Details
| Original language | English |
|---|---|
| Pages (from-to) | 6529-6552 |
| Number of pages | 24 |
| Journal | Management science |
| Volume | 67 |
| Issue number | 10 |
| Early online date | 18 Mar 2021 |
| Publication status | Published - Oct 2021 |
| Externally published | Yes |
Abstract
This paper assembles a toolkit for the assessment of model risk when model uncertainty sets are defined in terms of an F-divergence ball around a reference model. We propose a new family of F-divergences that are easy to implement and flexible enough to imply convincing uncertainty sets for broad classes of reference models. We use our theoretical results to construct concrete examples of divergences that allow for significant amounts of uncertainty about lognormal or heavy-tailed Weibull reference models without implying that the worst case is necessarily infinitely bad. We implement our tools in an open-source software package and apply them to three risk management problems from operations management, insurance, and finance.
Keywords
- F-divergence, Model risk, Risk management, Robustness
ASJC Scopus subject areas
- Business, Management and Accounting(all)
- Strategy and Management
- Decision Sciences(all)
- Management Science and Operations Research
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Management science, Vol. 67, No. 10, 10.2021, p. 6529-6552.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - A Toolkit for Robust Risk Assessment Using F-Divergences
AU - Kruse, Thomas
AU - Schneider, Judith C.
AU - Schweizer, Nikolaus
N1 - Publisher Copyright: Copyright: © 2021 The Author(s)
PY - 2021/10
Y1 - 2021/10
N2 - This paper assembles a toolkit for the assessment of model risk when model uncertainty sets are defined in terms of an F-divergence ball around a reference model. We propose a new family of F-divergences that are easy to implement and flexible enough to imply convincing uncertainty sets for broad classes of reference models. We use our theoretical results to construct concrete examples of divergences that allow for significant amounts of uncertainty about lognormal or heavy-tailed Weibull reference models without implying that the worst case is necessarily infinitely bad. We implement our tools in an open-source software package and apply them to three risk management problems from operations management, insurance, and finance.
AB - This paper assembles a toolkit for the assessment of model risk when model uncertainty sets are defined in terms of an F-divergence ball around a reference model. We propose a new family of F-divergences that are easy to implement and flexible enough to imply convincing uncertainty sets for broad classes of reference models. We use our theoretical results to construct concrete examples of divergences that allow for significant amounts of uncertainty about lognormal or heavy-tailed Weibull reference models without implying that the worst case is necessarily infinitely bad. We implement our tools in an open-source software package and apply them to three risk management problems from operations management, insurance, and finance.
KW - F-divergence
KW - Model risk
KW - Risk management
KW - Robustness
UR - http://www.scopus.com/inward/record.url?scp=85117249969&partnerID=8YFLogxK
U2 - 10.1287/mnsc.2020.3822
DO - 10.1287/mnsc.2020.3822
M3 - Article
AN - SCOPUS:85117249969
VL - 67
SP - 6529
EP - 6552
JO - Management science
JF - Management science
SN - 0025-1909
IS - 10
ER -