|
on Risk Management |
| By: | Hansjoerg Albrecher (University of Lausanne; Swiss Finance Institute); Michel M. Dacorogna (PRS Solutions) |
| Abstract: | This paper deals with the problem of assessing the risk related to time. On one hand, c}lassical regulatory rules for capital allocation of long-tailed insurance risks do not ask insurance companies to hold solvency capital early in the process. However, this may underestimate the risk of a deterioration of the credit state of the company until the time when the capital is needed. On the other hand, actually implemented actuarial capital management strategies of companies can often be interpreted as implicitly allocating that capital earlier than demanded. We propose a framework for quantifying risks associated with time. Using it, we evaluate strategies for capital allocation as a function of the time to ultimate, aiming to effectively manage long-tail business without impeding its growth. We model the impact of exogenous credit migration risk and the financial repercussions of overlooking it. We evaluate six different strategies, including the costs associated with potential company bankruptcy until the settlement of long-term claims. A detailed numerical implementation is provided through a simple example of a stand-alone heavy-tailed insurance risk anticipated in the future. We estimate a Markov chain credit migration model using insurance market data and analyze and interpret the liability values resulting from the various capital management strategies discussed in this paper. It appears that the actuarial practice of early capital raising is costly, even with appropriate penalties for avoided credit risk, unless the company's initial credit rating is poor. In such cases, purchasing protection through a credit derivative could be more efficient, provided such products are available in the market. |
| Keywords: | Insurance, solvency capital requirement, credit risk, bankruptcy, regulation |
| JEL: | G22 G31 |
| Date: | 2024–02 |
| URL: | https://d.repec.org/n?u=RePEc:chf:rpseri:rp2473 |
| By: | Peng Liu; Tiantian Mao; Ruodu Wang |
| Abstract: | Choquet capacities and integrals are central concepts in decision making under ambiguity or model uncertainty, pioneered by Schmeidler. Motivated by risk optimization problems for quantiles under ambiguity, we study the subclass of Choquet integrals, called Choquet quantiles, which generalizes the usual (probabilistic) quantiles, also known as Value-at-Risk in finance, from probabilities to capacities. Choquet quantiles share many features with probabilistic quantiles, in terms of axiomatic representation, optimization formulas, and risk sharing. We characterize Choquet quantiles via only one axiom, called ordinality. We prove that the inf-convolution of Choquet quantiles is again a Choquet quantile, leading to explicit optimal allocations in risk sharing problems for quantile agents under ambiguity. A new class of risk measures, Choquet Expected Shortfall, is introduced, which enjoys most properties of the coherent risk measure Expected Shortfall. Our theory is complemented by optimization algorithms, numerical examples, and a stylized illustration with financial data. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2412.19546 |
| By: | Gauch, Kevin |
| Abstract: | Companies are exposed to various risks in their day-to-day business that can affect their financial performance, competitiveness, and long-term profitability. Trends such as globalization and rapid technological development are changing the dynamics of and uncertainties faced by companies and increasing the likelihood of crises. The COVID-19 pandemic, the Wirecard scandal, and cyberattacks are just a few recent examples. Therefore, companies must deal with risks in a structured manner using a risk management system. However, the approaches used are not standardized. Although risk management standards guide how to structure them, they still need to be customized for each company. Common risk strategies range from risk reduction and risk transfer to the avoidance of certain business activities. For example, a risk can be transferred to insurance companies or reduced by voluntary assurance of the risk management system. With the introduction of the Financial Market Integrity Strengthening Act, risk management systems have become mandatory for listed companies in Germany. In the United States, a risk management system is not mandatory, although this is the case for an internal control system for financial reporting. Due to the high relevance of risk management systems, companies can voluntarily implement risk management system assurance to verify the effectiveness and appropriateness of the system. This can ensure that risks are adequately managed, while also sending a positive signal to stakeholders. However, it is not the mere implementation of the risk management system that is crucial, but also the communication of the risks and measures that the company intends to take to manage them. By disclosing risk-related information, managers can demonstrate their risk management capabilities and thus reduce information asymmetries between the company and its stakeholders. In addition, risk-related information is of major interest to stakeholders, as it enables them to more effectively assess the company’s risk exposure. In addition to mandatory risk disclosure and risk-related information, companies tend to supplement this with voluntary information. Given the relevance of risk disclosure and related assurance services, this dissertation deals with these topics in two main chapters. The first five studies deal with the spectrum of risk disclosure, whereas the last two address the impact of assurance services. The first study examines risk disclosure in the German capital market. For this purpose, the annual reports of HDAX companies from the 2018, 2019, and 2020 fiscal years were examined, using qualitative content analysis. The study focused on the volume of disclosure, the reported risk categories and individual risks over the period mentioned. The results indicate that the number of individual risks published increased significantly. Currency and cyber risks in particular were discussed frequently. Companies and stakeholders can use the results to identify best practices in risk disclosure. For legislators, the results offer guidance for further statutory regulation. The second study examines the determinants of risk disclosure using regression analysis. Again, the annual reports of HDAX companies between 2018 and 2020 were used as the data base. The determinants were identified for the volume of risk disclosure, individual risks, and risk management measures. The results contribute to recognizing the influencing factors, which can help investors make informed decisions. The third study examines textual dissimilarity in risk disclosures and its determinants in the US capital market from 2005 to 2022, with a sample of 29, 070 company-year observations. The results provide empirical evidence that risk disclosure is regularly updated only to a limited extent, except for unforeseen events such as the financial crisis or the COVID-19 pandemic. Concerning the determinants, it is evident that risk variables and audit-specific variables, in particular, influence textual dissimilarity. The fourth study describes a qualitative content analysis of HDAX companies for the 2019 fiscal year regarding disclosures on risk management systems. The results indicate rather heterogeneous reporting. An average of 6.52 of 8 basic components of the IDW assurance standard IDW AsS 981 were reported. However, only a few companies disclose that they have oriented towards a risk management standard. Notably, only four companies state that they have voluntarily assured their risk management system. Although the results indicate high reporting quality, best practices for reporting can also be identified, which also provides indications for statutory regulations. The fifth study is dedicated to the disclosure of IT risks. Due to increasing