategies, investigate the mechanics of learning within investment decisions, assess the value of expert opinions, and evaluate the effectiveness of stress test scenarios. Lastly, RSIM casts light on perilous strategies and practices that can lead to catastrophic losses for investment funds, such asexcessive betting or overbetting.

Designing Stress Tests for UK Fast-Growing Firms and Fintech

This paper captures advances in prudential regulation and supervision for challenger banks and fintech in the UK. It presents a critical analysis of the prudential supervisory approaches towards fintech. The focus is placed on fast-growing firms (FGFs), building on the review performed by the Prudential Regulation Authority (PRA) of the Bank of England (BoE) in 2019. Specifically, it comprises a critical examination of the underlying regulatory framework in relation to the robustness of stress testing practices, as part of the review of FGF risk management practices and the weakness identified in the Internal Capital Adequacy Assessment Process (ICAAP). The economic analysis of law comprises the underlying methodology, using economic theory to analyse regulation and its effectiveness regarding fintech regulation and supervision. Recommendations for enhancements towards supervisory practices about the prudential governance and management of FGFs and fintech are included, with advances to the underlying regulatory framework in the UK. Overall, this critical legal research examines the supervisory practices of FGFs and fintech in the UK, under the lens of prudential regulation and risk management approaches, focusing on the design, development and implementation of the stress testing tool and scenario practices.

Low-latency GNSS multipath simulation and building wall detection in urban environments

Precise navigation for fully autonomous driving—especially in dense urban areas—requires periodic precise position estimates. Global Navigation Satellite System (GNSS) technology has the potential to provide absolute positioning accuracy at a centimeter level. However, buildings in urban environments cause signal distortions and signal reflections—the so-called multipath—which are the most challenging parts in the GNSS error budget. Hence, we developed a scalable real-time multipath simulator for mitigating potential multipath receptions. The simulator uses three-dimensional (3D) building information, satellite, and user positions. The key drivers of latency are the calculation of reflection, diffraction, and line-of-sight, as well as the response time of the 3D building model database. The memory manager of the graphic processing units (GPUs) in combination with a dedicated load balancer enables fast and efficient multipath analysis. Selected case studies demonstrate the simulator’s potential to significantly improve the position accuracy of the processing engine. The use of the multipath simulator reduces the error in 61% of the error measurements in a stress test scenario to less than half of the non-multipath processing. The scalability of the simulator is demonstrated by combining the multipath simulator with a traffic simulator. Furthermore, we present a novel methodology for the detection of walls using GNSS signals to better account for incomplete or erroneous 3D building information in GNSS signal processing.

СИСТЕМНИЙ ПІДХІД ДО СТРЕС-ТЕСТУВАННЯ БАНКІВСЬКИХ РИЗИКІВ

The article reveals the essence of stress testing as a tool for assessing the reliability and stability of economic and technical systems. Based on a retrospective approach, the evolution of stress testing of banking risks in the world and in Ukraine is systematized. A comparative assessment of the definitions of stress testing of banking risks provided by international financial organizations and the national regulator was carried out, and their common and distinctive features were determined. It is shown that according to the level of stress testing of banking risks, it is divided into macro- and microeconomic, their comparison is made, advantages and disadvantages are determined. The main macro- and microeconomic risk factors recommended by the NBU for stress testing are systematized. The content of the banking risk stress testing system has been formalized, its elements have been identified: purpose, tasks, objects, methods, criteria for the effectiveness of stress testing scenarios, models of their formation, assumptions regarding the level of risk, components of the analysis, basic requirements for implementation, main stages. The main methods of stress testing banking risks and types of predicted losses are summarized, their content and main tools are disclosed. It is shown that the most common method of stress testing is the development of scenarios. Systematized NBU requirements for: development of internal bank documents regarding the organization and procedure of stress testing; to the functioning of the information system regarding risk management and reporting in stressful situations; responsibilities of the Chief Risk Manager in terms of communications with the bank's top management and the use of stress testing results for management decision-making. The directions for improving stress testing are substantiated: taking into account real shocks to the financial system, the speed of the spread of crises between countries, expanding the range of risks that are taken into account when building scenarios; regular and systematic implementation, constant review of methods and shock factors; ensuring proper practical use of results; adjusting strategies and plans to reduce banking risks.

