Paper: arXiv 2405.00606

Authors: Lars Holden

Abstract

The paper discusses capital allocation using the Euler formula and focuses on the risk measures Value-at-Risk (VaR) and Expected shortfall (ES). Some new results connected to this capital allocation is known. Two examples illustrate that capital allocation with VaR is not monotonous which may be surprising since VaR is monotonous. A third example illustrates why the same risk measure should be used in capital allocation as in the evaluation of the total portfolio. We show how simulation may be used in order to estimate the expected Return on risk adjusted capital in the commitment period of an asset. Finally, we show how Markov chain Monte Carlo may be used in the estimation of the capital allocation.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 3.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper involves advanced mathematical concepts like coherent risk measures, Euler’s theorem, and MCMC estimation, but lacks any empirical backtesting or dataset implementation, focusing instead on theoretical properties and illustrative examples.

Research Flowchart

  flowchart TD
  A["Research Goal: Investigate<br>Euler Capital Allocation<br>using VaR and ES"] --> B["Methodology: Theoretical Analysis<br>and Numerical Simulation"]
  
  B --> C["Key Analysis 1: Monotonicity<br>Test of VaR Allocation"]
  C --> D["Outcome 1: VaR capital allocation<br>can be non-monotonic<br>despite VaR being monotonic"]
  
  B --> E["Key Analysis 2: Consistency<br>of Risk Measures"]
  E --> F["Outcome 2: Allocation and evaluation<br>must use the same risk measure<br>for consistency"]
  
  B --> G["Key Analysis 3: Computational<br>Estimation via MCMC"]
  G --> H["Outcome 3: MCMC enables efficient<br>estimation of Expected Return on<br>Risk Adjusted Capital ERARC"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes using Markov chain Monte Carlo to estimate Euler capital allocation.

Source sentence
Finally, we show how Markov chain Monte Carlo may be used in the estimation of the capital allocation. 1.

Assumes 1

  • The risk measure is assumed to be positive homogeneous. model
    Source sentence
    In this paper we will assume the risk measure is positive homogeneous but do not require the other properties above.

Data

  • Synthetic portfolio simulated from normal distributions synthetic unknown
    Source sentence
    Assume Xj i = ai −exp(Y j i ) where Y j i ∼N(µi, σ2 i ) and all Y j i are uncorrelated in order to simplify the notation.

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • Euler capital allocation with VaR is not monotone even though VaR itself is monotone. theoretical
    Source sentence
    We have contributed to the research showing that: • Euler capital allocation with VaR is not monotone even though VaR is monotone.
  • The same risk measure should be used for the portfolio evaluation and in the capital allocation. theoretical
    Source sentence
    We have contributed to the research showing that: • Euler capital allocation with VaR is not monotone even though VaR is monotone.
  • Simulation may be used to estimate the expected RORAC over the commitment period of an asset. theoretical
    Source sentence
    We have contributed to the research showing that: • Euler capital allocation with VaR is not monotone even though VaR is monotone.
  • Markov chain Monte Carlo can improve sampling efficiency by a factor of about 100 compared to plain Monte Carlo simulation. performance
    Source sentence
    This indicates that that MCMC may give an improvement with a factor 100 compared to Monte Carlo simulation.
  • Importance sampling can improve estimation efficiency by a factor of about 10 compared to Monte Carlo simulation. performance
    Source sentence
    This implies that importance sampling may give an improvement with a factor 10 compared to Monte Carlo simulation.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 31 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.