digitalization and technological trends, considering new types of risks, such as IT risks, is of particular interest. A qualitative content analysis was used to evaluate the 2020 annual reports of DAX and MDAX companies. The results also demonstrate heterogeneous reporting. Notably, only 25 of the 90 companies follow international standards, while only twelve have been certified. Cyber insurance is rarely mentioned. This study also indicates best practices in reporting on IT risks and can serve as a basis for the regulator to initiate further standardization of risk disclosure. The sixth study examines the voluntary assurance of risk management systems with an experiment. For this purpose, 145 German bankers were asked whether or not they trust in the risk management system, loan granting, willingness to invest, and to recommend investing in a hypothetical company. For this purpose, the assurance itself, the assurance providers, and the assurance level were manipulated. The results indicate that voluntary assurance significantly increases trust in the risk management system, the probability of a loan being granted, and the willingness to invest and investment recommendations. However, neither the auditor provider nor the assurance level play a decisive role in the participants’ decision, so it can be stated that the mere presence of an assurance is sufficient. From a regulatory perspective, introducing a mandatory assurance of risk management systems could be considered. In addition, our results show that companies can benefit directly from voluntary assurance, as this can increase the chances of obtaining financing. Also using an experiment, the seventh study examines voluntary cybersecurity assurance and the purchase of cyber risk insurance. For this purpose, 100 non-professional investors were asked about their willingness to invest. The presence of assurance and the presence of cyber insurance were manipulated. An additional experiment varied the assurance provider. The experimental results indicate positive perceptions of a voluntary cybersecurity audit and cyber insurance. Non-professional investors are more willing to invest in a company if it has engaged an assurance or has purchased insurance against cyber risks. In contrast, the specific assurance provider is irrelevant to our participants, revealing that the mere existence of the assurance is considered sufficient. From a regulatory perspective, introducing a mandatory cybersecurity assurance and/or mandatory cyber risk insurance could be considered, due to the high relevance of cyber risks. The results also demonstrate that companies can benefit directly from voluntary assurance, as this could increase equity financing. |
| Date: | 2024–12–11 |
| URL: | https://d.repec.org/n?u=RePEc:dar:wpaper:150929 |
| By: | Nico Herrig |
| Abstract: | This work aims to implement Long Short-Term Memory mixture density networks (LSTM-MDNs) for Value-at-Risk forecasting and compare their performance with established models (historical simulation, CMM, and GARCH) using a defined backtesting procedure. The focus was on the neural network's ability to capture volatility clustering and its real-world applicability. Three architectures were tested: a 2-component mixture density network, a regularized 2-component model (Arimond et al., 2020), and a 3-component mixture model, the latter being tested for the first time in Value-at-Risk forecasting. Backtesting was performed on three stock indices (FTSE 100, S&P 500, EURO STOXX 50) over two distinct two-year periods (2017-2018 as a calm period, 2021-2022 as turbulent). Model performance was assessed through unconditional coverage and independence assumption tests. The neural network's ability to handle volatility clustering was validated via correlation analysis and graphical evaluation. Results show limited success for the neural network approach. LSTM-MDNs performed poorly for 2017/2018 but outperformed benchmark models in 2021/2022. The LSTM mechanism allowed the neural network to capture volatility clustering similarly to GARCH models. However, several issues were identified: the need for proper model initialization and reliance on large datasets for effective learning. The findings suggest that while LSTM-MDNs provide adequate risk forecasts, further research and adjustments are necessary for stable performance. |
| Date: | 2025–01 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2501.01278 |
| By: | Lorenzo Frattarolo |
| Abstract: | Financial crises are usually associated with increased cross-sectional dependence between asset returns, causing asymmetry between the lower and upper tail of return distribution. The detection of asymmetric dependence is now understood to be essential for market supervision, risk management, and portfolio allocation. I propose a non-parametric test procedure for the hypothesis of copula central symmetry based on the Cram\'er-von Mises distance of the empirical copula and its survival counterpart, deriving the asymptotic properties of the test under standard assumptions for stationary time series. I use the powerful tie-break bootstrap that, as the included simulation study implies, allows me to detect asymmetries with up to 25 series and the number of observations corresponding to one year of daily returns. Applying the procedure to US portfolio returns separately for each year in the sample shows that the amount of copula central asymmetry is time-varying and less present in the recent past. Asymmetry is more critical in portfolios based on size and less in portfolios based on book-to-market and momentum. In portfolios based on industry classification, asymmetry is present during market downturns, coherently with the financial contagion narrative. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2501.00634 |
| By: | Jie Cao (The Hong Kong Polytechnic University - School of Accounting and Finance); Amit Goyal (University of Lausanne; Swiss Finance Institute); Yajing (Stella) Wang (The Hong Kong Polytechnic University - School of Accounting and Finance); Xintong Zhan (Department of Finance, School of Management, Fudan University); Weiming Elaine Zhang (IE Business School - IE University) |
| Abstract: | We explore how the opioid crisis exposure affects firm downside tail risks implied from equity options. Using a large sample of U.S. public firms from 1999 to 2020, we find that firms headquartered in regions with higher opioid death rates face higher downside tail risks, i.e., the cost of option protection against left tail risks is higher. The effects are reversed following exogenous anti-opioid legislation, supporting a causal interpretation. Further analysis shows that the opioid crisis heightens firm risk by lowering labor productivity. We document more pronounced impacts among firms with higher reliance on labor, limited local labor supply, and lower workplace safety. |
| Keywords: | Opioid crisis, downside risk, equity options, labor productivity, labor supply |
| JEL: | G32 E24 J24 I18 |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:chf:rpseri:rp2474 |
| By: | Wenjie Lan |
| Abstract: | This paper distinguishes between risk resonance and risk diversification relationships in the cryptocurrency market based on the newly developed asymmetric breakpoint approach, and analyzes the risk propagation mechanism among cryptocurrencies under extreme events. In addition, through the lens of node association and network structure, this paper explores the dynamic evolutionary relationship of cryptocurrency risk association before and after the epidemic. In addition, the driving mechanism of the cryptocurrency risk movement is analyzed in a depth with the epidemic indicators. The findings show that the effect of propagation of risk among cryptocurrencies becomes more significant under the influence of the new crown outbreak. At the same time, the increase in the number of confirmed cases exacerbated the risk spillover effect among cryptocurrencies, while the risk resonance effect that exists between the crude oil market and the cryptocurrency market amplified the extent of the outbreak's impact on cryptocurrencies. However, other financial markets are relatively independent of the cryptocurrency market. This study proposes a strategy to deal with the spread of cryptocurrency risks from the perspective of a public health crisis, providing a useful reference basis for improving the regulatory mechanism of cryptocurrencies. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2412.19983 |
