<p>We propose a multilevel stochastic approximation (MLSA) scheme for the computation of the value-at-risk (VaR) and expected shortfall (ES) of a financial loss, which can only be computed via simulations conditionally on the realisation of future risk factors. Thus the problem of estimating its VaR and ES is nested in nature and can be viewed as an instance of stochastic approximation problems with biased innovations. In this framework, for a prescribed accuracy <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="780_2025_573_Article_IEq1.gif" Format="GIF" Height="10" Rendition="HTML" Resolution="72" Type="Linedraw" Width="11" /> </InlineMediaObject> <EquationSource Format="MATHML"><math> <mi>ε</mi> </math></EquationSource> <EquationSource Format="TEX">$\varepsilon $</EquationSource> </InlineEquation>, the optimal complexity of a nested stochastic approximation algorithm is shown to be of the order <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="780_2025_573_Article_IEq2.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="25" /> </InlineMediaObject> <EquationSource Format="MATHML"><math> <msup> <mi>ε</mi> <mrow> <mo>−</mo> <mn>3</mn> </mrow> </msup> </math></EquationSource> <EquationSource Format="TEX">$\varepsilon ^{-3}$</EquationSource> </InlineEquation>. To estimate the VaR, our MLSA algorithm attains an optimal complexity of the order <InlineEquation ID="IEq3"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="780_2025_573_Article_IEq3.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="41" /> </InlineMediaObject> <EquationSource Format="MATHML"><math> <msup> <mi>ε</mi> <mrow> <mo>−</mo> <mn>2</mn> <mo>−</mo> <mi>δ</mi> </mrow> </msup> </math></EquationSource> <EquationSource Format="TEX">$\varepsilon ^{-2-\delta }$</EquationSource> </InlineEquation>, where <InlineEquation ID="IEq4"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="780_2025_573_Article_IEq4.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="66" /> </InlineMediaObject> <EquationSource Format="MATHML"><math> <mi>δ</mi> <mo>∈</mo> <mo stretchy="false">(</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo stretchy="false">)</mo> </math></EquationSource> <EquationSource Format="TEX">$\delta \in (0,1)$</EquationSource> </InlineEquation> is some parameter depending on the integrability degree of the loss, while to estimate the ES, the algorithm achieves an optimal complexity of the order <InlineEquation ID="IEq5"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="780_2025_573_Article_IEq5.gif" Format="GIF" Height="20" Rendition="HTML" Resolution="72" Type="Linedraw" Width="67" /> </InlineMediaObject> <EquationSource Format="MATHML"><math> <msup> <mi>ε</mi> <mrow> <mo>−</mo> <mn>2</mn> </mrow> </msup> <mo stretchy="false">|</mo> <mo>ln</mo> <mi>ε</mi> <msup> <mo stretchy="false">|</mo> <mn>2</mn> </msup> </math></EquationSource> <EquationSource Format="TEX">$\varepsilon ^{-2}|\ln {\varepsilon }|^{2}$</EquationSource> </InlineEquation>. Numerical studies of the joint evolution of the error rate and the execution time demonstrate how our MLSA algorithm regains a significant amount of the performance lost due to the nested nature of the problem.</p>

A multilevel stochastic approximation algorithm for value-at-risk and expected shortfall estimation

期刊论文
发表日期：

2025年8月27日

查看全文

Stéphane Crépey,
Noufel Frikha,
Azar Louzi

摘要

We propose a multilevel stochastic approximation (MLSA) scheme for the computation of the value-at-risk (VaR) and expected shortfall (ES) of a financial loss, which can only be computed via simulations conditionally on the realisation of future risk factors. Thus the problem of estimating its VaR and ES is nested in nature and can be viewed as an instance of stochastic approximation problems with biased innovations. In this framework, for a prescribed accuracy $ε$ $\varepsilon $ , the optimal complexity of a nested stochastic approximation algorithm is shown to be of the order $ε^{- 3}$ $\varepsilon ^{-3}$ . To estimate the VaR, our MLSA algorithm attains an optimal complexity of the order $ε^{- 2 - δ}$ $\varepsilon ^{-2-\delta }$ , where $δ \in (0, 1)$ $\delta \in (0,1)$ is some parameter depending on the integrability degree of the loss, while to estimate the ES, the algorithm achieves an optimal complexity of the order $ε^{- 2} | ln ε |^{2}$ $\varepsilon ^{-2}|\ln {\varepsilon }|^{2}$ . Numerical studies of the joint evolution of the error rate and the execution time demonstrate how our MLSA algorithm regains a significant amount of the performance lost due to the nested nature of the problem.