Conditional value-at-risk forecasts of an optimal foreign currency portfolio

Dongwhan Kim; Kyu Ho Kang

doi:10.1016/j.ijforecast.2020.09.011

Conditional value-at-risk forecasts of an optimal foreign currency portfolio

Dongwhan Kim, Kyu Ho Kang

Research output: Contribution to journal › Article › peer-review

5 Citations (Scopus)

Abstract

This study provides daily conditional value-at-risk (C-VaR) forecasts for a foreign currency portfolio comprising the USD/EUR, USD/JPY, and USD/BRL currencies. To do so, we estimate multivariate stochastic volatility models with time-varying conditional correlations using a Bayesian Markov chain Monte Carlo algorithm. Then, given the model-specific currency return density forecasts, we make the optimal portfolio choice by minimizing the C-VaR through numerical optimization. According to out-of-sample experiment, including emerging markets into the currency basket is essential for downside risk management, and considering model uncertainty as well as the parameter uncertainty can improve the portfolio performance.

Original language	English
Pages (from-to)	838-861
Number of pages	24
Journal	International Journal of Forecasting
Volume	37
Issue number	2
DOIs	https://doi.org/10.1016/j.ijforecast.2020.09.011
Publication status	Published - 2021 Apr 1

Bibliographical note

Funding Information:
We thank two anonymous referees and participants at the Bank of Korea seminar and KEA-APEA 2017 conference for useful feedback. This work is supported by a Korea University Grant ( K2009031 )

Publisher Copyright:
© 2020 International Institute of Forecasters

Keywords

Bayesian MCMC method
Conditional correlation
Fat tail
Stochastic volatility
Time-varying

ASJC Scopus subject areas

Business and International Management

Access to Document

10.1016/j.ijforecast.2020.09.011

Cite this

@article{1351b80a657b4289b638a616aa76c0e8,

title = "Conditional value-at-risk forecasts of an optimal foreign currency portfolio",

abstract = "This study provides daily conditional value-at-risk (C-VaR) forecasts for a foreign currency portfolio comprising the USD/EUR, USD/JPY, and USD/BRL currencies. To do so, we estimate multivariate stochastic volatility models with time-varying conditional correlations using a Bayesian Markov chain Monte Carlo algorithm. Then, given the model-specific currency return density forecasts, we make the optimal portfolio choice by minimizing the C-VaR through numerical optimization. According to out-of-sample experiment, including emerging markets into the currency basket is essential for downside risk management, and considering model uncertainty as well as the parameter uncertainty can improve the portfolio performance.",

keywords = "Bayesian MCMC method, Conditional correlation, Fat tail, Stochastic volatility, Time-varying",

author = "Dongwhan Kim and Kang, {Kyu Ho}",

note = "Funding Information: We thank two anonymous referees and participants at the Bank of Korea seminar and KEA-APEA 2017 conference for useful feedback. This work is supported by a Korea University Grant ( K2009031 ) Publisher Copyright: {\textcopyright} 2020 International Institute of Forecasters",

year = "2021",

month = apr,

day = "1",

doi = "10.1016/j.ijforecast.2020.09.011",

language = "English",

volume = "37",

pages = "838--861",

journal = "International Journal of Forecasting",

issn = "0169-2070",

publisher = "Elsevier B.V.",

number = "2",

}

TY - JOUR

T1 - Conditional value-at-risk forecasts of an optimal foreign currency portfolio

AU - Kim, Dongwhan

AU - Kang, Kyu Ho

N1 - Funding Information: We thank two anonymous referees and participants at the Bank of Korea seminar and KEA-APEA 2017 conference for useful feedback. This work is supported by a Korea University Grant ( K2009031 ) Publisher Copyright: © 2020 International Institute of Forecasters

PY - 2021/4/1

Y1 - 2021/4/1

N2 - This study provides daily conditional value-at-risk (C-VaR) forecasts for a foreign currency portfolio comprising the USD/EUR, USD/JPY, and USD/BRL currencies. To do so, we estimate multivariate stochastic volatility models with time-varying conditional correlations using a Bayesian Markov chain Monte Carlo algorithm. Then, given the model-specific currency return density forecasts, we make the optimal portfolio choice by minimizing the C-VaR through numerical optimization. According to out-of-sample experiment, including emerging markets into the currency basket is essential for downside risk management, and considering model uncertainty as well as the parameter uncertainty can improve the portfolio performance.

AB - This study provides daily conditional value-at-risk (C-VaR) forecasts for a foreign currency portfolio comprising the USD/EUR, USD/JPY, and USD/BRL currencies. To do so, we estimate multivariate stochastic volatility models with time-varying conditional correlations using a Bayesian Markov chain Monte Carlo algorithm. Then, given the model-specific currency return density forecasts, we make the optimal portfolio choice by minimizing the C-VaR through numerical optimization. According to out-of-sample experiment, including emerging markets into the currency basket is essential for downside risk management, and considering model uncertainty as well as the parameter uncertainty can improve the portfolio performance.

KW - Bayesian MCMC method

KW - Conditional correlation

KW - Fat tail

KW - Stochastic volatility

KW - Time-varying

UR - http://www.scopus.com/inward/record.url?scp=85094595940&partnerID=8YFLogxK

U2 - 10.1016/j.ijforecast.2020.09.011

DO - 10.1016/j.ijforecast.2020.09.011

M3 - Article

AN - SCOPUS:85094595940

SN - 0169-2070

VL - 37

SP - 838

EP - 861

JO - International Journal of Forecasting

JF - International Journal of Forecasting

IS - 2

ER -

Conditional value-at-risk forecasts of an optimal foreign currency portfolio

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this