| By: | Eduardo Abi Jaber (CMAP - Centre de Mathématiques Appliquées de l'Ecole polytechnique - Inria - Institut National de Recherche en Informatique et en Automatique - X - École polytechnique - IP Paris - Institut Polytechnique de Paris - CNRS - Centre National de la Recherche Scientifique); Camille Illand (AXA Investment Managers, Multi Asset Client Solutions, Quantitative Research - AXA); Shaun Xiaoyuan Li (UP1 - Université Paris 1 Panthéon-Sorbonne, AXA Investment Managers, Multi Asset Client Solutions, Quantitative Research - AXA, CES - Centre d'économie de la Sorbonne - UP1 - Université Paris 1 Panthéon-Sorbonne - CNRS - Centre National de la Recherche Scientifique) |
| Abstract: | We consider the joint SPX-VIX calibration within a general class of Gaussian polynomial volatility models in which the volatility of the SPX is assumed to be a polynomial function of a Gaussian Volterra process defined as a stochastic convolution between a kernel and a Brownian motion. By performing joint calibration to daily SPX-VIX implied volatility surface data between 2012 and 2022, we compare the empirical performance of different kernels and their associated Markovian and non-Markovian models, such as rough and non-rough pathdependent volatility models. In order to ensure an efficient calibration and a fair comparison between the models, we develop a generic unified method in our class of models for fast and accurate pricing of SPX and VIX derivatives based on functional quantization and Neural Networks. For the first time, we identify a conventional one-factor Markovian continuous stochastic volatility model that is able to achieve remarkable fits of the implied volatility surfaces of the SPX and VIX together with the term structure of VIX futures. What is even more remarkable is that our conventional one-factor Markovian continuous stochastic volatility model outperforms, in all market conditions, its rough and non-rough path-dependent counterparts with the same number of parameters. |
| Keywords: | SPX and VIX modeling, Stochastic volatility, Gaussian Volterra processes, Quantization, Neural Networks |
| Date: | 2024–11 |
| URL: | https://d.repec.org/n?u=RePEc:hal:cesptp:hal-03902513 |
| By: | Radoslav Raykov |
| Abstract: | Derivatives exchanges often determine collateral requirements, which are fundamental to market safety, with dated risk models assuming normal returns. However, derivatives returns are heavy-tailed, which leads to the systematic under-collection of collateral (margin). This paper uses extreme value theory (EVT) to evaluate the cost of this margin inadequacy to market participants in the event of default. I find that the Canadian futures market was under-margined by about $1.6 billion during the Great Financial Crisis, and that the default of the highest-impact participant generates a cost of up to $302 million to be absorbed by surviving participants. I show that this cost can consume the market’s entire default fund and result in costly risk mutualization. I advocate for the adoption of EVT as a benchmarking tool and argue that the regulation of exchanges should be revised for financial products with heavy tails. |
| Keywords: | Financial institutions; Financial stability |
| JEL: | G10 G11 G20 |
| Date: | 2024–11 |
| URL: | https://d.repec.org/n?u=RePEc:bca:bocawp:24-46 |
| By: | Kagerer, B. |
| Abstract: | This paper studies the impact of conflict shocks, identified by event-day heteroskedasticity using EU-Russia sanction and countersanction announcements since 2014, on corporate credit spreads. Exploiting changes in the variance-covariance structure of financial and news variables with gas prices around (counter)sanction announcement days, the paper presents a new approach to the identification of geopolitical risk shocks. Conflict shocks raise credit spreads as both firm default risk and risk premia rise. The effect of conflict on credit risk is strongly heterogeneous across and within industries and countries, reflecting economic vulnerabilities. The borrowing cost of firms with high leverage levels and low earnings are more sensitive to conflict shocks, however, only for the former also risk premia rise, suggesting a collateral-based borrowing constraint. Heightened credit risk is also reflected in declining investment levels, rising bankruptcy rates, and elevated import prices due to conflict. |
| Keywords: | Conflict, Sanctions, Corporate Credit Market, Default Risk |
| JEL: | F50 F51 F40 G12 |
| Date: | 2024–12–13 |
| URL: | https://d.repec.org/n?u=RePEc:cam:camdae:2471 |
| By: | Chengyue Huang; Yahe Yang |
| Abstract: | With the widespread application of machine learning in financial risk management, conventional wisdom suggests that longer training periods and more feature variables contribute to improved model performance. This paper, focusing on mortgage default prediction, empirically discovers a phenomenon that contradicts traditional knowledge: in time series prediction, increased training data timespan and additional non-critical features actually lead to significant deterioration in prediction effectiveness. Using Fannie Mae's mortgage data, the study compares predictive performance across different time window lengths (2012-2022) and feature combinations, revealing that shorter time windows (such as single-year periods) paired with carefully selected key features yield superior prediction results. The experimental results indicate that extended time spans may introduce noise from historical data and outdated market patterns, while excessive non-critical features interfere with the model's learning of core default factors. This research not only challenges the traditional "more is better" approach in data modeling but also provides new insights and practical guidance for feature selection and time window optimization in financial risk prediction. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2501.00034 |
| By: | Xu, Yongdeng (Cardiff Business School) |
| Abstract: | This paper introduces an extended multivariate EGARCH model that overcomes the zero-return problem and allows for negative news and volatility spillover effects, making it an attractive tool for multivariate volatility modeling. Despite limitations, such as noninvertibility and unclear asymptotic properties of the QML estimator, our Monte Carlo simulations indicate that the standard QML estimator is consistent and asymptotically normal for larger sample sizes (i.e., T ≥ 2500). Two empirical examples demonstrate the model’s superior performance compared to multivariate GJR-GARCH and Log-GARCH models in volatility modeling. The first example analyzes the daily returns of three stocks from the DJ30 index, while the second example investigates volatility spillover effects among the bond, stock, crude oil, and gold markets. Overall, this extended multivariate EGARCH model offers a flexible and comprehensive framework for analyzing multivariate volatility and spillover effects in empirical finance research. |
| Keywords: | Multivariate EGARCH, QML Estimator, Volatility Spillovers, Zero Return |
| JEL: | C32 C58 G17 |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:cdf:wpaper:2024/24 |
| By: | Daniel Barth; Phillip J. Monin; Emil N. Siriwardane; Adi Sunderam |
