Prognozowanie Value-at-Risk dla portfeli inwestycyjnych
Value-at-Risk Forecasting for Investment Portfolios

Summary/Abstract: Aim: The aim of this paper is to identify the Value-at-Risk estimation method that generates the most accurate forecasts for investment portfolios. The paper assumes that the accuracy of Value-at-Risk forecasts is measured using the Kupiec and Christoffersen tests. Additionally, the impact of the portfolio construction on its vulnerability to risk was also analyzed. Methodology: The study was conducted using data on stock market indices obtained from Stooq.com. The historical simulation method and two GARCH(1,1) models with different theoretical distributions were used to estimate Value-at-Risk. Furthermore, the results were validated and verified, taking into account events such as Brexit, the COVID-19 pandemic, and Russia's attack on Ukraine. Results: The most accurate Value-at-Risk forecasts were obtained using the GARCH(1,1) model with a skewed t-student distribution. Additionally, in almost every case, the GARCH models generated more accurate forecasts than the historical simulation. Implications and recommendations: Assets and investment portfolios are often exposed to losses caused by an increase in market risk during periods of significant market changes. VaR forecasting is one of the key tools used in financial risk analysis, particularly in the context of investment portfolios. Investors should combine VaR with other analytical tools and regularly test and update models depending on changing market conditions. Originality/value: This issue is crucial from the perspective of risk management by banks and other financial institutions. Currently, there is no recommended Value-at-Risk estimation method that serves as an industry standard. Therefore, an attempt was made to develop and test an approach within the ways of determining the value at risk.

• Details
• Contents