The next meeting of the seminar of the Department of Operations Research, Probability and Statistics will be held on October 25th, 2022, at 2:00 p.m. (UTC+3).

Dragomir Nedelchev, a doctoral student, will deliver a talk on:

Value-at-Risk, Expected Shortfall, and expectile-based risk measure.

Abstract. The financial markets dynamics entails challenges for measuring the market risk. Hence, the agenda of both the market makers and regulators (like the Basel Committee on Bank Supervision) includes challenging and upgrading the market risk measure. Stakeholders have to reach a common understanding on the pros / cons of the available risk measures, and if necessary – design new measures.

Traditionally, the market risk is measured via the Value-at-Risk (VaR). Although marked with substantial advantages, the VaR does not meet some requirements for a coherent risk measures.

The VaR successor – the Expected Shortfall (ES) – is designed to satisfy the coherence checklist and thus embraced as the new risk measure. ES is subadditive; on the other hand, there are obstacles to backtest a ES model (i.e. there is elicitability gap).

On its way is the search for a risk measure that is both coherent and elicitable. While VaR is quantile-based and the ES reflects the tails structure beyond a certain quantile, the expectile summarizes the
features of both tails of the measured PDF.

When these 3 measures are considered together against real life data, the switch to an expectile-based risk measure becomes a justifiable conclusion.