| Abstract: | Since 2013, large U.S. hedge fund advisers have been required to report risk exposures in their regulatory filings. Using these data, we first establish that managers’ perceptions of risk contain useful information that is not embedded in fund returns. Investor flows do not respond to this information when managers perceive higher risk than what their past returns would indicate, suggesting managers strategically communicate their risk assessments with investors. During market downturns, investors withdraw capital from funds whose managers perceive higher risk, suggesting they find the performance of these funds in adverse market conditions surprising. These funds are identifiable ex-ante with information that is available to investors. |
| Keywords: | Delegated asset management; Hedge funds; Institutional investors; Investor behavior; Investor flows; Principal agent theory; Risk |
| JEL: | G23 G14 D82 G11 |
| Date: | 2024–12–20 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedgfe:2024-98 |
| By: | Laura Alfaro (Harvard Business School); Saleem Bahaj (University College London; Bank of England); Robert Czech (Bank of England); Jonathon Hazell (London School of Economics); Ioana Neamtu (Bank of England) |
| Abstract: | This paper studies a form of liquidity risk that we call Liquidity After Solvency Hedging or “LASH” risk. Financial institutions take LASH risk when they hedge against solvency risk, using strategies that require liquidity when the solvency of the institution improves. We focus on LASH risk relating to interest rate movements. Our framework implies that institutions with longer-duration liabilities than assets — e.g. pension funds and insurers—take more LASH risk as interest rates fall. Using UK regulatory data from 2019-22 on the universe of sterling repo and swap transactions, we measure, in real time and at the institution level, LASH risk for the non-bank sector. We find that at the peak level of LASH risk, a 100bps increase in interest rates would have led to liquidity needs close to the cash holdings of the pension fund and insurance sector. Using a cross-sectional identification strategy, we find that low interest rates caused increases in LASH risk. We then find that the pre-crisis LASH risk of non-banks predicts their bond sales during the 2022 UK bond market crisis, contributing to the yield spike in the market. |
| Keywords: | Liquidity, monetary policy, financial crisis, hedging |
| JEL: | E44 F30 G10 G22 G23 |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:cfm:wpaper:2443 |
| By: | Mukashov, A. |
| Abstract: | Even accurate and precisely specified models with excellent underlying data quality can be inhibited by parameter uncertainty that reflects uncertain factors. This paper suggests adopting portfolio management methods from finance in policy planning models as a practical tool for explicitly reckoning the parameter uncertainty when defining optimal policy. We demonstrate the approach in an economywide model that aims to find an optimal pro-poor agricultural value chain in Senegal under the world market uncertainty. We show that prioritizing the rice sector is the most effective policy in terms of expected policy return, but this policy is also associated with the highest risk, increasing poverty under unfavorable yet realistic scenarios. However, like diversified portfolios in finance, balancing rice promotion with support for other sectors can reduce risk at the cost of reduced expected policy return. The suggested methodology allows for explicit risk-based decision-making and can be used in policy prioritization models that cannot define robust optimal policy under standard parameter sensitivity analysis tests. |
| Date: | 2023 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:ifwkie:307023 |
| By: | Matias Quiroz (University of Technology Sydney, Australia); Laleh Tafakori (RMIT University, Australia); Hans Manner (University of Graz, Austria) |
| Abstract: | We investigate methods for forecasting multivariate realized covariances matrices applied to a set of 30 assets that were included in the DJ30 index at some point, including two novel methods that use existing (univariate) log of realized variance models that account for attenuation bias and time-varying parameters. We consider the implications of some modeling choices within the class of heterogeneous autoregressive models. The following are our key findings. First, modeling the logs of the marginal volatilities is strongly preferred over direct modeling of marginal volatility. Thus, our proposed model that accounts for attenuation bias (for the log-response) provides superior one-step-ahead forecasts over existing multivariate realized covariance approaches. Second, accounting for measurement errors in marginal realized variances generally improves multivariate forecasting performance, but to a lesser degree than previously found in the literature. Third, time-varying parameter models based on state-space models perform almost equally well. Fourth, statistical and economic criteria for comparing the forecasting performance lead to some differences in the model's rankings, which can partially be explained by the turbulent post-pandemic data in our out-of-sample validation dataset using sub-sample analyses. |
| Keywords: | State space model, Heterogeneous autoregressive, Realized measures, Volatility forecasting. |
| JEL: | C51 C53 G17 |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:grz:wpaper:2024-20 |
| By: | Fulvia Fringuellotti; Thomas Kroen |
| Abstract: | In June 2020, the Federal Reserve issued stringent payout restrictions for the largest banks in the United States as part of its policy response to the COVID-19 crisis. Similar curbs on share buybacks and dividend payments were adopted in other jurisdictions, including in the eurozone, the U.K., and Canada. Payout restrictions were aimed at enhancing banks’ resiliency amid heightened economic uncertainty and concerns about the risk of large losses. But besides being a tool to build capital buffers and preserve bank equity, payout restrictions may also prevent risk-shifting. This post, which is based on our recent research paper, attempts to answer whether and how payout restrictions reduce bank risk using the U.S. experience during the pandemic as a case study. |
| Keywords: | banking; payout restrictions; risk-shifting; prudential regulation |
| JEL: | G21 G28 G35 G38 |
| Date: | 2025–01–08 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fednls:99404 |
| By: | Isaak, Niklas (RWI); Jessen, Robin (RWI) |
| Abstract: | Women born later experience greater earnings growth volatility at given ages than older cohorts. This implies a welfare loss due to increased earnings risk. However, German registry data for the years 2001-2016 reveal a moderation in higher-order earnings risk: Men and women born later face higher skewness in earnings changes, indicating fewer large decreases than increases, and lower kurtosis at younger ages, implying fewer large earnings changes. These trends point at a welfare increase and persist for 5-year earnings changes, which are more reflective of persistent changes. During the Great Recession, males' skewness dropped sharply; younger women were unaffected. |
| Keywords: | wage risk, income dynamics, life cycle, business cycle |
| JEL: | D31 J31 E24 |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp17568 |
| By: | Ofelia Bonesini; Emilio Ferrucci; Ioannis Gasteratos; Antoine Jacquier |
| Abstract: | We introduce a canonical way of performing the joint lift of a Brownian motion $W$ and a low-regularity adapted stochastic rough path $\mathbf{X}$, extending [Diehl, Oberhauser and Riedel (2015). A L\'evy area between Brownian motion and rough paths with applications to robust nonlinear filtering and rough partial differential equations]. Applying this construction to the case where $\mathbf{X}$ is the canonical lift of a one-dimensional fractional Brownian motion (possibly correlated with $W$) completes the partial rough path of [Fukasawa and Takano (2024). A partial rough path space for rough volatility]. We use this to model rough volatility with the versatile toolkit of rough differential equations (RDEs), namely by taking the price and volatility processes to be the solution to a single RDE. We argue that our framework is already interesting when $W$ and $X$ are independent, as correlation between the price and volatility can be introduced in the dynamics. The lead-lag scheme of [Flint, Hambly, and Lyons (2016). Discretely sampled signals and the rough Hoff process] is extended to our fractional setting as an approximation theory for the rough path in the correlated case. Continuity of the solution map transforms this into a numerical scheme for RDEs. We numerically test this framework and use it to calibrate a simple new rough volatility model to market data. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2412.21192 |