How severe are the EBA macroeconomic scenarios for the Italian Economy? A joint probability approach

Measures of the severity of macroeconomic scenarios have been widely used in the literature, but a consistent methodology for their calculation has not been developed yet. Against this background, we provide a general method for calculating the joint probability of observing a macroeconomic scenario, which can be applied to various structural models. By doing so, we can attach probabilities to scenarios produced with multidimensional economic models to compare their severity and plausibility. We apply our methodology to the 2016 and 2018 EBA stress test scenarios and also provide reverse stress test applications. Our results show that for the Italian economy, the 2016 and 2018 EBA scenarios are unlikely, especially the 2016 one. The reverse stress tests allow us to identify the key variables that affect our probabilities.

Analisis Dampak Covid-19 terhadap Kinerja Perbankan

This study aimed to analyze the impact of the Corona Virus Disease-19 (Covid-19) pandemic on the performance of PT. Bank Pembangunan Daerah Jawa Timur Tbk. Cabang Magetan, and analyzed the scheme design which contains the most neededstrategic steps to correct the weakening of its performance. The method used in this research is descriptive statistics. The data used are financial statements in financial ratios, including BOPO, NIM, NPL, and CKPN. This study indicates that there has been a change and a slowdown in the financial performance of PT. Bank Pembangunan Daerah Jawa Timur Tbk. Cabang Magetan, both in terms of CKPN, NPL, BOPO, and NIM. Meanwhile, the policy that has been taken to correct these weaknesses is the implementation of a stress test preparation for the fixed effects industry level with the estimation of the ordinary least square model and the balance sheet approach, which can capture the dynamics of individual banks in surviving the shocks given in the stress test scenario (PD, LGD, and EAD) to see industrial resilience.
 Keywords: Covid-19; Banking; Financial Performance

Corporate Vulnerabilities in the Middle East, North Africa, and Pakistan in the Wake of COVID-19 Pandemic

This paper analyzes corporate vulnerabilities in the Middle East, North Africa and Pakistan (MENAP hereafter) in the wake of the COVID-19 pandemic shock. Using a sample of nearly 700 firms from eleven countries in MENAP, we assess the non-financial corporate (NFC) sector’s liquidity and solvency risk and viability over the medium term under different stress test scenarios. Our findings suggest that the health crisis has exacerbated vulnerabilities in the corporate sector, though the effects are heterogenous across the region. Small firms, which entered the pandemic in a more vulnerable position, would remain under high liquidity stress over the medium term, putting a substantial share of these firms’ debt at risk of default. Similarly, liquidity needs of firms in contact-intensive sectors have also worsened and would remain elevated in 2022-23. We also show that an adverse scenario of subdued growth and premature withdrawal of policy support would impair the capacity to service interest expenses, especially among small firms, resulting in higher insolvency risk. Overall, our results indicate that some segments of the MENAP corporate sector could remain reliant on policy support during the recovery phase and that structural reforms are critical to save distressed but viable firms from bankruptcy and ensure an efficient liquidation of “zombie” firms.

Machine learning for liquidity risk modelling: A supervisory perspective

The purpose of an effective liquidity risk assessment policy is to ensure that any given credit institution can meet its cash flow obligations, even factoring in the uncertainty caused by external factors. As part of the Supervisory Review and Evaluation Process (SREP), the European Central Bank (ECB) has determined this assessment should take into consideration both the institution’s ability to meet its short-term obligations and its long-term funding strategy. Due to the fast pace of financial markets and more demanding regulations, there is a structural need for a precise and widely accepted risk assessment methodology. Furthermore, the ability to foresee alternative scenarios by stressing the involved key risk indicators is of the utmost importance. This work investigates whether machine learning techniques can successfully model liquidity risk, thus providing insights for stress-testing scenarios. We have applied the Risk Assessment System (RAS) methodology to classify credit institutions from the Portuguese banking sector according to their liquidity risk, using real supervisory data (from 2014 until March 2021). We then studied the ability to model this risk classification, by comparing a series of well-established machine learning algorithms to a traditional statistical model for benchmarking. The results show that extreme gradient boosting (XGBoost) outperforms other methods for this classification problem. The resulting model can be set up for a production environment and provide scenarios for stress-testing, or as an early warning system (EWS), thus supporting the overall SREP exercise.

