Switching latent factor value-at-risk models for conditionally heteroskedastic portfolios: A comparative approach

Saidane, Mohamed

doi:10.1080/23737484.2022.2031346

Cited by 2 publications

(2 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To evaluate the accuracy of the VaR and ES predictions, in the out-of-sample period, we used a Monte Carlobased simulation approach similar to that described in Saidane (2022b). In each backtesting step, the observed failure rates given by the different models for the significance levels 1%, 2%, 5% and 10% were recorded.…”

Section: Monte Carlo Var and Es Analysismentioning

confidence: 99%

Risk assessment in cryptocurrency portfolios: a composite hidden Markov factor analysis framework

Saidane

2024

Stat., optim. inf. comput.

View full text Add to dashboard Cite

In this paper, we deal with the estimation of two widely used risk measures such as Value-at-Risk (VaR) and Expected Shortfall (ES) in a cryptocurrency context. To face the presence of regime switching in the cryptocurrency volatilities and the dynamic interconnection between them, we propose a Monte Carlo-based approach using heteroskedastic factor analysis and hidden Markov models (HMM) combined with a structured variational Expectation-Maximization (EM) learning approach. This composite approach allows the construction of a diversified portfolio and determines an optimal allocation strategy making it possible to minimize the conditional risk of the portfolio and maximize the return. The out-of-sample prediction experiments show that the composite factorial HMM approach performs better, in terms of prediction accuracy, than some other baseline methods presented in the literature. Moreover, our results show that the proposed methodology provides the best performing crypto-asset allocation strategies and it is also clearly superior to the existing methods in VaR and ES predictions.

show abstract

Section: Monte Carlo Var and Es Analysismentioning

confidence: 99%

Risk assessment in cryptocurrency portfolios: a composite hidden Markov factor analysis framework

Saidane

2024

Stat., optim. inf. comput.

View full text Add to dashboard Cite

show abstract

“…Results showed that the Modified AKFs perform on par with the benchmark methods, even when considering the adaptive noise covariance assumptions, suggesting that the proposed techniques offer a viable and effective approach for estimating β and VaR in financial applications. Saidane (2022) presents a computationally efficient Monte Carlobased latent factor modelling approach for estimating portfolio VaR using a Kalman filter with maximum likelihood estimation. The methodology allows for the calculation of model parameters and inferences about unobservable factors, their volatilities, and the hidden state sequence of the Markov process.…”

Section: Literature Reviewmentioning

confidence: 99%

Estimating Value at Risk and Expected Shortfall: A Kalman Filter Approach

Lecq,

Vuuren

2024

IJEFI

View full text Add to dashboard Cite

Value at Risk (VaR) estimates the maximum loss a portfolio may incur at a given confidence level over a specified time, while expected shortfall (ES) determines the probability weighted losses greater than VaR. VaR has recently been replaced by (but remains a crucial step in the computation of) ES by the Basel Committee on Banking Supervision (BCBS) as the primary metric for banks to forecast market risk and allocate the relevant amount of regulatory market risk capital. The aim of the study is to introduce a more accurate approach of measuring VaR and hence ES determined using loss forecast accuracy. VaR (hence ES) is unobservable and depends on subjective measures like volatility, more accurate (loss forecast) estimates of both are constantly sought. Modelling the volatility of asset returns as a stochastic process, so a Kalman filter (which distinguishes and isolates noise from data using Bayesian statistics and variance reduction) is used to estimate both market risk metrics. A variety of volatility estimates, including the Kalman filter's recursive approach, are used to measure VaR and ES. Loss forecast accuracy is then computed and compared. The Kalman filter produces the most accurate loss forecast estimates in periods of both calm and volatile markets. The Kalman filter provides the most accurate forecasts of future market risk losses compared with standard methods which results in more accurate provision of regulatory market risk capital.

show abstract

Switching latent factor value-at-risk models for conditionally heteroskedastic portfolios: A comparative approach

Cited by 2 publications

References 21 publications

Risk assessment in cryptocurrency portfolios: a composite hidden Markov factor analysis framework

Risk assessment in cryptocurrency portfolios: a composite hidden Markov factor analysis framework

Estimating Value at Risk and Expected Shortfall: A Kalman Filter Approach

Contact Info

Product

Resources

About