| By: | Lee, Heungmin |
| Abstract: | The rapid advancements in large language models (LLMs) have ushered in a new era of transformative potential for the finance industry. This paper explores the latest developments in the application of LLMs across key areas of the finance domain, highlighting their significant impact and future implications. In the realm of financial analysis and modelling, LLMs have demonstrated the ability to outperform traditional models in tasks such as stock price prediction, portfolio optimization, and risk assessment. By processing vast amounts of financial data and leveraging their natural language understanding capabilities, these models can generate insightful analyses, identify patterns, and provide data-driven recommendations to support decision-making processes. The conversational capabilities of LLMs have also revolutionized the customer service landscape in finance. LLMs can engage in natural language dialogues, addressing customer inquiries, providing personalized financial advice, and even handling complex tasks like loan applications and investment planning. This integration of LLMs into financial institutions has the potential to enhance customer experiences, improve response times, and reduce the workload of human customer service representatives. Furthermore, LLMs are making significant strides in the realm of risk management and compliance. These models can analyze complex legal and regulatory documents, identify potential risks, and suggest appropriate remedial actions. By automating routine compliance tasks, such as anti-money laundering (AML) checks and fraud detection, LLMs can help financial institutions enhance their risk management practices and ensure better compliance, mitigating the risk of costly penalties or reputational damage. As the finance industry continues to embrace the transformative potential of LLMs, it will be crucial to address the challenges surrounding data privacy, algorithmic bias, and the responsible development of these technologies. By navigating these considerations, the finance sector can harness the full capabilities of LLMs to drive innovation, improve efficiency, and ultimately, enhance the overall financial ecosystem. |
| Date: | 2025–01–03 |
| URL: | https://d.repec.org/n?u=RePEc:osf:osfxxx:ahkd3 |
| By: | Fengler, Matthias; Koeniger, Winfried; Minger, Stephan |
| Abstract: | We analyze the transmission of monetary policy to the costs of hedging using options order book data. Monetary policy transmits to hedging costs both by changing the relevant state variables, such as the value of the underlying, its volatility and tail risk, and by affecting option market liquidity, including the bid-ask spread and market depth. Our estimates suggest that during the peak of the pandemic crisis in March 2020, monetary policy decisions resulted in substantial changes in hedging costs even within short intraday time windows around the decisions, amounting approximately to the annual expenses of a typical equity mutual fund. |
| Keywords: | Liquidity, Monetary policy, Option order books, Option markets, COVID-19 pandemic |
| JEL: | G13 G14 D52 E52 |
| Date: | 2025–01 |
| URL: | https://d.repec.org/n?u=RePEc:usg:econwp:2025:01 |
| By: | Francesco Audrino (University of St. Gallen; Swiss Finance Institute); Jessica Gentner (University of St. Gallen; Swiss National Bank); Simon Stalder (Swiss National Bank; University of Lugano) |
| Abstract: | This paper presents an innovative method for measuring uncertainty using Large Language Models (LLMs), offering enhanced precision and contextual sensitivity compared to the conventional methods used to construct prominent uncertainty indices. By analyzing newspaper texts with state-of-the-art LLMs, our approach captures nuances often missed by conventional methods. We develop indices for various types of uncertainty, including geopolitical risk, economic policy, monetary policy, and financial market uncertainty. Our findings show that shocks to these LLM-based indices exhibit stronger associations with macroeconomic variables, shifts in investor behaviour, and asset return variations than conventional indices, underscoring their potential for more accurately reflecting uncertainty. |
| Keywords: | Large Language Models, Economic policy, Geopolitical risk, Monetary policy, Financial markets, Uncertainty measurment |
| JEL: | C45 C55 E44 G12 |
| Date: | 2024–08 |
| URL: | https://d.repec.org/n?u=RePEc:chf:rpseri:rp2468 |
| By: | Herbst, Tobias; Plaasch, Jannick; Stammwitz, Florian |
| Abstract: | We apply the growth-at-risk model of Adrian et al. (2019) to the German commercial real estate (CRE) market. We derive a distribution for CRE price growth four quarters ahead conditional on macro-financial variables. This approach allows us to make probability statements about the downside risk to future CRE price growth, which serve as an input to financial stability analyses. We find that the conditional distribution has shifted strongly to the left since the COVID-19 pandemic, in line with deteriorating macroeconomic conditions, an increase in long-term interest rates and a decline in the net initial yield, resulting in lower expected CRE price growth rates across the entire distribution. |
| Keywords: | Commercial Real Estate, Quantile Regression, Growth-at-Risk, Germany |
| JEL: | C32 E37 G01 R33 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:bubtps:308094 |
| By: | Lorenzo Bretscher (Swiss Finance Institute - HEC Lausanne; Centre for Economic Policy Research (CEPR)); Aytek Malkhozov (McGill University); Andrea Tamoni (University of Notre Dame - Mendoza College of Business); Haoxi Yang (USun Yat-sen University (SYSU) - Lingnan (University) College) |
| Abstract: | We investigate the role of distorted beliefs in the stock market, particularly their impact on risk premia. We identify the bias in investors' expectations stemming from belief distortions and decompose the predictable component of market returns into investors' beliefs about future returns and their bias. We then show that shocks to this bias, because it manifests itself as discount-rate risk in the data but represents cash-flow risk from investors' perspective, emerges as a priced risk factor. Our findings indicate that distorted beliefs impact both the time series and cross-section of expected returns, helping to explain observed deviations from theoretical predictions under rational expectations. |
| Keywords: | distorted beliefs, return predictability, ICAPM, cross-section of stock returns |
| JEL: | G12 |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:chf:rpseri:rp2466 |
| By: | Jens Hilscher (University of California, Davis); Alon Raviv (Bar-Ilan University); Ricardo Reis (London School of Economics) |
| Abstract: | Long-dated inflation swap contracts provide widely-used estimates of expected inflation. We develop methods to estimate complementary tail probabilities for persistently very high or low inflation using inflation options prices. We show that three new adjustments to conventional methods are crucial: inflation, horizon, and risk. An application of these methods finds: (i) US deflation risk in 2011-14 has been overstated, (ii) ECB unconventional policies lowered the deflation disaster probability, (iii) inflation expectations deanchored in 2021-22, (iv) and reanchored as policy tightened, (v) but the 2021-24 disaster left scars, (vi) US expectations are less sensitive to inflation realizations than in the EZ. |
| Keywords: | option prices, inflation derivatives, Arrow-Debreu securities |
