Estimating the Value-at-Risk by Temporal VAE
Estimation of the value-at-risk (VaR) of a large portfolio of assets is an important task for financial institutions. As the joint log-returns of asset prices can often be projected to a latent space of a much smaller dimension, the use of a variational autoencoder (VAE) for estimating the VaR is a...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Risks |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-9091/11/5/79 |
| _version_ | 1797598451246039040 |
|---|---|
| author | Robert Buch Stefanie Grimm Ralf Korn Ivo Richert |
| author_facet | Robert Buch Stefanie Grimm Ralf Korn Ivo Richert |
| author_sort | Robert Buch |
| collection | DOAJ |
| description | Estimation of the value-at-risk (VaR) of a large portfolio of assets is an important task for financial institutions. As the joint log-returns of asset prices can often be projected to a latent space of a much smaller dimension, the use of a variational autoencoder (VAE) for estimating the VaR is a natural suggestion. To ensure the bottleneck structure of autoencoders when learning sequential data, we use a temporal VAE (TempVAE) that avoids the use of an autoregressive structure for the observation variables. However, the low signal-to-noise ratio of financial data in combination with the auto-pruning property of a VAE typically makes use of a VAE prone to posterior collapse. Therefore, we use annealing of the regularization to mitigate this effect. As a result, the auto-pruning of the TempVAE works properly, which also leads to excellent estimation results for the VaR that beat classical GARCH-type, multivariate versions of GARCH and historical simulation approaches when applied to real data. |
| first_indexed | 2024-03-11T03:21:25Z |
| format | Article |
| id | doaj.art-60aa3f331baa4894861dc1aceaec543d |
| institution | Directory Open Access Journal |
| issn | 2227-9091 |
| language | English |
| last_indexed | 2024-03-11T03:21:25Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Risks |
| spelling | doaj.art-60aa3f331baa4894861dc1aceaec543d2023-11-18T03:09:04ZengMDPI AGRisks2227-90912023-04-011157910.3390/risks11050079Estimating the Value-at-Risk by Temporal VAERobert Buch0Stefanie Grimm1Ralf Korn2Ivo Richert3Department of Financial Mathematics, Fraunhofer ITWM, Fraunhofer-Platz 1, 67663 Kaiserslautern, GermanyDepartment of Financial Mathematics, Fraunhofer ITWM, Fraunhofer-Platz 1, 67663 Kaiserslautern, GermanyDepartment of Mathematics, RPTU Kaiserslautern-Landau, Gottlieb-Daimler-Straße 48, 67663 Kaiserslautern, GermanyDepartment of Financial Mathematics, Fraunhofer ITWM, Fraunhofer-Platz 1, 67663 Kaiserslautern, GermanyEstimation of the value-at-risk (VaR) of a large portfolio of assets is an important task for financial institutions. As the joint log-returns of asset prices can often be projected to a latent space of a much smaller dimension, the use of a variational autoencoder (VAE) for estimating the VaR is a natural suggestion. To ensure the bottleneck structure of autoencoders when learning sequential data, we use a temporal VAE (TempVAE) that avoids the use of an autoregressive structure for the observation variables. However, the low signal-to-noise ratio of financial data in combination with the auto-pruning property of a VAE typically makes use of a VAE prone to posterior collapse. Therefore, we use annealing of the regularization to mitigate this effect. As a result, the auto-pruning of the TempVAE works properly, which also leads to excellent estimation results for the VaR that beat classical GARCH-type, multivariate versions of GARCH and historical simulation approaches when applied to real data.https://www.mdpi.com/2227-9091/11/5/79value-at-risk estimationvariational autoencodersrecurrent neural networksrisk-managementauto-pruningposterior collapse |
| spellingShingle | Robert Buch Stefanie Grimm Ralf Korn Ivo Richert Estimating the Value-at-Risk by Temporal VAE Risks value-at-risk estimation variational autoencoders recurrent neural networks risk-management auto-pruning posterior collapse |
| title | Estimating the Value-at-Risk by Temporal VAE |
| title_full | Estimating the Value-at-Risk by Temporal VAE |
| title_fullStr | Estimating the Value-at-Risk by Temporal VAE |
| title_full_unstemmed | Estimating the Value-at-Risk by Temporal VAE |
| title_short | Estimating the Value-at-Risk by Temporal VAE |
| title_sort | estimating the value at risk by temporal vae |
| topic | value-at-risk estimation variational autoencoders recurrent neural networks risk-management auto-pruning posterior collapse |
| url | https://www.mdpi.com/2227-9091/11/5/79 |
| work_keys_str_mv | AT robertbuch estimatingthevalueatriskbytemporalvae AT stefaniegrimm estimatingthevalueatriskbytemporalvae AT ralfkorn estimatingthevalueatriskbytemporalvae AT ivorichert estimatingthevalueatriskbytemporalvae |