Towards climate resilient municipal water supply in Bangkok: A collaborative risk informed analysis

Typical top-down approaches to climate change adaptation in the water sector rely on climate models to inform the adaptation design. In view of the inherent uncertainty associated with these models, this study uses a combined top-down and bottom-up approach called Collaborative Risk Informed Decision Analysis (CRIDA). Using the case of the Metropolitan Waterworks Authority, Bangkok, this study demonstrates the application of the approach, which includes (a) identifying critical thresholds that impact the system’s performance, (b) unearthing the system vulnerabilities through a stress test (scenario analysis), and (c) identifying feasible adaptation interventions. Through a stakeholder consultation process, it was found that streamflow, salinity, and turbidity are the key performance metrics of the water supply system. A series of stress tests to the system was conducted by adjusting input variables assuming a wide range of possible future climatic conditions. Frequency curves of the various input variables were developed to facilitate stakeholders’ planning for impending risks and for improving the overall robustness of the water supply system. Based on the stress test results and past studies of the study area, the raw water source of the Metropolitan Waterworks Authority is expected to be more saline in the future with more frequent and severe low flow conditions, for which a number of adaptation actions were formulated and proposed. It is suggested to develop adaptation pathways that can address a range of expected impacts of future climatic conditions on the Bangkok water supply system. The methodology presented is useful for study areas where a model of the system is unavailable, and data limitation is a challenge.

Monte Carlo simulated stress test scenario on the case of Macedonian logistic companies

Monte Carlo simulated stress test scenario on the case of Macedonian logistic companies

Insolvency Risk. Application of Altman Z-score to the auto parts sector in Romania

This paper intends to apply the Altman Z-score model to all the companies active in the wholesale of motor vehicle parts and accessories (NACE 4531), with extended financial statements. Using the panel data model over the time series for 2008-2016 on the companies of this sector, we conclude that 99% of the Z-score is explained by the independent variables (working capital, capital structure, turnover, earnings before interest and tax), with estimated parameters very close to the models classical values. The sample description of the paper and the corresponding results highlights the Z-score evolution by turnover clusters and principal components, with the largest companies performing the best (the only cluster with Z-score median above 3). We notice a tendency for decreasing high risk companies and increase in the medium risk companies, whereas the low risk companies are relatively stable. This improvement is mostly due to increasing capitalization rate and less external debt, despite the deteriorating working capital and operating margin. We believe that future research to evaluate Z-score sensitivity under stress test scenarios would be very useful to provide an insight of companies’ insolvency risk amid increasing interest rates and different fiscal tax on dividend.

Predicting economists: Generating scenarios for stress testing future loss reserves

Stress testing under the US Comprehensive Capital Analysis and Review (CCAR) regulations and those of many other countries seeks to assess the full possible financial position of a lender through an economic crisis. The introduction of lifetime loan loss reserves under FASB’s Current Expected Credit Loss (CECL) and IASB’s International Financial Reporting Standards 9 (IFRS 9) rules complicates the task of stress testing, because lenders need to estimate future losses using scenarios that are contingent on the stress testing scenario, but without perfect foresight of the future stress test scenario. This work casts the CECL and IFRS 9 stress testing problem as one of generating future economic scenarios that are consistent with how future economists would create scenarios. To that end, we obtained historic consensus economic scenarios for testing. The results here demonstrate that a second-order Ornstein–Uhlenbeck model fits historic scenarios well and could be used to generate future scenarios that would be a realistic representation of what economists would predict given economic conditions up to that point. This approach was tested for US real gross domestic product (RGDP) and unemployment rate scenarios through the 2009 recession. The RGDP modeling was straight-forward, but we discovered that consensus economic scenarios for unemployment rate appear to be conditional on the phase of the economy.