| JEL: | E31 E44 G13 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cfm:wpaper:2437 |
| By: | Carol C. Bertaut; Stephanie E. Curcuru; Ester Faia; Pierre-Olivier Gourinchas |
| Abstract: | We provide new estimates of the return on US external claims and liabilities using confidential, high-quality, security-level data. The excess return is positive on average, since claims are tilted toward higher return equities. The excess return is large and positive in normal times but large and negative during global crises, reflecting the global insurance role of the US external balance sheet. Controlling for issuer’s nationality, we find that US investors have a larger exposure to equity issued by Asia-headquartered corporations than reported in the aggregate statistics. Finally, equity portfolios are concentrated in ’superstar’ firms, but for US liabilities foreign holdings are less concentrated than the overall market. |
| Keywords: | Exorbitant privilege; excess return; portfolio composition; crises; offshore centers; superstar firms; equity portfolio; higher-return equities; equity share; Bonds; Exchange rates; Securities; Global financial crisis of 2008-2009; Global; Caribbean |
| Date: | 2024–11–22 |
| URL: | https://d.repec.org/n?u=RePEc:imf:imfwpa:2024/241 |
| By: | Wolf Wagner; Jing Zeng |
| Abstract: | We show that the too-many-to-fail problem can be resolved through an appropriate design of the bailout regime. In our model, optimal investment balances benefits from more banks investing in high-return projects against higher systemic costs due to more banks failing simultaneously. Under a standard bailout regime, banks herd, anticipating that simultaneous failures trigger bailouts. However, a policy that prioritizes bailing out a predesignated group of banks eliminates herding and achieves the first-best. If such a policy is not feasible, its benefits can be attained by decentralizing bailout decisions to two regulators each responsible for a separate group of banks. |
| Keywords: | systemic risk, optimal investment, too-many-to-fail, time-consistency, bailouts |
| JEL: | G1 G2 |
| Date: | 2024–11 |
| URL: | https://d.repec.org/n?u=RePEc:bon:boncrc:crctr224_2024_613 |
| By: | Abdollah Rida |
| Abstract: | Credit Scoring is one of the problems banks and financial institutions have to solve on a daily basis. If the state-of-the-art research in Machine and Deep Learning for finance has reached interesting results about Credit Scoring models, usage of such models in a heavily regulated context such as the one in banks has never been done so far. Our work is thus a tentative to challenge the current regulatory status-quo and introduce new BASEL 2 and 3 compliant techniques, while still answering the Federal Reserve Bank and the European Central Bank requirements. With the help of Gradient Boosting Machines (mainly XGBoost) we challenge an actual model used by BANK A for scoring through the door Auto Loan applicants. We prove that the usage of such algorithms for Credit Scoring models drastically improves performance and default capture rate. Furthermore, we leverage the power of Shapley Values to prove that these relatively simple models are not as black-box as the current regulatory system thinks they are, and we attempt to explain the model outputs and Credit Scores within the BANK A Model Design and Validation framework |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2412.20225 |
| By: | Tomer Ifergane (Ben-Gurion University of the Negev; London School of Economics; Centre For Macroeconomics) |
| Abstract: | This paper uncovers a novel interaction between production efficiency and economic stability. Using a tractable heterogeneous firms model, I prove the existence of an efficiency-stability trade-off in granular economies. Specifically, reducing misallocation increases business cycle volatility. This trade-off originates because firms choose their optimal size without internalizing their effect on aggregate consumption risk. Utilizing approximations and results on order statistics, I propose a tractable method to quantify this effect and show that commonly studied misallocation counterfactuals involve a sizeable increase in business cycle volatility. I discuss how different assumptions on the nature of misallocation and factor mobility influence this result. |
| Keywords: | Business cycles, misallocation, granularity, stabilization policies, size-dependent policies |
| JEL: | E32 D24 O47 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:cfm:wpaper:2438 |
| By: | Yijia Xiao; Edward Sun; Di Luo; Wei Wang |
| Abstract: | Significant progress has been made in automated problem-solving using societies of agents powered by large language models (LLMs). In finance, efforts have largely focused on single-agent systems handling specific tasks or multi-agent frameworks independently gathering data. However, multi-agent systems' potential to replicate real-world trading firms' collaborative dynamics remains underexplored. TradingAgents proposes a novel stock trading framework inspired by trading firms, featuring LLM-powered agents in specialized roles such as fundamental analysts, sentiment analysts, technical analysts, and traders with varied risk profiles. The framework includes Bull and Bear researcher agents assessing market conditions, a risk management team monitoring exposure, and traders synthesizing insights from debates and historical data to make informed decisions. By simulating a dynamic, collaborative trading environment, this framework aims to improve trading performance. Detailed architecture and extensive experiments reveal its superiority over baseline models, with notable improvements in cumulative returns, Sharpe ratio, and maximum drawdown, highlighting the potential of multi-agent LLM frameworks in financial trading. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2412.20138 |
| By: | Bernardus Van Doornik; Armando Gomes; David Schoenherr; Janis Skrastins |
| Abstract: | In this paper, we present a Savings-and-Credit Contract (SCC) design that mandates a savings period with a default penalty before providing credit. We demonstrate that SCCs mitigate adverse selection and can outperform traditional loan contracts amidst information frictions, thereby expanding access to credit. Empirical evidence from a financial product incorporating an SCC design supports our theory. While appearing riskier on observables, we observe lower realized default rates for product participants than for bank borrowers. Further consistent with the theory, a reform that reduces the default penalty during the savings period induces worse selection and higher realized default rates. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:bcb:wpaper:610 |
| By: | Jorge P. Zubelli; Kuldeep Singh; Vinicius Albani; Ioannis Kourakis |
| Abstract: | The Black-Scholes framework is crucial in pricing a vast number of financial instruments that permeate the complex dynamics of world markets. Associated with this framework, we consider a second-order differential operator $L(x, {\partial_x}) := v^2(x, t) (\partial_x^2 -\partial_x)$ that carries a variable volatility term $v(x, t)$ and which is dependent on the underlying log-price $x$ and a time parameter $t$ motivated by the celebrated Dupire local volatility model. In this context, we ask and answer the question of whether one can find a non-linear evolution equation derived from a zero-curvature condition for a time-dependent deformation of the operator $L$. The result is a variant of the Harry Dym equation for which we can then find a family of travelling wave solutions. This brings in extensive machinery from soliton theory and integrable systems. As a by-product, it opens up the way to the use of coherent structures in financial-market volatility studies. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2412.19020 |
| By: | Ponthiere, Gregory |
| Abstract: | This paper examines the potential role of higher education subsidies as an insurance device against the risk of having a short life, that is, as a device reducing the variance in lifetime well-being due to unequal longevities. We use a two-period dynamic OLG economy with human capital and risky lifetime to study the impact of a subsidy on higher education (financed by taxing labor earnings at older ages) on the distribution of lifetime well-being between long-lived and short-lived individuals. It is shown that, whereas the subsidy on higher education improves necessarily the lot of short-lived individuals in comparison to the laissez-faire, it is only when the subsidy is higher than a critical threshold that this reduces inequalities in lifetime well-being between long-lived and short-lived individuals. Whether one adopts the utilitarian or the ex post egalitarian social welfare function, the optimal subsidy on higher education lies above the critical threshold, but is larger under the latter social objective. |
| Keywords: | higher education, mortality risk, insurance, longevity, human capital |
| JEL: | I25 I28 I31 J17 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:glodps:1528 |
| By: | Paulo M.M. Rodrigues; Nicolau João |
| Abstract: | This paper introduces a flexible framework for the estimation of the conditional tail index of heavy tailed distributions. In this framework, the tail index is computed from an auxiliary linear regression model that facilitates estimation and inference based on established econometric methods, such as ordinary least squares (OLS), least absolute deviations, or M-estimation. We show theoretically and via simulations that OLS provides interesting results. Our Monte Carlo results highlight the adequate finite sample properties of the OLS tail index estimator computed from the proposed new framework and contrast its behavior to that of tail index estimates obtained by maximum likelihood estimation of exponential regression models, which is one of the approaches currently in use in the literature. An empirical analysis of the impact of determinants of the conditional left- and right-tail indexes of commodities’ return distributions highlights the empirical relevance of our proposed approach. The novel framework’s flexibility allows for extensions and generalizations in various directions, empowering researchers andpractitioners to straightforwardly explore a wide range of research questions. |
| JEL: | C12 C22 G17 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:ptu:wpaper:w202412 |
| By: | De Polis, Andrea (University of Strathclyde & ESCOE); Melosi, Leonardo (University of Warwick, European University Institute, DNB & CEPR); Petrella, Ivan (Collegio Carlo Alberto, University of Turin & CEPR) |
| Abstract: | We document that inflation risk in the U.S. varies significantly over time and is often asymmetric. To analyze the first-order macroeconomic effects of these asymmetric risks within a tractable framework, we construct the beliefs representation of a general equilibrium model with skewed distribution of markup shocks. Optimal policy requires shifting agents’ expectations counter to the direction of inflation risks. We perform counterfactual analyses using a quantitative general equilibrium model to evaluate the implications of incorporating real-time estimates of the balance of inflation risks into monetary policy communications and decisions. |
| Keywords: | Asymmetric risks ; optimal monetary policy ; balance of inflation risks ; risk-adjusted inflation targeting ; flexible average inflation targeting JEL Codes: E52 ; E31 ; C53 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:wrk:warwec:1530 |
| By: | Dan Li; Phillip J. Monin; Lubomir Petrasek |
| Abstract: | Constraints on the supply of credit by prime brokers affect hedge funds' leverage and performance. Using dealer surveys and hedge fund regulatory filings, we identify individual funds' credit supply from the availability of credit under agreements currently in place between a hedge fund and its prime brokers. We find that hedge funds connected to prime brokers that make more credit available to their hedge fund clients increase their borrowing and generate higher returns and alphas. These effects are more pronounced among hedge funds that rely on a small number of prime brokers, and those that rely on borrowing rather than derivatives for their leverage. Credit supply matters more for hedge fund performance during periods of financial market stress and when trading opportunities are abundant. |
| Keywords: | Hedge funds; Dealers; Leverage; Prime brokerage; Financing; Surveys |
| JEL: | G23 |
| Date: | 2024–11–21 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedgfe:2024-89 |
| By: | Kramer, Berber; Pattnaik, Subhransu; Ward, Patrick S.; Xu, Yingchen |
| Abstract: | Smallholder farmers often lack documented land rights to serve as collateral for formal loans, with livelihoods inextricably linked to weather conditions. Resulting credit and risk constraints prevent them from investing in their farms. We implemented a randomized evaluation of KhetScore, an innovative credit scoring approach that uses remote sensing to unlock credit and insurance for smallholders including landless farmers in Odisha, a state in eastern India. In our treatment group, where we offered KhetScore loans and insurance, farmers - and especially women - were more likely to be insured and borrow from formal sources without substituting formal for informal loans. Despite increased borrowing, treated households faced less difficulty in repaying loans, suggesting that insured KhetScore loans transferred risk and eased the burden of repayment. Moreover, the treatment enhanced agricultural profitability by increasing revenues during the monsoon season and reducing costs in the dry season. Positive and significant effects are found among both farmers with unconstrained baseline credit access, and quantity rationed farmers, suggesting that KhetScore helps address supply-side credit constraints. Finally, the treatment significantly enhanced women’s empowerment and mental health. In conclusion, remote sensing-enabled financial products can substantially improve landless farmers’ access to agricultural credit, risk management, resilience, and well-being. |
| Keywords: | smallholders; land rights; loans; livelihoods; weather; credit; remote sensing; access to finance; gender; impact assessment; insurance; India; Asia |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:fpr:ifprid:2288 |
| By: | Denisa Millo; Blerina Vika; Nevila Baci |
| Abstract: | The financial sector, a pivotal force in economic development, increasingly uses the intelligent technologies such as natural language processing to enhance data processing and insight extraction. This research paper through a review process of the time span of 2018-2023 explores the use of text mining as natural language processing techniques in various components of the financial system including asset pricing, corporate finance, derivatives, risk management, and public finance and highlights the need to address the specific problems in the discussion section. We notice that most of the research materials combined probabilistic with vector-space models, and text-data with numerical ones. The most used technique regarding information processing is the information classification technique and the most used algorithms include the long-short term memory and bidirectional encoder models. The research noticed that new specific algorithms are developed and the focus of the financial system is mainly on asset pricing component. The research also proposes a path from engineering perspective for researchers who need to analyze financial text. The challenges regarding text mining perspective such as data quality, context-adaption and model interpretability need to be solved so to integrate advanced natural language processing models and techniques in enhancing financial analysis and prediction. Keywords: Financial System (FS), Natural Language Processing (NLP), Software and Text Engineering, Probabilistic, Vector-Space, Models, Techniques, TextData, Financial Analysis. |