Techniques for constructing climate scenarios for stress test applications

In this review, we provide guidance on the construction of climate scenarios for stress tests—scenarios that represent disruptive climatic events and can be used to assess the impacts of climate and weather risks at the level of detail that is necessary to identify specific adaptation actions or strategies. While there is a wealth of guidance on scenario-based climate adaptation planning, this guidance typically assumes the selection and use of decadal to century-long time segments of downscaled climate model projections, rather than the creation of a customized scenario depicting a specific extreme event. We address this gap by synthesizing a variety of data sources and analytical techniques for constructing climate scenarios for stress tests that are customized to address specific end-users’ needs. We then illustrate the development and application of climate scenarios with a case study that explores water sustainability under changing climate in the Truckee and Carson River basins of California and Nevada. Finally, we assess the potential advantages and disadvantages of the different data sources and analytical techniques described to provide guidance on which are best suited for an intended application based on the system of study, the stakeholders involved, and the resources available. Ultimately, this work is intended to provide the building blocks with which scientist-stakeholder teams can produce their own stress test scenarios to explore place-based weather and climate risks.

Limit Theorems for Default Contagion and Systemic Risk

We consider a general tractable model for default contagion and systemic risk in a heterogeneous financial network, subject to an exogenous macroeconomic shock. We show that, under some regularity assumptions, the default cascade model could be transferred to a death process problem represented by balls-and-bins model. We also reduce the dimension of the problem by classifying banks according to different types, in an appropriate type space. These types may be calibrated to real-world data by using machine learning techniques. We then state various limit theorems regarding the final size of default cascade over different types. In particular, under suitable assumptions on the degree and threshold distributions, we show that the final size of default cascade has asymptotically Gaussian fluctuations. We next state limit theorems for different system-wide wealth aggregation functions and show how the systemic risk measure, in a given stress test scenario, could be related to the structure and heterogeneity of financial networks. We finally show how these results could be used by a social planner to optimally target interventions during a financial crisis, with a budget constraint and under partial information of the financial network.

Using stress testing model in measuring the impact of the financial crisis on the Jordanian commercial banks

This paper aims to answer multiple questions regarding Jordanian bank stability under credit, market, and liquidity shock scenarios in order to evaluate bank solvency and liquidity. The paper will use stress testing model to measure the effect of the financial crisis in 2008 on the Jordanian commercial banks. It will conduct various stress testing scenarios on 13 Jordanian banks with the aim of better understanding the risk profile of the Jordanian banking system. Two spreadsheet models are utilised to conduct credit, market and liquidity shocks to assess the performance of the banks under each shock. The findings of this paper demonstrate that the Jordanian banking system is generally able to withstand reasonable shocks of each type, with some addressable weaknesses such as collateral under-provisioning and liquid asset shortcomings.

Modelling of market risk stress-testing scenarios using methods of vector-regression analysis

Modelling of market risk stress-testing scenarios using methods of vector-regression analysis

Stress Testing U.S. Leveraged Corporates in a COVID-19 World

This paper analyzes a group of 755 firms, with aggregate indebtedness of US$6.2 trillion, to assess the solvency risks and liquidity needs facing the U.S. corporate sector based on projections of net income, availability and cost of funding, and debt servicing flows under different stress test scenarios. The paper finds that leveraged corporates account for most of the potential losses arising from the macroeconomic stresses associated with the COVID-19 crisis, with a concentration of these losses in the oil and gas, auto, and capital and durable goods manufacturing sectors. However, potential losses from corporate debt write-downs appear to be a fraction of banks’ capital buffers and, given the size of the leveraged segment and the relatively long duration of that sector’s debt, the near-term liquidity needs of these corporates appear modest. Corporate stresses could, however, amplify the current economic downturn—as firms cut investment spending and reduce employment—potentially giving rise to significant indirect losses for the financial system.