| Date: | 2024–12 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2412.20438 |
| By: | Gordon Anderson; Oliver Linton |
| Abstract: | For the last 60 years, Expected Utility Theory, Rational Expectations, and a tacit presumption of symmetry in outcome distributions have been the micro and macro foundations of decision-making paradigms which seek optimum risk tempered expected outcomes. The Sharpe ratio, in its use in evaluating portfolio performance and its focus on average returns, epitomizes the practice. When outcome distributions are symmetric unimodal, expected and most likely outcomes coincide, and choices can be construed as being made on the basis of either. However, when outcome distributions are asymmetric or multi-modal, expected outcomes are not the most likely and, in contradiction of rational expectations assumptions, expectations-based choice will engender systematic information laden surprises raising questions as to whether choice should be most likely or expected outcome based. Here, the impact of switching to a Most Likely view of the world is examined and “Most Likely†focused versions of the Sharpe and Sortino Ratios are introduced. Simple exercises performed on commonly used benchmark portfolio and stock returns data demonstrate that portfolio orderings change substantially when the focus is switched to most likely outcomes, all of which gives some pause for thought. |
| Keywords: | Portfolio choice, expected outcomes, most likely outcomes, Sharpe Ratio, Sortino Ratio |
| JEL: | G11 C14 C18 |
| Date: | 2024–12–19 |
| URL: | https://d.repec.org/n?u=RePEc:tor:tecipa:tecipa-787 |
| By: | Jourdan, Sara |
| Abstract: | The increasing availability of large amounts of valuable data and the development of ever more powerful machine learning (ML) algorithms enable ML systems to quickly and independently identify complex relationships in data. As a result, ML systems not only generate new knowledge, but also offer significant potential to augment human capabilities and assist decision makers in challenging tasks. In high-risk areas such as aviation or healthcare, humans retain final decision-making responsibility, but will increasingly collaborate with ML systems to improve decision-making processes. However, since ML systems rely on statistical approaches, they are susceptible to error, and the complexity of modern algorithms often renders the output of ML systems opaque to humans. While initial approaches from the field of explainable artificial intelligence (XAI) aim to make the output of ML systems more understandable and comprehensible to humans, current research investigating the impact of ML systems on human decision makers is limited and lacks approaches on how humans can improve their capabilities through collaboration to make better decisions in the long run. To fully exploit the potential of ML systems in high-risk areas, both humans and ML systems should be able to learn from each other to enhance their performance in the context of collaboration. Furthermore, it is essential to design effective collaboration that considers the unique characteristics of ML systems and enables humans to critically assess system decisions. This dissertation comprises five published papers that use a mixed-methods study, two quantitative experiments and two qualitative design science research (DSR) studies to explore the collaboration and bilateral influences between humans and ML systems in decision-making contexts within high-risk areas from three perspectives: (1) the human perspective, (2) the ML system perspective, and (3) the collaborative perspective. From a human perspective, this dissertation examines how humans can learn from ML systems in collaboration to enhance their own capabilities and avoid the risk of false learning due to erroneous ML output. In a mixed-methods study, radiologists segmented 690 brain tumors in MRI scans supported by either high-performing or low-performing ML systems, which provided explainable or non-explainable output design. The study shows that human decision makers can learn from ML systems to improve their decision performance and confidence. However, incorrect system outputs also lead to false learning and pose risks for decision makers. Explanations from the XAI field can significantly improve the learning success of radiologists and prevent false learning in the case of incorrect ML system output. In fact, some radiologists were even able to learn from mistakes made by low-performing ML systems when local explanations were provided with the system output. This study provides first empirical insights into the human learning potential in the context of collaborating with ML systems. The finding that explainable design of ML systems enables radiologists to identify erroneous output may facilitate earlier adoption of explainable ML systems that can improve their performance over time. The ML system perspective, on the other hand, examines how ML systems must be designed to respond flexibly to changes in human problem perception and their dynamic deployment environment. This allows the systems to also learn from humans and ensures reliable system performance in dynamic collaborative environments. Through 15 qualitative interviews with data science and ML experts in the context of a DSR study, challenges for the long-term deployment of ML systems are identified. The results show that the requirements for flexible adaptation of systems in long-term use must be established in the early phases of the ML development process. Tangible design requirements and principles for ML systems that can learn from their environment and humans are derived for all phases of the CRISP-ML(Q) process model for the development and deployment of ML models. Implementing these principles allows ML systems to maintain or even improve their performance in the long run despite occurring changes, thus creating the prerequisites for a sustainable lifecycle of ML systems. Finally, the collaborative perspective examines how the collaboration between humans and ML systems should be designed to account for the unique characteristics of ML systems, such as error proneness and opacity, as well as the cognitive biases that are inherent to human decision making. In this context, pilots were provided with different ML systems for the visual detection of other aircraft in the airspace during 222 recorded flight simulations. The experiment examines the influence of different ML error types and XAI approaches in collaboration, and shows that an explainable output design can significantly reduce ML error-induced pilot trust and performance degradation for individual error types. However, processing explanations from the XAI field increases the pilot’s mental workload. While ML errors erode the trust of human decision makers, a DSR study is conducted to derive design principles for acceptance-promoting artifacts for collaboration between humans and ML systems. Finally, the last part of the analysis shows how cognitive biases such as the IKEA effect cause humans to overvalue the results of collaboration with ML systems when a high level of personal effort is invested in the collaboration. The findings provide a broad foundation for designing effective human-AI collaboration in organizations, especially in high-risk areas where humans will be involved in decision making for the long term. Overall, the papers show how by designing effective collaboration, both humans and ML systems can benefit from each other in the long run and enhance their own capabilities. The explainable design of ML system outputs can serve as a catalyst for the adoption of ML systems, especially in high-risk areas. This dissertation defines novel requirements for the collaboration between humans and ML systems and provides guidance for ML developers, scientists, and organizations that aspire to involve both human decision makers and ML systems in decision-making processes and ensure high and robust performance in the long term. |
| Date: | 2024–11–28 |
| URL: | https://d.repec.org/n?u=RePEc:dar:wpaper:150768 |